Article Pan-cancer analysis of whole genomes https://doi.org/10.1038/s41586-020-1969-6 The ICGC/TCGA Pan-Cancer Analysis of Whole Genomes Consortium Received: 29 July 2018 Accepted: 11 December 2019 Cancer is driven by genetic change, and the advent of massively parallel sequencing has Published online: 5 February 2020 enabled systematic documentation of this variation at the whole-genome scale1–3. Here we report the integrative analysis of 2,658 whole-cancer genomes and their matching Open access normal tissues across 38 tumour types from the Pan-Cancer Analysis of Whole Genomes (PCAWG) Consortium of the International Cancer Genome Consortium (ICGC) and The Cancer Genome Atlas (TCGA). We describe the generation of the PCAWG resource, facilitated by international data sharing using compute clouds. On average, cancer genomes contained 4–5 driver mutations when combining coding and non-coding genomic elements; however, in around 5% of cases no drivers were identified, suggesting that cancer driver discovery is not yet complete. Chromothripsis, in which many clustered structural variants arise in a single catastrophic event, is frequently an early event in tumour evolution; in acral melanoma, for example, these events precede most somatic point mutations and affect several cancer-associated genes simultaneously. Cancers with abnormal telomere maintenance often originate from tissues with low replicative activity and show several mechanisms of preventing telomere attrition to critical levels. Common and rare germline variants affect patterns of somatic mutation, including point mutations, structural variants and somatic retrotransposition. A collection of papers from the PCAWG Consortium describes non-coding mutations that drive cancer beyond those in the TERT promoter4; identifies new signatures of mutational processes that cause base substitutions, small insertions and deletions and structural variation5,6; analyses timings and patterns of tumour evolution7; describes the diverse transcriptional consequences of somatic mutation on splicing, expression levels, fusion genes and promoter activity8,9; and evaluates a range of more-specialized features of cancer genomes8,10–18. Cancer is the second most-frequent cause of death worldwide, on clones during the competition to escape the tight physiological killing more than 8 million people every year; the incidence of cancer controls wired into somatic cells. Mutations that provide a selective is expected to increase by more than 50% over the coming decades19,20. advantage to the clone are termed driver mutations, as opposed to ‘Cancer’ is a catch-all term used to denote a set of diseases characterized selectively neutral passenger mutations. by autonomous expansion and spread of a somatic clone. To achieve Initial studies using massively parallel sequencing demonstrated the this behaviour, the cancer clone must co-opt multiple cellular pathways feasibility of identifying every somatic point mutation, copy-number that enable it to disregard the normal constraints on cell growth, modify change and structural variant (SV) in a given cancer1–3. In 2008, recog- the local microenvironment to favour its own proliferation, invade nizing the opportunity that this advance in technology provided, the through tissue barriers, spread to other organs and evade immune sur- global cancer genomics community established the ICGC with the veillance21. No single cellular program directs these behaviours. Rather, goal of systematically documenting the somatic mutations that drive there is a large pool of potential pathogenic abnormalities from which common tumour types22. individual cancers draw their own combinations: the commonalities of macroscopic features across tumours belie a vastly heterogeneous landscape of cellular abnormalities. The pan-cancer analysis of whole genomes This heterogeneity arises from the stochastic nature of Darwinian The expansion of whole-genome sequencing studies from individual evolution. There are three preconditions for Darwinian evolution: ICGC and TCGA working groups presented the opportunity to under- characteristics must vary within a population; this variation must be take a meta-analysis of genomic features across tumour types. To heritable from parent to offspring; and there must be competition for achieve this, the PCAWG Consortium was established. A Technical survival within the population. In the context of somatic cells, heritable Working Group implemented the informatics analyses by aggregating variation arises from mutations acquired stochastically throughout the raw sequencing data from different working groups that studied life, notwithstanding additional contributions from germline and individual tumour types, aligning the sequences to the human genome epigenetic variation. A subset of these mutations alter the cellular and delivering a set of high-quality somatic mutation calls for down- phenotype, and a small subset of those variants confer an advantage stream analysis (Extended Data Fig. 1). Given the recent meta-analysis A list of members and their affiliations appears in the online version of the paper and lists of working groups appear in the Supplementary Information. 82 | Nature | Vol 578 | 6 February 2020 Box 1 Online resources for data access, visualization and analysis The PCAWG landing page (http://docs.icgc.org/pcawg) provides UCSC Xena links to several data resources for interactive online browsing, UCSC Xena87 (https://pcawg.xenahubs.net) visualizes all PCAWG analysis and download of PCAWG data and results (Supplementary primary results, including copy-number, gene-expression, gene-fusion Table 4). and promoter-usage alterations, simple somatic mutations, large Direct download of PCAWG data somatic structural variations, mutational signatures and phenotypic Aligned PCAWG read data in BAM format are also available at data. These open-access data are available through a public Xena the European Genome Phenome Archive (EGA; https://www. hub, and consensus simple somatic mutations can be loaded to the ebi.ac.uk/ega/search/site/pcawg under accession number local computer of a user via a private Xena hub. Kaplan–Meier plots, EGAS00001001692). In addition, all open-tier PCAWG genomics histograms, box plots, scatter plots and transcript-specific views offer data, as well as reference datasets used for analysis, can be additional visualization options and statistical analyses. downloaded from the ICGC Data Portal at http://docs.icgc.org/ The Expression Atlas pcawg/data/. Controlled-tier genomic data, including SNVs and The Expression Atlas (https://www.ebi.ac.uk/gxa/home) contains indels that originated from TCGA projects (in VCF format) and RNA-sequencing and expression microarray data for querying aligned reads (in BAM format) can be downloaded using the gene expression across tissues, cell types, developmental stages Score (https://www.overture.bio/) software package, which has and/or experimental conditions88. Two different views of the data accelerated and secure file transfer, as well as BAM slicing facilities are provided: summarized expression levels for each tumour type to selectively download defined regions of genomic alignments. and gene expression at the level of individual samples, including PCAWG computational pipelines reference-gene expression datasets for matching normal tissues. The core alignment, somatic variant-calling, quality-control and PCAWG Scout variant consensus-generation pipelines used by PCAWG have each PCAWG Scout (http://pcawgscout.bsc.es/) provides a framework for been packaged into portable cross-platform images using the -omics workflow and website templating to generate on-demand, Dockstore system84 and released under an Open Source licence that in-depth analyses of the PCAWG data that are openly available to the enables unrestricted use and redistribution. All PCAWG Dockstore whole research community. Views of protected data are available images are available to the public at https://dockstore.org/ that still safeguard sensitive data. Through the PCAWG Scout web organizations/PCAWG/collections/PCAWG. interface, users can access an array of reports and visualizations ICGC Data Portal that leverage on-demand bioinformatic computing infrastructure The ICGC Data Portal85 (https://dcc.icgc.org) serves as the main to produce results in real time, allowing users to discover trends as entry point for accessing PCAWG datasets with a single uniform web well as form and test hypotheses. interface and a high-performance data-download client. This uniform Chromothripsis Explorer interface provides users with easy access to the myriad of PCAWG Chromothripsis Explorer (http://compbio.med.harvard.edu/ sequencing data and variant calls that reside in many repositories chromothripsis/) is a portal that allows structural variation in the and compute clouds worldwide. Streaming technology86 provides PCAWG dataset to be explored on an individual patient basis users with high-level visualizations in real time of BAM and VCF files through the use of circos plots. Patterns of chromothripsis can also stored remotely on the Cancer Genome Collaboratory. be explored in aggregated formats. of exome data from the TCGA Pan-Cancer Atlas23–25, scientific working processed to call transcriptomic alterations. Germline variants identi- groups concentrated their efforts on analyses best-informed by whole- fied by the three separate pipelines included single-nucleotide poly- genome sequencing data. morphisms, indels, SVs and mobile-element insertions (Supplementary We collected genome data from 2,834 donors (Extended Data Table 2). Table 1), of which 176 were excluded after quality assurance. A further The requirement to uniformly realign and call variants on approxi- 75 had minor issues that could affect some of the analyses (grey-listed mately 5,800 whole genomes presented considerable computational donors) and 2,583 had data of optimal quality (white-listed donors) challenges, and raised ethical issues owing to the use of data from dif- (Supplementary Table 1). Across the 2,658 white- and grey-listed donors, ferent jurisdictions (Extended Data Table 2). We used cloud comput- whole-genome sequencing data were available from 2,605 primary ing26,27 to distribute alignment and variant calling across 13 data centres tumours and 173 metastases or local recurrences. Mean read coverage on 3 continents (Supplementary Table 3). Core pipelines were pack- was 39× for normal samples, whereas tumours had a bimodal cover- aged into Docker containers28 as reproducible, stand-alone packages, age distribution with modes at 38× and 60× (Supplementary Fig. 1). which we have made available for download. Data repositories for raw RNA-sequencing data were available for 1,222 donors. The final cohort and derived datasets, together with portals for data visualization and comprised 1,469 men (55%) and 1,189 women (45%), with a mean age of exploration, have also been created (Box 1 and Supplementary Table 4). 56 years (range, 1–90 years) across 38 tumour types (Extended Data Table 1 and Supplementary Table 1). To identify somatic mutations, we analysed all 6,835 samples using Benchmarking of genetic variant calls a uniform set of algorithms for alignment, variant calling and quality To benchmark mutation calling, we ran the 3 core pipelines, together control (Extended Data Fig. 1, Supplementary Fig. 2 and Supplementary with 10 additional pipelines, on 63 representative tumour–normal Methods 2). We used three established pipelines to call somatic single- genome pairs (Supplementary Note 1). For 50 of these cases, we per- nucleotide variations (SNVs), small insertions and deletions (indels), formed validation by hybridization of tumour and matched normal DNA copy-number alterations (CNAs) and SVs. Somatic retrotransposition to a custom bait set with deep sequencing29. The 3 core somatic variant- events, mitochondrial DNA mutations and telomere lengths were also calling pipelines had individual estimates of sensitivity of 80–90% called by bespoke algorithms. RNA-sequencing data were uniformly to detect a true somatic SNV called by any of the 13 pipelines; more Nature | Vol 578 | 6 February 2020 | 83 Article a 1.00 OICR-bl SGA DKFZ Sanger b 1.00 Adiscan LOH c BETA T H complete m e MuSE 0.9 M 0. Tier0 W WUSTL MuTect Mu SGA GA Sanger ng Pan-cancer burden of somatic mutations DKFZ DK 0.75 F1 = 0 0.8 0.75 U L WUSTL F1 = 0.8 Across the 2,583 white-listed PCAWG donors, we called 43,778,859 MuTect2 M t2 F1 = 0 0.7 SMuFin M F1 = 0.7 somatic SNVs, 410,123 somatic multinucleotide variants, 2,418,247 F1 = 0 Precision 0.6 Precision 0.50 F1 = 0.6 somatic indels, 288,416 somatic SVs, 19,166 somatic retrotransposition F1 = 0 0.5 0.50 F1 = 0.5 F1 = 0 0.4 F1 = 0.4 events and 8,185 de novo mitochondrial DNA mutations (Supplemen- tary Table 1). There was considerable heterogeneity in the burden of novobreak o r indel de CRG Clindel F1 = 0 0.3 F1 = 0 0.3 0.25 0.25 F1 = 0 0.2 F1 = 0 0.2 somatic mutations across patients and tumour types, with a broad F1 = 0.1 F1 = 0.1 correlation in mutation burden among different classes of somatic 0 0 0 0.25 0.50 0.75 1.00 0 0.25 0.50 0.75 1.00 variation (Extended Data Fig. 3). Analysed at a per-patient level, this Sensitivity Sensitivity correlation held, even when considering tumours with similar purity c DKFZ Sanger two_plus d DKFZ SMuFin two_plus and ploidy (Supplementary Fig. 3). Why such correlation should apply MuTect Logistic regression Sanger Logistic regression on a pan-cancer basis is unclear. It is likely that age has some role, as we 1.0 1.00 0.8 observe a correlation between most classes of somatic mutation and F1 F1 0.50 0.6 0 age at diagnosis (around 190 SNVs per year, P = 0.02; about 22 indels 1.0 Accuracy 1.00 Accuracy Precision Sensitivity Precision Sensitivity 0.8 0.50 per year, P = 5 × 10−5; 1.5 SVs per year, P < 2 × 10−16; linear regression 0.6 1.0 0 with likelihood ratio tests; Supplementary Fig. 4). Other factors are 1.00 0.8 0.50 also likely to contribute to the correlations among classes of somatic 0.6 0 mutation, as there is evidence that some DNA-repair defects can cause multiple types of somatic mutation30, and a single carcinogen can cause ] ] ] ] ] ] ] ] ] ] .1 .2 .3 .5 ,1 .1 .2 .3 .5 ,1 ,0 ,0 ,0 ,0 .5 ,0 ,0 ,0 ,0 .5 [0 .1 .2 .3 (0 [0 .1 .2 .3 (0 (0 (0 (0 (0 (0 (0 VAF VAF a range of DNA lesions31. Fig. 1 | Validation of variant-calling pipelines in PCAWG. a, Scatter plot of estimated sensitivity and precision for somatic SNVs across individual Panorama of driver mutations in cancer algorithms assessed in the validation exercise across n = 63 PCAWG samples. We extracted the subset of somatic mutations in PCAWG tumours Core algorithms included in the final PCAWG call set are shown in blue. b, Sensitivity and precision estimates across individual algorithms for that have high confidence to be driver events on the basis of current somatic indels. c, Accuracy (precision, sensitivity and F1 score, defined as knowledge. One challenge to pinpointing the specific driver muta- 2 × sensitivity × precision/(sensitivity + precision)) of somatic SNV calls across tions in an individual tumour is that not all point mutations in recur- variant allele fractions (VAFs) for the core algorithms. The accuracy of two rently mutated cancer-associated genes are drivers32. For genomic methods of combining variant calls (two-plus, which was used in the final elements significantly mutated in PCAWG data, we developed a ‘rank- dataset, and logistic regression) is also shown. d, Accuracy of indel calls and-cut’ approach to identify the probable drivers (Supplementary across variant allele fractions. Methods 8.1). This approach works by ranking the observed mutations in a given genomic element based on recurrence, estimated functional consequence and expected pattern of drivers in that element. We then than 95% of SNV calls made by each of the core pipelines were genu- estimate the excess burden of somatic mutations in that genomic ine somatic variants (Fig. 1a). For indels—a more-challenging class of element above that expected for the background mutation rate, and cut variants to identify with short-read sequencing—the 3 core algorithms the ranked mutations at this level. Mutations in each element with the had individual sensitivity estimates in the range of 40–50%, with pre- highest driver ranking were then assigned as probable drivers; those cision of 70–95% (Fig. 1b). For individual SV algorithms, we estimated below the threshold will probably have arisen through chance and were precision to be in the range 80–95% for samples in the 63-sample pilot assigned as probable passengers. Improvements to features that are dataset. used to rank the mutations and the methods used to measure them Next, we defined a strategy to merge results from the three pipelines will contribute to further development of the rank-and-cut approach. into one final call-set to be used for downstream scientific analyses We also needed to account for the fact that some bona fide cancer (Methods and Supplementary Note 2). Sensitivity and precision of genomic elements were not rediscovered in PCAWG data because consensus somatic variant calls were 95% (90% confidence interval, of low statistical power. We therefore added previously known 88–98%) and 95% (90% confidence interval, 71–99%), respectively, for cancer-associated genes to the discovery set, creating a ‘compendium SNVs (Extended Data Fig. 2). For somatic indels, sensitivity and preci- of mutational driver elements’ (Supplementary Methods 8.2). Then, sion were 60% (34–72%) and 91% (73–96%), respectively (Extended Data using stringent rules to nominate driver point mutations that affect Fig. 2). Regarding somatic SVs, we estimate the sensitivity of merged these genomic elements on the basis of prior knowledge33, we separated calls to be 90% for true calls generated by any one pipeline; precision probable driver from passenger point mutations. To cover all classes was estimated as 97.5%. The improvement in calling accuracy from of variant, we also created a compendium of known driver SVs, using combining different pipelines was most noticeable in variants with analogous rules to identify which somatic CNAs and SVs are most likely low variant allele fractions, which probably originate from tumour to act as drivers in each tumour. For probable pathogenic germline subclones (Fig. 1c, d). Germline variant calls, phased using a haplotype- variants, we identified all truncating germline point mutations and reference panel, displayed a precision of more than 99% and a sensitivity SVs that affect high-penetrance germline cancer-associated genes. of 92–98% (Supplementary Note 2). This analysis defined a set of mutations that we could confidently assert, based on current knowledge, drove tumorigenesis in the more than 2,500 tumours of PCAWG. We found that 91% of tumours had at Analysis of PCAWG data least one identified driver mutation, with an average of 4.6 drivers per The uniformly generated, high-quality set of variant calls across more tumour identified, showing extensive variation across cancer types than 2,500 donors provided the springboard for a series of scientific (Fig. 2a). For coding point mutations, the average was 2.6 drivers per working groups to explore the biology of cancer. A comprehensive tumour, similar to numbers estimated in known cancer-associated suite of companion papers that describe the analyses and discoveries genes in tumours in the TCGA using analogous approaches32. across these thematic areas is copublished with this paper4–18 (Extended To address the frequency of non-coding driver point mutations, Data Table 3). we combined promoters and enhancers that are known targets of 84 | Nature | Vol 578 | 6 February 2020 Mutations Proportion of patients a b Coding Promoter 5′ UTR 0 0.1 0.3 0.5 1.0 Intron splicing 3′ UTR SoftTissue–Leiomyo ColoRect–AdenoCA Stomach–AdenoCA SCNA and SV Uterus–AdenoCA Breast–AdenoCa Biliary–AdenoCA Bone–Osteosarc Ovary–AdenoCA Panc–Endocrine Prost–AdenoCA Skin–Melanoma Panc–AdenoCA Lung–AdenoCA Kidney–ChRCC CNS–PiloAstro Amplified oncogene Deleted TSG Thy–AdenoCA Eso–AdenoCa Lymph–BNHL CNS–Medullo Bladder–TCC Kidney–RCC Lymph–CLL Truncated TSG Fusion gene Head–SCC CNS–GBM Lung–SCC Liver–HCC cis-activating GR Number of 19 38 patients 101176 48 107 5 6 109 20 23 82 6 3 31 30 31 1 42 32 6 13 16 21 8 12 10 954 TP53 37 153 3 15 9 1 55 13 48 4 11 23 241 26 7 6 8 13 11 475 CDKN2A 63 79 1 35 3 1 32 2 13 19 1 21 4 3 20 2 9 6 316 ARID1A 1 206 1 4 3 13 1 8 1 28 14 8 1 4 1 287 KRAS 14 1 60 33 9 5 16 21 5 16 6 8 4 9 10 18 10 4 3 5 269 PTEN 53 7 22 74 1 19 1 11 3 2 27 6 3 3 17 263 TERT 8 94 10 5 1 27 8 16 3 7 9 10 22 4 4 6 9 11 258 CDKN2B 3 115 3 1 1 21 2 9 3 11 2 7 181 SMAD4 3 4 2 63 1 4 2 2 1 7 1 11 12 19 6 18 4 1 1 2 3 2 177 PIK3CA 42 8 2 30 23 2 4 1 1 4 2 5 5 1 4 2 11 8 4 8 167 RB1 3 4 2 1 1 52 1 1 4 75 2 2 10 2 0 1 1 162 BRAF 80 1 3 3 1 4 1 7 9 5 4 118 CTNNB1 1.1 107 107 ERG Germline 17 39 4 29 21 1 1 6 1 2 106 MYC 3.4 SCNA 73 2 1 16 1 29 10 1 5 2 3 8 6 2 2 7 3 1 1 103 NF1 1.3 1 12 8 25 12 10 4 10 8 4 1 95 CCNE1 Rearrangements 26 1.2 85 1 2 1 1 90 VHL Non-coding point muts 25 2.6 35 53 90 19p13.3a Coding point muts 76 2 3 4 3 12 2 35 1 3 3 2 10 1 1 1 5 89 KMT2D 4.6 All 91 8 4 4 3 2 1 12 5 44 1 2 1 1 88 APC 2 5 3 57 5 5 1 1 1 1 1 3 1 85 PBRM1 20 60 100 0 2.5 5.0 7.5 Patients with Number of drivers 23 30 8 8 1 6 4 2 2 84 MCL1 drivers (%) 21 30 15 4 1 2 1 6 83 CCND1 4 25 17 5 2 5 8 1 2 4 1 76 MAP2K4 WG duplications 1 1 14 35 2 1 11 4 5 74 CREBBP Germline susceptibility variants 2 9 6 2 7 1 2 2 3 11 1 3 3 10 1 2 1 1 2 1 71 ATM Somatic non-coding drivers 0 0.25 0.50 0.75 1.00 Somatic coding drivers c 400 600 Somatic/somatic Germline/somatic SGR drivers 80 Number of patients Deletion/deletion Deletion/deletion SCNA drivers 0 10 60 Deletion/GR(break) Mutation/deletion 400 200 Deletion/mutation Mutation/mutation 40 200 20 0 0 0 FAT1 ATM 0.28 MAP3K1 0.70 MAP2K4 STK11 KDM6A FAS B2M KDM5C CDKN2B SMAD4 PBRM1 BAP1 CDKN2A RB1 MEN1 BRCA1 MSR1 CDH1 0.33 SMARCA4 BRCA2 CIC APC DCC NF1 NF2 TSC1 TP53 VHL NCOR1 PTCH1 PIK3R1 AXIN1 ARID1A EME2 RNF43 BRD7 POLR2L IRF2 0.22 CREBBP 0.57 TGFBR2 FBXW7 SETD2 PTEN ATRX 0.71 TNFRSF14 DAXX DDX3X 0.77 0.80 1.00 0.63 0.82 0.91 0.46 0.76 0.17 0.47 0.48 0.75 0.86 0.42 0.83 0.77 1.00 0.76 0.43 0.69 0.92 1.00 0.75 0.53 0.66 0.36 0.57 0.38 1.00 0.57 0.38 0.52 0.58 0.47 0.67 0.86 1.00 0.52 0.25 0.33 0.73 0.71 Proportion of patients with the gene altered as biallelic Fig. 2 | Panorama of driver mutations in PCAWG. a, Top, putative driver patients. b, Genomic elements targeted by different types of mutations in the mutations in PCAWG, represented as a circos plot. Each sector represents a cohort altered in more than 65 tumours. Both germline and somatic variants tumour in the cohort. From the periphery to the centre of the plot the are included. Left, the heat map shows the recurrence of alterations across concentric rings represent: (1) the total number of driver alterations; (2) the cancer types. The colour indicates the proportion of mutated tumours and the presence of whole-genome (WG) duplication; (3) the tumour type; (4) the number indicates the absolute count of mutated tumours. Right, the number of driver CNAs; (5) the number of driver genomic rearrangements; proportion of each type of alteration that affects each genomic element. (6) driver coding point mutations; (7) driver non-coding point mutations; and c, Tumour-suppressor genes with biallelic inactivation in 10 or more patients. (8) pathogenic germline variants. Bottom, snapshots of the panorama of driver The values included under the gene labels represent the proportions of mutations. The horizontal bar plot (left) represents the proportion of patients patients who have biallelic mutations in the gene out of all patients with a with different types of drivers. The dot plot (right) represents the mean somatic mutation in that gene. GR, genomic rearrangement; SCNA, somatic number of each type of driver mutation across tumours with at least one event copy-number alteration; SGR, somatic genome rearrangement; TSG, tumour (the square dot) and the standard deviation (grey whiskers), based on n = 2,583 suppressor gene; UTR, untranslated region. non-coding drivers34–37 with those newly discovered in PCAWG data; a larger contribution in colorectal adenocarcinomas (2.4 ± 1.4 SVs this is reported in a companion paper4. Using this approach, only compared with 7.4 ± 7.0 point mutations; P = 4 × 10−10) and mature 13% (785 out of 5,913) of driver point mutations were non-coding B cell lymphomas (2.2 ± 1.3 SVs compared with 6 ± 3.8 point muta- in PCAWG. Nonetheless, 25% of PCAWG tumours bear at least one tions; P < 1 × 10−16), as previously shown38. Across tumour types, there putative non-coding driver point mutation, and one third (237 out are differences in which classes of mutation affect a given genomic of 785) affected the TERT promoter (9% of PCAWG tumours). Overall, element (Fig. 2b). non-coding driver point mutations are less frequent than coding We confirmed that many driver mutations that affect tumour- driver mutations. With the exception of the TERT promoter, indi- suppressor genes are two-hit inactivation events (Fig. 2c). For exam- vidual enhancers and promoters are only infrequent targets of driver ple, of the 954 tumours in the cohort with driver mutations in TP53, mutations4. 736 (77%) had both alleles mutated, 96% of which (707 out of 736) Across tumour types, SVs and point mutations have different rela- combined a somatic point mutation that affected one allele with tive contributions to tumorigenesis. Driver SVs are more prevalent somatic deletion of the other allele. Overall, 17% of patients had in breast adenocarcinomas (6.4 ± 3.7 SVs (mean ± s.d.) compared rare germline protein-truncating variants (PTVs) in cancer-predis- with 2.2 ± 1.3 point mutations; P < 1 × 10−16, Mann–Whitney U-test) position genes39, DNA-damage response genes40 and somatic driver and ovary adenocarcinomas (5.8 ± 2.6 SVs compared with 1.9 ± 1.0 genes. Biallelic inactivation due to somatic alteration on top of a point mutations; P < 1 × 10−16), whereas driver point mutations have germline PTV was observed in 4.5% of patients overall, with 81% of Nature | Vol 578 | 6 February 2020 | 85 Article Chromosome 5: 1,259,228 a c 15 1 Tumour-in-normal estimate (%) 10 Detection sensitivity 5 0 0 b Chromosome 5: 1,259,250 detection sensitivity 1 1 Average 0 0 Kidney–RCC Lymph–BNHL Breast–AdenoCA Bone–Epith Kidney–ChRCC Biliary–AdenoCA Bone–Cart Eso–AdenoCA Liver–HCC Lung–AdenoCa Lymph–CLL Myeloid–AML Myeloid–MDS Myeloid–MPN Panc–AdenoCA Panc–Endocrine Prost–AdenoCA Skin–Melanoma Stomach–AdenoCA Thy–AdenoCA Biliary–AdenoCA CNS–Medullo Liver–HCC Skin–Melanoma Thy–AdenoCA CNS–Medullo Head–SCC Head–SCC Bone–Osteosarc d 1 e Chromosome loss Chromosome gain 2 2q37.3 (80 genes) 19 21 3 3p21.31 SETD2 (13 genes) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 20 22 Kidney–ChRCC 4 5 5q35.2 (287 genes) 6 Chromosome 7 8p23.1 PCM1 (187 genes) 8 9 10 Panc–Endocrine 10q26.13 FGFR2 (151 genes) 11 12 13 14 15 16 16q24.3 FANCA (40 genes) 17 17p13.3 TP53 (289 genes) 18 19 20 21 22 0.25 10–2 10–4 10–7 10–20 q value Fig. 3 | Analysis of patients with no detected driver mutations. a, Individual sensitivity for TERT promoter hotspots in tumour types in which TERT is estimates of the percentage of tumour-in-normal contamination across frequently mutated. Coloured areas are estimated density distributions. patients with no driver mutations in PCAWG (n = 181). No data were available for d, Significant copy-number losses identified by two-sided hypothesis testing myelodysplastic syndromes and acute myeloid leukaemia. Points represent using GISTIC2.0, corrected for multiple-hypothesis testing. Numbers in estimates for individual patients, and the coloured areas are estimated density parentheses indicate the number of genes in significant regions when distributions (violin plots). Abbreviations of the tumour types are defined in analysing medulloblastomas without known drivers (n = 42). Significant Extended Data Table 1. b, Average detection sensitivity by tumour type for regions with known cancer-associated genes are labelled with the tumours without known drivers (n = 181). Each dot represents a given sample representative cancer-associated gene. e, Aneuploidy in chromophobe renal and is the average sensitivity of detecting clonal substitutions across the cell carcinomas and pancreatic neuroendocrine tumours without known genome, taking into account purity and ploidy. Coloured areas are estimated drivers. Patients are ordered on the y axis by tumour type and then by presence density distributions, shown for cohorts with at least five cases. c, Detection of whole-genome duplication (bottom) or not (top). these affecting known cancer-predisposition genes (such as BRCA1, in adequately sequenced genomes, lack of read depth at specific BRCA2 and ATM). driver loci can impair mutation detection. For example, only around 50% of PCAWG tumours had sufficient coverage to call a mutation (≥90% power) at the two TERT promoter hotspots, probably because PCAWG tumours with no apparent drivers the high GC content of this region causes biased coverage (Fig. 3c). Although more than 90% of PCAWG cases had identified drivers, we In fact, 6 hepatocellular carcinomas and 2 biliary cholangiocarcinomas found none in 181 tumours (Extended Data Fig. 4a). Reasons for miss- among the 181 cases with no known drivers actually did contain TERT ing drivers have not yet been systematically evaluated in a pan-cancer mutations, which were discovered after deep targeted sequencing42. cohort, and could arise from either technical or biological causes. Finally, technical reasons for missing driver mutations include fail- Technical explanations could include poor-quality samples, inad- ures in the bioinformatic algorithms. This affected 35 myeloprolif- equate sequencing or failures in the bioinformatic algorithms used. erative neoplasms in PCAWG, in which the JAK2V617F driver mutation We assessed the quality of the samples and found that 4 of the 181 should have been called. Our somatic variant-calling algorithms rely cases with no known drivers had more than 5% tumour DNA contami- on ‘panels of normals’, typically from blood samples, to remove recur- nation in their matched normal sample (Fig. 3a). Using an algorithm rent sequencing artefacts. As 2–5% of healthy individuals carry occult designed to correct for this contamination41, we identified previously haematopoietic clones43, recurrent driver mutations in these clones missed mutations in genes relevant to the respective cancer types. can enter panels of normals. Similarly, if the fraction of tumour cells in the cancer sample is low With regard to biological causes, tumours may be driven by muta- through stromal contamination, the detection of driver mutations tions in cancer-associated genes that are not yet described for that can be impaired. Most tumours with no known drivers had an aver- tumour type. Using driver discovery algorithms on tumours with no age power to detect mutations close to 100%; however, a few had known drivers, no individual genes reached significance for point muta- power in the 70–90% range (Fig. 3b and Extended Data Fig. 4b). Even tions. However, we identified a recurrent CNA that spanned SETD2 in 86 | Nature | Vol 578 | 6 February 2020 136 + SV – SV a Punctuated events across PCAWG b Kataegis Alt. C deamin. Pol η C01 LIN G2 interfocal distance TM XB IGL 1.0 PIM CX C F3 P2 36A Fraction APOBEC3 C[T>N]T Uncertain AF U 44 3 DM BT SB P1 AK M E K 2 IR 100 M IF2 D 4 IG 4 11 R 0.5 X 15 S Kataegis I CRL2L 102 5H D S1 G B 1 0 0 MI DNMPR2 104 TN L1 R4 T RF SBP860 No. foci 100 1 106 BC322 O NF Z IT L2 10 FH V M MIR P1 LPP6GAL1 1 BZR 142 ST L6 AP1 BC F595 Chromoplexy ZN L3 1.0 IRF8 SEL1 Fraction Chromoplexy SMIM20 Balanced translocations RMI2 RHOH 0.5 CIITA SOCS1 0 1.0 Multiple chrom. IGH Fraction Small TCL61A1 L Chromothripsis Amplified ZFP3 IR C F1 0.5 I D74 Far from telomere CRF4 D4 H D8 0 Liposarcoma-like C1 6L HI IST 3 TB 022O1 LT ST 1H Classic single I A A X O A BPIM B 1H 2B 2A C Events K F L7X1 SGAC 1 C BCDT G1 K1 H2 BT AC E C P AI RM BI TSDA TB NE C31 IMRRN L 1 L AT MP 3 R is m −S C e − C ne ph osa l −O −B rc er o L So Pa He ix− arc Sk is −A − C −M e en C Es no iom A Lu o− ma yo ea −A no t O st− de CA −A n A ea C d en C A Bi st− S− oCA C ma y−A ula M oR h− e A t− en A Li en CA Ly er− oCA Pr Bo ph− CC Ki teru t−A −E L ey −A no h C no A Pa C C− CA Ki idn End Olig r dn ey o o Th −RC hR e y− C CC C ne− noC p C S− Ben A M −Pi du n M elo As o ye id− tro −M L PN Bo ym Lip cra Br ng Ade −Cu K − − a n C ste NH U os ne CL dn s de pit N M ig id M y lo ll e a el Le oC r y de C ol c d rC ec Ad noC −R de C ftT ano ng TC su a C ftT nc ad SC in su d SC MYC ZCCHC7 2L o r b B nc NS Cle PAX5 Bo Ad −P ey −C cri PALM2 AKAP2 va A no o Ad o v s lo A m H St lia Lo G L − A So el Lu er− S e d N v − ad a N Bl −M Br in Sk c Chromoplexy d Homozygous deletion Amplification Rearrangement TM E NX X1 interfootprint distance PR RG 1 A 75 distance (bp) No. losses No. gains MDM2 (36) AD RU M SS 10 ERBB2 (30) CDK4 (30) TH 2 102 50 CCND1 (30) NF1 (11) 104 SOX2 (12)TERT (22) EGFR (9) RB1 (7) 106 25 BC L2 0 RTN4 RL1 25 CDKN2A (15) RPA1 106 Interbreakpoint 105 104 103 102 IGH 101 100 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 20 0 22 2 X B2 SO BT 1 S X RC RBR6 MI RSF 4 PA R3 3 L 92 5 IG RA CR T BS F2 2A 17 TB 14 W 44 BP ZB ST BR 3 AF RUN MYC 11 CASC KDM4C X1T1 Fig. 4 | Patterns of clustered mutational processes in PCAWG. a, Kataegis. the position of consecutive chromoplexy and reciprocal translocation Top, prevalence of different types of kataegis and their association with SVs footprints across PCAWG. Lymphoid, prostate and thyroid cancers exhibit (≤1 kb from the focus). Bottom, the distribution of the number of foci of recurrent events (≥2 footprints with distance ≤10 kb; pale red zone) that are kataegis per sample. Chromoplexy. Prevalence of chromoplexy across cancer likely to be driver SVs and are annotated with nearby genes and associated SVs, types, subdivided into balanced translocations and more complex events. which are shown as bold and thin arcs for chromoplexy and reciprocal Chromothripsis. Top, frequency of chromothripsis across cancer types. translocations, respectively (colours as in a). d, Effect of chromothripsis along Bottom, for each cancer type a column is shown, in which each row is a the genome and involvement of PCAWG driver genes. Top, number of chromothripsis region represented by five coloured rectangles relating to its chromothripsis-induced gains or losses (grey) and amplifications (blue) or categorization. b, Circos rainfall plot showing the distances between deletions (red). Within the identified chromothripsis regions, selected consecutive kataegis events across PCAWG compared with their genomic recurrently rearranged (light grey), amplified (blue) and homozygously position. Lymphoid tumours (khaki, B cell non-Hodgkin’s lymphoma; orange, deleted (magenta) driver genes are indicated. Bottom, interbreakpoint chronic lymphocytic leukaemia) have hypermutation hot spots (≥3 foci with distance between all subsequent breakpoints within chromothripsis regions distance ≤1 kb; pale red zone), many of which are near known cancer-associated across cancer types, coloured by cancer type. Regions with an average genes (red annotations) and have associated SVs (≤10 kb from the focus; shown interbreakpoint distance <10 kb are highlighted. C[T>N]T, kataegis with a as arcs in the centre). c, Circos rainfall plot as in b that shows the distance versus pattern of thymine mutations in a Cp TpT context. medulloblastomas that lacked known drivers (Fig. 3d), indicating that profile of chromosomal aneuploidy—patterns that have previously restricting hypothesis testing to missing-driver cases can improve been reported45,46 (Fig. 3e). The absence of other identified driver muta- power if undiscovered genes are enriched in such tumours. Inactivation tions in these patients raises the possibility that certain combinations of SETD2 in medulloblastoma significantly decreased gene expres- of whole-chromosome gains and losses may be sufficient to initiate sion (P = 0.002) (Extended Data Fig. 4c). Notably, SETD2 mutations a cancer in the absence of more-targeted driver events such as point occurred exclusively in medulloblastoma group-4 tumours (P < 1 × 10−4). mutations or fusion genes of focal CNAs. Group-4 medulloblastomas are known for frequent mutations in other Even after accounting for technical issues and novel drivers, 5.3% of chromatin-modifying genes44, and our results suggest that SETD2 loss PCAWG tumours still had no identifiable driver events. In a research of function is an additional driver that affects chromatin regulators in setting, in which we are interested in drawing conclusions about popu- this subgroup. lations of patients, the consequences of technical issues that affect Two tumour types had a surprisingly high fraction of patients with- occasional samples will be mitigated by sample size. In a clinical setting, out identified driver mutations: chromophobe renal cell carcinoma in which we are interested in the driver mutations in a specific patient, (44%; 19 out of 43) and pancreatic neuroendocrine cancers (22%; these issues become substantially more important. Careful and critical 18 out of 81) (Extended Data Fig. 4a). A notable feature of the miss- appraisal of the whole pipeline—including sample acquisition, genome ing-driver cases in both tumour types was a remarkably consistent sequencing, mapping, variant calling and driver annotation, as done Nature | Vol 578 | 6 February 2020 | 87 Article here—should be required for laboratories that offer clinical sequenc- increased with whole-genome duplications in most cancer types ing of cancer genomes. (Extended Data Fig. 7a), as previously shown in medulloblastoma58. The most recurrently associated driver was TP5352 (pan-cancer odds ratio = 3.22; pan-cancer P = 8.3 × 10−35; q < 0.05 in breast lobular (odds Patterns of clustered mutations and SVs ratio = 13), colorectal (odds ratio = 25), prostate (odds ratio = 2.6) and Some somatic mutational processes generate multiple mutations in a hepatocellular (odds ratio = 3.9) cancers; Fisher–Boschloo tests). In single catastrophic event, typically clustered in genomic space, leading two cancer types (osteosarcoma and B cell lymphoma), women had a to substantial reconfiguration of the genome. Three such processes higher incidence of chromothripsis than men (Extended Data Fig. 7b). have previously been described: (1) chromoplexy, in which repair of In prostate cancer, we observed a higher incidence of chromothripsis co-occurring double-stranded DNA breaks—typically on different chro- in patients with late-onset than early-onset disease59 (Extended Data mosomes—results in shuffled chains of rearrangements47,48 (Extended Fig. 7c). Data Fig. 5a); (2) kataegis, a focal hypermutation process that leads to Chromothripsis regions coincided with 3.6% of all identified driv- locally clustered nucleotide substitutions, biased towards a single DNA ers in PCAWG and around 7% of copy-number drivers (Fig. 4d). These strand49–51 (Extended Data Fig. 5b); and (3) chromothripsis, in which proportions are considerably enriched compared to expectation if tens to hundreds of DNA breaks occur simultaneously, clustered on selection were not acting on these events (Extended Data Fig. 7d). The one or a few chromosomes, with near-random stitching together of majority of coinciding driver events were amplifications (58%), followed the resulting fragments52–55 (Extended Data Fig. 5c). We characterized by homozygous deletions (34%) and SVs within genes or promoter the PCAWG genomes for these three processes (Fig. 4). regions (8%). We frequently observed a ≥2-fold increase or decrease in Chromoplexy events and reciprocal translocations were identified expression of amplified or deleted drivers, respectively, when these loci in 467 (17.8%) samples (Fig. 4a, c). Chromoplexy was prominent in were part of a chromothripsis event, compared with samples without prostate adenocarcinoma and lymphoid malignancies, as previously chromothripsis (Extended Data Fig. 7e). described47,48, and—unexpectedly—thyroid adenocarcinoma. Differ- Chromothripsis manifested in diverse patterns and frequencies ent genomic loci were recurrently rearranged by chromoplexy across across tumour types, which we categorized on the basis of five charac- the three tumour types, mediated by positive selection for particu- teristics (Fig. 4a). In liposarcoma, for example, chromothripsis events lar fusion genes or enhancer-hijacking events. Of 13 fusion genes or often involved multiple chromosomes, with universal MDM2 ampli- enhancer hijacking events in 48 thyroid adenocarcinomas, at least fication60 and co-amplification of TERT in 4 of 19 cases (Fig. 4d). By 4 (31%) were caused by chromoplexy, with a further 4 (31%) part of com- contrast, in glioblastoma the events tended to affect a smaller region plexes that contained chromoplexy footprints (Extended Data Fig. 5a). on a single chromosome that was distant from the telomere, resulting These events generated fusion genes that involved RET (two cases) and in focal amplification of EGFR and MDM2 and loss of CDKN2A. Acral NTRK3 (one case)56, and the juxtaposition of the oncogene IGF2BP3 melanomas frequently exhibited CCND1 amplification, and lung squa- with regulatory elements from highly expressed genes (five cases). mous cell carcinomas SOX2 amplifications. In both cases, these drivers Kataegis events were found in 60.5% of all cancers, with particularly were more-frequently altered by chromothripsis compared with other high abundance in lung squamous cell carcinoma, bladder cancer, drivers in the same cancer type and to other cancer types for the same acral melanoma and sarcomas (Fig. 4a, b). Typically, kataegis com- driver (Fig. 4d and Extended Data Fig. 7f). Finally, in chromophobe renal prises C > N mutations in a TpC context, which are probably caused cell carcinoma, chromothripsis nearly always affected chromosome by APOBEC activity49–51, although a T > N conversion in a TpT or CpT 5 (Supplementary Fig. 7): these samples had breakpoints immediately process (the affected T is highlighted in bold) attributed to error-prone adjacent to TERT, increasing TERT expression by 80-fold on average polymerases has recently been described57. The APOBEC signature compared with samples without rearrangements (P = 0.0004; Mann– accounted for 81.7% of kataegis events and correlated positively with Whitney U-test). APOBEC3B expression levels, somatic SV burden and age at diagnosis (Supplementary Fig. 5). Furthermore, 5.7% of kataegis events involved the T > N error-prone polymerase signature and 2.3% of events, most Timing clustered mutations in evolution notably in sarcomas, showed cytidine deamination in an alternative An unanswered question for clustered mutational processes is whether GpC or CpC context. they occur early or late in cancer evolution. To address this, we used Kataegis events were frequently associated with somatic SV break- molecular clocks to define broad epochs in the life history of each points (Fig. 4a and Supplementary Fig. 6a), as previously described50,51. tumour49,61. One transition point is between clonal and subclonal muta- Deletions and complex rearrangements were most-strongly associ- tions: clonal mutations occurred before, and subclonal mutations after, ated with kataegis, whereas tandem duplications and other simple the emergence of the most-recent common ancestor. In regions with SV classes were only infrequently associated (Supplementary Fig. 6b). copy-number gains, molecular time can be further divided according Kataegis inducing predominantly T > N mutations in CpTpT context to whether mutations preceded the copy-number gain (and were them- was enriched near deletions, specifically those in the 10–25-kilobase selves duplicated) or occurred after the gain (and therefore present on (kb) range (Supplementary Fig. 6c). only one chromosomal copy)7. Samples with extreme kataegis burden (more than 30 foci) comprise Chromothripsis tended to have greater relative odds of being clonal four types of focal hypermutation (Extended Data Fig. 6): (1) off-target than subclonal, suggesting that it occurs early in cancer evolution, somatic hypermutation and foci of T > N at CpTpT, found in B cell non- especially in liposarcomas, prostate adenocarcinoma and squamous Hodgkin lymphoma and oesophageal adenocarcinomas, respectively; cell lung cancer (Fig. 5a). As previously reported, chromothripsis was (2) APOBEC kataegis associated with complex rearrangements, notably especially common in melanomas62. We identified 89 separate chromo- found in sarcoma and melanoma; (3) rearrangement-independent thripsis events that affected 66 melanomas (61%); 47 out of 89 events APOBEC kataegis on the lagging strand and in early-replicating regions, affected genes known to be recurrently altered in melanoma63 (Sup- mainly found in bladder and head and neck cancer; and (4) a mix of plementary Table 6). Involvement of a region on chromosome 11 that the last two types. Kataegis only occasionally led to driver mutations includes the cell-cycle regulator CCND1 occurred in 21 cases (10 out (Supplementary Table 5). of 86 cutaneous, and 11 out of 21 acral or mucosal melanomas), typi- We identified chromothripsis in 587 samples (22.3%), most fre- cally combining chromothripsis with amplification (19 out of 21 cases) quently among sarcoma, glioblastoma, lung squamous cell carci- (Extended Data Fig. 8). Co-involvement of other cancer-associated noma, melanoma and breast adenocarcinoma18. Chromothripsis genes in the same chromothripsis event was also frequent, including 88 | Nature | Vol 578 | 6 February 2020 a No. samples 198 146 111 317 107 113 239 210 34 23 16 10 38 13 18 41 18 89 60 98 57 45 33 38 48 95 13 23 85 20 86 15 19 75 48 51 Relative odds Relative odds Fraction of (early/late) (clonal/subclonal) samples 3 2 2 1 Chromoplexy Chromothripsis 100 Kataegis 10 1 0.1 0.01 100 10 1 0.1 0.01 Biliary−AdenoCA Bladder−TCC Bone−Benign Bone−Epith Bone−Osteosarc Breast−AdenoCA Breast−DCIS Breast−LobularCA Cervix−AdenoCA Cervix−SCC CNS−GBM CNS−Medullo CNS−Oligo CNS−PiloAstro ColoRect−AdenoCA Eso−AdenoCA Head−SCC Kidney−ChRCC Kidney−RCC−Clear Kidney−RCC−Pap Liver−HCC Lung−AdenoCA Lung−SCC Lymph−BNHL Lymph−CLL Myeloid−AML Myeloid−MDS Myeloid−MPN Ovary−AdenoCA Panc−AdenoCA Panc−Endocrine Prost−AdenoCA Skin−Melanoma−Acral Skin−Melanoma−Cut Skin−Melanoma−Mucosal SoftTissue−Leiomyo SoftTissue−Liposarc Stomach−AdenoCA Thy−AdenoCA Uterus−AdenoCA b TERT CCND1 Sample: SA557318 Acral melanoma 40 C>A number C>G Copy 20 C>T 0 1.0 T>A VAF 0.5 T>C 0 T>G Sample: SA557322 Acral melanoma 120 number Copy 80 40 0 1.0 VAF 0.5 0 Sample: SA557416 Acral melanoma 40 number Copy 20 0 1.0 VAF 0.5 0 0 10 20 30 40 50 55 65 75 85 95 105 115 Chr. 5 position (Mb) Chr. 11 position (Mb) Fig. 5 | Timing of clustered events in PCAWG. a, Extent and timing of sizes (number of patients) are shown across the top. b, Three representative chromothripsis, kataegis and chromoplexy across PCAWG. Top, stacked bar patients with acral melanoma and chromothripsis-induced amplification that charts illustrate co-occurrence of chromothripsis, kataegis and chromoplexy simultaneously affects TERT and CCND1. The black points (top) represent in the samples. Middle, relative odds of clustered events being clonal or sequence coverage from individual genomic bins, with SVs shown as coloured subclonal are shown with bootstrapped 95% confidence intervals. Point arcs (translocation in black, deletion in purple, duplication in brown, tail-to-tail estimates are highlighted when they do not overlap odds of 1:1. Bottom, inversion in cyan and head-to-head inversion in green). Bottom, the variant relative odds of the events being early or late clonal are shown as above. Sample allele fractions of somatic point mutations. TERT (five cases), CDKN2A (three cases), TP53 (two cases) and MYC allele fraction. Regions of CCND1 amplification had few—sometimes (two cases) (Fig. 5b). In these co-amplifications, a chromothripsis zero—mutations at high variant allele fraction in acral melanomas, in event involving multiple chromosomes initiated the process, creat- contrast to later CCND1 amplifications in cutaneous melanomas, in ing a derivative chromosome in which hundreds of fragments were which hundreds to thousands of mutations typically predated ampli- stitched together in a near-random order (Fig. 5b). This derivative fication (Fig. 5b and Extended Data Fig. 9a, b). Thus, both chromoth- then rearranged further, leading to massive co-amplification of the ripsis and the subsequent amplification generally occurred very early multiple target oncogenes together with regions located nearby on during the evolution of acral melanoma. By comparison, in lung squa- the derivative chromosome. mous cell carcinomas, similar patterns of chromothripsis followed by In these cases of amplified chromothripsis, we can use the inferred SOX2 amplification are characterized by many amplified SNVs, sug- number of copies bearing each SNV to time the amplification process. gesting a later event in the evolution of these cancers (Extended Data SNVs present on the chromosome before amplification will them- Fig. 9c). selves be amplified and are therefore reported in a high fraction of Notably, in cancer types in which the mutational load was sufficiently sequence reads (Fig. 5b and Extended Data Fig. 8). By contrast, late high, we could detect a larger-than-expected number of SNVs on an SNVs that occur after the amplification has concluded will be present intermediate number of DNA copies, suggesting that they appeared on only one chromosome copy out of many, and thus have a low variant during the amplification process (Supplementary Fig. 8). Nature | Vol 578 | 6 February 2020 | 89 Article a c 8 10 9 22q13.1 7 BRCA2 8 6 MBD4 7 –log10(Pobs) –log10(P) 5 6 5 4 4 3 3 2 2 1 1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 181920 2122 0 0.5 1.0 1.5 2.0 2.5 3.0 Chromosomes –log10(Pexp) Prost–AdenoCA b (DO51965) d 10 22 X (1) 1 Long read (kb) Volcano 21 size 20 (2) (3) 2 5 1 2 3a 3b Strombolian Plinian Not hot 0 1 5 ≥10 Contribution (%) 19 (4) 18 3 0 1 22 17 Chr. 2: 59,279,205–59,289,368 Chr. 5: 14,8202,017–148,202,805 2 21 16 1 2 3 3 20 4 15 4 19 reads Short Chromosome 5 18 5 14 6 17 13 6 12 7 16 7 Germline Tumour 8 15 11 10 9 8 Chr. 2 Chr. 2 9 14 Deletion Interchromosomal Chr. 5 Chr. 5 10 13 Duplication Inversion (tail-to-tail) 11 12 Inversion (head-to-head) X Y Fig. 6 | Germline determinants of the somatic mutation landscape. breakpoints. c, Association between rare germline PTVs (MAF < 0.5%) and a, Association between common (MAF > 5%) germline variants and somatic somatic CpG mutagenesis (approximately with signature 1) in individuals of APOBEC3B-like mutagenesis in individuals of European ancestry (n = 1,201). European ancestry (n = 1,201). Genes highlighted in blue or red were associated Two-sided hypothesis testing was performed with PLINK v.1.9. To mitigate with lower or higher somatic mutation rates. Two-sided hypothesis testing was multiple-hypothesis testing, the significance threshold was set to genome- performed using linear-regression models with sex, age at diagnosis and wide significance (P < 5 × 10 −8). b, Templated insertion SVs in a BRCA1- cancer project as variables. To mitigate multiple-hypothesis testing, the associated prostate cancer. Left, chromosome bands (1); SVs ≤ 10 megabases significance threshold was set to exome-wide significance (P < 2.5 × 10 −6). (Mb) (2); 1-kb read depth corrected to copy number 0–6 (3); inter- and The black line represents the identity line that would be followed if the intrachromosomal SVs > 10 Mb (4). Right, a complex somatic SV composed of a observed P values followed the null expectation; the shaded area shows 2.2-kb tandem duplication on chromosome 2 together with a 232-base-pair the 95% confidence intervals. d, Catalogue of polymorphic germline L1 source (bp) inverted templated insertion SV that is derived from chromosome 5 and elements that are active in cancer. The chromosomal map shows germline inserted inbetween the tandem duplication (bottom). Consensus sequence source L1 elements as volcano symbols. Each volcano is colour-coded alignment of locally assembled Oxford Nanopore Technologies long according to the type of source L1 activity. The contribution of each source sequencing reads to chromosomes 2 and 5 of the human reference genome locus (expressed as a percentage) to the total number of transductions (top). Breakpoints are circled and marked as 1 (beginning of tandem identified in PCAWG tumours is represented as a gradient of volcano size, with duplication), 2 (end of tandem duplication) or 3 (inverted templated insertion). top contributing elements exhibiting larger sizes. For each breakpoint, the middle panel shows Illumina short reads at SV to increase breast cancer risk and APOBEC mutagenesis in breast can- Germline effects on somatic mutations cer genomes66,67. Here, we found that rs12628403 reduces APOBEC3B- We integrated the set of 88 million germline genetic variant calls like mutagenesis specifically in cancer types with low levels of APOBEC with somatic mutations in PCAWG, to study germline determinants mutagenesis (βlow = −0.50, Plow = 1 × 10−8; βhigh = 0.17, Phigh = 0.2), and of somatic mutation rates and patterns. First, we performed a genome- increases APOBEC3A-like mutagenesis in cancer types with high lev- wide association study of somatic mutational processes with common els of APOBEC mutagenesis (βhigh = 0.44, Phigh = 8 × 10−4; βlow = −0.21, germline variants (minor allele frequency (MAF) > 5%) in individuals Plow = 0.02). Moreover, we identified a second, novel locus at 22q13.1 with inferred European ancestry. An independent genome-wide associ- that was associated with APOBEC3B-like mutagenesis across cancer ation study was performed in East Asian individuals from Asian cancer types (rs2142833, β = 0.23, P = 1.3 × 10−8). We independently validated the genome projects. We focused on two prevalent endogenous muta- association between both loci and APOBEC3B-like mutagenesis using tional processes: spontaneous deamination of 5-methylcytosine at East Asian individuals from Asian cancer genome projects CpG dinucleotides5 (signature 1) and activity of the APOBEC3 family of (βrs12628403 = 0.57, Prs12628403 = 4.2 × 10−12; βrs2142833 = 0.58, Prs2142833 = 8 × 10−15) cytidine deaminases64 (signatures 2 and 13). No locus reached genome- (Extended Data Fig. 10d). Notably, in a conditional analysis that wide significance (P < 5 × 10−8) for signature 1 (Extended Data Fig. 10a, accounted for rs12628403, we found that rs2142833 and rs12628403 b). However, a locus at 22q13.1 predicted an APOBEC3B-like mutagen- are inherited independently in Europeans (r2<0.1), and rs2142833 esis at the pan-cancer level65 (Fig. 6a). The strongest signal at 22q13.1 remained significantly associated with APOBEC3B-like mutagenesis was driven by rs12628403, and the minor (non-reference) allele was in Europeans (βEUR = 0.17, PEUR = 3 × 10−5) and East Asians (βASN = 0.25, protective against APOBEC3B-like mutagenesis (β = −0.43, P = 5.6 × 10−9, PASN = 2 × 10−3) (Extended Data Fig. 10e, f). Analysis of donor-matched MAF = 8.2%, n = 1,201 donors) (Extended Data Fig. 10c). This variant expression data further suggests that rs2142833 is a cis-expression tags a common, approximately 30-kb germline SV that deletes the quantitative trait locus (eQTL) for APOBEC3B at the pan-cancer level APOBEC3B coding sequence and fuses the APOBEC3B 3′ untranslated (β = 0.19, P = 2 × 10−6) (Extended Data Fig. 10g, h), consistent with region with the coding sequence of APOBEC3A. The deletion is known cis-eQTL studies in normal cells68,69. 90 | Nature | Vol 578 | 6 February 2020 a b 100 long-read sequencing data generated for a BRCA1-deficient PCAWG O prostate tumour verified the small tandem-duplication and templated- insertion SV phenotypes (Fig. 6b). Almost all (20 out of 21) of BRCA1- Distribution of clusters (%) 75 associated tumours with a templated-insertion SV phenotype displayed combined germline and somatic hits in the gene. Together, these data t-SNE dimension 2 Cluster 50 T1 T2 suggest that biallelic inactivation of BRCA1 is a driver of the templated- T3 insertion SV phenotype. T4 N1 25 Third, rare-variant association analysis revealed that patients with N2 N3 germline MBD4 PTVs had increased rates of somatic C > T mutation N4 0 rates at CpG dinucleotides (P < 2.5 × 10−6) (Fig. 6c and Extended Data SoftTissue−Leiomyo Bone−Osteosarc CNS−Medullo Thy−AdenoCA CNS−GBM Bone−Epith Panc−Endocrine Skin−Melanoma Liver−HCC Eso−AdenoCA Fig. 11f, g). Analysis of previously published whole-exome sequencing CNS−LGG Kidney−RCC samples from the TCGA (n = 8,134) replicated the association between germline MBD4 PTVs and increased somatic CpG mutagenesis at the t-SNE dimension 1 pan-cancer level (P = 7.1 × 10−4) (Extended Data Fig. 11h). Moreover, c d Cluster 3 gene-expression profiling revealed a significant but modest correlation CNV loss SV SNV Truncating 0.8 ALT features, cluster 2 No TMM mutations between MBD4 expression and somatic CpG mutation rates between ALT features, cluster 1 Cluster 1 (47 samples) and within PCAWG tumour types (Extended Data Fig. 11i–k). MBD4 30 Multiple mutations 20 TERT promoter SNV TMM encodes a DNA-repair gene that removes thymidines from T:G mis- 10 0.6 TERT SV hijacking TERT amplification mutations matches within methylated CpG sites71, a functionality that would be 0 ATRX DAXX RB1 TERT ATRX or DAXX consistent with a CpG mutational signature in cancer. Fraction of patients 50 Cluster 2 (42 samples) Fourth, we assessed long interspersed nuclear elements (LINE-1; L1 40 30 0.4 hereafter) that mediate somatic retrotransposition events72–74. We iden- 20 tified 114 germline source L1 elements capable of active somatic retro- Mutated (%) 10 0 ATRX DAXX RB1 TERT transposition, including 70 that represent insertions with respect to the Cluster 3 (33 samples) 30 human reference genome (Fig. 6d and Supplementary Table 7), and 53 0.2 20 that were tagged by single-nucleotide polymorphisms in strong linkage 10 0 disequilibrium (Supplementary Table 7). Only 16 germline L1 elements ATRX DAXX RB1 TERT accounted for 67% (2,440 out of 3,669) of all L1-mediated transduc- Cluster 4 (2396 samples) 0 tions10 detected in the PCAWG dataset (Extended Data Fig. 12a). These CNS−Medullo CNS−PiloAstro Skin−Melanoma Lymph−BNHL Panc−Endocrine Lymph−CLL Thy−AdenoCA Liver−HCC Uterus−AdenoCA ColoRect−AdenoCA Ovary−AdenoCA Eso−AdenoCA Stomach−AdenoCA Breast−AdenoCA Prost−AdenoCA Panc−AdenoCA Head−SCC Lung−SCC Kidney−ChRCC Kidney−RCC 9 6 16 hot-L1 elements followed two broad patterns of somatic activity (8 3 0 of each), which we term Strombolian and Plinian in analogy to patterns ATRX DAXX RB1 TERT Gene of volcanic activity. Strombolian L1s are frequently active in cancer, but mediate only small-to-modest eruptions of somatic L1 activity in Fig. 7 | Telomere sequence patterns across PCAWG. a, Scatter plot of the cancer samples (Extended Data Fig. 12b). By contrast, Plinian L1s are clusters of telomere patterns identified across PCAWG using t-distributed more rarely seen, but display aggressive somatic activity. Whereas stochastic neighbour embedding (t-SNE), based on n = 2,518 tumour samples Strombolian elements are typically relatively common (MAF > 2%) and and their matched normal samples. Axes have arbitrary dimensions such that samples with similar telomere profiles are clustered together and samples with sometimes even fixed in the human population, all Plinian elements dissimilar telomere profiles are far apart with high probability. b, Distribution were infrequent (MAF ≤ 2%) in PCAWG donors (Extended Data Fig. 12c; of the four tumour-specific clusters of telomere patterns in selected tumour P = 0.001, Mann–Whitney U-test). This dichotomous pattern of activ- types from PCAWG. c, Distribution of relevant driver mutations associated ity and allele frequency may reflect differences in age and selective with alternative lengthening of telomere and normal telomere maintenance pressures, with Plinian elements potentially inserted into the human across the four clusters. d, Distribution of telomere maintenance germline more recently. PCAWG donors bear on average between 50 abnormalities across tumour types with more than 40 patients in PCAWG. and 60 L1 source elements and between 5 and 7 elements with hot Samples were classified as tumour clusters 1–3 if they fell into a relevant cluster activity (Extended Data Fig. 12d), but only 38% (1,075 out of 2,814) of without mutations in TERT, ATRX or DAXX and had no ALT phenotype. TMM, PCAWG donors carried ≥1 Plinian element. Some L1 germline source telomere maintenance mechanisms. loci caused somatic loss of tumour-suppressor genes (Extended Data Fig. 12e). Many are restricted to individual continental population ancestries (Extended Data Fig. 12f–j). Second, we performed a rare-variant association study (MAF <0.5%) to investigate the relationship between germline PTVs and somatic DNA rearrangements in individuals with European ancestry (Extended Replicative immortality Data Fig. 11a–c). Germline BRCA2 and BRCA1 PTVs were associated One of the hallmarks of cancer is the ability of cancer to evade cellular with an increased burden of small (less than 10 kb) somatic SV dele- senescence21. Normal somatic cells typically have finite cell division tions (P = 1 × 10−8) and tandem duplications (P = 6 × 10−13), respectively, potential; telomere attrition is one mechanism to limit numbers of corroborating recent studies in breast and ovarian cancer30,70. In mitoses75. Cancers enlist multiple strategies to achieve replicative PCAWG data, this pattern also extends to other tumour types, includ- immortality. Overexpression of the telomerase gene, TERT, which main- ing adenocarcinomas of the prostate and pancreas6, typically in the tains telomere lengths, is especially prevalent. This can be achieved setting of biallelic inactivation. In addition, tumours with high lev- through point mutations in the promoter that lead to de novo tran- els of small SV tandem duplications frequently exhibited a novel and scription factor binding34,37; hitching TERT to highly active regulatory distinct class of SVs termed ‘cycles of templated insertions’6. These elements elsewhere in the genome46,76; insertions of viral enhancers complex SV events consist of DNA templates that are copied from upstream of the gene77,78; and increased dosage through chromosomal across the genome, joined into one contiguous sequence and inserted amplification, as we have seen in melanoma (Fig. 5b). In addition, there is into a single derivative chromosome. We found a significant associa- an ‘alternative lengthening of telomeres’ (ALT) pathway, in which telom- tion between germline BRCA1 PTVs and templated insertions at the eres are lengthened through homologous recombination, mediated by pan-cancer level (P = 4 × 10−15) (Extended Data Fig. 11d, e). Whole-genome loss-of-function mutations in the ATRX and DAXX genes79. Nature | Vol 578 | 6 February 2020 | 91 Article As reported in a companion paper13, 16% of tumours in the PCAWG dataset exhibited somatic mutations in at least one of ATRX, DAXX Conclusions and future perspectives and TERT. TERT alterations were detected in 270 samples, whereas The resource reported in this paper and its companion papers has 128 tumours had alterations in ATRX or DAXX, of which 71 were protein- yielded insights into the nature and timing of the many mutational truncating. In the companion paper, which focused on describing pat- processes that shape large- and small-scale somatic variation in the terns of ALT and TERT-mediated telomere maintenance13, 12 features cancer genome; the patterns of selection that act on these varia- of telomeric sequence were measured in the PCAWG cohort. These tions; the widespread effect of somatic variants on transcription; included counts of nine variants of the core hexameric sequence, the complementary roles of the coding and non-coding genome for the number of ectopic telomere-like insertions within the genome, both germline and somatic mutations; the ubiquity of intratumoral the number of genomic breakpoints and telomere length as a ratio heterogeneity; and the distinctive evolutionary trajectory of each between tumour and normal. Here we used the 12 features as an over- cancer type. Many of these insights can be obtained only from an view of telomere integrity across all tumours in the PCAWG dataset. integrated analysis of all classes of somatic mutation on a whole- On the basis of these 12 features, tumour samples formed 4 dis- genome scale, and would not be accessible with, for example, targeted tinct subclusters (Fig. 7a and Extended Data Fig. 13a), suggesting that exome sequencing. telomere-maintenance mechanisms are more diverse than the well- The promise of precision medicine is to match patients to targeted established TERT and ALT dichotomy. Clusters C1 (47 tumours) and therapies using genomics. A major barrier to its evidence-based imple- C2 (42 tumours) were enriched for traits of the ALT pathway—having mentation is the daunting heterogeneity of cancer chronicled in these longer telomeres, more genomic breakpoints, more ectopic telomere papers, from tumour type to tumour type, from patient to patient, from insertions and variant telomere sequence motifs (Supplementary clone to clone and from cell to cell. Building meaningful clinical predic- Fig. 9). C1 and C2 were distinguished from one another by the latter tors from genomic data can be achieved, but will require knowledge having a considerable increase in the number of TTCGGG and TGAGGG banks comprising tens of thousands of patients with comprehensive variant motifs among the telomeric hexamers. Thyroid adenocarci- clinical characterization83. As these sample sizes will be too large for nomas were markedly enriched among C3 samples (26 out of 33 C3 any single funding agency, pharmaceutical company or health system, samples; P < 10−16); the C1 cluster (ALT subtype 1) was common among international collaboration and data sharing will be required. The next sarcomas; and both pancreatic endocrine neoplasms and low-grade phase of ICGC, ICGC-ARGO (https:// www.icgc-argo.org/), will bring gliomas had a high proportion of samples in the C2 cluster (ALT sub- the cancer genomics community together with healthcare providers, type 2) (Fig. 7b). Notably, some of the thyroid adenocarcinomas and pharmaceutical companies, data science and clinical trials groups to pancreatic neuroendocrine tumours that cluster together (cluster C3) build comprehensive knowledge banks of clinical outcome and treat- had matched normal samples that also cluster together (normal cluster ment data from patients with a wide variety of cancers, matched with N3) (Extended Data Fig. 13a) and which share common properties. For detailed molecular profiling. example, the GTAGGG repeat was overrepresented among samples in Extending the story begun by TCGA, ICGC and other cancer genom- this group (Supplementary Fig. 10). ics projects, the PCAWG has brought us closer to a comprehensive Somatic driver mutations were also unevenly distributed across the narrative of the causal biological changes that drive cancer phenotypes. four clusters (Fig. 7c). C1 tumours were enriched for RB1 mutations or We must now translate this knowledge into sustainable, meaningful SVs (P = 3 × 10−5), as well as frequent SVs that affected ATRX (P = 6 × 10−14), clinical treatments. but not DAXX. RB1 and ATRX mutations were largely mutually exclusive (Extended Data Fig. 13b). By contrast, C2 tumours were enriched for somatic point mutations in ATRX and DAXX (P = 6 × 10−5), but not RB1. Online content The enrichment of RB1 mutations in C1 remained significant when Any methods, additional references, Nature Research reporting sum- only leiomyosarcomas and osteosarcomas were considered, confirm- maries, source data, extended data, supplementary information, ing that this enrichment is not merely a consequence of the different acknowledgements, peer review information; details of author con- distribution of tumour types across clusters. C3 samples had frequent tributions and competing interests; and statements of data and code TERT promoter mutations (30%; P = 2 × 10−6). availability are available at https://doi.org/10.1038/s41586-020-1969-6. There was a marked predominance of RB1 mutations in C1. Nearly a third of the samples in C1 contained an RB1 alteration, which were 1. Pleasance, E. D. et al. A comprehensive catalogue of somatic mutations from a human evenly distributed across truncating SNVs, SVs and shallow dele- cancer genome. Nature 463, 191–196 (2010). tions (Extended Data Fig. 13c). Previous research has shown that RB1 2. Pleasance, E. D. et al. A small-cell lung cancer genome with complex signatures of mutations are associated with long telomeres in the absence of TERT tobacco exposure. Nature 463, 184–190 (2010). 3. Ley, T. J. et al. DNA sequencing of a cytogenetically normal acute myeloid leukaemia mutations and ATRX inactivation80, and studies using mouse models genome. Nature 456, 66–72 (2008). have shown that knockout of Rb-family proteins causes elongated 4. Rheinbay, E. et al. Analyses of non-coding somatic drivers in 2,693 cancer whole telomeres81. The association with the C1 cluster here suggests that RB1 genomes. Nature https://doi.org/10.1038/s41586-020-1965-x (2020). 5. Alexandrov, L. B. et al. The repertoire of mutational signatures in human cancer. Nature mutations can represent another route to activating the ALT pathway, https://doi.org/10.1038/s41586-020-1943-3 (2020). which has subtly different properties of telomeric sequence com- 6. Li, Y. et al. Patterns of somatic structural variation in human cancer genomes. Nature pared with the inactivation of DAXX—these fall almost exclusively in https://doi.org/10.1038/s41586-019-1913-9 (2020). 7. Gerstung, M. et al. The evolutionary history of 2,658 cancers. Nature https://doi.org/ cluster C2. 10.1038/s41586-019-1907-7 (2020). Tumour types with the highest rates of abnormal telomere mainte- 8. PCAWG Transcriptome Core Group et al. Genomic basis of RNA alterations in cancer. nance mechanisms often originate in tissues that have low endogenous Nature https://doi.org/10.1038/s41586-020-1970-0 (2020). 9. Zhang, Y. et al. High-coverage whole-genome analysis of 1,220 cancers reveals replicative activity (Fig. 7d). In support of this, we found an inverse cor- hundreds of genes deregulated by rearrangement-mediated cis-regulatory relation between previously estimated rates of stem cell division across alterations. Nat. Commun. https://doi.org/10.1038/s41467-019-13885-w (2020). tissues82 and the frequency of telomere maintenance abnormalities 10. Rodriguez-Martin, B. et al. Pan-cancer analysis of whole genomes identifies driver rearrangements promoted by LINE-1 retrotransposition. Nat. Genet. https://doi. (P = 0.01, Poisson regression) (Extended Data Fig. 13d). This suggests org/10.1038/s41588-019-0562-0 (2020). that restriction of telomere maintenance is an important tumour- 11. Zapatka, M. et al. The landscape of viral associations in human cancers. Nat. Genet. suppression mechanism, particularly in tissues with low steady-state https://doi.org/10.1038/s41588-019-0558-9 (2020). 12. Jiao, W. et al. A deep learning system can accurately classify primary and metastatic cellular proliferation, in which a clone must overcome this constraint cancers based on patterns of passenger mutations. Nat. Commun. https://doi.org/ to achieve replicative immortality. 10.1038/s41467-019-13825-8 (2020). 92 | Nature | Vol 578 | 6 February 2020 13. Sieverling, L. et al. Genomic footprints of activated telomere maintenance mechanisms 57. Supek, F. & Lehner, B. Clustered mutation signatures reveal that error-prone DNA repair in cancer. Nat. Commun. https://doi.org/10.1038/s41467-019-13824-9 (2020). targets mutations to active genes. Cell 170, 534–547 (2017). 14. Yuan, Y. et al. Comprehensive molecular characterization of mitochondrial genomes in 58. Mardin, B. R. et al. A cell-based model system links chromothripsis with hyperploidy. Mol. human cancers. Nat. Genet. https://doi.org/10.1038/s41588-019-0557-x (2020). Syst. Biol. 11, 828 (2015). 15. Akdemir, K. C. et al. Chromatin folding domains disruptions by somatic genomic 59. Weischenfeldt, J. et al. Integrative genomic analyses reveal an androgen-driven somatic rearrangements in human cancers. Nat. Genet. https://doi.org/10.1038/s41588-019- alteration landscape in early-onset prostate cancer. Cancer Cell 23, 159–170 (2013). 0564-y (2020). 60. Garsed, D. W. et al. The architecture and evolution of cancer neochromosomes. Cancer 16. Reyna, M. A. et al. Pathway and network analysis of more than 2,500 whole cancer Cell 26, 653–667 (2014). genomes. Nat. Commun. https://doi.org/10.1038/s41467-020-14351-8 (2020). 61. Durinck, S. et al. Temporal dissection of tumorigenesis in primary cancers. Cancer Discov. 17. Bailey, M. H. et al. Retrospective evaluation of whole exome and genome mutation calls 1, 137–143 (2011). in 746 cancer samples. Nat. Commun. (2020). 62. Hayward, N. K. et al. Whole-genome landscapes of major melanoma subtypes. Nature 18. Cortes-Ciriano, I. et al. Comprehensive analysis of chromothripsis in 2,658 human 545, 175–180 (2017). cancers using whole-genome sequencing. Nat. Genet. https://doi.org/10.1038/s41588- 63. The Cancer Genome Atlas Network. Genomic classification of cutaneous melanoma. Cell 019-0576-7 (2020). 161, 1681–1696 (2015). 19. Bray, F., Ren, J.-S., Masuyer, E. & Ferlay, J. Global estimates of cancer prevalence for 27 64. Alexandrov, L. B. et al. Signatures of mutational processes in human cancer. Nature 500, sites in the adult population in 2008. Int. J. Cancer 132, 1133–1145 (2013). 415–421 (2013). 20. Tarver, T. Cancer Facts & Figures 2012. American Cancer Society (ACS). J. Consum. Health 65. Chan, K. et al. An APOBEC3A hypermutation signature is distinguishable from the Internet 16, 366–367 (2012). signature of background mutagenesis by APOBEC3B in human cancers. Nat. Genet. 47, 21. Hanahan, D. & Weinberg, R. A. Hallmarks of cancer: the next generation. Cell 144, 1067–1072 (2015). 646–674 (2011). 66. Nik-Zainal, S. et al. Association of a germline copy number polymorphism of APOBEC3A 22. International Cancer Genome Consortium. International network of cancer genome and APOBEC3B with burden of putative APOBEC-dependent mutations in breast cancer. projects. Nature 464, 993–998 (2010). Nat. Genet. 46, 487–491 (2014). 23. Bailey, M. H. et al. Comprehensive characterization of cancer driver genes and mutations. 67. Middlebrooks, C. D. et al. Association of germline variants in the APOBEC3 region with Cell 173, 371–385 (2018). cancer risk and enrichment with APOBEC-signature mutations in tumors. Nat. Genet. 48, 24. Sanchez-Vega, F. et al. Oncogenic signaling pathways in The Cancer Genome Atlas. Cell 1330–1338 (2016). 173, 321–337 (2018). 68. Westra, H.-J. et al. Systematic identification of trans eQTLs as putative drivers of known 25. Hoadley, K. A. et al. Cell-of-origin patterns dominate the molecular classification of disease associations. Nat. Genet. 45, 1238–1243 (2013). 10,000 tumors from 33 types of cancer. Cell 173, 291–304 (2018). 69. Stranger, B. E. et al. Population genomics of human gene expression. Nat. Genet. 39, 26. Stein, L. D., Knoppers, B. M., Campbell, P., Getz, G. & Korbel, J. O. Data analysis: create a 1217–1224 (2007). cloud commons. Nature 523, 149–151 (2015). 70. Menghi, F. et al. The tandem duplicator phenotype as a distinct genomic configuration in 27. Phillips, M. et al. Genomics: data sharing needs international code of conduct. Nature cancer. Proc. Natl Acad. Sci. USA 113, E2373–E2382 (2016). https://doi.org/10.1038/d41586-020-00082-9 (2020). 71. Hendrich, B., Hardeland, U., Ng, H. H., Jiricny, J. & Bird, A. The thymine glycosylase MBD4 28. Krochmalski, J. Developing with Docker (Packt Publishing, 2016). can bind to the product of deamination at methylated CpG sites. Nature 401, 301–304 29. Welch, J. S. et al. The origin and evolution of mutations in acute myeloid leukemia. Cell (1999). 150, 264–278 (2012). 72. Lee, E. et al. Landscape of somatic retrotransposition in human cancers. Science 337, 30. Nik-Zainal, S. et al. Landscape of somatic mutations in 560 breast cancer whole-genome 967–971 (2012). sequences. Nature 534, 47–54 (2016). 73. Tubio, J. M. C. et al. Extensive transduction of nonrepetitive DNA mediated by L1 31. Meier, B. et al. C. elegans whole-genome sequencing reveals mutational signatures retrotransposition in cancer genomes. Science 345, 1251343–1251343 (2014). related to carcinogens and DNA repair deficiency. Genome Res. 24, 1624–1636 (2014). 74. Helman, E. et al. Somatic retrotransposition in human cancer revealed by whole-genome 32. Martincorena, I. et al. Universal patterns of selection in cancer and somatic tissues. Cell and exome sequencing. Genome Res. 24, 1053–1063 (2014). 171, 1029–1041 (2017). 75. Shay, J. W. & Wright, W. E. Hayflick, his limit, and cellular ageing. Nat. Rev. Mol. Cell Biol. 1, 33. Tamborero, D. et al. Cancer Genome Interpreter annotates the biological and clinical 72–76 (2000). relevance of tumor alterations. Genome Med. 10, 25 (2018). 76. Peifer, M. et al. Telomerase activation by genomic rearrangements in high-risk 34. Huang, F. W. et al. Highly recurrent TERT promoter mutations in human melanoma. neuroblastoma. Nature 526, 700–704 (2015). Science 339, 957–959 (2013). 77. Totoki, Y. et al. Trans-ancestry mutational landscape of hepatocellular carcinoma 35. Rheinbay, E. et al. Recurrent and functional regulatory mutations in breast cancer. Nature genomes. Nat. Genet. 46, 1267–1273 (2014). 547, 55–60 (2017). 78. Paterlini-Bréchot, P. et al. Hepatitis B virus-related insertional mutagenesis occurs 36. Fredriksson, N. J., Ny, L., Nilsson, J. A. & Larsson, E. Systematic analysis of noncoding frequently in human liver cancers and recurrently targets human telomerase gene. somatic mutations and gene expression alterations across 14 tumor types. Nat. Genet. Oncogene 22, 3911–3916 (2003). 46, 1258–1263 (2014). 79. Heaphy, C. M. et al. Prevalence of the alternative lengthening of telomeres telomere 37. Horn, S. et al. TERT promoter mutations in familial and sporadic melanoma. Science 339, maintenance mechanism in human cancer subtypes. Am. J. Pathol. 179, 1608–1615 959–961 (2013). (2011). 38. Ciriello, G. et al. Emerging landscape of oncogenic signatures across human cancers. 80. Barthel, F. P. et al. Systematic analysis of telomere length and somatic alterations in 31 Nat. Genet. 45, 1127–1133 (2013). cancer types. Nat. Genet. 49, 349–357 (2017). 39. Rahman, N. Realizing the promise of cancer predisposition genes. Nature 505, 302–308 81. García-Cao, M., Gonzalo, S., Dean, D. & Blasco, M. A. A role for the Rb family of proteins in (2014). controlling telomere length. Nat. Genet. 32, 415–419 (2002). 40. Pearl, L. H., Schierz, A. C., Ward, S. E., Al-Lazikani, B. & Pearl, F. M. G. Therapeutic 82. Tomasetti, C. & Vogelstein, B. Variation in cancer risk among tissues can be explained by opportunities within the DNA damage response. Nat. Rev. Cancer 15, 166–180 (2015). the number of stem cell divisions. Science 347, 78–81 (2015). 41. Taylor-Weiner, A. et al. DeTiN: overcoming tumor-in-normal contamination. Nat. Methods 83. Gerstung, M. et al. Precision oncology for acute myeloid leukemia using a knowledge 15, 531–534 (2018). bank approach. Nat. Genet. 49, 332–340 (2017). 42. Fujimoto, A. et al. Whole-genome mutational landscape and characterization of 84. O’Connor, B. D. et al. The Dockstore: enabling modular, community-focused sharing of noncoding and structural mutations in liver cancer. Nat. Genet. 48, 500–509 (2016). Docker-based genomics tools and workflows. F1000Res. 6, 52 (2017). 43. Shlush, L. I. Age-related clonal hematopoiesis. Blood 131, 496–504 (2018). 85. Zhang, J. et al. The International Cancer Genome Consortium Data Portal. 44. Northcott, P. A. et al. The whole-genome landscape of medulloblastoma subtypes. Nat. Biotechnol. 37, 367–369 (2019). Nature 547, 311–317 (2017). 86. Miller, C. A., Qiao, Y., DiSera, T., D’Astous, B. & Marth, G. T. bam.iobio: a web-based, 45. Scarpa, A. et al. Whole-genome landscape of pancreatic neuroendocrine tumours. real-time, sequence alignment file inspector. Nat. Methods 11, 1189–1189 (2014). Nature 543, 65–71 (2017). 87. Goldman, M. et al. The UCSC Xena platform for public and private cancer genomics data 46. Davis, C. F. et al. The somatic genomic landscape of chromophobe renal cell carcinoma. visualization and interpretation. Preprint at https://www.biorxiv.org/content/10.1101/ Cancer Cell 26, 319–330 (2014). 326470v6 (2019) 47. Berger, M. F. et al. The genomic complexity of primary human prostate cancer. Nature 88. Papatheodorou, I. et al. Expression Atlas: gene and protein expression across multiple 470, 214–220 (2011). studies and organisms. Nucleic Acids Res. 46, D246–D251 (2018). 48. Baca, S. C. et al. Punctuated evolution of prostate cancer genomes. Cell 153, 666–677 (2013). 49. Nik-Zainal, S. et al. The life history of 21 breast cancers. Cell 149, 994–1007 (2012). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in 50. Nik-Zainal, S. et al. Mutational processes molding the genomes of 21 breast cancers. Cell published maps and institutional affiliations. 149, 979–993 (2012). 51. Roberts, S. A. et al. Clustered mutations in yeast and in human cancers can arise from Open Access This article is licensed under a Creative Commons Attribution damaged long single-strand DNA regions. Mol. Cell 46, 424–435 (2012). 4.0 International License, which permits use, sharing, adaptation, distribution 52. Rausch, T. et al. Genome sequencing of pediatric medulloblastoma links catastrophic and reproduction in any medium or format, as long as you give appropriate DNA rearrangements with TP53 mutations. Cell 148, 59–71 (2012). credit to the original author(s) and the source, provide a link to the Creative Commons license, 53. Stephens, P. J. et al. Massive genomic rearrangement acquired in a single catastrophic and indicate if changes were made. The images or other third party material in this article are event during cancer development. Cell 144, 27–40 (2011). included in the article’s Creative Commons license, unless indicated otherwise in a credit line 54. Korbel, J. O. & Campbell, P. J. Criteria for inference of chromothripsis in cancer genomes. to the material. If material is not included in the article’s Creative Commons license and your Cell 152, 1226–1236 (2013). intended use is not permitted by statutory regulation or exceeds the permitted use, you will 55. Zhang, C.-Z. et al. Chromothripsis from DNA damage in micronuclei. Nature 522, 179–184 need to obtain permission directly from the copyright holder. To view a copy of this license, (2015). visit http://creativecommons.org/licenses/by/4.0/. 56. The Cancer Genome Atlas Research Network. Integrated genomic characterization of papillary thyroid carcinoma. Cell 159, 676–690 (2014). © The Author(s) 2020 Nature | Vol 578 | 6 February 2020 | 93 Article The ICGC/TCGA Pan-Cancer Analysis of Whole Genomes Consortium Michael R. Stratton1, Takashi Yugawa228, Giampaolo Tortora229,230, K. VijayRaghavan218, Jean C. Zenklusen231, David Townend232, Bartha M. Knoppers130, Brice Aminou15, Javier Peter J. Campbell1,2,745*, Gad Getz3,4,5,6,745*, Jan O. Korbel7,8,745*, Joshua M. Stuart9,745*, Bartolome40, Keith A. Boroevich81,233, Rich Boyce7, Alex Buchanan38, Niall J. Byrne15, Jennifer L. Jennings10,11,745, Lincoln D. Stein12,13,745*, Marc D. Perry14,15, Hardeep K. Nahal- Zhaohong Chen234, Sunghoon Cho235, Wan Choi236, Peter Clapham1, Michelle T. Dow234, Bose15, B. F. Francis Ouellette16,17, Constance H. Li12,18, Esther Rheinbay3,6,19, G. Petur Lewis Jonathan Dursi12,25, Juergen Eils142,143, Claudiu Farcas234, Nodirjon Fayzullaev15, Paul Nielsen19, Dennis C. Sgroi19, Chin-Lee Wu19, William C. Faquin19, Vikram Deshpande19, Paul Flicek7, Allison P. Heath237, Oliver Hofmann238, Jongwhi H. Hong239, Thomas J. Hudson240,241, C. Boutros12,18,20,21, Alexander J. Lazar22, Katherine A. Hoadley23,24, David N. Louis19, L. Daniel Hübschmann30,120,142,242,243, Sinisa Ivkovic244, Seung-Hyup Jeon236, Wei Jiao12, Rolf Jonathan Dursi12,25, Christina K. Yung15, Matthew H. Bailey26,27, Gordon Saksena3, Keiran M. Kabbe28, Andre Kahles112,113,114,115,174, Jules N. A. Kerssemakers28, Hyunghwan Kim236, Jihoon Raine1, Ivo Buchhalter28,29,30, Kortine Kleinheinz28,30, Matthias Schlesner28,31, Junjun Zhang15, Kim245, Michael Koscher246, Antonios Koures234, Milena Kovacevic244, Chris Lawerenz143, Jia Wenyi Wang32, David A. Wheeler33,34, Li Ding26,27,35, Jared T. Simpson12,36, Brian D. Liu247, Sanja Mijalkovic244, Ana Mijalkovic Mijalkovic-Lazic244, Satoru Miyano39, Mia O’Connor15,37, Sergei Yakneen8, Kyle Ellrott38, Naoki Miyoshi39, Adam P. Butler1, Romina Nastic244, Jonathan Nicholson1, David Ocana7, Kazuhiro Ohi39, Lucila Ohno-Machado234, Royo40, Solomon I. Shorser12, Miguel Vazquez40,41, Tobias Rausch8, Grace Tiao3, Sebastian Todd D. Pihl248, Manuel Prinz28, Petar Radovic244, Charles Short7, Heidi J. Sofia217, Jonathan M. Waszak8, Bernardo Rodriguez-Martin42,43,44, Suyash Shringarpure45, Dai-Ying Wu46, Spring162, Adam J. Struck38, Nebojsa Tijanic244, David Vicente40, Zhining Wang231, Ashley German M. Demidov47,48,49, Olivier Delaneau50,51,52, Shuto Hayashi39, Seiya Imoto39, Nina Williams234, Youngchoon Woo236, Adam J. Wright12, Liming Yang231, Mark P. Hamilton249, Habermann8, Ayellet V. Segre3,53, Erik Garrison1, Andy Cafferkey7, Eva G. Alvarez42,43,44, José Todd A. Johnson233, Abdullah Kahraman250,251,252, Manolis Kellis3,253, Paz Polak3,4,6, Richard María Heredia-Genestar54, Francesc Muyas47,48,49, Oliver Drechsel47,49, Alicia L. Sallari3, Nasa Sinnott-Armstrong3,45, Christian von Mering252,254, Sergi Beltran49,74, Daniela Bruzos42,43,44, Javier Temes42,43, Jorge Zamora1,42,43,44, Adrian Baez-Ortega55, Hyung-Lae S. Gerhard255, Marta Gut49,74, Jean-Rémi Trotta74, Justin P. Whalley74, Beifang Niu256, Kim56, R. Jay Mashl27,57, Kai Ye58,59, Anthony DiBiase60, Kuan-lin Huang27,61, Ivica Letunic62, Shadrielle M. G. Espiritu12, Shengjie Gao179, Yi Huang157,257, Christopher M. Lalansingh12, Jon Michael D. McLellan26,27,35, Steven J. Newhouse7, Tal Shmaya46, Sushant Kumar63,64, David C. W. Teague1, Michael C. Wendl27,158,159, Federico Abascal1, Gary D. Bader13, Pratiti Wedge1,65,66, Mark H. Wright45, Venkata D. Yellapantula67,68, Mark Gerstein63,64,69, Ekta Bandopadhayay3,258,259, Jonathan Barenboim12, Søren Brunak260,261, Joana Carlevaro- Khurana70,71,72,73, Tomas Marques-Bonet74,75,76,77, Arcadi Navarro74,75,76, Carlos D. Fita262,263,264, Dimple Chakravarty265,266, Calvin Wing Yiu Chan28,128, Jung Kyoon Choi267, Klev Bustamante45,78, Reiner Siebert79,80, Hidewaki Nakagawa81, Douglas F. Easton82,83, Stephan Diamanti268, J. Lynn Fink40,269, Joan Frigola204, Carlo Gambacorti-Passerini270, Dale W. Ossowski47,48,49, Jose M. C. Tubio42,43,44, Francisco M. De La Vega45,46,78, Xavier Estivill47,84, Garsed271, Nicholas J. Haradhvala3,19, Arif O. Harmanci64,272, Mohamed Helmy170, Carl Denis Yuen12, George L. Mihaiescu15, Larsson Omberg85, Vincent Ferretti15,86, Herrmann28,30,273, Asger Hobolth153,205, Ermin Hodzic155, Chen Hong127,128, Keren Isaev12,18, Radhakrishnan Sabarinathan87,88,89, Oriol Pich87,89, Abel Gonzalez-Perez87,89, Amaro Taylor- Jose M. G. Izarzugaza260, Rory Johnson263,274, Randi Istrup Juul152, Jaegil Kim3, Jong K. Weiner90, Matthew W. Fittall91, Jonas Demeulemeester91,92, Maxime Tarabichi1,91, Nicola D. Kim275, Jan Komorowski268,276, Andrés Lanzós263,264,274, Erik Larsson112, Donghoon Lee64, Roberts1, Peter Van Loo91,92, Isidro Cortés-Ciriano93,94,95, Lara Urban7,8, Peter Park94,95, Bin Shantao Li64, Xiaotong Li64, Ziao Lin3,277, Eric Minwei Liu71,73,278, Lucas Lochovsky63,64,185, Zhu96, Esa Pitkänen8, Yilong Li1, Natalie Saini97, Leszek J. Klimczak98, Joachim Shaoke Lou63,64, Tobias Madsen152, Kathleen Marchal279,280, Alexander Martinez- Weischenfeldt8,99,100, Nikos Sidiropoulos100, Ludmil B. Alexandrov1,101, Raquel Fundichely71,72,73, Patrick D. McGillivray63, William Meyerson64,281, Marta Paczkowska12, Rabionet47,49,102, Georgia Escaramis47,103,104, Mattia Bosio40,47,49, Aliaksei Z. Holik47, Hana Keunchil Park282,283, Kiejung Park284, Tirso Pons285, Sergio Pulido-Tamayo279,280, Iker Reyes- Susak47,49, Aparna Prasad49, Serap Erkek8, Claudia Calabrese7,8, Benjamin Raeder8, Eoghan Salazar87, Matthew A. Reyna286, Mark A. Rubin274,287,288,289,290, Leonidas Salichos63,64, Chris Harrington105, Simon Mayes106, Daniel Turner106, Sissel Juul105, Steven A. Roberts107, Lei Sander112,156,291,292, Steven E. Schumacher3,203, Mark Shackleton271, Ciyue Shen292,293, Raunak Song96, Roelof Koster108, Lisa Mirabello96, Xing Hua96, Tomas J. Tanskanen109, Marta Tojo44, Shrestha146, Shimin Shuai12,13, Tatsuhiko Tsunoda233,294,295,296, Husen M. Umer268,297, Liis Jieming Chen64,110, Lauri A. Aaltonen111, Gunnar Rätsch112,113,114,115,116,117, Roland F. Uusküla-Reimand298,299, Lieven P. C. Verbeke280,300, Claes Wadelius301, Lina Wadi12, Jonathan Schwarz7,118,119,120, Atul J. Butte121, Alvis Brazma7, Stephen J. Chanock96, Nilanjan Warrell63,64, Guanming Wu302, Jun Yu303, Jing Zhang64, Xuanping Zhang157,304, Yan Chatterjee122,123, Oliver Stegle7,8,124, Olivier Harismendy125, G. Steven Bova126, Dmitry A. Zhang64,305,306, Zhongming Zhao307, Lihua Zou308, Michael S. Lawrence3,19,233, Benjamin J. Gordenin97, David Haan9, Lina Sieverling127,128, Lars Feuerbach127, Don Chalmers129, Yann Raphael286, Peter J. Bailey186, David Craft3,309, Mary J. Goldman132, Hiroyuki Aburatani310, Joly130, Bartha Knoppers130, Fruzsina Molnár-Gábor131, Mark Phillips130, Adrian Hans Binder311,312, Huy Q. Dinh313, Simon C. Heath49,74, Steve Hoffmann311,312,314,315, Charles Thorogood130, David Townend130, Mary Goldman132, Nuno A. Fonseca7,133, Qian Xiang15, David Imbusch127, Helene Kretzmer312,315, Peter W. Laird316, Jose I. Martin-Subero76,317, Genta Brian Craft132, Elena Piñeiro-Yáñez134, Alfonso Muñoz7, Robert Petryszak7, Anja Füllgrabe7, Nagae310,318, Hui Shen319, Qi Wang246, Dieter Weichenhan320, Wanding Zhou319, Benjamin P. Fatima Al-Shahrour134, Maria Keays7, David Haussler132,135, John Weinstein136,137, Wolfgang Berman313,321,322, Benedikt Brors127,149,323, Christoph Plass320, Kadir C. Akdemir140, David D. L. Huber8, Alfonso Valencia40,76, Irene Papatheodorou7, Jingchun Zhu132, Yu Fan32, David Bowtell271, Kathleen H. Burns324,325, John Busanovich3,326, Kin Chan327, Ana Dueso-Barroso40, Torrents40,76, Matthias Bieg138,139, Ken Chen140, Zechen Chong141, Kristian Cibulskis3, Roland Paul A. Edwards328,329, Dariush Etemadmoghadam271, James E. Haber330, David T. W. Eils28,30,142,143, Robert S. Fulton26,27,35, Josep L. Gelpi40,144, Santiago Gonzalez7,8, Ivo G. Gut49,74, Jones331,332, Young Seok Ju1,267, Marat D. Kazanov333,334,335, Youngil Koh336,337, Kiran Kumar3, Faraz Hach145,146, Michael Heinold28,30, Taobo Hu147, Vincent Huang12, Barbara Hutter139,148,149, Eunjung Alice Lee338, Jake June-Koo Lee94,95, Andy G. Lynch328,329,339, Geoff Macintyre328, Natalie Jäger28, Jongsun Jung150, Yogesh Kumar147, Christopher Lalansingh12, Ignaty Florian Markowetz328,329, Fabio C. P. Navarro63, John V. Pearson340,341, Karsten Rippe120, Leshchiner3, Dimitri Livitz3, Eric Z. Ma147, Yosef E. Maruvka3,19,151, Ana Milovanovic40, Morten Ralph Scully342, Izar Villasante40, Nicola Waddell340,341, Lixing Yang343, Xiaotong Yao165,344, Muhlig Nielsen152, Nagarajan Paramasivam28,139, Jakob Skou Pedersen152,153, Montserrat Sung-Soo Yoon337, Cheng-Zhong Zhang3,6,156, Erik N. Bergstrom345, Arnoud Boot195,346, Kyle Puiggròs40, S. Cenk Sahinalp146,154,155, Iman Sarrafi146,155, Chip Stewart3, Miranda D. Covington34, Akihiro Fujimoto81, Mi Ni Huang195,346, S. M. Ashiqul Islam101, John R. Stobbe49,74, Jeremiah A. Wala3,6,156, Jiayin Wang27,58,157, Michael Wendl27,158,159, Johannes McPherson195,346, Sandro Morganella1, Ville Mustonen347,348,349, Alvin Wei Tian Ng350, Werner28,160, Zhenggang Wu147, Hong Xue147, Takafumi N. Yamaguchi12, Venkata Stephenie D. Prokopec12, Ignacio Vázquez-García1,67,351,352, Yang Wu195,346, Fouad Yousif12, Yellapantula67,68, Brandi N. Davis-Dusenbery161, Robert L. Grossman162, Youngwook Willie Yu353, Steven G. Rozen195,196,346, Vasilisa A. Rudneva8, Suyash S. Shringarpure45, Daniel Kim163,164, Michael C. Heinold28,30, Jonathan Hinton1, David R. Jones1, Andrew Menzies1, Lucy J. Turner106, Tian Xia354, Gurnit Atwal12,13,171, David K. Chang186,355, Susanna L. Cooke186, Stebbings1, Julian M. Hess3,151, Mara Rosenberg3,19, Andrew J. Dunford3, Manaswi Gupta3, Bishoy M. Faltas117, Syed Haider12, Vera B. Kaiser356, Rosa Karlić357, Mamoru Kato358, Kirsten Marcin Imielinski165,166, Matthew Meyerson3,6,156, Rameen Beroukhim3,6,167, Jüri Reimand12,18, Kübler3,6,19, Adam Margolin38, Sancha Martin1,359, Serena Nik-Zainal1,360,361,362, Christine Priyanka Dhingra71,73, Francesco Favero168, Stefan Dentro1,65,91, Jeff Wintersinger169,170,171, P’ng12, Colin A. Semple356, Jaclyn Smith38, Ren X. Sun12, Kevin Thai15, Derek W. Wright363,364, Vasilisa Rudneva8, Ji Wan Park172, Eun Pyo Hong172, Seong Gu Heo172, André Ke Yuan328,359,365, Andrew V. Biankin186,355,366,367, Levi Garraway156, Sean M. Grimmond368, Kahles112,113,114,115,116, Kjong-Van Lehmann112,114,115,173,174, Cameron M. Soulette37, Yuichi David J. Adams1, Pavana Anur369, Shaolong Cao32, Elizabeth L. Christie271, Marek Shiraishi39, Fenglin Liu175,176, Yao He175, Deniz Demircioğlu177,178, Natalie R. Cmero370,371,372, Yupeng Cun373, Kevin J. Dawson1, Stefan C. Dentro1,65,91, Amit G. Deshwar374, Davidson112,114,115,117,173, Liliana Greger7, Siliang Li179,180, Dongbing Liu179,180, Stefan G. Nilgun Donmez146,155, Ruben M. Drews328, Moritz Gerstung7,8, Gavin Ha3, Kerstin Haase91, Stark115,173,181,182, Fan Zhang175, Samirkumar B. Amin183,184,185, Peter Bailey186, Aurélien Lara Jerman8,375, Yuan Ji376,377, Clemency Jolly91, Juhee Lee378, Henry Lee-Six1, Salem Chateigner15, Milana Frenkel-Morgenstern187, Yong Hou179,180, Matthew R. Huska118, Helena Malikic146,155, Thomas J. Mitchell1,329,379, Quaid D. Morris171,380, Layla Oesper381, Martin Kilpinen188, Fabien C. Lamaze12, Chang Li179,180, Xiaobo Li179,180, Xinyue Li179, Xingmin Liu179,180, Peifer373, Myron Peto382, Daniel Rosebrock3, Yulia Rubanova36,171, Adriana Salcedo12, Maximillian G. Marin37, Julia Markowski118, Tannistha Nandi189, Akinyemi I. Ojesina190,191,192, Subhajit Sengupta383, Ruian Shi380, Seung Jun Shin182, Oliver Spiro3, Shankar Vembu380,384, Qiang Pan-Hammarström179,193, Peter J. Park94,95, Chandra Sekhar Pedamallu3,6,167, Hong Jeffrey A. Wintersinger169,170,171, Tsun-Po Yang373, Kaixian Yu385, Hongtu Zhu386,387, Paul T. Su179,180, Patrick Tan189,194,195,196, Bin Tean Teh194,195,196,197,198, Jian Wang179, Heng Xiong179,180, Spellman388, John N. Weinstein136,137, Yiwen Chen32, Masashi Fujita81, Leng Han304, Takanori Chen Ye179,180, Christina Yung15, Xiuqing Zhang179, Liangtao Zheng175, Shida Zhu179,180, Philip Hasegawa39, Mitsuhiro Komura39, Jun Li32, Shinichi Mizuno389, Eigo Shimizu39, Yumeng Awadalla12,13, Chad J. Creighton199, Kui Wu179,180, Huanming Yang179, Jonathan Göke177,200, Wang32,390, Yanxun Xu391, Rui Yamaguchi39, Fan Yang380, Yang Yang304, Christopher J. Zemin Zhang175,201, Angela N. Brooks3,37,156, Matthew W Fittall91, Iñigo Martincorena1, Carlota Yoon267, Yuan Yuan32, Han Liang32, Malik Alawi392,393, Ivan Borozan12, Daniel S. Brewer394,395, Rubio-Perez87,89,202, Malene Juul152, Steven Schumacher3,203, Ofer Shapira3,156, David Colin S. Cooper395,396,397, Nikita Desai15, Adam Grundhoff392,398, Murat Iskar399, Xiaoping Tamborero87,89, Loris Mularoni87,89, Henrik Hornshøj152, Jordi Deu-Pons89,204, Ferran Su400, Marc Zapatka399, Peter Lichter148,399, Kathryn Alsop271, Timothy J. C. Bruxner269, Muiños87,89, Johanna Bertl152,205, Qianyun Guo153, Abel Gonzalez-Perez87,89,206, Qian Xiang207, Angelika N. Christ269, Stephen M. Cordner401, Prue A. Cowin402, Ronny Drapkin403, Sian Wojciech Bazant7, Elisabet Barrera7, Sultan T. Al-Sedairy208, Axel Aretz209, Cindy Bell210, Fereday402, Joshy George185, Anne Hamilton402, Oliver Holmes340,341, Jillian A. Hung404,405, Miguel Betancourt211, Christiane Buchholz212, Fabien Calvo213, Christine Chomienne214, Karin S. Kassahn269,406, Stephen H. Kazakoff340,341, Catherine J. Kennedy407,408, Michael Dunn215, Stuart Edmonds216, Eric Green217, Shailja Gupta218, Carolyn M. Hutter217, Conrad R. Leonard340,341, Linda Mileshkin271, David K. Miller269,355,409, Gisela Mir Arnau402, Karine Jegalian219, Nic Jones220, Youyong Lu221,222,223, Hitoshi Nakagama224, Gerd Chris Mitchell402, Felicity Newell340,341, Katia Nones340,341, Ann-Marie Patch340,341, Nettekoven225, Laura Planko225, David Scott220, Tatsuhiro Shibata226,227, Kiyo Shimizu228, Michael C. Quinn340,341, Darrin F. Taylor269, Heather Thorne402, Nadia Traficante402, Ravikiran Vedururu402, Nick M. Waddell341, Paul M. Waring410, Scott Wood340,341, Qinying Fumie Hosoda226, Hiromi Nakamura226, Hidenori Ojima583, Takuji Okusaka584, Yasushi Xu340,341, Anna deFazio411,412,413, Matthew J. Anderson269, Davide Antonello414, Andrew P. Totoki226, Tomoko Urushidate227, Masashi Fukayama585, Shumpei Ishikawa586, Hitoshi Barbour415,416, Claudio Bassi414, Samantha Bersani417, Ivana Cataldo417,418, Lorraine A. Katai587, Hiroto Katoh586, Daisuke Komura586, Hirofumi Rokutan358, Mihoko Saito-Adachi358, Chantrill355,419, Yoke-Eng Chiew411, Angela Chou355,420, Sara Cingarlini229, Nicole Cloonan421, Akihiro Suzuki310,588, Hirokazu Taniguchi589, Kenji Tatsuno310, Tetsuo Ushiku585, Shinichi Vincenzo Corbo418,422, Maria Vittoria Davi423, Fraser R. Duthie186,424, Anthony J. Gill355,420, Yachida226,590, Shogo Yamamoto310, Hiroshi Aikata591, Koji Arihiro591, Shun-ichi Ariizumi592, Janet S. Graham186,425, Ivon Harliwong269, Nigel B. Jamieson186,367,426, Amber L. Johns355,409, Kazuaki Chayama591, Mayuko Furuta81, Kunihito Gotoh593, Shinya Hayami594, Satoshi James G. Kench355,420,427, Luca Landoni414, Rita T. Lawlor418, Andrea Mafficini418, Neil D. Hirano595, Yoshiiku Kawakami591, Kazuhiro Maejima81, Toru Nakamura595, Kaoru Nakano81, Merrett414,428, Marco Miotto414, Elizabeth A. Musgrove186, Adnan M. Nagrial355, Karin A. Hideki Ohdan591, Aya Sasaki-Oku81, Hiroko Tanaka39, Masaki Ueno594, Masakazu Oien410,429, Marina Pajic355, Mark Pinese430, Alan J. Robertson269, Ilse Rooman355, Borislav C. Yamamoto592, Hiroki Yamaue594, Su Pin Choo596, Ioana Cutcutache195,346, Narong Rusev418, Jaswinder S. Samra414,420, Maria Scardoni417, Christopher J. Scarlett355,431, Aldo Khuntikeo414,597, Choon Kiat Ong598, Chawalit Pairojkul410, Irinel Popescu599, Keun Soo Scarpa418, Elisabetta Sereni414, Katarzyna O. Sikora418, Michele Simbolo422, Morgan L. Ahn600, Marta Aymerich601, Armando Lopez-Guillermo602, Carlos López-Otín603, Xose S. Taschuk15, Christopher W. Toon355, Caterina Vicentini418, Jianmin Wu355, Nikolajs Zeps432,433, Puente603, Elias Campo604,605, Fernanda Amary606, Daniel Baumhoer607, Sam Behjati1, Bodil Andreas Behren434, Hazel Burke435, Jonathan Cebon434, Rebecca A. Dagg436, Ricardo De Bjerkehagen607,608, P. A. Futreal505, Ola Myklebost516, Nischalan Pillay609, Patrick Tarpey610, Paoli-Iseppi437, Ken Dutton-Regester340, Matthew A. Field438, Anna Fitzgerald439, Peter Roberto Tirabosco611, Olga Zaikova612, Adrienne M. Flanagan613, Jacqueline Boultwood614, Hersey435, Valerie Jakrot435, Peter A. Johansson340, Hojabr Kakavand437, Richard F. David T. Bowen1, Mario Cazzola615, Anthony R. Green329, Eva Hellstrom-Lindberg616, Luca Kefford440, Loretta M. S. Lau441, Georgina V. Long442, Hilda A. Pickett441, Antonia L. Malcovati615, Jyoti Nangalia617, Elli Papaemmanuil1, Paresh Vyas340,618, Yeng Ang619, Hugh Pritchard340, Gulietta M. Pupo443, Robyn P. M. Saw442, Sarah-Jane Schramm444, Catherine A. Barr620, Duncan Beardsmore621, Matthew Eldridge328, James Gossage622, Nicola Grehan361, Shang439, Ping Shang442, Andrew J. Spillane442, Jonathan R. Stretch442, Varsha Tembe411,444, George B. Hanna623, Stephen J. Hayes624,625, Ted R. Hupp626, David Khoo627, Jesper John F. Thompson442, Ricardo E. Vilain445, James S. Wilmott442, Jean Y. Yang446, Nicholas K. Lagergren616,628, Laurence B. Lovat188, Shona MacRae136, Maria O’Donovan361, J. Robert Hayward340,435, Graham J. Mann411,447, Richard A. Scolyer412,442,445,448, John Bartlett449,450, O’Neill629, Simon L. Parsons630, Shaun R. Preston631, Sonia Puig632, Tom Roques633, Grant Prashant Bavi451, Dianne E. Chadwick452, Michelle Chan-Seng-Yue451, Sean Cleary451,453, Sanders24, Sharmila Sothi634, Simon Tavaré328, Olga Tucker635, Richard Turkington636, Ashton A. Connor453,454, Karolina Czajka241, Robert E. Denroche451, Neesha C. Dhani455, Timothy J. Underwood637, Ian Welch638, Rebecca C. Fitzgerald361, Daniel M. Berney639, Jenna Eagles241, Steven Gallinger451,453,454, Robert C. Grant451,454, David Hedley455, Michael Johann S. De Bono396, Declan Cahill640, Niedzica Camacho396, Nening M. Dennis640, Tim A. Hollingsworth456, Gun Ho Jang451, Jeremy Johns241, Sangeetha Kalimuthu451, Sheng-Ben Dudderidge640,641, Sandra E. Edwards396, Cyril Fisher640, Christopher S. Foster642,643, Liang457, Ilinca Lungu451,458, Xuemei Luo12, Faridah Mbabaali241, Treasa A. McPherson454, Mohammed Ghori1, Pelvender Gill618, Vincent J. Gnanapragasam379,644, Gunes Gundem278, Jessica K. Miller241, Malcolm J. Moore455, Faiyaz Notta451,459, Danielle Pasternack241, Gloria Freddie C. Hamdy645, Steve Hawkins328, Steven Hazell640, William Howat379, William B. M. Petersen460, Michael H. A. Roehrl18,451,461,462,463, Michelle Sam241, Iris Selander454, Stefano Isaacs646, Katalin Karaszi618, Jonathan D. Kay188, Vincent Khoo640, Zsofia Kote-Jarai396, Serra410, Sagedeh Shahabi457, Sarah P. Thayer456, Lee E. Timms241, Gavin W. Wilson12,451, Julie Barbara Kremeyer1, Pardeep Kumar640, Adam Lambert618, Daniel A. Leongamornlert1,396, M. Wilson451, Bradly G. Wouters464, John D. McPherson241,451,465, Timothy A. Beck15,466, Naomi Livni640, Yong-Jie Lu639,647, Hayley J. Luxton188, Luke Marsden618, Charlie E. Massie328, Vinayak Bhandari12, Colin C. Collins146, Neil E. Fleshner467, Natalie S. Fox12, Michael Fraser12, Lucy Matthews396, Erik Mayer640,648, Ultan McDermott1, Sue Merson396, David E. Neal328,379, Lawrence E. Heisler468, Emilie Lalonde12, Julie Livingstone12, Alice Meng469, Veronica Y. Anthony Ng649, David Nicol640, Christopher Ogden640, Edward W. Rowe640, Nimish C. Sabelnykova12, Yu-Jia Shiah12, Theodorus Van der Kwast470, Robert G. Bristow18,471,472,473,474, Shah379, Sarah Thomas640, Alan Thompson640, Clare Verrill618,650, Tapio Visakorpi126, Anne Y. Shuai Ding475, Daiming Fan476, Lin Li179, Yongzhan Nie476,477, Xiao Xiao157, Rui Xing222,478, Warren379,651, Hayley C. Whitaker188, Hongwei Zhang647, Nicholas van As640, Rosalind A. Shanlin Yang475, Yingyan Yu479, Yong Zhou179, Rosamonde E. Banks480, Guillaume Eeles396,640, Adam Abeshouse278, Nishant Agrawal162, Rehan Akbani361,652, Hikmat Al- Bourque481,482, Paul Brennan483, Louis Letourneau484, Yasser Riazalhosseini482, Ghislaine Ahmadie278, Monique Albert450, Kenneth Aldape400,653, Adrian Ally654, Elizabeth L. Scelo483, Naveen Vasudev480,485, Juris Viksna486, Mark Lathrop482, Jörg Tost487, Sung-Min Appelbaum27,188, Joshua Armenia655, Sylvia Asa630,656, J. Todd Auman657, Miruna Ahn488, Samuel Aparicio489, Laurent Arnould490, M. R. Aure491, Shriram G. Bhosle1, Ewan Balasundaram654, Saianand Balu24, Jill Barnholtz-Sloan658,659, Oliver F. Bathe660,661, Stephen Birney7, Ake Borg492, Sandrine Boyault493, Arie B. Brinkman494, Jane E. Brock495, Annegien B. Baylin123,641, Christopher Benz662, Andrew Berchuck663, Mario Berrios664, Darell Bigner665, Broeks496, Anne-Lise Børresen-Dale491, Carlos Caldas497,498, Suet-Feung Chin497,498, Helen Michael Birrer19, Tom Bodenheimer24, Lori Boice632, Moiz S. Bootwalla664, Marcus Davies1,360,361, Christine Desmedt499,500, Luc Dirix501, Serge Dronov1, Anna Ehinger502, Jorunn Bosenberg666, Reanne Bowlby654, Jeffrey Boyd667, Russell R. Broaddus400, Malcolm E. Eyfjord503, Aquila Fatima203, John A. Foekens504, P. Andrew Futreal505, Øystein Brock668, Denise Brooks654, Susan Bullman3,167, Samantha J. Caesar-Johnson231, Thomas E. Garred506,507, Dilip D. Giri508, Dominik Glodzik1, Dorthe Grabau509, Holmfridur Carey669, Rebecca Carlsen654, Robert Cerfolio670, Vishal S. Chandan671, Hsiao-Wei Hilmarsdottir503, Gerrit K. Hooijer510, Jocelyne Jacquemier511, Se Jin Jang512, Jon G. Chen619,655, Andrew D. Cherniack3,156,167, Jeremy Chien672, Juok Cho3, Eric Chuah654, Carrie Jonasson503, Jos Jonkers513, Hyung-Yong Kim511, Tari A. King514,515, Stian Knappskog1,516, Gu Cibulskis3, Leslie Cope673, Matthew G. Cordes27,633, Erin Curley674, Bogdan Czerniak400,627, Kong511, Savitri Krishnamurthy517, Sunil R. Lakhani518, Anita Langerød491, Denis Larsimont519, Ludmila Danilova673, Ian J. Davis675, Timothy Defreitas3, John A. Demchok231, Noreen Hee Jin Lee512, Jeong-Yeon Lee520, Ming Ta Michael Lee505, Ole Christian Lingjærde521, Dhalla654, Rajiv Dhir676, HarshaVardhan Doddapaneni34, Adel El-Naggar400,627, Ina Felau231, Gaetan MacGrogan522, John W. M. Martens504, Sarah O’Meara1, Iris Pauporté214, Sarah Martin L. Ferguson677, Gaetano Finocchiaro678, Kwun M. Fong679, Scott Frazer3, William Pinder523, Xavier Pivot524, Elena Provenzano525, Colin A. Purdie526, Manasa Ramakrishna1, Friedman680, Catrina C. Fronick27,633, Lucinda A. Fulton27, Stacey B. Gabriel3, Jianjiong Kamna Ramakrishnan1, Jorge Reis-Filho508, Andrea L. Richardson203, Markus Ringnér492, Gao655, Nils Gehlenborg3,681, Jeffrey E. Gershenwald682,683, Ronald Ghossein508, Nasra H. Javier Bartolomé Rodriguez40, F. Germán Rodríguez-González261, Gilles Romieu527, Roberto Giama684, Richard A. Gibbs34, Carmen Gomez685, Ramaswamy Govindan26, D. Neil Salgado410, Torill Sauer521, Rebecca Shepherd1, Anieta M. Sieuwerts504, Peter T. Simpson518, Hayes24,686,687, Apurva M. Hegde136,137, David I. Heiman3, Zachary Heins278, Austin J. Marcel Smid504, Christos Sotiriou234, Paul N. Span528, Ólafur Andri Stefánsson529, Alasdair Hepperla24, Andrea Holbrook664, Robert A. Holt654, Alan P. Hoyle24, Ralph H. Hruban673, Stenhouse530, Henk G. Stunnenberg180,531, Fred Sweep532, Benita Kiat Tee Tan533, Gilles Jianhong Hu34, Mei Huang632, David Huntsman688, Jason Huse278, Christine A. Iacobuzio- Thomas534, Alastair M. Thompson530, Stefania Tommasi535, Isabelle Treilleux536,537, Andrew Donahue508, Michael Ittmann689,690, Joy C. Jayaseelan34, Stuart R. Jefferys24, Corbin D. Tutt203, Naoto T. Ueno387, Steven Van Laere501, Gert G. Van den Eynden501, Peter Jones691, Steven J. M. Jones692, Hartmut Juhl693, Koo Jeong Kang694, Beth Karlan695, Vermeulen501, Alain Viari418, Anne Vincent-Salomon531, Bernice H. Wong538, Lucy Yates1, Katayoon Kasaian692, Electron Kebebew696,697, Hark Kyun Kim698, Viktoriya Korchina34, Xueqing Zou1, Carolien H. M. van Deurzen539, Marc J. van de Vijver410, Laura van’t Veer540, Ritika Kundra619,655, Phillip H. Lai664, Eric Lander3, Xuan Le699, Darlene Lee654, Douglas A. Ole Ammerpohl541,542,543, Sietse Aukema542,543,544, Anke K. Bergmann545, Stephan H. Levine278,700, Lora Lewis34, Tim Ley701, Haiyan Irene Li654, Pei Lin3, W. M. Linehan702, Fei Fei Bernhart311,312,315, Arndt Borkhardt546, Christoph Borst547, Birgit Burkhardt548, Alexander Liu380, Yiling Lu137, Lisa Lype703, Yussanne Ma654, Dennis T. Maglinte664,704, Elaine R. Claviez549, Maria Elisabeth Goebler550, Andrea Haake541, Siegfried Haas547, Martin Mardis27,667,705, Jeffrey Marks414,706, Marco A. Marra654, Thomas J. Matthew37, Michael Hansmann551, Jessica I. Hoell546, Michael Hummel552, Dennis Karsch553, Wolfram Mayo654, Karen McCune707, Samuel R. Meier3, Shaowu Meng24, Piotr A. Mieczkowski23, Tom Klapper544, Michael Kneba553, Markus Kreuz554, Dieter Kube555, Ralf Küppers556, Dido Mikkelsen708, Christopher A. Miller27, Gordon B. Mills709, Richard A. Moore654, Carl Lenze552, Markus Loeffler554, Cristina López80,541, Luisa Mantovani-Löffler557, Peter Möller558, Morrison410,710, Lisle E. Mose24, Catherine D. Moser684, Andrew J. Mungall654, Karen German Ott559, Bernhard Radlwimmer399, Julia Richter541,544, Marius Rohde560, Philip C. Mungall654, David Mutch711, Donna M. Muzny712, Jerome Myers713, Yulia Newton37, Michael S. Rosenstiel561, Andreas Rosenwald562, Markus B. Schilhabel561, Stefan Schreiber563, Peter F. Noble3, Peter O’Donnell714, Brian Patrick O’Neill715, Angelica Ochoa278, Joong-Won Park716, Stadler311,312,315, Peter Staib564, Stephan Stilgenbauer565, Stephanie Sungalee8, Monika Joel S. Parker717, Harvey Pass718, Alessandro Pastore112, Nathan A. Pennell719, Charles M. Szczepanowski544, Umut H. Toprak30,566, Lorenz H. P. Trümper555, Rabea Wagener80,541, Perou720, Nicholas Petrelli721, Olga Potapova722, Janet S. Rader723, Suresh Ramalingam724, W. Thorsten Zenz149, Volker Hovestadt399, Christof von Kalle120, Marcel Kool246,331, Andrey Kimryn Rathmell725, Victor Reuter508, Sheila M. Reynolds703, Matthew Ringel726, Jeffrey Korshunov246, Pablo Landgraf567,568, Hans Lehrach569, Paul A. Northcott570, Stefan M. Roach727, Lewis R. Roberts684, A. Gordon Robertson654, Sara Sadeghi654, Charles Saller728, Pfister246,331,571, Guido Reifenberger568, Hans-Jörg Warnatz569, Stephan Wolf572, Marie-Laure Francisco Sanchez-Vega619,655, Dirk Schadendorf148,729, Jacqueline E. Schein654, Heather K. Yaspo569, Yassen Assenov573, Clarissa Gerhauser320, Sarah Minner574, Thorsten Schmidt27, Nikolaus Schultz655, Raja Seethala730, Yasin Senbabaoglu112, Troy Shelton674, Yan Schlomm99,575, Ronald Simon576, Guido Sauter576, Holger Sültmann149,577, Nidhan K. Shi24, Juliann Shih3,167, Ilya Shmulevich703, Craig Shriver731, Sabina Signoretti167,263,732, Janae Biswas578, Arindam Maitra578, Partha P. Majumder578, Rajiv Sarin579, Stefano Barbi422, Giada V. Simons24, Samuel Singer414,733, Payal Sipahimalani654, Tara J. Skelly23, Karen Smith- Bonizzato418, Cinzia Cantù418, Angelo P. Dei Tos580, Matteo Fassan581, Sonia Grimaldi418, McCune707, Nicholas D. Socci112, Matthew G. Soloway717, Anil K. Sood734, Angela Tam654, Claudio Luchini417, Giuseppe Malleo414, Giovanni Marchegiani414, Michele Milella229, Donghui Tan23, Roy Tarnuzzer231, Nina Thiessen654, R. Houston Thompson735, Leigh B. Salvatore Paiella414, Antonio Pea414, Paolo Pederzoli414, Andrea Ruzzenente414, Roberto Thorne632, Ming Tsao630,656, Christopher Umbricht324,621,736, David J. Van Den Berg664, Erwin Salvia414, Nicola Sperandio418, Yasuhito Arai226, Natsuko Hama226, Nobuyoshi Hiraoka582, G. Van Meir737, Umadevi Veluvolu23, Douglas Voet3, Linghua Wang34, Paul Weinberger738, Article Daniel J. Weisenberger664, Dennis Wigle739, Matthew D. Wilkerson23, Richard K. Wilson27,740, 69 Department of Computer Science, Yale University, New Haven, CT, USA. 70Sandra and Boris Winterhoff741, Maciej Wiznerowicz742,743, Tina Wong27,654, Winghing Wong744, Liu Xi34, Edward Meyer Cancer Center, Weill Cornell Medicine, New York, NY, USA. 71Department of Christina Yau662, Hailei Zhang3, Hongxin Zhang655 & Jiashan Zhang231 Physiology and Biophysics, Weill Cornell Medicine, New York, NY, USA. 72Englander Institute for Precision Medicine, Weill Cornell Medicine, New York, NY, USA. 73Institute for 1 Wellcome Sanger Institute, Hinxton, UK. 2Department of Haematology, University of Computational Biomedicine, Weill Cornell Medicine, New York, NY, USA. 74CNAG-CRG, Cambridge, Cambridge, UK. 3Broad Institute of MIT and Harvard, Cambridge, MA, USA. Centre for Genomic Regulation (CRG), Barcelona Institute of Science and Technology (BIST), 4 Center for Cancer Research, Massachusetts General Hospital, Boston, MA, USA. Barcelona, Spain. 75Department of Experimental and Health Sciences, Institute of 5 Department of Pathology, Massachusetts General Hospital, Boston, MA, USA. 6Harvard Evolutionary Biology (UPF-CSIC), Universitat Pompeu Fabra (UPF), Barcelona, Spain. Medical School, Boston, MA, USA. 7European Molecular Biology Laboratory (EMBL), 76 Institució Catalana de Recerca i Estudis Avançats (ICREA), Barcelona, Spain. 77Institut Català European Bioinformatics Institute (EMBL-EBI), Hinxton, UK. 8European Molecular Biology de Paleontologia Miquel Crusafont, Universitat Autònoma de Barcelona, Barcelona, Spain. Laboratory (EMBL), Genome Biology Unit, Heidelberg, Germany. 9Biomolecular Engineering 78 Department of Biomedical Data Science, Stanford University School of Medicine, Stanford, Department, University of California Santa Cruz, Santa Cruz, CA, USA. 10Adaptive Oncology CA, USA. 79Human Genetics, University of Kiel, Kiel, Germany. 80Institute of Human Genetics, Initiative, Ontario Institute for Cancer Research, Toronto, Ontario, Canada. 11International Ulm University and Ulm University Medical Center, Ulm, Germany. 81RIKEN Center for Cancer Genome Consortium (ICGC)/ICGC Accelerating Research in Genomic Oncology Integrative Medical Sciences, Yokohama, Japan. 82Department of Oncology, Centre for (ICGC-ARGO) Secretariat, Toronto, Ontario, Canada. 12Computational Biology Program, Cancer Genetic Epidemiology, University of Cambridge, Cambridge, UK. 83Department of Ontario Institute for Cancer Research, Toronto, Ontario, Canada. 13Department of Molecular Public Health and Primary Care, Centre for Cancer Genetic Epidemiology, University of Genetics, University of Toronto, Toronto, Ontario, Canada. 14Department of Radiation Cambridge, Cambridge, UK. 84Quantitative Genomics Laboratories (qGenomics), Barcelona, Oncology, University of California San Francisco, San Francisco, CA, USA. 15Genome Spain. 85Sage Bionetworks, Seattle, WA, USA. 86Department of Biochemistry and Molecular Informatics Program, Ontario Institute for Cancer Research, Toronto, Ontario, Canada. Medicine, University of Montreal, Montreal, Quebec, Canada. 87Institute for Research in 16 Department of Cell and Systems Biology, University of Toronto, Toronto, Ontario, Canada. Biomedicine (IRB Barcelona), Barcelona, Spain. 88National Centre for Biological Sciences, 17 Genome Informatics, Ontario Institute for Cancer Research, Toronto, Ontario, Canada. Tata Institute of Fundamental Research, Bangalore, India. 89Research Program on Biomedical 18 Department of Medical Biophysics, University of Toronto, Toronto, Ontario, Canada. Informatics, Universitat Pompeu Fabra (UPF), Barcelona, Spain. 90Broad Institute of Harvard 19 Massachusetts General Hospital, Boston, MA, USA. 20Department of Pharmacology, and MIT, Cambridge, MA, USA. 91The Francis Crick Institute, London, UK. 92University of University of Toronto, Toronto, Ontario, Canada. 21University of California Los Angeles, Los Leuven, Leuven, Belgium. 93Centre for Molecular Science Informatics, Department of Angeles, CA, USA. 22Department of Pathology, Department of Genomic Medicine and Chemistry, University of Cambridge, Cambridge, UK. 94Department of Biomedical Department of Translational Molecular Pathology, The University of Texas MD Anderson Informatics, Harvard Medical School, Boston, MA, USA. 95Ludwig Center at Harvard Medical Cancer Center, Houston, TX, USA. 23Department of Genetics, University of North Carolina at School, Boston, MA, USA. 96Division of Cancer Epidemiology and Genetics, National Cancer Chapel Hill, Chapel Hill, NC, USA. 24Lineberger Comprehensive Cancer Center, University of Institute, National Institutes of Health, Bethesda, MD, USA. 97Genome Integrity and Structural North Carolina at Chapel Hill, Chapel Hill, NC, USA. 25The Hospital for Sick Children, Toronto, Biology Laboratory, National Institute of Environmental Health Sciences (NIEHS), Durham, Ontario, Canada. 26Alvin J. Siteman Cancer Center, Washington University School of NC, USA. 98Integrative Bioinformatics Support Group, National Institute of Environmental Medicine, St Louis, MO, USA. 27The McDonnell Genome Institute, Washington University, St Health Sciences (NIEHS), Durham, NC, USA. 99Department of Urology, Charité Louis, MO, USA. 28Division of Theoretical Bioinformatics, German Cancer Research Center Universitätsmedizin Berlin, Berlin, Germany. 100Finsen Laboratory and Biotech Research & (DKFZ), Heidelberg, Germany. 29Heidelberg Center for Personalized Oncology (DKFZ-HIPO), Innovation Centre (BRIC), University of Copenhagen, Copenhagen, Denmark. 101Department German Cancer Research Center, Heidelberg, Germany. 30Institute of Pharmacy and of Bioengineering and Department of Cellular and Molecular Medicine, Moores Cancer Molecular Biotechnology, and BioQuant, Heidelberg University, Heidelberg, Germany. Center, University of California San Diego, La Jolla, CA, USA. 102Department of Genetics, 31 Bioinformatics and Omics Data Analytics, German Cancer Research Center (DKFZ), Microbiology and Statistics, University of Barcelona, IRSJD, IBUB, Barcelona, Spain. 103CIBER Heidelberg, Germany. 32Department of Bioinformatics and Computational Biology, The Epidemiología y Salud Pública (CIBERESP), Madrid, Spain. 104Research Group on Statistics, University of Texas MD Anderson Cancer Center, Houston, TX, USA. 33Department of Econometrics and Health (GRECS), UdG, Barcelona, Spain. 105Oxford Nanopore Molecular and Human Genetics, Baylor College of Medicine, Houston, TX, USA. 34Human Technologies, New York, NY, USA. 106Applications Department, Oxford Nanopore Genome Sequencing Center, Baylor College of Medicine, Houston, TX, USA. 35Department of Technologies, Oxford, UK. 107School of Molecular Biosciences and Center for Reproductive Genetics and Department of Medicine, Washington University in St Louis, St Louis, MO, USA. Biology, Washington State University, Pullman, WA, USA. 108Laboratory of Translational 36 Department of Computer Science, University of Toronto, Toronto, Ontario, Canada. Genomics, Division of Cancer Epidemiology and Genetics, National Cancer Institute, 37 University of California Santa Cruz, Santa Cruz, CA, USA. 38Computational Biology Program, National Institutes of Health, Bethesda, MD, USA. 109Department of Medical and Clinical Oregon Health & Science University, Portland, OR, USA. 39The Institute of Medical Science, Genetics, Genome-Scale Biology Research Program, University of Helsinki, Helsinki, Finland. The University of Tokyo, Tokyo, Japan. 40Barcelona Supercomputing Center (BSC), Barcelona, 110 Integrated Graduate Program in Physical and Engineering Biology, Yale University, New Spain. 41Department of Clinical and Molecular Medicine, Faculty of Medicine and Health Haven, CT, USA. 111Applied Tumor Genomics Research Program, Research Programs Unit, Sciences, Norwegian University of Science and Technology, Trondheim, Norway. 42Centre for University of Helsinki, Helsinki, Finland. 112Computational Biology Center, Memorial Sloan Research in Molecular Medicine and Chronic Diseases (CiMUS), Universidade de Santiago de Kettering Cancer Center, New York, NY, USA. 113Department of Biology, ETH Zurich, Zurich, Compostela, Santiago de Compostela, Spain. 43Department of Zoology, Genetics and Switzerland. 114Department of Computer Science, ETH Zurich, Zurich, Switzerland. 115SIB Physical Anthropology, Centre for Research in Molecular Medicine and Chronic Diseases Swiss Institute of Bioinformatics, Lausanne, Switzerland. 116University Hospital Zurich, Zurich, (CiMUS), Universidade de Santiago de Compostela, Santiago de Compostela, Spain. 44The Switzerland. 117Weill Cornell Medical College, New York, NY, USA. 118Berlin Institute for Biomedical Research Centre (CINBIO), Universidade de Vigo, Vigo, Spain. 45Department of Medical Systems Biology, Max Delbrück Center for Molecular Medicine, Berlin, Germany. Genetics, Stanford University School of Medicine, Stanford, CA, USA. 46Annai Systems, 119 German Cancer Consortium (DKTK), Partner site Berlin, Berlin, Germany. 120German Cancer Carlsbad, CA, USA. 47Centre for Genomic Regulation (CRG), The Barcelona Institute of Research Center (DKFZ), Heidelberg, Germany. 121Bakar Computational Health Sciences Science and Technology (BIST), Barcelona, Spain. 48Institute of Medical Genetics and Institute and Department of Pediatrics, University of California, San Francisco, CA, USA. Applied Genomics, University of Tübingen, Tübingen, Germany. 49Universitat Pompeu Fabra 122 Department of Biostatistics, Bloomberg School of Public Health, Johns Hopkins University, (UPF), Barcelona, Spain. 50Department of Computational Biology, University of Lausanne, Baltimore, MD, USA. 123Department of Oncology, The Johns Hopkins School of Medicine, The Lausanne, Switzerland. 51Department of Genetic Medicine and Development, University of Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins University, Baltimore, MD, Geneva Medical School, Geneva, Switzerland. 52Swiss Institute of Bioinformatics, University USA. 124Division of Computational Genomics and Systems Genetics, German Cancer of Geneva, Geneva, Switzerland. 53Department of Ophthalmology, Ocular Genomics Research Center (DKFZ), Heidelberg, Germany. 125Department of Medicine and Moores Institute, Massachusetts Eye and Ear, Harvard Medical School, Boston, MA, USA. Cancer Center, Division of Biomedical Informatics, UC San Diego School of Medicine, San 54 Department of Experimental and Health Sciences, Institute of Evolutionary Biology (UPF- Diego, CA, USA. 126Faculty of Medicine and Health Technology, Tampere University and Tays CSIC), Universitat Pompeu Fabra (UPF), Barcelona, Spain. 55Department of Veterinary Cancer Center, Tampere University Hospital, Tampere, Finland. 127Division of Applied Medicine, Transmissible Cancer Group, University of Cambridge, Cambridge, UK. Bioinformatics, German Cancer Research Center (DKFZ), Heidelberg, Germany. 128Faculty of 56 Department of Biochemistry, College of Medicine, Ewha Womans University, Seoul, South Biosciences, Heidelberg University, Heidelberg, Germany. 129Centre for Law and Genetics, Korea. 57Division of Oncology, Washington University School of Medicine, St Louis, MO, USA. University of Tasmania, Hobart, Tasmania, Australia. 130Centre of Genomics and Policy, McGill 58 School of Electronic and Information Engineering, Xi’an Jiaotong University, Xi’an, China. University and Génome Québec Innovation Centre, Montreal, Quebec, Canada. 131Heidelberg 59 The First Affiliated Hospital, Xi’an Jiaotong University, Xi’an, China. 60Independent Academy of Sciences and Humanities, Heidelberg, Germany. 132UC Santa Cruz Genomics Consultant, Wellesley, MA, USA. 61Icahn School of Medicine at Mount Sinai, New York, NY, Institute, University of California Santa Cruz, Santa Cruz, CA, USA. 133CIBIO/InBIO, Research USA. 62Biobyte Solutions, Heidelberg, Germany. 63Department of Molecular Biophysics and Center in Biodiversity and Genetic Resources, Universidade do Porto, Vairão, Portugal. Biochemistry, Yale University, New Haven, CT, USA. 64Program in Computational Biology and 134 Bioinformatics Unit, Spanish National Cancer Research Centre (CNIO), Madrid, Spain. Bioinformatics, Yale University, New Haven, CT, USA. 65Big Data Institute, Li Ka Shing Centre, 135 Howard Hughes Medical Institute, University of California Santa Cruz, Santa Cruz, CA, USA. University of Oxford, Oxford, UK. 66Oxford NIHR Biomedical Research Centre, University of 136 Cancer Unit, MRC University of Cambridge, Cambridge, UK. 137Department of Oxford, Oxford, UK. 67Department of Epidemiology and Biostatistics, Memorial Sloan Bioinformatics and Computational Biology and Department of Systems Biology, The Kettering Cancer Center, New York, NY, USA. 68The McDonnell Genome Institute at University of Texas MD Anderson Cancer Center, Houston, TX, USA. 138Center for Digital Washington University School of Medicine, and Department of Genetics and Department of Health, Berlin Institute of Health (BIH) and Charitè–Universitätsmedizin Berlin, Berlin, Medicine, Siteman Cancer Center, Washington University in St Louis, St Louis, MO, USA. Germany. 139Heidelberg Center for Personalized Oncology (DKFZ-HIPO), German Cancer Research Center (DKFZ), Heidelberg, Germany. 140Department of Bioinformatics and Agency for Medical Research and Development, Chiyoda-ku, Tokyo, Japan. 229Medical Computational Biology, University of Texas MD Anderson Cancer Center, Houston, TX, USA. Oncology, University and Hospital Trust of Verona, Verona, Italy. 230University of Verona, 141 Department of Genetics and Informatics Institute, University of Alabama at Birmingham, Verona, Italy. 231National Cancer Institute, National Institutes of Health, Bethesda, MD, USA. Birmingham, AL, USA. 142Heidelberg University, Heidelberg, Germany. 143New BIH Digital 232 CAPHRI Research School, Maastricht University, Maastricht, The Netherlands. 233Laboratory Health Center, Berlin Institute of Health (BIH) and Charité–Universitätsmedizin Berlin, Berlin, for Medical Science Mathematics, RIKEN Center for Integrative Medical Sciences, Yokohama, Germany. 144Department of Biochemistry and Molecular Biomedicine, University of Japan. 234University of California San Diego, San Diego, CA, USA. 235PDXen Biosystems, Seoul, Barcelona, Barcelona, Spain. 145Department of Urologic Sciences, University of British South Korea. 236Electronics and Telecommunications Research Institute, Daejeon, South Columbia, Vancouver, British Columbia, Canada. 146Vancouver Prostate Centre, Vancouver, Korea. 237Children’s Hospital of Philadelphia, Philadelphia, PA, USA. 238University of British Columbia, Canada. 147Division of Life Science and Applied Genomics Center, Hong Melbourne Centre for Cancer Research, Melbourne, Victoria, Australia. 239Syntekabio, Kong University of Science and Technology, Hong Kong, China. 148German Cancer Daejon, South Korea. 240AbbVie, North Chicago, IL, USA. 241Genomics Research Program, Consortium (DKTK), Heidelberg, Germany. 149National Center for Tumor Diseases (NCT) Ontario Institute for Cancer Research, Toronto, Ontario, Canada. 242Department of Pediatric Heidelberg, Heidelberg, Germany. 150Genome Integration Data Center, Syntekabio, Daejon, Immunology, Hematology and Oncology, University Hospital, Heidelberg, Germany. South Korea. 151Massachusetts General Hospital Center for Cancer Research, Charlestown, 243 Heidelberg Institute for Stem Cell Technology and Experimental Medicine (HI-STEM), MA, USA. 152Department of Molecular Medicine (MOMA), Aarhus University Hospital, Aarhus, Heidelberg, Germany. 244Seven Bridges, Charlestown, MA, USA. 245Health Sciences Denmark. 153Bioinformatics Research Centre (BiRC), Aarhus University, Aarhus, Denmark. Department of Biomedical Informatics, University of California San Diego, La Jolla, CA, USA. 154 Indiana University, Bloomington, IN, USA. 155Simon Fraser University, Burnaby, British 246 Functional and Structural Genomics, German Cancer Research Center (DKFZ), Heidelberg, Columbia, Canada. 156Dana-Farber Cancer Institute, Boston, MA, USA. 157School of Computer Germany. 247Leidos Biomedical Research, McLean, VA, USA. 248CSRA Incorporated, Fairfax, Science and Technology, Xi’an Jiaotong University, Xi’an, China. 158Department of Genetics, VA, USA. 249Department of Internal Medicine, Stanford University, Stanford, CA, USA. Washington University School of Medicine, St Louis, MO, USA. 159Department of Mathematics, 250 Clinical Bioinformatics, Swiss Institute of Bioinformatics, Geneva, Switzerland. 251Institute Washington University in St Louis, St Louis, MO, USA. 160Department of Biological for Pathology and Molecular Pathology, University Hospital Zurich, Zurich, Switzerland. Oceanography, Leibniz Institute of Baltic Sea Research, Rostock, Germany. 161Seven Bridges 252 Institute of Molecular Life Sciences, University of Zurich, Zurich, Switzerland. 253MIT Genomics, Charlestown, MA, USA. 162University of Chicago, Chicago, IL, USA. 163Department Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of of Health Sciences and Technology, Sungkyunkwan University School of Medicine, Seoul, Technology, Cambridge, MA, USA. 254Institute of Molecular Life Sciences and Swiss Institute South Korea. 164Samsung Genome Institute, Seoul, South Korea. 165New York Genome Center, of Bioinformatics, University of Zurich, Zurich, Switzerland. 255Office of Cancer Genomics, New York, NY, USA. 166Weill Cornell Medicine, New York, NY, USA. 167Department of Medical National Cancer Institute, National Institutes of Health, Bethesda, MD, USA. 256Computer Oncology, Dana-Farber Cancer Institute, Boston, MA, USA. 168Rigshospitalet, Copenhagen, Network Information Center, Chinese Academy of Sciences, Beijing, China. 257Geneplus- Denmark. 169Department of Computer Science, University of Toronto, Toronto, Ontario, Shenzhen, Shenzhen, China. 258Dana-Farber/Boston Children’s Cancer and Blood Disorders Canada. 170The Donnelly Centre, University of Toronto, Toronto, Ontario, Canada. 171Vector Center, Boston, MA, USA. 259Department of Pediatrics, Harvard Medical School, Boston, MA, Institute, Toronto, Ontario, Canada. 172Department of Medical Genetics, College of Medicine, USA. 260Technical University of Denmark, Lyngby, Denmark. 261University of Copenhagen, Hallym University, Chuncheon, South Korea. 173Department of Biology, ETH Zurich, Zurich, Copenhagen, Denmark. 262Department for BioMedical Research, University of Bern, Bern, Switzerland. 174University Hospital Zurich, Zurich, Switzerland. 175Peking University, Beijing, Switzerland. 263Department of Medical Oncology, Inselspital, University Hospital and China. 176School of Life Sciences, Peking University, Beijing, China. 177Computational and University of Bern, Bern, Switzerland. 264Graduate School for Cellular and Biomedical Systems Biology, Genome Institute of Singapore, Singapore, Singapore. 178School of Sciences, University of Bern, Bern, Switzerland. 265Department of Genitourinary Medical Computing, National University of Singapore, Singapore, Singapore. 179BGI-Shenzhen, Oncology - Research, Division of Cancer Medicine, The University of Texas MD Anderson Shenzhen, China. 180China National GeneBank-Shenzhen, Shenzhen, China. 181Computational Cancer Center, Houston, TX, USA. 266Department of Urology, Icahn School of Medicine at & Systems Biology Program, Memorial Sloan Kettering Cancer Center, New York, NY, USA. Mount Sinai, New York, NY, USA. 267Korea Advanced Institute of Science and Technology, 182 Korea University, Seoul, South Korea. 183Department of Genomic Medicine, The University Daejeon, South Korea. 268Science for Life Laboratory, Department of Cell and Molecular of Texas MD Anderson Cancer Center, Houston, TX, USA. 184Quantitative & Computational Biology, Uppsala University, Uppsala, Sweden. 269Queensland Centre for Medical Genomics, Biosciences Graduate Program, Baylor College of Medicine, Houston, TX, USA. 185The Institute for Molecular Bioscience, The University of Queensland, Brisbane, Queensland, Jackson Laboratory for Genomic Medicine, Farmington, CT, USA. 186Wolfson Wohl Cancer Australia. 270University of Milano Bicocca, Monza, Italy. 271Sir Peter MacCallum Department of Research Centre, Institute of Cancer Sciences, University of Glasgow, Bearsden, UK. 187The Oncology, Peter MacCallum Cancer Centre, University of Melbourne, Melbourne, Victoria, Azrieli Faculty of Medicine, Bar-Ilan University, Safed, Israel. 188University College London, Australia. 272Center for Precision Health, School of Biomedical Informatics, The University of London, UK. 189Genome Institute of Singapore, Singapore, Singapore. 190Department of Texas Health Science Center, Houston, TX, USA. 273Health Data Science Unit, University Epidemiology, University of Alabama at Birmingham, Birmingham, AL, USA. 191HudsonAlpha Clinics, Heidelberg, Germany. 274Department for Biomedical Research, University of Bern, Institute for Biotechnology, Huntsville, AL, USA. 192O’Neal Comprehensive Cancer Center, Bern, Switzerland. 275Research Core Center, National Cancer Centre Korea, Goyang-si, South University of Alabama at Birmingham, Birmingham, AL, USA. 193Department of Biosciences Korea. 276Institute of Computer Science, Polish Academy of Sciences, Warsawa, Poland. and Nutrition, Karolinska Institutet, Stockholm, Sweden. 194Cancer Science Institute of 277 Harvard University, Cambridge, MA, USA. 278Memorial Sloan Kettering Cancer Center, New Singapore, National University of Singapore, Singapore, Singapore. 195Programme in Cancer York, NY, USA. 279Department of Information Technology, Ghent University, Ghent, Belgium. & Stem Cell Biology, Duke-NUS Medical School, Singapore, Singapore. 196SingHealth, Duke- 280 Department of Plant Biotechnology and Bioinformatics, Ghent University, Ghent, Belgium. NUS Institute of Precision Medicine, National Heart Centre Singapore, Singapore, Singapore. 281 Yale School of Medicine, Yale University, New Haven, CT, USA. 282Division of Hematology- 197 Institute of Molecular and Cell Biology, Singapore, Singapore. 198Laboratory of Cancer Oncology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Epigenome, Division of Medical Science, National Cancer Centre Singapore, Singapore, South Korea. 283Samsung Advanced Institute for Health Sciences and Technology, Singapore. 199Department of Medicine, Baylor College of Medicine, Houston, TX, USA. Sungkyunkwan University School of Medicine, Seoul, South Korea. 284Cheonan Industry- 200 National Cancer Centre Singapore, Singapore, Singapore. 201BIOPIC, ICG and College of Academic Collaboration Foundation, Sangmyung University, Cheonan, South Korea. Life Sciences, Peking University, Beijing, China. 202Vall d’Hebron Institute of Oncology (VHIO), 285 Spanish National Cancer Research Centre, Madrid, Spain. 286Department of Computer Barcelona, Spain. 203Department of Cancer Biology, Dana-Farber Cancer Institute, Boston, Science, Princeton University, Princeton, NJ, USA. 287Bern Center for Precision Medicine, MA, USA. 204Institute for Research in Biomedicine (IRB Barcelona), The Barcelona Institute of University Hospital of Bern, University of Bern, Bern, Switzerland. 288Englander Institute for Science and Technology (BIST), Barcelona, Spain. 205Department of Mathematics, Aarhus Precision Medicine, Weill Cornell Medicine and New York Presbyterian Hospital, New York, University, Aarhus, Denmark. 206Institut Hospital del Mar d’Investigacions Mèdiques (IMIM), NY, USA. 289Meyer Cancer Center, Weill Cornell Medicine, New York, NY, USA. 290Pathology Barcelona, Spain. 207Ontario Institute for Cancer Research, Toronto, Ontario, Canada. 208King and Laboratory, Weill Cornell Medical College, New York, NY, USA. 291cBio Center, Dana- Faisal Specialist Hospital and Research Centre, Riyadh, Saudi Arabia. 209DLR Project Farber Cancer Institute, Harvard Medical School, Boston, MA, USA. 292Department of Cell Management Agency, Bonn, Germany. 210Genome Canada, Ottawa, Ontario, Canada. Biology, Harvard Medical School, Boston, MA, USA. 293cBio Center, Dana-Farber Cancer 211 Instituto Carlos Slim de la Salud, Mexico City, Mexico. 212Federal Ministry of Education and Institute, Boston, MA, USA. 294CREST, Japan Science and Technology Agency, Tokyo, Japan. Research, Berlin, Germany. 213Institut Gustave Roussy, Villejuif, France. 214Institut National du 295 Department of Medical Science Mathematics, Medical Research Institute, Tokyo Medical Cancer (INCA), Boulogne-Billancourt, France. 215The Wellcome Trust, London, UK. 216Prostate and Dental University, Bunkyo-ku, Tokyo, Japan. 296Laboratory for Medical Science Cancer Canada, Toronto, Ontario, Canada. 217National Human Genome Research Institute, Mathematics, Department of Biological Sciences, Graduate School of Science, The National Institutes of Health, Bethesda, MD, USA. 218Department of Biotechnology, Ministry of University of Tokyo, Bunkyo-ku, Tokyo, Japan. 297Science for Life Laboratory, Department of Science & Technology, Government of India, New Delhi, Delhi, India. 219Science Writer, Oncology-Pathology, Karolinska Institutet, Stockholm, Sweden. 298Department of Gene Garrett Park, MD, USA. 220Cancer Research UK, London, UK. 221Chinese Cancer Genome Technology, Tallinn University of Technology, Tallinn, Estonia. 299Genetics & Genome Biology Consortium, Shenzhen, China. 222Laboratory of Molecular Oncology, Key Laboratory of Program, SickKids Research Institute, The Hospital for Sick Children, Toronto, Ontario, Carcinogenesis and Translational Research (Ministry of Education), Peking University Cancer Canada. 300Department of Information Technology, Ghent University, Interuniversitair Micro- Hospital & Institute, Beijing, China. 223Key Laboratory of Carcinogenesis and Translational Electronica Centrum (IMEC), Ghent, Belgium. 301Science for Life Laboratory, Department of Research (Ministry of Education), Peking University Cancer Hospital & Institute, Beijing, Immunology, Genetics and Pathology, Uppsala University, Uppsala, Sweden. 302Oregon China. 224National Cancer Center, Tokyo, Japan. 225German Cancer Aid, Bonn, Germany. Health & Sciences University, Portland, OR, USA. 303Department of Medicine and 226 Division of Cancer Genomics, National Cancer Center Research Institute, National Cancer Therapeutics, The Chinese University of Hong Kong, Shatin, Hong Kong, China. 304The Center, Tokyo, Japan. 227Laboratory of Molecular Medicine, Human Genome Center, The University of Texas Health Science Center at Houston, Houston, TX, USA. 305Department of Institute of Medical Science, The University of Tokyo, Minato-ku, Tokyo, Japan. 228Japan Biomedical Informatics, College of Medicine, The Ohio State University, Columbus, OH, USA. Article 306 The Ohio State University Comprehensive Cancer Center (OSUCCC – James), Columbus, of Statistics, University of California Santa Cruz, Santa Cruz, CA, USA. 379Cambridge OH, USA. 307The University of Texas School of Biomedical Informatics (SBMI) at Houston, University Hospitals NHS Foundation Trust, Cambridge, UK. 380University of Toronto, Toronto, Houston, TX, USA. 308Department of Biochemistry and Molecular Genetics, Feinberg School Ontario, Canada. 381Department of Computer Science, Carleton College, Northfield, MN, of Medicine, Northwestern University, Chicago, IL, USA. 309Physics Division, Optimization and USA. 382Molecular and Medical Genetics, Oregon Health & Science University, Portland, OR, Systems Biology Lab, Massachusetts General Hospital, Boston, MA, USA. 310Genome Science USA. 383Center for Psychiatric Genetics, NorthShore University HealthSystem, Evanston, IL, Division, Research Center for Advanced Science and Technology, The University of Tokyo, USA. 384Argmix Consulting, North Vancouver, British Columbia, Canada. 385Department of Tokyo, Japan. 311Bioinformatics Group, Department of Computer Science, University of Biostatistics, The University of Texas MD Anderson Cancer Center, Houston, TX, USA. Leipzig, Leipzig, Germany. 312Interdisciplinary Center for Bioinformatics, University of Leipzig, 386 Department of Biostatistics, University of North Carolina at Chapel Hill, Chapel Hill, NC, Leipzig, Germany. 313Center for Bioinformatics and Functional Genomics, Cedars-Sinai USA. 387The University of Texas MD Anderson Cancer Center, Houston, TX, USA. 388Molecular Medical Center, Los Angeles, CA, USA. 314Computational Biology, Leibniz Institute on Aging - and Medical Genetics, Knight Cancer Institute, Oregon Health & Science University, Portland, Fritz Lipmann Institute (FLI), Jena, Germany. 315Transcriptome Bioinformatics, LIFE Research OR, USA. 389Department of Health Sciences, Faculty of Medical Sciences, Kyushu University, Center for Civilization Diseases, University of Leipzig, Leipzig, Germany. 316Center for Fukuoka, Japan. 390Baylor College of Medicine, Houston, TX, USA. 391Department of Applied Epigenetics, Van Andel Research Institute, Grand Rapids, MI, USA. 317Institut d’Investigacions Mathematics and Statistics, Johns Hopkins University, Baltimore, MD, USA. 392Heinrich Pette Biomèdiques August Pi i Sunyer (IDIBAPS), Barcelona, Spain. 318Research Center for Institute, Leibniz Institute for Experimental Virology, Hamburg, Germany. 393University Advanced Science and Technology, The University of Tokyo, Minato-ku, Tokyo, Japan. 319Van Medical Center Hamburg-Eppendorf, Bioinformatics Core, Hamburg, Germany. 394Earlham Andel Research Institute, Grand Rapids, MI, USA. 320Cancer Epigenomics, German Cancer Institute, Norwich, UK. 395Norwich Medical School, University of East Anglia, Norwich, UK. Research Center (DKFZ), Heidelberg, Germany. 321Department of Biomedical Sciences, 396 The Institute of Cancer Research, London, UK. 397University of East Anglia, Norwich, UK. Cedars-Sinai Medical Center, Los Angeles, CA, USA. 322The Hebrew University Faculty of 398 German Center for Infection Research (DZIF), Partner Site Hamburg-Borstel-Lübeck-Riems, Medicine, Jerusalem, Israel. 323German Cancer Consortium (DKTK), German Cancer Research Hamburg, Germany. 399Division of Molecular Genetics, German Cancer Research Center Center (DKFZ), Heidelberg, Germany. 324Department of Pathology, Johns Hopkins University (DKFZ), Heidelberg, Germany. 400Department of Pathology, The University of Texas MD School of Medicine, Baltimore, MD, USA. 325McKusick-Nathans Institute of Genetic Medicine, Anderson Cancer Center, Houston, TX, USA. 401Victorian Institute of Forensic Medicine, Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins University School of Medicine, Southbank, Victoria, Australia. 402Peter MacCallum Cancer Centre, University of Melbourne, Baltimore, MD, USA. 326Foundation Medicine, Cambridge, MA, USA. 327Department of Melbourne, Victoria, Australia. 403University of Pennsylvania, Philadelphia, PA, USA. 404Centre Biochemistry, Microbiology and Immunology, Faculty of Medicine, University of Ottawa, for Cancer Research, The Westmead Institute for Medical Research, Sydney, New South Ottawa, Ontario, Canada. 328Cancer Research UK Cambridge Institute, University of Wales, Australia. 405Department of Gynaecological Oncology, Westmead Hospital, Sydney, Cambridge, Cambridge, UK. 329University of Cambridge, Cambridge, UK. 330Brandeis New South Wales, Australia. 406Genetics and Molecular Pathology, SA Pathology, Adelaide, University, Waltham, MA, USA. 331Hopp Children’s Cancer Center (KiTZ), Heidelberg, South Australia, Australia. 407Centre for Cancer Research, The Westmead Institute for Medical Germany. 332Pediatric Glioma Research Group, German Cancer Research Center (DKFZ), Research, The University of Sydney, Sydney, New South Wales, Australia. 408Department of Heidelberg, Germany. 333A. A. Kharkevich Institute of Information Transmission Problems, Gynaecological Oncology, Westmead Hospital, Sydney, New South Wales, Australia. Moscow, Russia. 334Oncology and Immunology, Dmitry Rogachev National Research Center 409 Garvan Institute of Medical Research, Darlinghurst, New South Wales, Australia. of Pediatric Hematology, Moscow, Russia. 335Skolkovo Institute of Science and Technology, 410 Department of Clinical Pathology, University of Melbourne, Melbourne, Victoria, Australia. Moscow, Russia. 336Center for Medical Innovation, Seoul National University Hospital, Seoul, 411 Centre for Cancer Research, The Westmead Institute for Medical Research, The University South Korea. 337Department of Internal Medicine, Seoul National University Hospital, Seoul, of Sydney, Sydney, New South Wales, Australia. 412Department of Gynaecological Oncology, South Korea. 338Division of Genetics and Genomics, Boston Children’s Hospital, Harvard Westmead Hospital, Sydney, New South Wales, Australia. 413Westmead Clinical School, The Medical School, Boston, MA, USA. 339School of Medicine/School of Mathematics and Westmead Institute for Medical Research, Sydney, New South Wales, Australia. 414Department Statistics, University of St Andrews, St Andrews, UK. 340Department of Genetics and of Surgery, Pancreas Institute, University and Hospital Trust of Verona, Verona, Italy. Computational Biology, QIMR Berghofer Medical Research Institute, Brisbane, Queensland, 415 Department of Surgery, Princess Alexandra Hospital, Brisbane, Queensland, Australia. Australia. 341Institute for Molecular Bioscience, University of Queensland, Brisbane, 416 Surgical Oncology Group, Diamantina Institute, The University of Queensland, Brisbane, Queensland, Australia. 342Cancer Research Institute, Beth Israel Deaconess Medical Center, Queensland, Australia. 417Department of Diagnostics and Public Health, University and Boston, MA, USA. 343Ben May Department for Cancer Research, Department of Human Hospital Trust of Verona, Verona, Italy. 418ARC-Net Centre for Applied Research on Cancer, Genetics, The University of Chicago, Chicago, IL, USA. 344Tri-Institutional PhD Program in University and Hospital Trust of Verona, Verona, Italy. 419Illawarra Shoalhaven Local Health Computational Biology and Medicine, Weill Cornell Medicine, New York, NY, USA. District L3 Illawarra Cancer Care Centre, Wollongong Hospital, Wollongong, New South 345 Department of Bioengineering, and Department of Cellular and Molecular Medicine, Wales, Australia. 420Department of Pathology, University of Sydney, Sydney, New South Moores Cancer Center, University of California, San Diego, La Jolla, CA, USA. 346Centre for Wales, Australia. 421School of Biological Sciences, The University of Auckland, Auckland, New Computational Biology, Duke-NUS Medical School, Singapore, Singapore. 347Department of Zealand. 422Department of Pathology and Diagnostics, University and Hospital Trust of Computer Science, University of Helsinki, Helsinki, Finland. 348Institute of Biotechnology, Verona, Verona, Italy. 423Department of Medicine, Section of Endocrinology, University and University of Helsinki, Helsinki, Finland. 349Organismal and Evolutionary Biology Research Hospital Trust of Verona, Verona, Italy. 424Department of Pathology, Queen Elizabeth Programme, University of Helsinki, Helsinki, Finland. 350Programme in Cancer & Stem Cell University Hospital, Glasgow, UK. 425Department of Medical Oncology, Beatson West of Biology, Centre for Computational Biology, Duke-NUS Medical School, Singapore, Scotland Cancer Centre, Glasgow, UK. 426Academic Unit of Surgery, School of Medicine, Singapore. 351Department of Applied Mathematics and Theoretical Physics, Centre for College of Medical, Veterinary and Life Sciences, University of Glasgow, Glasgow Royal Mathematical Sciences, University of Cambridge, Cambridge, UK. 352Department of Infirmary, Glasgow, UK. 427Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Statistics, Columbia University, New York, NY, USA. 353Duke-NUS Medical School, Singapore, Hospital, Camperdown, New South Wales, Australia. 428Discipline of Surgery, Western Sydney Singapore. 354School of Electronic Information and Communications, Huazhong University of University, Penrith, New South Wales, Australia. 429Institute of Cancer Sciences, College of Science and Technology, Wuhan, China. 355The Kinghorn Cancer Centre, Cancer Division, Medical Veterinary and Life Sciences, University of Glasgow, Glasgow, UK. 430The Kinghorn Garvan Institute of Medical Research, University of New South Wales, Sydney, New South Cancer Centre, Cancer Division, Garvan Institute of Medical Research, University of New Wales, Australia. 356MRC Human Genetics Unit, MRC IGMM, University of Edinburgh, South Wales, Sydney, New South Wales, Australia. 431School of Environmental and Life Edinburgh, UK. 357Bioinformatics Group, Division of Molecular Biology, Department of Sciences, Faculty of Science, The University of Newcastle, Ourimbah, New South Wales, Biology, Faculty of Science, University of Zagreb, Zagreb, Croatia. 358Department of Australia. 432Eastern Clinical School, Monash University, Melbourne, Victoria, Australia. Bioinformatics, Division of Cancer Genomics, National Cancer Center Research Institute, 433 Epworth HealthCare, Richmond, Victoria, Australia. 434Olivia Newton-John Cancer National Cancer Center, Tokyo, Japan. 359University of Glasgow, Glasgow, UK. 360Academic Research Institute, La Trobe University, Heidelberg, Victoria, Australia. 435Melanoma Institute Department of Medical Genetics, University of Cambridge, Addenbrooke’s Hospital, Australia, The University of Sydney, Wollstonecraft, New South Wales, Australia. 436Children’s Cambridge, UK. 361MRC Cancer Unit, University of Cambridge, Cambridge, UK. 362The Hospital at Westmead, The University of Sydney, Sydney, New South Wales, Australia. University of Cambridge School of Clinical Medicine, Cambridge, UK. 363MRC-University of 437 Melanoma Institute Australia, The University of Sydney, Sydney, New South Wales, Glasgow Centre for Virus Research, Glasgow, UK. 364Wolfson Wohl Cancer Research Centre, Australia. 438Australian Institute of Tropical Health and Medicine, James Cook University, Institute of Cancer Sciences, University of Glasgow, Bearsden, UK. 365School of Computing Douglas, Queensland, Australia. 439Bioplatforms Australia, North Ryde, New South Wales, Science, University of Glasgow, Glasgow, UK. 366South Western Sydney Clinical School, Australia. 440Melanoma Institute Australia, Macquarie University, Wollstonecraft, New South Faculty of Medicine, University of New South Wales, Liverpool, New South Wales, Australia. Wales, Australia. 441Children’s Medical Research Institute, Sydney, New South Wales, 367 West of Scotland Pancreatic Unit, Glasgow Royal Infirmary, Glasgow, UK. 368University of Australia. 442Melanoma Institute Australia, The University of Sydney, Wollstonecraft, New Melbourne Centre for Cancer Research, Melbourne, Victoria, Australia. 369Molecular and South Wales, Australia. 443Centre for Cancer Research, The Westmead Millennium Institute Medical Genetics, Oregon Health & Science University, Portland, OR, USA. 370Department of for Medical Research, University of Sydney, Westmead Hospital, Sydney, New South Wales, Surgery, University of Melbourne, Parkville, Victoria, Australia. 371The Murdoch Children’s Australia. 444Translational Cancer Research Centre, The University of Sydney at the Research Institute, Royal Children’s Hospital, Parkville, Victoria, Australia. 372Walter + Eliza Westmead Institute, Sydney, New South Wales, Australia. 445Discipline of Pathology, Sydney Hall Institute, Parkville, Victoria, Australia. 373University of Cologne, Cologne, Germany. 374The Medical School, The University of Sydney, Sydney, New South Wales, Australia. 446School of Edward S. Rogers Sr Department of Electrical and Computer Engineering, University of Mathematics and Statistics, The University of Sydney, Sydney, New South Wales, Australia. Toronto, Toronto, Ontario, Canada. 375University of Ljubljana, Ljubljana, Slovenia. 447 Melanoma Institute Australia, The University of Sydney, Wollstonecraft, New South Wales, 376 Department of Public Health Sciences, The University of Chicago, Chicago, IL, USA. Australia. 448Royal Prince Alfred Hospital, Sydney, New South Wales, Australia. 449Diagnostic 377 Research Institute, NorthShore University HealthSystem, Evanston, IL, USA. 378Department Development, Ontario Institute for Cancer Research, Toronto, Ontario, Canada. 450Ontario Tumour Bank, Ontario Institute for Cancer Research, Toronto, Ontario, Canada. 451PanCuRx France. 523Department of Research Oncology, Guy’s Hospital, King’s Health Partners AHSC, Translational Research Initiative, Ontario Institute for Cancer Research, Toronto, Ontario, King’s College London School of Medicine, London, UK. 524University Hospital of Minjoz, Canada. 452BioSpecimen Sciences Program, University Health Network, Toronto, Ontario, INSERM UMR 1098, Besançon, France. 525Cambridge Breast Unit, Addenbrooke’s Hospital, Canada. 453Hepatobiliary/Pancreatic Surgical Oncology Program, University Health Network, Cambridge University Hospital NHS Foundation Trust and NIHR Cambridge Biomedical Toronto, Ontario, Canada. 454Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, Research Centre, Cambridge, UK. 526East of Scotland Breast Service, Ninewells Hospital, Toronto, Ontario, Canada. 455Division of Medical Oncology, Princess Margaret Cancer Centre, Aberdeen, UK. 527Oncologie Sénologie, ICM Institut Régional du Cancer, Montpellier, France. Toronto, Ontario, Canada. 456University of Nebraska Medical Center, Omaha, NE, USA. 528 Department of Radiation Oncology, Radboud University Nijmegen Medical Centre, 457 BioSpecimen Sciences Program, University Health Network, Toronto, Ontario, Canada. Nijmegen, The Netherlands. 529University of Iceland, Reykjavik, Iceland. 530Dundee Cancer 458 Transformative Pathology, Ontario Institute for Cancer Research, Toronto, Ontario, Centre, Ninewells Hospital, Dundee, UK. 531Institut Curie, INSERM Unit 830, Paris, France. Canada. 459University Health Network, Princess Margaret Cancer Centre, Toronto, Ontario, 532 Department of Laboratory Medicine, Radboud University Nijmegen Medical Centre, Canada. 460Department of Health Sciences Research, Mayo Clinic, Rochester, MN, USA. Nijmegen, The Netherlands. 533Department of General Surgery, Singapore General Hospital, 461 BioSpecimen Sciences, Laboratory Medicine (Toronto), Medical Biophysics, PanCuRX, Singapore, Singapore. 534INCa-Synergie, Centre Léon Bérard, Universite Lyon, Lyon, France. Toronto, Ontario, Canada. 462Department of Laboratory Medicine and Pathobiology, 535 Giovanni Paolo II/I.R.C.C.S. Cancer Institute, Bari, Italy. 536Department of Biopathology, University of Toronto, Toronto, Ontario, Canada. 463Department of Pathology, Human Centre Léon Bérard, Lyon, France. 537Université Claude Bernard Lyon 1, Villeurbanne, France. Oncology and Pathogenesis Program, Memorial Sloan Kettering Cancer Center, New York, 538 NCCS-VARI Translational Research Laboratory, National Cancer Centre Singapore, NY, USA. 464Department of Medical Biophysics, University of Toronto, Toronto, Ontario, Singapore, Singapore. 539Department of Pathology, Erasmus Medical Center Rotterdam, Canada. 465Department of Biochemistry and Molecular Medicine, University California at Rotterdam, The Netherlands. 540Division of Molecular Carcinogenesis, The Netherlands Davis, Sacramento, CA, USA. 466Human Longevity, San Diego, CA, USA. 467Department of Cancer Institute, Amsterdam, The Netherlands. 541Institute of Human Genetics, Christian- Surgical Oncology, Princess Margaret Cancer Centre, Toronto, Ontario, Canada. 468Genome Albrechts-University, Kiel, Germany. 542Institute of Human Genetics, University of Ulm, Ulm, Informatics Program, Ontario Institute for Cancer Research, Toronto, Ontario, Canada. Germany. 543University Hospital of Ulm, Ulm, Germany. 544Hematopathology Section, Institute 469 STTARR Innovation Facility, Princess Margaret Cancer Centre, Toronto, Ontario, Canada. of Pathology, Christian-Albrechts-University, Kiel, Germany. 545Department of Human 470 Department of Pathology, Toronto General Hospital, Toronto, Ontario, Canada. 471CRUK Genetics, Hannover Medical School, Hannover, Germany. 546Department of Pediatric Manchester Institute and Centre, Manchester, UK. 472Department of Radiation Oncology, Oncology, Hematology and Clinical Immunology, Heinrich-Heine-University, Düsseldorf, University of Toronto, Toronto, Ontario, Canada. 473Manchester Cancer Research Centre, Germany. 547Department of Internal Medicine/Hematology, Friedrich-Ebert-Hospital, Cancer Division, FBMH, University of Manchester, Manchester, UK. 474Radiation Medicine Neumünster, Germany. 548Pediatric Hematology and Oncology, University Hospital Muenster, Program, Princess Margaret Cancer Centre, Toronto, Ontario, Canada. 475Hefei University of Muenster, Germany. 549Department of Pediatrics, University Hospital Schleswig-Holstein, Technology, Anhui, China. 476State Key Laboratory of Cancer Biology and Xijing Hospital of Kiel, Germany. 550Department of Medicine II, University of Würzburg, Würzburg, Germany. Digestive Diseases, Fourth Military Medical University, Shaanxi, China. 477Fourth Military 551 Senckenberg Institute of Pathology, University of Frankfurt Medical School, Frankfurt, Medical University, Shaanxi, China. 478Laboratory of Molecular Oncology, Key Laboratory of Germany. 552Institute of Pathology, Charité–University Medicine Berlin, Berlin, Germany. Carcinogenesis and Translational Research (Ministry of Education), Peking University Cancer 553 Department for Internal Medicine II, University Hospital Schleswig-Holstein, Kiel, Germany. Hospital & Institute, Beijing, China. 479Department of Surgery, Ruijin Hospital, Shanghai 554 Institute for Medical Informatics Statistics and Epidemiology, University of Leipzig, Leipzig, Jiaotong University School of Medicine, Shanghai, China. 480Leeds Institute of Medical Germany. 555Department of Hematology and Oncology, Georg-Augusts-University of Research, University of Leeds, St James’s University Hospital, Leeds, UK. 481Canadian Center Göttingen, Göttingen, Germany. 556Institute of Cell Biology (Cancer Research), University of for Computational Genomics, McGill University, Montreal, Quebec, Canada. 482Department Duisburg-Essen, Essen, Germany. 557MVZ Department of Oncology, PraxisClinic am of Human Genetics, McGill University, Montreal, Quebec, Canada. 483International Agency for Johannisplatz, Leipzig, Germany. 558Institute of Pathology, Ulm University and University Research on Cancer, Lyon, France. 484McGill University and Genome Quebec Innovation Hospital of Ulm, Ulm, Germany. 559Department of Pathology, Robert-Bosch-Hospital, Centre, Montreal, Quebec, Canada. 485St James Institute of Oncology, University of Leeds, St Stuttgart, Germany. 560Pediatric Hematology and Oncology, University Hospital Giessen, James’s University Hospital, Leeds, UK. 486Institute of Mathematics and Computer Science, Giessen, Germany. 561Institute of Clinical Molecular Biology, Christian-Albrechts-University, University of Latvia, Riga, Latvia. 487Centre National de Génotypage, CEA - Institute de Kiel, Germany. 562Institute of Pathology, University of Wuerzburg, Wuerzburg, Germany. Génomique, Evry, France. 488Department of Oncology, Gil Medical Center, Gachon University, 563 Department of General Internal Medicine, University Kiel, Kiel, Germany. 564Clinic for Incheon, South Korea. 489Department of Molecular Oncology, BC Cancer Agency, Vancouver, Hematology and Oncology, St-Antonius-Hospital, Eschweiler, Germany. 565Department for British Columbia, Canada. 490Los Alamos National Laboratory, Los Alamos, NM, USA. Internal Medicine III, University of Ulm and University Hospital of Ulm, Ulm, Germany. 491 Department of Genetics, Institute for Cancer Research, Oslo University Hospital, The 566 Neuroblastoma Genomics, German Cancer Research Center (DKFZ), Heidelberg, Germany. Norwegian Radium Hospital, Oslo, Norway. 492Lund University, Lund, Sweden. 493Translational 567 Department of Pediatric Oncology and Hematology, University of Cologne, Cologne, Research Lab, Centre Léon Bérard, Lyon, France. 494Department of Molecular Biology, Faculty Germany. 568University of Düsseldorf, Düsseldorf, Germany. 569Department of Vertebrate of Science, Radboud Institute for Molecular Life Sciences, Radboud University, Nijmegen, Genomics/Otto Warburg Laboratory Gene Regulation and Systems Biology of Cancer, Max The Netherlands. 495Department of Pathology, Brigham and Women’s Hospital, Harvard Planck Institute for Molecular Genetics, Berlin, Germany. 570St Jude Children’s Research Medical School, Boston, MA, USA. 496Department of Molecular Pathology, The Netherlands Hospital, Memphis, TN, USA. 571Heidelberg University Hospital, Heidelberg, Germany. Cancer Institute, Amsterdam, The Netherlands. 497Li Ka Shing Centre, Cancer Research UK 572 Genomics and Proteomics Core Facility High Throughput Sequencing Unit, German Cambridge Institute, University of Cambridge, Cambridge, UK. 498Department of Oncology, Cancer Research Center (DKFZ), Heidelberg, Germany. 573Epigenomics and Cancer Risk University of Cambridge, Cambridge, UK. 499Breast Cancer Translational Research Laboratory Factors, German Cancer Research Center (DKFZ), Heidelberg, Germany. 574University Medical J. C. Heuson, Institut Jules Bordet, Brussels, Belgium. 500Laboratory for Translational Breast Center Hamburg-Eppendorf, Hamburg, Germany. 575Martini-Clinic, Prostate Cancer Center, Cancer Research, Department of Oncology, KU Leuven, Leuven, Belgium. 501Translational University Medical Center Hamburg-Eppendorf, Hamburg, Germany. 576Institute of Pathology, Cancer Research Unit, GZA Hospitals St-Augustinus, Center for Oncological Research, University Medical Center Hamburg-Eppendorf, Hamburg, Germany. 577Division of Cancer Faculty of Medicine and Health Sciences, University of Antwerp, Antwerp, Belgium. Genome Research, German Cancer Research Center (DKFZ), Heidelberg, Germany. 502 Department of Gynecology & Obstetrics and Department of Clinical Sciences, Skåne 578 National Institute of Biomedical Genomics, Kalyani, India. 579Advanced Centre for University Hospital, Lund University, Lund, Sweden. 503Icelandic Cancer Registry, Icelandic Treatment Research & Education in Cancer, Tata Memorial Centre, Navi Mumbai, India. Cancer Society, Reykjavik, Iceland. 504Department of Medical Oncology, Josephine Nefkens 580 Department of Pathology, General Hospital of Treviso, Department of Medicine, University Institute and Cancer Genomics Centre, Erasmus Medical Center, Rotterdam, The of Padua, Treviso, Italy. 581Department of Medicine (DIMED), Surgical Pathology Unit, Netherlands. 505National Genotyping Center, Institute of Biomedical Sciences, Academia University of Padua, Padua, Italy. 582Department of Hepatobiliary and Pancreatic Oncology, Sinica, Taipei, Taiwan. 506Department of Pathology, Oslo University Hospital Ulleval, Oslo, Hepatobiliary and Pancreatic Surgery Division, Division of Pathology and Clinical Norway. 507Faculty of Medicine and Institute of Clinical Medicine, University of Oslo, Oslo, Laboratories, National Cancer Center Hospital, Chuo-ku, Tokyo, Japan. 583Department of Norway. 508Department of Pathology, Memorial Sloan Kettering Cancer Center, New York, NY, Pathology, Keio University School of Medicine, Tokyo, Japan. 584Department of Hepatobiliary USA. 509Department of Pathology, Skåne University Hospital, Lund University, Lund, Sweden. and Pancreatic Oncology, National Cancer Center Hospital, Tokyo, Japan. 585Department of 510 Department of Pathology, Academic Medical Center, Amsterdam, The Netherlands. Pathology, Graduate School of Medicine, The University of Tokyo, Bunkyo-ku, Tokyo, Japan. 511 Department of Pathology, College of Medicine, Hanyang University, Seoul, South Korea. 586 Preventive Medicine, Graduate School of Medicine, The University of Tokyo, Tokyo, Japan. 512 Department of Pathology, Asan Medical Center, College of Medicine, Ulsan University, 587 Gastric Surgery Division, Division of Pathology and Clinical Laboratories, National Cancer Songpa-gu, Seoul, South Korea. 513The Netherlands Cancer Institute, Amsterdam, The Center Hospital, Tokyo, Japan. 588Department of Gastroenterology and Hepatology, Netherlands. 514Department of Surgery, Dana-Farber Cancer Institute, Brigham and Women’s Yokohama City University Graduate School of Medicine, Kanagawa, Japan. 589Laboratory of Hospital, Boston, MA, USA. 515Department of Surgery, Memorial Sloan Kettering Cancer Molecular Medicine, Human Genome Center, The Institute of Medical Science, University of Center, New York, NY, USA. 516Department of Clinical Science, University of Bergen, Bergen, Tokyo, Tokyo, Japan. 590Department of Cancer Genome Informatics, Graduate School of Norway. 517Morgan Welch Inflammatory Breast Cancer Research Program and Clinic, The Medicine, Osaka University, Osaka, Japan. 591Hiroshima University, Hiroshima, Japan. 592Tokyo University of Texas MD Anderson Cancer Center, Houston, TX, USA. 518The University of Women’s Medical University, Tokyo, Japan. 593Osaka International Cancer Center, Osaka, Queensland Centre for Clinical Research, The Royal Brisbane & Women’s Hospital, Herston, Japan. 594Wakayama Medical University, Wakayama, Japan. 595Hokkaido University, Sapporo, Queensland, Australia. 519Department of Pathology, Institut Jules Bordet, Brussels, Belgium. Japan. 596Division of Medical Oncology, National Cancer Centre, Singapore, Singapore. 520 Institute for Bioengineering and Biopharmaceutical Research (IBBR), Hanyang University, 597 Cholangiocarcinoma Screening and Care Program and Liver Fluke and Seoul, South Korea. 521University of Oslo, Oslo, Norway. 522Institut Bergonié, Bordeaux, Cholangiocarcinoma Research Centre, Faculty of Medicine, Khon Kaen University, Article Khon Kaen, Thailand. 598Lymphoma Genomic Translational Research Laboratory, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA. 678Department of Cancer Centre, Singapore, Singapore. 599Center of Digestive Diseases and Liver Neuro-Oncology, Istituto Neurologico Besta, Milan, Italy. 679University of Queensland Transplantation, Fundeni Clinical Institute, Bucharest, Romania. 600Division of Hepatobiliary Thoracic Research Centre, The Prince Charles Hospital, Brisbane, Queensland, Australia. and Pancreatic Surgery, Department of Surgery, School of Medicine, Keimyung University 680 Department of Neurosurgery, University of Florida, Gainesville, FL, USA. 681Center for Dongsan Medical Center, Daegu, South Korea. 601Pathology, Hospital Clinic, Institut Biomedical Informatics, Harvard Medical School, Boston, MA, USA. 682Department of Cancer d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), University of Barcelona, Biology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA. Barcelona, Spain. 602Hematology, Hospital Clinic, Institut d’Investigacions Biomèdiques 683 Department of Surgical Oncology, The University of Texas MD Anderson Cancer Center, August Pi i Sunyer (IDIBAPS), University of Barcelona, Barcelona, Spain. 603Department of Houston, TX, USA. 684Division of Gastroenterology and Hepatology, Mayo Clinic, Rochester, Biochemistry and Molecular Biology, Faculty of Medicine, University Institute of Oncology- MN, USA. 685Sylvester Comprehensive Cancer Center, University of Miami, Miami, FL, USA. IUOPA, Oviedo, Spain. 604Anatomia Patológica, Hospital Clinic, Institut d’Investigacions 686 Department of Internal Medicine, Division of Medical Oncology, Lineberger Biomèdiques August Pi i Sunyer (IDIBAPS), University of Barcelona, Barcelona, Spain. Comprehensive Cancer Center, University of North Carolina at Chapel Hill, Chapel Hill, NC, 605 Spanish Ministry of Science and Innovation, Madrid, Spain. 606Royal National Orthopaedic USA. 687University of Tennessee Health Science Center for Cancer Research, Memphis, TN, Hospital (Bolsover), London, UK. 607Department of Pathology, Oslo University Hospital, The USA. 688Centre for Translational and Applied Genomics, British Columbia Cancer Agency, Norwegian Radium Hospital, Oslo, Norway. 608Institute of Clinical Medicine and Institute of Vancouver, British Columbia, Canada. 689Department of Pathology & Immunology, Baylor Oral Biology, University of Oslo, Oslo, Norway. 609Research Department of Pathology, College of Medicine, Houston, TX, USA. 690Michael E. DeBakey Veterans Affairs Medical University College London Cancer Institute, London, UK. 610East Anglian Medical Genetics Center, Houston, TX, USA. 691Carolina Center for Genome Sciences, University of North Service, Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK. 611Royal Carolina at Chapel Hill, Chapel Hill, NC, USA. 692Canada’s Michael Smith Genome Sciences National Orthopaedic Hospital (Stanmore), London, UK. 612Division of Orthopaedic Surgery, Centre, BC Cancer Agency, Vancouver, British Columbia, Canada. 693Indivumed, Hamburg, Oslo University Hospital, Oslo, Norway. 613Department of Pathology (Research), University Germany. 694Division of Hepatobiliary and Pancreatic Surgery, Department of Surgery, School College London Cancer Institute, London, UK. 614Radcliffe Department of Medicine, of Medicine, Keimyung University Dong-san Medical Center, Daegu, South Korea. 695Women’s University of Oxford, Oxford, UK. 615University of Pavia, Pavia, Italy. 616Karolinska Institute, Cancer Program at the Samuel Oschin Comprehensive Cancer Institute, Cedars-Sinai Stockholm, Sweden. 617Wellcome Sanger Institute, Hinxton, UK. 618University of Oxford, Medical Center, Los Angeles, CA, USA. 696Department of Surgery, School of Medicine and Oxford, UK. 619Salford Royal NHS Foundation Trust, Salford, UK. 620Gloucester Royal Hospital, Health Science, The George Washington University, Washington, DC, USA. 697Endocrine Gloucester, UK. 621Royal Stoke University Hospital, Stoke-on-Trent, UK. 622St Thomas’s Oncology Branch, Center for Cancer Research, National Cancer Institute, National Institutes Hospital, London, UK. 623Imperial College NHS Trust, Imperial College London, London, UK. of Health, Bethesda, MD, USA. 698National Cancer Center, Gyeonggi, South Korea. 699ILSbio, 624 Department of Histopathology, Salford Royal NHS Foundation Trust, Salford, UK. 625Faculty LLC Biobank, Chestertown, MD, USA. 700Gynecologic Oncology, NYU Laura and Isaac of Biology, Medicine and Health, The University of Manchester, Manchester, UK. 626Edinburgh Perlmutter Cancer Center, New York University, New York, NY, USA. 701Division of Oncology, Royal Infirmary, Edinburgh, UK. 627Barking Havering and Redbridge University Hospitals NHS Stem Cell Biology Section, Washington University School of Medicine, St Louis, MO, USA. Trust, Romford, UK. 628King’s College London and Guy’s and St Thomas’ NHS Foundation 702 Urologic Oncology Branch, Center for Cancer Research, National Cancer Institute, Trust, London, UK. 629Cambridge Oesophagogastric Centre, Cambridge University Hospitals National Institutes of Health, Bethesda, MD, USA. 703Institute for Systems Biology, Seattle, NHS Foundation Trust, Cambridge, UK. 630Nottingham University Hospitals NHS Trust, WA, USA. 704Center for Personalized Medicine, Department of Pathology and Laboratory Nottingham, UK. 631St Luke’s Cancer Centre, Royal Surrey County Hospital NHS Foundation Medicine, Children’s Hospital Los Angeles, Los Angeles, CA, USA. 705Institute for Genomic Trust, Guildford, UK. 632University of North Carolina at Chapel Hill, Chapel Hill, NC, USA. Medicine, Nationwide Children’s Hospital, Columbus, OH, USA. 706Department of Surgery, 633 Norfolk and Norwich University Hospital NHS Trust, Norwich, UK. 634University Hospitals Duke University, Durham, NC, USA. 707Department of Obstetrics, Gynecology and Coventry and Warwickshire NHS Trust, Coventry, UK. 635University Hospitals Birmingham NHS Reproductive Services, University of California San Francisco, San Francisco, CA, USA. Foundation Trust, Birmingham, UK. 636Centre for Cancer Research and Cell Biology, Queen’s 708 Department of Neurology and Department of Neurosurgery, Henry Ford Hospital, Detroit, University, Belfast, UK. 637School of Cancer Sciences, Faculty of Medicine, University of MI, USA. 709Knight Cancer Institute, Oregon Health & Science University, Portland, OR, USA. Southampton, Southampton, UK. 638Wythenshawe Hospital, Manchester, UK. 639Barts Cancer 710 Department of Pathology, Roswell Park Cancer Institute, Buffalo, NY, USA. 711Department of Institute, Barts and the London School of Medicine and Dentistry, Queen Mary University of Obstetrics and Gynecology, Division of Gynecologic Oncology, Washington University London, London, UK. 640Royal Marsden NHS Foundation Trust, London and Sutton, London, School of Medicine, St Louis, MO, USA. 712Department of Palliative, Rehabilitation and UK. 641University Hospital Southampton NHS Foundation Trust, Southampton, UK. 642HCA Integrative Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX, USA. Laboratories, London, UK. 643University of Liverpool, Liverpool, UK. 644Academic Urology 713 Penrose St Francis Health Services, Colorado Springs, CO, USA. 714The University of Group, Department of Surgery, University of Cambridge, Cambridge, UK. 645University of Chicago, Chicago, IL, USA. 715Department of Neurology, Mayo Clinic, Rochester, MN, USA. Oxford, Oxford, Oxford, UK. 646Department of Urology, James Buchanan Brady Urological 716 Center for Liver Cancer, Research Institute and Hospital, National Cancer Center, Institute, Johns Hopkins University School of Medicine, Baltimore, MD, USA. 647Second Gyeonggi, South Korea. 717Department of Genetics and Lineberger Comprehensive Cancer Military Medical University, Shanghai, China. 648Department of Surgery and Cancer, Imperial Center, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA. 718NYU Langone College London, London, UK. 649The Chinese University of Hong Kong, Shatin, Hong Kong, Medical Center, New York, NY, USA. 719Department of Hematology and Medical Oncology, China. 650Nuffield Department of Surgical Sciences, John Radcliffe Hospital, University of Cleveland Clinic, Cleveland, OH, USA. 720Department of Genetics, Department of Pathology Oxford, Headington, Oxford, UK. 651Department of Histopathology, Cambridge University and Laboratory Medicine, School of Medicine, University of North Carolina at Chapel Hill, Hospitals NHS Foundation Trust, Cambridge, UK. 652Department of Bioinformatics and Chapel Hill, NC, USA. 721Helen F. Graham Cancer Center at Christiana Care Health Systems, Computational Biology and Department of Systems Biology, The University of Texas MD Newark, DE, USA. 722Cureline, South San Francisco, CA, USA. 723Department of Obstetrics and Anderson Cancer Center, Houston, TX, USA. 653Laboratory of Pathology, Center for Cancer Gynecology, Medical College of Wisconsin, Milwaukee, WI, USA. 724Hematology and Medical Research, National Cancer Institute, Bethesda, MD, USA. 654Canada’s Michael Smith Genome Oncology, Winship Cancer Institute of Emory University, Atlanta, GA, USA. 725Vanderbilt Sciences Center, BC Cancer Agency, Vancouver, British Columbia, Canada. 655Center for Ingram Cancer Center, Vanderbilt University, Nashville, TN, USA. 726Ohio State University Molecular Oncology, Memorial Sloan Kettering Cancer Center, New York, NY, USA. College of Medicine and Arthur G. James Comprehensive Cancer Center, Columbus, OH, 656 University Health Network, Toronto, Ontario, Canada. 657Department of Pathology and USA. 727Research Computing Center, University of North Carolina at Chapel Hill, Chapel Hill, Laboratory Medicine, School of Medicine, University of North Carolina at Chapel Hill, Chapel NC, USA. 728Analytical Biological Services, Wilmington, DE, USA. 729Department of Hill, NC, USA. 658Department of Population and Quantitative Health Sciences, Case Western Dermatology, University Hospital Essen, Westdeutsches Tumorzentrum and German Cancer Reserve University School of Medicine, Cleveland, OH, USA. 659Research Health Analytics Consortium, Essen, Germany. 730University of Pittsburgh, Pittsburgh, PA, USA. 731Murtha and Informatics, University Hospitals Cleveland Medical Center, Cleveland, OH, USA. 660Arnie Cancer Center, Walter Reed National Military Medical Center, Bethesda, MD, USA. 732Brigham Charbonneau Cancer Institute, University of Calgary, Calgary, Alberta, Canada. and Women’s Hospital, Harvard Medical School, Boston, MA, USA. 733Department of Surgery, 661 Department of Surgery and Department of Oncology, University of Calgary, Calgary, Memorial Sloan Kettering Cancer Center, New York, NY, USA. 734Department of Gynecologic Alberta, Canada. 662Buck Institute for Research on Aging, Novato, CA, USA. 663Duke University Oncology and Reproductive Medicine, and Center for RNA Interference and Non-Coding Medical Center, Durham, NC, USA. 664USC Norris Comprehensive Cancer Center, University RNA, The University of Texas MD Anderson Cancer Center, Houston, TX, USA. 735Department of Southern California, Los Angeles, CA, USA. 665The Preston Robert Tisch Brain Tumor of Urology, Mayo Clinic, Rochester, MN, USA. 736Department of Surgery, Johns Hopkins Center, Duke University Medical Center, Durham, NC, USA. 666Department of Dermatology University School of Medicine, Baltimore, MD, USA. 737Department of Neurosurgery, and Department of Pathology, Yale University, New Haven, CT, USA. 667Fox Chase Cancer Department of Hematology and Department of Medical Oncology, Winship Cancer Institute Center, Philadelphia, PA, USA. 668Department of Surgery, Division of Thoracic Surgery, The and School of Medicine, Emory University, Atlanta, GA, USA. 738Georgia Regents University Johns Hopkins University School of Medicine, Baltimore, MD, USA. 669University of Michigan Cancer Center, Augusta, GA, USA. 739Thoracic Oncology Laboratory, Mayo Clinic, Rochester, Comprehensive Cancer Center, Ann Arbor, MI, USA. 670University of Alabama at Birmingham, MN, USA. 740Institute for Genomic Medicine, Nationwide Children’s Hospital, Columbus, OH, Birmingham, AL, USA. 671Division of Anatomic Pathology, Mayo Clinic, Rochester, MN, USA. USA. 741Department of Obstetrics & Gynecology, Division of Gynecologic Oncology, Mayo 672 Division of Experimental Pathology, Mayo Clinic, Rochester, MN, USA. 673Department of Clinic, Rochester, MN, USA. 742International Institute for Molecular Oncology, Poznań, Poland. Oncology, The Johns Hopkins School of Medicine, The Sidney Kimmel Comprehensive 743 Poznan University of Medical Sciences, Poznań, Poland. 744Edison Family Center for Cancer Center at Johns Hopkins University, Baltimore, MD, USA. 674International Genomics Genome Sciences and Systems Biology, Washington University, St Louis, MO, USA. 745These Consortium, Phoenix, AZ, USA. 675Department of Pediatrics and Department of Genetics, authors jointly supervised this work: Peter J. Campbell, Gad Getz, Jan O. Korbel, Joshua M. University of North Carolina at Chapel Hill, Chapel Hill, NC, USA. 676Department of Stuart, Lincoln D. Stein. *e-mail: pc8@sanger.ac.uk; gadgetz@broadinstitute.org; Pathology, UPMC Shadyside, Pittsburgh, PA, USA. 677Center for Cancer Genomics, National korbel@embl.de; jstuart@ucsc.edu; lincoln.stein@gmail.com Methods them—consisted of multiple software packages for calling somatic SNVs, small indels, CNAs and somatic SVs (with intrachromosomal SVs Samples defined as those >100 bp). Two additional variant algorithms102,103 were We compiled an inventory of matched tumour–normal whole-cancer included to further improve accuracy across a broad range of clonal genomes in the ICGC Data Coordinating Centre. Most samples came and subclonal mutations. We tested different merging strategies using from treatment-naive, primary cancers, although a small number of validation data, and choses the optimal method for each variant type donors had multiple samples of primary, metastatic and/or recurrent to generate a final consensus set of mutation calls (Supplementary tumours. Our inclusion criteria were: (1) matched tumour and normal Methods S2.4). specimen pair; (2) a minimal set of clinical fields; and (3) characteri- Somatic retrotransposition events, including Alu and LINE-1 inser- zation of tumour and normal whole genomes using Illumina HiSeq tions72, L1-mediated transductions73 and pseudogene formation104, were paired-end sequencing reads. called using a dedicated pipeline73. We removed these retrotransposi- We collected genome data from 2,834 donors, representing all ICGC tion events from the somatic SV call-set. Mitochondrial DNA mutations and TCGA donors that met these criteria at the time of the final data were called using a published algorithm105. RNA-sequencing data were freeze in autumn 2014 (Extended Data Table 1). After quality assurance uniformly processed to quantify normalized gene-level expression, (Supplementary Methods 2.5), data from 176 donors were excluded splicing variation and allele-specific expression, and to identify fusion as unusable, 75 had minor issues that could affect some analyses transcripts, alternative promoter usage and sites of RNA editing8. (grey-listed donors) and 2,583 had data of optimal quality (white-listed donors) (Supplementary Table 1). Across the 2,658 white- and grey- Integration, phasing and validation of germline variant call-sets listed donors, whole-genome sequences were available from 2,605 Calls of common (≥1% frequency in PCAWG) and rare (<1%) germline primary tumours and 173 metastases or local recurrences. Matching variants including single-nucleotide polymorphisms, indels, SVs and normal samples were obtained from blood (2,064 donors), tissue mobile-element insertions (MEIs) were generated using a population- adjacent to the primary tumour (87 donors) or from distant sites (507 scale genetic polymorphism-detection approach91,106. The uniform donors). Whole-genome sequencing data were available for tumour germline data-processing workflow comprised variant identification and normal DNA for the entire cohort. The mean read coverage was using six different variant-calling algorithms96,107,108 and was orches- 39× for normal samples, whereas tumours had a bimodal coverage trated using the Butler workflow system109. distribution with modes at 38× and 60× (Supplementary Fig. 1). The We performed call-set benchmarking, merging, variant genotyp- majority of specimens (65.3%) were sequenced using 101-bp paired- ing and statistical haplotype-block phasing91 (Supplementary Meth- end reads. An additional 28% were sequenced with 100-bp paired-end ods 3.4). Using this strategy, we identified 80.1 million germline reads. Of the remaining specimens, 4.7% were sequenced with read single-nucleotide polymorphisms, 5.9 million germline indels, 1.8 mil- lengths longer than 101 bp, and 1.9% with read lengths shorter than lion multi-allelic short (<50 bp) germline variants, as well as germline 100 bp. The distribution of read lengths by tumour cohort is shown in SVs ≥ 50 bp in size including 29,492 biallelic deletions and 27,254 MEIs Supplementary Fig. 11. Median read length for whole-genome (Supplementary Table 2). We statistically phased this germline variant sequencing paired-end reads was 101 bp (mean = 106.2, s.d. = 16.7; set using haplotypes from the 1000 Genomes Project91 as a reference minimum–maximum = 50–151). RNA-sequencing data were collected panel, yielding an N50-phased block length of 265 kb based on haploid and re-analysed centrally for 1,222 donors, including 1,178 primary chromosomes from donor-matched tumour genomes. Precision esti- tumours, 67 metastases or local recurrences and 153 matched normal mates for germline SNVs and indels were >99% for the phased merged tissue samples adjacent to the primary tumour. call-set, and sensitivity estimates ranged from 92% to 98%. Demographically, the cohort included 1,469 men (55%) and 1,189 women (45%), with a mean age of 56 years (range, 1–90 years) Core alignment and variant calling by cloud computing (Supplementary Table 1). Using population ancestry-differentiated The requirement to uniformly realign and call variants on nearly 5,800 single nucleotide polymorphisms, the ancestry distribution was heavily whole genomes (tumour plus normal) presented considerable com- weighted towards donors of European descent (77% of total) followed putational challenges, and raised ethical issues owing to the use of by East Asians (16%), as expected for large contributions from European, data from different jurisdictions (Extended Data Table 2). To process North American and Australian projects (Supplementary Table 1). the data, we adopted a cloud-computing architecture26 in which the We consolidated histopathology descriptions of the tumour sam- alignment and variant calling was spread across 13 data centres on 3 ples, using the ICD-0-3 tumour site controlled vocabulary89. Overall, continents, representing a mixture of commercial, infrastructure-as- the PCAWG dataset comprises 38 distinct tumour types (Extended a-service, academic cloud compute and traditional academic high- Data Table 1 and Supplementary Table 1). Although the most common performance computer clusters (Supplementary Table 3). Together, tumour types are included in the dataset, their distribution does not the effort used 10 million CPU-core hours. match the relative population incidences, largely owing to differences To generate reproducible variant calling across the 13 data centres, among contributing ICGC/TCGA groups in the numbers of sequenced we built the core pipelines into Docker containers28, in which the work- samples. flow description, required code and all associated dependencies were packaged together in stand-alone packages. These heavily tested, exten- Uniform processing and somatic variant calling sively validated workflows are available for download (Box 1). To generate a consistent set of somatic mutation calls that could be used for cross-tumour analyses, we analysed all 6,835 samples using a Validation, benchmarking and merging of somatic variant calls uniform set of algorithms for alignment, variant calling and quality con- To evaluate the performance of each of the mutation-calling pipelines trol (Extended Data Fig. 1, Supplementary Fig. 2, Supplementary Table 3 and determine an integration strategy, we performed a large-scale and Supplementary Methods 2). We used the BWA-MEM algorithm90 deep-sequencing validation experiment (Supplementary Notes 1). We to align each tumour and normal sample to human reference build selected a pilot set of 63 representative tumour–normal pairs, on which hs37d5 (as used in the 1000 Genomes Project91). Somatic mutations we ran the 3 core pipelines, together with a set of 10 additional somatic were identified in the aligned data using three established pipelines, variant-calling pipelines contributed by members of the PCAWG SNV which were run independently on each tumour–normal pair. Each of the Calling Methods Working Group. Sufficient DNA remained for 50 of three pipelines—labelled ‘Sanger’92–95, ‘EMBL/DKFZ’96,97 and ‘Broad’98–101 the 63 cases for validation, which was performed by hybridization of after the computational biology groups that created or assembled tumour and matched normal DNA to a custom RNA bait set, followed Article by deep sequencing, as previously described29. Although performed cgi?page=login) for access to the TCGA portion of the dataset, and to using the same sequencing chemistry as the original whole-genome the ICGC Data Access Compliance Office (DACO; http://icgc.org/daco) sequencing analyses, the considerably greater depth achieved in the for the ICGC portion. In addition, to access somatic single nucleotide validation experiment enabled accurate assessment of sensitivity and variants derived from TCGA donors, researchers will also need to obtain precision of variant calls. Variant calls in repeat-masked regions were dbGaP authorization. not tested, owing to the challenge of designing reliable validation Beyond the core sequence data and variant call-sets, the analyses in probes in these areas. this paper used a number of datasets that were derived from the variant The 3 core pipelines had individual estimates of sensitivity of 80–90% calls (Supplementary Table 4). The individual datasets are available at to detect a true somatic SNV called by any of the 13 pipelines; with >95% Synapse (https://www.synapse.org/), and are denoted with synXXXXX of SNV calls made by each of the core pipelines being genuine somatic accession numbers; all these datasets are also mirrored at https://dcc. variants (Fig. 1a). For indels—a more-challenging class of variants to icgc.org, with full links, filenames, accession numbers and descriptions identify in short-read sequencing data—the 3 core algorithms had indi- detailed in Supplementary Table 4. The datasets encompass: clinical vidual sensitivity estimates in the range of 40–50%, with precision data from each patient including demographics, tumour stage and vital 70–95% (Fig. 1b). Validation of SV calls is inherently more difficult, status (syn10389158); harmonized tumour histopathology annotations as methods based on PCR or hybridization to RNA baits often fail to using a standardised hierarchical ontology (syn1038916); inferred isolate DNA that spans the breakpoint. To assess the accuracy of SV purity and ploidy values for each tumour sample (syn8272483); driver calls, we therefore used the property that an SV must either generate mutations for each patient from their cancer genome spanning all a copy-number change or be balanced, whereas artefactual calls will classes of variant, and coding versus non-coding drivers (syn11639581); not respect this property. For individual SV-calling algorithms, we mutational signatures inferred from PCAWG donors (syn11804065), estimated precision to be in the range of 80–95% for samples in the including APOBEC mutagenesis (syn7437313); and transcriptional data 63-sample pilot dataset. from RNA sequencing, including gene expression levels (syn5553985, Next, we examined multiple methods for merging calls made by syn5553991, syn8105922) and gene fusions (syn10003873, syn7221157). several algorithms into a single definitive call-set to be used for down- stream analysis. The final consensus calls for SNVs were based on a sim- ple approach that required two or more methods to agree on a call. For Code availability indels, because methods were less concordant, we used stacked logistic Computational pipelines for calling somatic mutations are available to regression110,111 to integrate the calls. The merged SV set includes all calls the public at https://dockstore.org/organizations/PCAWG/collections/ made by two or more of the four primary SV-calling algorithms96,100,112,113. PCAWG. A range of data-visualization and -exploration tools are also Consensus CNA calls were obtained by joining the outputs of six indi- available for the PCAWG data (Box 1). vidual CNA-calling algorithms with SV consensus breakpoints to obtain base-pair resolution CNAs (Supplementary Methods 2.4.3). Consensus 89. NCI SEER. ICD-O-3 Coding Materials (2018). purity and ploidy were derived, and a multitier system was developed 90. Li, H. & Durbin, R. Fast and accurate long-read alignment with Burrows–Wheeler for consensus copy-number calls (Supplementary Methods 2.4.3, and transform. Bioinformatics 26, 589–595 (2010). 91. 1000 Genomes Project Consortium. A global reference for human genetic variation. described in detail elsewhere7). Nature 526, 68–74 (2015). Overall, the sensitivity and precision of the consensus somatic vari- 92. Raine, K. M. et al. ascatNgs: identifying somatically acquired copy-number alterations ant calls were 95% (90% confidence interval, 88–98%) and 95% (90% from whole-genome sequencing data. Curr. Protoc. Bioinformatics 56, 15.9.1–15.9.17 (2016). confidence interval, 71–99%), respectively, for SNVs (Extended Data 93. Jones, D. et al. cgpCaVEManWrapper: simple execution of CaVEMan in order to detect Fig. 2). For somatic indels, sensitivity and precision were 60% (90% con- somatic single nucleotide variants in NGS data. Curr. Protoc. Bioinformatics 56, fidence interval, 34–72%) and 91% (90% confidence interval, 73–96%), 15.10.1–15.10.18 (2016). 94. Raine, K. M. et al. cgpPindel: identifying somatically acquired insertion and deletion respectively. Regarding SVs, we estimate the sensitivity of the merging events from paired end sequencing. Curr. Protoc. Bioinformatics 52, 15.7.1–15.7.12 algorithm to be 90% for true calls generated by any one calling pipeline; (2015). precision was estimated to be 97.5%. That is, 97.5% of SVs in the merged 95. Ye, K., Schulz, M. H., Long, Q., Apweiler, R. & Ning, Z. Pindel: a pattern growth approach to detect break points of large deletions and medium sized insertions from paired-end short SV call-set had an associated copy-number change or balanced partner reads. Bioinformatics 25, 2865–2871 (2009). rearrangement. The improvement in calling accuracy from combining 96. Rausch, T. et al. DELLY: structural variant discovery by integrated paired-end and split- different pipelines was most noticeable in variants that had low variant read analysis. Bioinformatics 28, i333–i339 (2012). 97. Rimmer, A. et al. Integrating mapping-, assembly- and haplotype-based approaches allele fractions, which are likely to originate from subclonal popula- for calling variants in clinical sequencing applications. Nat. Genet. 46, 912–918 (2014). tions of the tumour (Fig. 1c, d). There remains much work to be done 98. Cibulskis, K. et al. Sensitive detection of somatic point mutations in impure and to improve indel calling software; we still lack sensitivity for calling heterogeneous cancer samples. Nat. Biotechnol. 31, 213–219 (2013). 99. Carter, S. L. et al. Absolute quantification of somatic DNA alterations in human cancer. even fully clonal complex indels from short-read sequencing data. Nat. Biotechnol. 30, 413–421 (2012). 100. Drier, Y. et al. Somatic rearrangements across cancer reveal classes of samples with Reporting summary distinct patterns of DNA breakage and rearrangement-induced hypermutability. Genome Res. 23, 228–235 (2013). Further information on research design is available in the Nature 101. Ramos, A. H. et al. Oncotator: cancer variant annotation tool. Hum. Mutat. 36, E2423– Research Reporting Summary linked to this paper. E2429 (2015). 102. Moncunill, V. et al. Comprehensive characterization of complex structural variations in cancer by directly comparing genome sequence reads. Nat. Biotechnol. 32, 1106–1112 (2014). Data availability 103. Fan, Y. et al. MuSE: accounting for tumor heterogeneity using a sample-specific error The PCAWG-generated alignments, somatic variant calls, annotations model improves sensitivity and specificity in mutation calling from sequencing data. Genome Biol. 17, 178 (2016). and derived datasets are available for general research use for browsing 104. Cooke, S. L. et al. Processed pseudogenes acquired somatically during cancer and download at http://dcc.icgc.org/pcawg/ (Box 1 and Supplementary development. Nat. Commun. 5, 3644 (2014). Table 4). In accordance with the data access policies of the ICGC and 105. Ju, Y. S. et al. Origins and functional consequences of somatic mitochondrial DNA mutations in human cancer. eLife 3, e02935 (2014). TCGA projects, most molecular, clinical and specimen data are in an 106. Sudmant, P. H. et al. An integrated map of structural variation in 2,504 human genomes. open tier which does not require access approval. To access poten- Nature 526, 75–81 (2015). 107. Garrison, E. & Marth, G. Haplotype-based variant detection from short-read sequencing. tially identifying information, such as germline alleles and underly- Preprint at https://arxiv.org/abs/1207.3907 (2012). ing read data, researchers will need to apply to the TCGA Data Access 108. DePristo, M. A. et al. A framework for variation discovery and genotyping using next- Committee (DAC) via dbGaP (https://dbgap.ncbi.nlm.nih.gov/aa/wga. generation DNA sequencing data. Nat. Genet. 43, 491–498 (2011). 109. Yakneen, S., Waszak, S. M., Gertz, M. & Korbel, J. O. & PCAWG Consortium. Butler enables Yogesh Kumar, Christopher Lalansingh, Ignaty Leshchiner, Ivica Letunic, Dimitri Livitz, Eric Z. rapid cloud-based analysis of thousands of human genomes. Nat. Biotechnol. https://doi. Ma, Yosef E. Maruvka, R. Jay Mashl, Michael D. McLellan, Ana Milovanovic, Morten Muhlig org/10.1038/s41587-019-0360-3 (2020). Nielsen, Brian D. O’Connor, Stephan Ossowski, Nagarajan Paramasivam, Jakob Skou Pedersen, 110. Kim, S. Y., Jacob, L. & Speed, T. P. Combining calls from multiple somatic mutation- Marc D. Perry, Montserrat Puiggros, Romina Royo, Esther Rheinbay, S. Cenk Sahinalp, Iman callers. BMC Bioinformatics 15, 154 (2014). Sarrafi, Chip Stewart, Miranda D. Stobbe, Grace Tiao, Jeremiah A. Wala, Jiayin Wang, Wenyi 111. Breiman, L. Stacked regressions. Mach. Learn. 24, 49–64 (1996). Wang, Sebastian M. Waszak, Joachim Weischenfeldt, Michael Wendl, Johannes Werner, 112. Campbell, P. J. et al. Identification of somatically acquired rearrangements in cancer Zhenggang Wu, Hong Xue, Sergei Yakneen, Takafumi N. Yamaguchi, Kai Ye, Venkata using genome-wide massively parallel paired-end sequencing. Nat. Genet. 40, 722–729 Yellapantula, Junjun Zhang, David A. Wheeler; led by Li Ding, Jared T. Simpson. Processing of (2008). validation data: major contributions from Christina K. Yung, Brian D. O’Connor, Sergei Yakneen, 113. Wala, J. A. et al. SvABA: genome-wide detection of structural variants and indels by local Junjun Zhang; further contributions from Kyle Ellrott, Kortine Kleinheinz, Naoki Miyoshi, Keiran assembly. Genome Res. 28, 581–591 (2018). M. Raine, Romina Royo, Gordon Saksena, Matthias Schlesner, Solomon I. Shorser, Miguel Vazquez, Joachim Weischenfeldt, Denis Yuen, Adam P. Butler, Brandi N. Davis-Dusenbery, Roland Eils, Vincent Ferretti, Robert L. Grossman, Olivier Harismendy, Youngwook Kim, Acknowledgements We thank research participants who donated samples and data, the Hidewaki Nakagawa, Steven J. Newhouse, David Torrents; led by Lincoln D. Stein. Whole- physicians and clinical staff who contributed to sample annotation and collection, and the genome sequencing somatic variant calling: major contribution from Junjun Zhang; further numerous funding agencies that contributed to the collection and analysis of this dataset. contributions from Christina K. Yung, Solomon I. Shorser. Whole-genome alignment: Keiran M. Raine, Junjun Zhang, Brian D. O’Connor. DKFZ pipeline: Kortine Kleinheinz, Tobias Rausch, Jan Author contributions Writing committee leads: Peter J. Campbell, Gad Getz, Jan O. Korbel, O. Korbel, Ivo Buchhalter, Michael C. Heinold, Barbara Hutter, Natalie Jager, Nagarajan Joshua M. Stuart, Jennifer L. Jennings, Lincoln D. Stein. Head of project management: Jennifer Paramasivam, Matthias Schlesner. EMBL pipeline: Joachim Weischenfeldt. Sanger pipeline: L. Jennings. Sample collection: major contributions from Marc D. Perry, Hardeep K. Nahal- Keiran M. Raine, Jonathan Hinton, David R. Jones, Andrew Menzies, Lucy Stebbings, Adam P. Bose; led by B. F. Francis Ouellette. Histopathology harmonization: major contribution from Butler. Broad pipeline: Gordon Saksena, Dimitri Livitz, Esther Rheinbay, Julian M. Hess, Ignaty Constance H. Li; further contributions from Esther Rheinbay, G. Petur Nielsen, Dennis C. Sgroi, Leshchiner, Chip Stewart, Grace Tiao, Jeremiah A. Wala, Amaro Taylor-Weiner, Mara Chin-Lee Wu, William C. Faquin, Vikram Deshpande, Paul C. Boutros, Alexander J. Lazar, Rosenberg, Andrew J. Dunford, Manaswi Gupta, Marcin Imielinski, Matthew Meyerson, Rameen Katherine A. Hoadley; led by Lincoln D. Stein, David N. Louis. Uniform processing, somatic, Beroukhim, Gad Getz. MuSE Pipeline: Yu Fan, Wenyi Wang. Consensus somatic SNV/indel germline variant calling: major contribution from L. Jonathan Dursi; further contributions from annotation: Andrew Menzies, Matthias Schlesner, Juri Reimand, Priyanka Dhingra, Ekta Christina K. Yung, Matthew H. Bailey, Gordon Saksena, Keiran M. Raine, Ivo Buchhalter, Kortine Khurana. Somatic SNV, indel merging: major contribution from L. Jonathan Dursi; further Kleinheinz, Matthias Schlesner, Junjun Zhang, Wenyi Wang, David A. Wheeler; led by Li Ding, contributions from Christina K. Yung, Matthew H. Bailey, Gordon Saksena, Keiran M. Raine, Ivo Jared T. Simpson. Core alignment, variant calling by cloud computing: major contributions Buchhalter, Kortine Kleinheinz, Matthias Schlesner, Yu Fan, David Torrents, Matthias Bieg, Paul from Christina K. Yung, Brian D. O’Connor, Sergei Yakneen, Junjun Zhang; further contributions C. Boutros, Ken Chen, Zechen Chong, Kristian Cibulskis, Oliver Drechsel, Roland Eils, Robert S. from Kyle Ellrott, Kortine Kleinheinz, Naoki Miyoshi, Keiran M. Raine, Adam P. Butler, Romina Fulton, Josep L. Gelpi, Mark Gerstein, Santiago Gonzalez, Gad Getz, Ivo G. Gut, Faraz Hach, Royo, Gordon Saksena, Matthias Schlesner, Solomon I. Shorser, Miguel Vazquez. Integration, Michael Heinold, Taobo Hu, Vincent Huang, Barbara Hutter, Hyung-Lae Kim, Natalie Jager, phasing, validation of germline variant callsets: major contributions from Tobias Rausch, Jongsun Jung, Sushant Kumar, Yogesh Kumar, Christopher Lalansingh, Ignaty Leshchiner, Ivica Grace Tiao, Sebastian M. Waszak, Bernardo Rodriguez-Martin, Suyash Shringarpure, Dai-Ying Letunic, Dimitri Livitz, Eric Z. Ma, Yosef E. Maruvka, R. Jay Mashl, Michael D. McLellan, Ana Wu; further contributions from Sergei Yakneen, German M. Demidov, Olivier Delaneau, Shuto Milovanovic, Morten Muhlig Nielsen, Brian D. O’Connor, Stephan Ossowski, Nagarajan Hayashi, Seiya Imoto, Nina Habermann, Ayellet V. Segre, Erik Garrison, Andy Cafferkey, Eva G. Paramasivam, Jakob Skou Pedersen, Marc D. Perry, Montserrat Puiggros, Romina Royo, Esther Alvarez, José María Heredia-Genestar, Francesc Muyas, Oliver Drechsel, Alicia L. Bruzos, Javier Rheinbay, S. Cenk Sahinalp, Iman Sarrafi, Chip Stewart, Miranda D. Stobbe, Grace Tiao, Temes, Jorge Zamora, L. Jonathan Dursi, Adrian Baez-Ortega, Hyung-Lae Kim, Matthew H. Jeremiah A. Wala, Jiayin Wang, Wenyi Wang, Sebastian M. Waszak, Joachim Weischenfeldt, Bailey, R. Jay Mashl, Kai Ye, Ivo Buchhalter, Anthony DiBiase, Kuan-lin Huang, Ivica Letunic, Michael Wendl, Johannes Werner, Zhenggang Wu, Hong Xue, Sergei Yakneen, Takafumi N. Michael D. McLellan, Steven J. Newhouse, Matthias Schlesner, Tal Shmaya, Sushant Kumar, Yamaguchi, Kai Ye, Venkata Yellapantula, Junjun Zhang, David A. Wheeler; major contributions David C. Wedge, Mark H. Wright, Venkata D. Yellapantula, Mark Gerstein, Ekta Khurana, Tomas from Li Ding, Jared T. Simpson. Somatic SV merging: Joachim Weischenfeldt, Francesco Marques-Bonet, Arcadi Navarro, Carlos D. Bustamante, Jared T. Simpson, Li Ding, Reiner Favero, Yilong Li. Somatic CNA merging: Stefan Dentro, Jeff Wintersinger, Ignaty Leshchiner. Siebert, Hidewaki Nakagawa, Douglas F. Easton; led by Stephan Ossowski, Jose M. C. Tubio, Oxidative artefact filtration: Dimitri Livitz, Ignaty Leshchiner, Chip Stewart, Esther Rheinbay, Gad Getz, Francisco M. De La Vega, Xavier Estivill, Jan O. Korbel. Validation, benchmarking, Gordon Saksena, Gad Getz. Strand bias filtration: Matthias Bieg, Ivo Buchhalter, Johannes merging of somatic variant calls: major contribution from L. Jonathan Dursi; further Werner, Matthias Schlesner. miniBAM generation: Jeremiah A. Wala, Gordon Saksena, Rameen contributions from David A. Wheeler, Christina K. Yung; led by Li Ding, Jared T. Simpson. Data Beroukhim, Gad Getz. Germline variant identification from whole-genome sequencing: major and code availability: major contribution from Junjun Zhang; further contributions from contributions from Tobias Rausch, Grace Tiao, Sebastian M. Waszak, Bernardo Rodriguez- Christina K. Yung, Sergei Yakneen, Denis Yuen, George L. Mihaiescu, Larsson Omberg; led by Martin, Suyash Shringarpure, Dai-Ying Wu; further contributions from Sergei Yakneen, German Vincent Ferretti. Pan-cancer burden of somatic mutations: major contribution from Junjun M. Demidov, Olivier Delaneau, Shuto Hayashi, Seiya Imoto, Nina Habermann, Ayellet V. Segre, Zhang; led by Peter J. Campbell. Panorama of driver mutations in human cancer: led by Erik Garrison, Andy Cafferkey, Eva G. Alvarez, Alicia L. Bruzos, Jorge Zamora, José María Radhakrishnan Sabarinathan, Oriol Pich, Abel Gonzalez-Perez. PCAWG tumours with no Heredia-Genestar, Francesc Muyas, Oliver Drechsel, L. Jonathan Dursi, Adrian Baez-Ortega, apparent driver mutations: major contribution from Esther Rheinbay; further contributions Hyung-Lae Kim, Matthew H. Bailey, R. Jay Mashl, Kai Ye, Ivo Buchhalter, Vasilisa Rudneva, Ji from Amaro Taylor-Weiner, Radhakrishnan Sabarinathan; led by Peter J. Campbell, Gad Getz. Wan Park, Eun Pyo Hong, Seong Gu Heo, Anthony DiBiase, Kuan-lin Huang, Ivica Letunic, Patterns, oncogenicity of kataegis, chromoplexy: major contributions from Matthew W. Fittall, Michael D. McLellan, Steven J. Newhouse, Matthias Schlesner, Tal Shmaya, Sushant Kumar, Jonas Demeulemeester, Maxime Tarabichi; further contributions from Nicola D. Roberts, Peter David C. Wedge, Mark H. Wright, Venkata D. Yellapantula, Mark Gerstein, Ekta Khurana, Tomas J. Campbell, Jan O. Korbel; led by Peter Van Loo. Patterns, oncogenicity of chromothripsis: Marques-Bonet, Arcadi Navarro, Carlos D. Bustamante, Jared T. Simpson, Li Ding, Reiner major contributions from Maxime Tarabichi, Jonas Demeulemeester, Matthew W. Fittall; further Siebert, Hidewaki Nakagawa, Douglas F. Easton; led by Stephan Ossowski, Jose M. C. Tubio, contributions from Isidro Cortes-Ciriano, Lara Urban, Peter J. Park, Peter J. Campbell, Jan O. Gad Getz, Francisco M. De La Vega, Xavier Estivill, Jan O. Korbel. RNA-sequencing analysis: Korbel; led by Peter Van Loo. Timing-clustered mutational processes during tumour evolution: major contributions from Nuno A. Fonseca, Andre Kahles, Kjong-Van Lehmann, Lara Urban, major contributions from Jonas Demeulemeester, Maxime Tarabichi, Matthew W. Fittall; further Cameron M. Soulette, Yuichi Shiraishi, Fenglin Liu, Yao He, Deniz Demircioglu, Natalie R. contributions from Jan O. Korbel, Peter J. Campbell; led by Peter Van Loo. Germline effects on Davidson, Claudia Calabrese, Junjun Zhang, Marc D. Perry, Qian Xiang; further contributions somatic mutation: major contributions from Sebastian M. Waszak, Bin Zhu, Bernardo from Liliana Greger, Siliang Li, Dongbing Liu, Stefan G. Stark, Fan Zhang, Samirkumar B. Amin, Rodriguez-Martin, Esa Pitkanen, Tobias Rausch; further contributions from Yilong Li, Natalie Peter Bailey, Aurelien Chateigner, Isidro Cortes-Ciriano, Brian Craft, Serap Erkek, Milana Saini, Leszek J. Klimczak, Joachim Weischenfeldt, Nikos Sidiropoulos, Ludmil B. Alexandrov, Frenkel-Morgenstern, Mary Goldman, Katherine A. Hoadley, Yong Hou, Matthew R. Huska, Ekta Francesc Muyas, Raquel Rabionet, Georgia Escaramis, Adrian Baez-Ortega, Mattia Bosio, Khurana, Helena Kilpinen, Jan O. Korbel, Fabien C. Lamaze, Chang Li, Xiaobo Li, Xinyue Li, Aliaksei Z. Holik, Hana Susak, Eva G. Alvarez, Alicia L. Bruzos, Javier Temes, Aparna Prasad, Xingmin Liu, Maximillian G. Marin, Julia Markowski, Tannistha Nandi, Morten Muhlig Nielsen, Nina Habermann, Serap Erkek, Lara Urban, Claudia Calabrese, Benjamin Raeder, Eoghan Akinyemi I. Ojesina, Qiang Pan-Hammarstrom, Peter J. Park, Chandra Sekhar Pedamallu, Jakob Harrington, Simon Mayes, Daniel Turner, Sissel Juul, Steven A. Roberts, Lei Song, Roelof Koster, Skou Pedersen, Reiner Siebert, Hong Su, Patrick Tan, Bin Tean Teh, Jian Wang, Sebastian M. Lisa Mirabello, Xing Hua, Tomas J. Tanskanen, Marta Tojo, David C. Wedge, Jorge Zamora, Waszak, Heng Xiong, Sergei Yakneen, Chen Ye, Christina Yung, Xiuqing Zhang, Liangtao Jieming Chen, Lauri A. Aaltonen, Gunnar Ratsch, Roland F. Schwarz, Atul J. Butte, Alvis Brazma, Zheng, Jingchun Zhu, Shida Zhu, Philip Awadalla, Chad J. Creighton, Matthew Meyerson, B. F. Peter J. Campbell, Stephen J. Chanock, Nilanjan Chatterjee, Oliver Stegle, Olivier Harismendy; Francis Ouellette, Kui Wu, Huanming Yang; led by Jonathan Goke, Roland F. Schwarz, Oliver led by G. Steven Bova, Dmitry A. Gordenin, Jose M. C. Tubio, Douglas F. Easton, Xavier Estivill, Stegle, Zemin Zhang, Alvis Brazma, Gunnar Ratsch, Angela N. Brooks. Clustering of tumour Jan O. Korbel. Replicative immortality: major contribution from David Haan; further genomes based on telomere maintenance-related features: major contribution from David contributions from Lina Sieverling, Lars Feuerbach; led by Lincoln D. Stein, Joshua M. Stuart. Haan; led by Lincoln D. Stein, Joshua M. Stuart. Clustered mutational processes in PCAWG: Ethical considerations of genomic cloud computing: led by Don Chalmers, Yann Joly, Bartha major contributions from Jonas Demeulemeester, Maxime Tarabichi, Matthew W. Fittall; led by Knoppers, Fruzsina Molnar-Gabor, Jan O. Korbel, Mark Phillips, Adrian Thorogood, David Peter J. Campbell, Jan O. Korbel, Peter Van Loo. Tumours without detected driver mutations: Townend. Online resources for data access, visualization, exploration and analysis: major Esther Rheinbay, Amaro Taylor-Weiner, Radhakrishnan Sabarinathan, Peter J. Campbell, Gad contributions from Mary Goldman, Junjun Zhang, Nuno A. Fonseca; further contributions from Getz. Panorama of driver mutations in human cancer: major contributions from Radhakrishnan Qian Xiang, Brian Craft, Elena Pineiro-Yanez, Alfonso Munoz, Robert Petryszak, Anja Fullgrabe, Sabarinathan, Oriol Pich; further contributions from Inigo Martincorena, Carlota Rubio-Perez, Fatima Al-Shahrour, Maria Keays, David Haussler, John Weinstein, Wolfgang Huber, Alfonso Malene Juul, Jeremiah A. Wala, Steven Schumacher, Ofer Shapira, Nikos Sidiropoulos, Valencia, Irene Papatheodorou, Jingchun Zhu; led by Brian D. O’Connor, Lincoln D. Stein, Alvis Sebastian M. Waszak, David Tamborero, Loris Mularoni, Esther Rheinbay, Henrik Hornshoj, Jordi Brazma, Vincent Ferretti, Miguel Vazquez. The 63-sample pilot-analysis validation process: Deu-Pons, Ferran Muinos, Johanna Bertl, Qianyun Guo, Chad J. Creighton, Joachim major contribution from L. Jonathan Dursi; further contributions from Christina K. Yung, Weischenfeldt, Jan O. Korbel, Gad Getz, Peter J. Campbell, Jakob Skou Pedersen, Rameen Matthew H. Bailey, Gordon Saksena, Keiran M. Raine, Ivo Buchhalter, Kortine Kleinheinz, Beroukhim; led by Abel Gonzalez-Perez. Pilot benchmarking, variant consensus development Matthias Schlesner, Yu Fan, David Torrents, Matthias Bieg, Paul C. Boutros, Ken Chen, Zechen and validation: major contribution from L. Jonathan Dursi; further contributions from Christina Chong, Kristian Cibulskis, Oliver Drechsel, Roland Eils, Robert S. Fulton, Josep Gelpi, Mark K. Yung, Matthew H. Bailey, Gordon Saksena, Keiran M. Raine, Ivo Buchhalter, Kortine Gerstein, Santiago Gonzalez, Gad Getz, Ivo G. Gut, Faraz Hach, Michael Heinold, Taobo Hu, Kleinheinz, Matthias Schlesner, Yu Fan, David Torrents, Matthias Bieg, Paul C. Boutros, Ken Vincent Huang, Barbara Hutter, Hyung-Lae Kim, Natalie Jager, Jongsun Jung, Sushant Kumar, Chen, Zechen Chong, Kristian Cibulskis, Oliver Drechsel, Roland Eils, Robert S. Fulton, Josep Article Gelpi, Mark Gerstein, Santiago Gonzalez, Gad Getz, Ivo G. Gut, Faraz Hach, Michael Heinold, number of patent applications that encompass the use of mutational signatures. Eoghan Taobo Hu, Vincent Huang, Barbara Hutter, Hyung-Lae Kim, Natalie Jager, Jongsun Jung, Harrington is a full-time employee of Oxford Nanopore Technologies and is a stock holder. Sushant Kumar, Yogesh Kumar, Christopher Lalansingh, Ignaty Leshchiner, Ivica Letunic, Yann Joly is responsible for the Data Access Compliance Office (DACO) of ICGC 2009-2018. Dimitri Livitz, Eric Z. Ma, Yosef E. Maruvka, R. Jay Mashl, Michael D. McLellan, Ana Milovanovic, Sissel Juul is a full-time employee of Oxford Nanopore Technologies and is a stock holder. Morten Muhlig Nielsen, Brian D. O’Connor, Stephan Ossowski, Nagarajan Paramasivam, Jakob Vincent Khoo has received personal fees and non-financial support from Accuray, Astellas, Skou Pedersen, Marc D. Perry, Montserrat Puiggros, Romina Royo, Esther Rheinbay, S. Cenk Bayer, Boston Scientific and Janssen. Stian Knappskog is a coprincipal investigator on a clinical Sahinalp, Iman Sarrafi, Chip Stewart, Miranda D. Stobbe, Grace Tiao, Jeremiah A. Wala, Jiayin trial that receives research funding from AstraZeneca and Pfizer. Ignaty Leshchiner is a Wang, Wenyi Wang, Sebastian M. Waszak, Joachim Weischenfeldt, Michael Wendl, Johannes consultant for PACT Pharma. Carlos López-Otín has ownership interest (including stock and Werner, Zhenggang Wu, Hong Xue, Sergei Yakneen, Takafumi N. Yamaguchi, Kai Ye, Venkata patents) in DREAMgenics. Matthew Meyerson is a scientific advisory board chair of, and Yellapantula, Junjun Zhang, David A. Wheeler; led by Li Ding, Jared T. Simpson. Production consultant for, OrigiMed, has obtained research funding from Bayer and Ono Pharma and somatic variant calling on the PCAWG compute cloud: major contributions from Christina K. receives patent royalties from LabCorp. Serena Nik-Zainal is an inventor on a number of patent Yung, Brian D. O’Connor, Sergei Yakneen, Junjun Zhang; further contributions from Kyle Ellrott, applications that encompass the use of mutational signatures. Nathan Pennell has done Kortine Kleinheinz, Naoki Miyoshi, Keiran M. Raine, Romina Royo, Gordon Saksena, Matthias consulting work with Merck, Astrazeneca, Eli Lilly and Bristol-Myers Squibb. Xose S. Puente has Schlesner, Solomon I. Shorser, Miguel Vazquez, Joachim Weischenfeldt, Denis Yuen, Adam P. ownership interest (including stock and patents in DREAMgenics. Benjamin J. Raphael is a Butler, Brandi N. Davis-Dusenbery, Roland Eils, Vincent Ferretti, Robert L. Grossman, Olivier consultant for and has ownership interest (including stock and patents) in Medley Genomics. Harismendy, Youngwook Kim, Hidewaki Nakagawa, Steven J Newhouse, David Torrents; led by Jorge Reis-Filho is a consultant for Goldman Sachs and REPARE Therapeutics, member of the Lincoln D. Stein. PCAWG data portals: major contributions from Mary Goldman, Junjun Zhang, scientific advisory board of Volition RX and Paige.AI and an ad hoc member of the scientific Nuno A. Fonseca, Isidro Cortes-Ciriano; further contributions from Qian Xiang, Brian Craft, advisory board of Ventana Medical Systems, Roche Tissue Diagnostics, InVicro, Roche, Elena Pineiro-Yanez, Brian D O’Connor, Wojciech Bazant, Elisabet Barrera, Alfonso Munoz, Genentech and Novartis. Lewis R. Roberts has received grant support from ARIAD Robert Petryszak, Anja Fullgrabe, Fatima Al-Shahrour, Maria Keays, David Haussler, John Pharmaceuticals, Bayer, BTG International, Exact Sciences, Gilead Sciences, Glycotest, Weinstein, Wolfgang Huber, Alfonso Valencia, Irene Papatheodorou, Jingchun Zhu; led by RedHill Biopharma, Target PharmaSolutions and Wako Diagnostics and has provided Vincent Ferretti, Miguel Vazquez. advisory services to Bayer, Exact Sciences, Gilead Sciences, GRAIL, QED Therapeutics and TAVEC Pharmaceuticals. Richard A. Scolyer has received fees for professional services from Competing interests Gad Getz receives research funds from IBM and Pharmacyclics and is an Merck Sharp & Dohme, GlaxoSmithKline Australia, Bristol-Myers Squibb, Dermpedia, inventor on patent applications related to MuTect, ABSOLUTE, MutSig, MSMuTect, MSMutSig Novartis Pharmaceuticals Australia, Myriad, NeraCare GmbH and Amgen. Tal Shmaya is and POLYSOLVER. Hikmat Al-Ahmadie is consultant for AstraZeneca and Bristol-Myers Squibb. employed at Annai Systems. Reiner Siebert has received speaker honoraria from Roche and Samuel Aparicio is a founder and shareholder of Contextual Genomics. Pratiti Bandopadhayay AstraZeneca. Sabina Signoretti is a consultant for Bristol-Myers Squibb, AstraZeneca, receives grant funding from Novartis for an unrelated project. Rameen Beroukhim owns equity Merck, AACR and NCI and has received funding from Bristol-Myers Squibb, AstraZeneca, in Ampressa Therapeutics. Andrew Biankin receives grant funding from Celgene, AstraZeneca Exelixis and royalties from Biogenex. Jared Simpson has received research funding and and is a consultant for or on advisory boards of AstraZeneca, Celgene, Elstar Therapeutics, travel support from Oxford Nanopore Technologies. Anil K. Sood is a consultant for Merck Clovis Oncology and Roche. Ewan Birney is a consultant for Oxford Nanopore, Dovetail and and Kiyatec, has received research funding from M-Trap and is a shareholder in BioPath. GSK. Marcus Bosenberg is a consultant for Eli Lilly. Atul Butte is a cofounder of and consultant Simon Tavaré is on the scientific advisory board of Ipsen and a consultant for Kallyope. John for Personalis, NuMedii, a consultant for Samsung, Geisinger Health, Mango Tree Corporation, F. Thompson has received honoraria and travel support for attending advisory board Regenstrief Institute and in the recent past a consultant for 10x Genomics and Helix, a meetings of GlaxoSmithKline and Provectus and has received honoraria for participation in shareholder in Personalis, a minor shareholder in Apple, Twitter, Facebook, Google, Microsoft, advisory boards for MSD Australia and BMS Australia. Daniel Turner is a full-time employee Sarepta, 10x Genomics, Amazon, Biogen, CVS, Illumina, Snap and Sutro and has received of Oxford Nanopore Technologies and is a stock holder. Naveen Vasudev has received honoraria and travel reimbursement for invited talks from Genentech, Roche, Pfizer, Optum, speaker honoraria and/or consultancy fees from Bristol-Myers Squibb, Pfizer, EUSA pharma, AbbVie and many academic institutions and health systems. Carlos Caldas has served on the MSD and Novartis. Jeremiah A. Wala is a consultant for Nference. Daniel J. Weisenberger is a Scientific Advisory Board of Illumina. Lorraine Chantrill acted on an advisory board for AMGEN consultant for Zymo Research. Dai-Ying Wu is employed at Annai Systems. Cheng-Zhong Australia in the past 2 years. Andrew D. Cherniack receives research funding from Bayer. Helen Zhang is a cofounder and equity holder of Pillar Biosciences, a for-profit company that Davies is an inventor on a number of patent applications that encompass the use of mutational specializes in the development of targeted sequencing assays. The other authors declare signatures. Francisco De La Vega was employed at Annai Systems during part of the project. no competing interests. Ronny Drapkin serves on the scientific advisory board of Repare Therapeutics and Siamab Therapeutics. Rosalind Eeles has received an honorarium for the GU-ASCO meeting in San Additional information Francisco in January 2016 as a speaker, a honorarium and support from Janssen for the RMH FR Supplementary information is available for this paper at https://doi.org/10.1038/s41586-020- meeting in November 2017 as a speaker (title: genetics and prostate cancer), a honorarium for 1969-6. an University of Chicago invited talk in May 2018 as speaker and an educational honorarium Correspondence and requests for materials should be addressed to P.J.C., G.G., J.O.K., J.M.S. paid by Bayer & Ipsen to attend GU Connect ‘Treatment sequencing for mCRPC patients within or L.D.S. the changing landscape of mHSPC’ at a venue at ESMO, Barcelona, on 28 September 2019. Peer review information Nature thanks Arul Chinnaiyan, Ben Lehner, Nicolas Robine and the Paul Flicek is a member of the scientific advisory boards of Fabric Genomics and Eagle other, anonymous, reviewer(s) for their contribution to the peer review of this work. Genomics. Ronald Ghossein is a consultant for Veracyte. Dominik Glodzik is an inventor on a Reprints and permissions information is available at http://www.nature.com/reprints. Extended Data Fig. 1 | Flow-chart showing key steps in the analysis of PCAWG for downstream scientific analyses. Subs, substitutions; DKFZ/EMBL, the genomes. After alignment to the genome, somatic mutations were identified German Cancer Research Centre (DKFZ) and Europen Molecular Biology by three pipelines, with subsequent merging into a consensus variant set used Laboratory (EMBL). Article Extended Data Fig. 2 | Distribution of accuracy estimates across algorithms dataset. b, F1 accuracy, precision and sensitivity estimates for somatic indels and samples from validation data. a, F1 accuracy, precision and sensitivity (n = 50 samples). SVM, support vector machine; union, calls made by all variant- estimates for somatic SNVs across the core algorithms and different calling algorithms; intersect2, calls made by any combination of two variant- approaches to merging the call-sets. The box plots demarcate the interquartile calling algorithms; intersect3, calls made by any three variant-calling range and median of estimates across the n = 50 samples in the validation algorithms. Extended Data Fig. 3 | Distribution of numbers of somatic mutations of indicate substitutions; indels are taken as insertions or deletions <100 bp in different classes across tumour types. The y axis is on a log scale. The 2,583 size; retrotranspositions are the combined counts of somatic retrotransposon donors with the highest quality metrics (white-listed donors) are plotted. SNVs insertions, transductions and somatic pseudogene insertions. Article Extended Data Fig. 4 | Patients with no detected driver mutations in PCAWG. sensitivity values for that tumour type (the width proportional is to density). a, Number (red) of patients without detected driver mutations distributed c, SETD2 expression levels across different medulloblastoma subtypes. Points across the different tumour types studied. b, Estimated sensitivity for represent individual patients, coloured by whether the gene exhibited focal detecting somatic point mutations genome-wide across tumour types (total copy number (CN) loss or a truncating point mutation, or was the wild-type sample size: n = 2,583 patients). Each point represents the estimate for a single gene. The coloured areas are violin plots showing the estimated density patient, layered on violin plots that show the estimated density distribution of distribution of expression values for that medulloblastoma subtype. Extended Data Fig. 5 | See next page for caption. Article Extended Data Fig. 5 | Examples of clustered mutational processes. prototypical APOBEC3A/B type with C > T (signature 2) and/or C>G/A a, Chromoplexy example in a thyroid adenocarcinoma. Genes at the (signature 13) substitutions at TpC. Right, an alternative cytidine deaminase(s) breakpoints are schematically depicted in their normal genomic context and with a preference for substitutions at C/GpC. Most of the SNVs in each of these again in the reconstructed derivative chromosomes below. b, Distinct kataegis foci can be phased to the same allele and no evidence of anti-phasing is signatures in the genome of a pancreatic adenocarcinoma sample. SVs and observed. c, Example of a chromothripsis event in a melanoma. The black their classification are shown above the main rainfall plot, as well as the total points (top) represent copy-number estimates from individual genomic bins, and minor allele copy number. Tra, translocation; del, deletion; dup, with SVs shown as coloured arcs (translocation in black, deletion in purple, duplication; t2tInv, tail-to-tail inversion; h2hInv, head-to-head inversion. duplication in brown, tail-to-tail inversion in cyan, head-to-head inversion in Magnifications of the three foci on chromosomes 1, 8 and 12, respectively, green) that mostly demarcate copy-number changes. The mate chromosomes highlight distinct manifestations of kataegis. Left, a novel process similar to are displayed above translocations. Bottom, the variant allele fractions of signature 17 with T > N mutations at CT or TT dinucleotides. Middle, the somatic mutations distributed along the relevant chromosomal region. Extended Data Fig. 6 | Patterns of intense kataegis. a, Distribution of the replication timing bias (c) for rearrangement-(in)dependent APOBEC kataegis, tumour types (colour-coded as in Extended Data Fig. 3) of the samples in the based on n = 2,583 tumours. P values were derived using a two-sided Mann– top 5% of kataegis intensity in each of the four identified genome-wide Whittney U-test. d, Example rainfall plots for each of the four identified patterns: non-APOBEC, replication stress, rearrangement-associated and the kataegis patterns. combination of the last two. b, c, Distribution of leading/lagging strand (b) and Article Extended Data Fig. 7 | See next page for caption. Extended Data Fig. 7 | Association of chromothripsis with covariates and plotted as a function of the number of times those drivers were hit. e, For each driver events. a, Odds ratios per cancer type of containing chromothripsis in sample in which chromothripsis coincided with a driver event in those genes, whole-genome duplicated versus diploid samples (n = 2,583 patients). we show the fold change in gene expression compared to the median ***q < 0.001; **q < 0.01; *q < 0.05. Two-sided hypothesis testing was performed expression of the gene in non-chromothripsis samples of the same cancer type, using Fisher–Boschloo tests, corrected for multiple-hypothesis testing. coloured by cancer type and shaped by the type of driver event. We show with b, Same as a for female versus male. c, Proportion of mutations explained by added transparency the fold changes calculated the same way for samples with single-base substitution signature 1 and age at diagnosis in prostate cancer driver mutations hitting the same driver genes, but that had no evidence of samples (n = 210 patients) with or without chromothripsis (q < 0.05). The early- chromothripsis. Analysis is based on n = 1,222 patients with RNA-sequencing onset prostate cancer project drives the signal and was sequenced at lower data. f, Enrichment of co-occurrence of chromothripsis with driver events. The depth. For the box-and-whisker plots, the box denotes the interquartile range, x axis shows the association of chromothripsis with a driver in a given cancer with the median marked as a horizontal line. The whiskers extend as far as the type compared with its rate of association with that driver in all other cancer range or 1.5× the interquartile range, whichever is less. Two-sided hypothesis types. The y axis shows the association of chromothripsis with a driver in a testing was performed using Mann–Whitney U-tests. d, Counts of co- given cancer type compared with its rate of association with all other drivers in occurrence of chromothripsis with amplification (blue) and homozygous that type. Exact binomial tests are used and P values are corrected for multiple deletions (red) in driver regions: observed (thick line) versus randomized testing according to the Benjamini–Hochberg method. (shaded area and thin line). The cumulative number of drivers that were hit is Article Extended Data Fig. 8 | See next page for caption. Extended Data Fig. 8 | Further examples of chromothripsis-induced along the relevant chromosomal region. The paucity of somatic mutations at amplification targeting multiple cancer-associated genes simultaneously high variant allele fractions in the most-heavily amplified regions indicates that in melanoma. a, Examples of amplifications that occurred early in the these amplifications began very early in tumour evolution, before the lineage development of melanoma. The black points (top) represent copy-number had had opportunity to acquire many SNVs. b, Example of an amplification that estimates from individual genomic bins, with SVs shown as coloured arcs occurred late in melanoma development. The large numbers of somatic (translocation in black, deletion in purple, duplication in brown, tail-to-tail mutations at high variant allele fractions in the most-heavily amplified regions inversion in cyan and head-to-head inversion in green) that mostly demarcate indicate that these amplifications began late in tumour evolution, after the copy-number changes. Bottom, the variant allele fractions of SNVs distributed lineage had already acquired many SNVs. Article Extended Data Fig. 9 | See next page for caption. Extended Data Fig. 9 | Timing the amplifications after chromothripsis in represents the molecular timing of the amplification (red bar; high is early, low molecular time for 10 representative cases. a, Copy-number plot of is late) and is coloured by the fraction of total SNVs assigned to the following chromothriptic regions categorized as ‘liposarc-like’ in five acral melanomas timing categories: clonal [early], clonal mutations that occurred before with CCND1 amplification. Segments indicate the copy number of the major duplications involving the relevant chromosome (including whole-genome allele. Points represent SNV multiplicities, that is, the estimated number of duplications); clonal [late], clonal mutations that occurred after such copies carrying each SNV, coloured by base change and shaped by strand. Small duplications; and clonal [NA], mutations that occurred when no duplication vertical arrows link SNVs to their corresponding copy-number segment. was observed. b, Same as a in two cutaneous melanomas, one shows early Kataegis foci are shown within black boxes and show typical strand amplification, the other late amplification. c, Same as a, b, for three lung specificities (all triangles or all circles), similar multiplicities and base changes squamous cell carcinomas and late amplification of SOX2. of signatures 2 and 13 (red and black, respectively). A coloured bar (top right) Article Extended Data Fig. 10 | See next page for caption. Extended Data Fig. 10 | Association between common germline variants and respectively). Locuszoom plot for somatic APOBEC3B-like mutagenesis endogenous mutational processes. Genome-wide association of somatic CpG association results around the 22q13.1 locus in individuals with European (e) mutagenesis in individuals of European ancestry (n = 1,201 patients) based on and East Asian (f) ancestry after conditioning on rs12628403. g, h, Association mutational signature analysis (a) and NpCpG motif analysis (b). Two-sided between rs2142833 and expression of APOBEC3 genes in PCAWG tumour hypothesis testing was performed using PLINK v.1.9. To mitigate multiple- samples (adjusted for sex, age at diagnosis, histology and population structure hypothesis testing, the significance threshold was set to genome-wide in linear-regression models with two-sided hypothesis testing not corrected significance (P < 5 × 10 −8). c, d, Locuszoom plot for somatic APOBEC3B-like for multiple tests). For the box-and-whisker plot, the box denotes the mutagenesis association results, linkage disequilibrium and recombination interquartile range, with the median marked as a horizontal line. The whiskers rates around the genome-wide significant 22q13.1 locus in individuals with extend as far as the range or 1.5× the interquartile range, whichever is less. European (c) and East Asian (d) ancestry (n = 1,201 and 318 patients, Outliers are shown as points. Article Extended Data Fig. 11 | See next page for caption. Extended Data Fig. 11 | Association between rare germline PTVs in protein- interquartile range, with the median marked as a white point. The whiskers coding genes and somatic mutational phenotypes. a–d, f, Data are based on extend as far as the range or 1.5× the interquartile range, whichever is less. Two- two-sided rare-variant association testing across n = 2,583 patients, with a sided hypothesis testing, not corrected for multiple testing, was performed stringent P value threshold of P < 2.5 × 10 −6 used to mitigate multiple- using linear regression models. h, Replication of germline MBD4 and BRCA2 hypothesis testing (significant genes marked with coloured circles). Blue/red PTV associations with somatic CpG mutagenesis in TCGA whole-exome circles mark genes that decrease/increase somatic mutation rates. The black sequencing donors. Violin plots show the estimated density of the proportion line represents the identity line that would be followed if the observed P values of somatic CpG mutations in TCGA exomes with germline MBD4 and BRCA2 followed the null expectation, with the shaded area showing the 95% PTVs. The box denotes the interquartile range, with the median marked as a confidence intervals. a, QQ plots for the proportion of somatic SV deletions, white point. The whiskers extend as far as the range or 1.5× the interquartile tandem duplications, inversions and translocation in cancer genomes. b, QQ range, whichever is less. Two-sided hypothesis testing, not corrected for plots for the proportion of somatic SV deletions in cancer genomes stratified multiple testing, was performed using linear-regression models. i, Correlation by four size groups (1–10 kb, 10–100 kb, 100–1,000 kb and >1,000 kb). c, QQ between MBD4 expression and somatic CpG mutagenesis in primary solid plots for the proportion of somatic SV tandem duplications in cancer genomes PCAWG tumours. Hypothesis testing was two-sided and not corrected for stratified by four size groups (1–10 kb, 10–100 kb, 100–1,000 kb and >1,000 kb). multiple testing, using linear-regression models. The box denotes the d, QQ plot for the presence or absence of somatic SV templated insertion interquartile range, with the median marked as a horizontal line. The whiskers (cycles) in cancer genomes. e, Number of SV-templated insertion cycles in extend as far as the range or 1.5× the interquartile range, whichever is less. PCAWG tumours with germline BRCA1 PTVs. Only histological samples with at j, Data are mean ± s.e.m. across n = 20 tumour types. The dashed black line least one germline BRCA1 PTV carrier are shown (n = 1,095 patients combined). shows the fitted line to the data, estimated using linear-regression models. The box denotes the interquartile range, with the median marked as a Hypothesis testing was two-sided and not corrected for multiple testing, using horizontal line. The whiskers extend as far as the range or 1.5× the interquartile Spearman’s rank correlations. k, MBD4 effect sizes (open circles) with 95% range, whichever is less. Outliers are shown as points. f, QQ plot for somatic confidence intervals (error bars) for individual cancer types were estimated CpG mutagenesis in cancer genomes based on NpCpG motif analysis. g, Violin using linear-regression analysis after (if available) accounting for sex, age at plots show estimated densities of the proportion of somatic CpG mutations in diagnosis (young/old) and ICGC project. Hypothesis testing was two-sided and PCAWG donors with germline MBD4 and BRCA2 PTVs. The box denotes the not corrected for multiple testing. Article Extended Data Fig. 12 | See next page for caption. Extended Data Fig. 12 | Germline MEI call set. a, Left, dots show the number of AMR, admixed American; EAS, East Asian; EUR, European; SAS, South Asian. transductions promoted by each hot element in individual samples. Arrows Sample sizes are shown under each axis label. g, For each type of MEI (L1, Alu highlight retrotransposition burst. Right, the contribution of each hot locus is and SVA) identified both in PCAWG and in the 1000 Genomes Project (1KGP), represented. The total number of transductions mediated by each source the correlations between allele frequency estimates per ancestry derived from element is shown on the right. b, Source L1 activity rate (that is, measured as the both projects are displayed in a blue (0) to red (1) coloured gradient. n = 2,583 average number of transductions mediated by an element) versus the PCAWG patients. Two-sided hypothesis testing was performed using percentage of samples with retrotransposition activity in which the germline Spearman’s rank correlations without correction for multiple tests. h, Example element is active. For visualization purposes, extreme points observed for a correlation between MEI allele frequencies derived from PCAWG and the 1000 source L1 with an activity rate of 49 and for a L1 active in 31% of the samples are Genomes Project for individuals with European ancestry (n = 1,201 patients in shown at ≥20 and ≥10, respectively. c, Contrasting allele frequencies for PCAWG). Two-sided hypothesis testing was performed using Spearman’s rank Strombolian and Plinian source loci (sample sizes shown under each axis label). correlations without correction for multiple tests. i, Evaluation of TraFiC-mem The box denotes the interquartile range, with the median marked as a white false-discovery rate on a liver hepatocellular carcinoma sample (DO50807) and point. The whiskers extend as far as the range or 1.5× the interquartile range, a cell line (NCI-BL2087) sequenced using single-molecule sequencing with whichever is less. Hypothesis testing was performed using two-sided Mann– MinION (Oxford Nanopore). For each allele frequency bin (common, >5%; low Whitney U-tests without correction for multiple tests. d, Numbers of active and frequency, 1–5%; rare, <1%), the percentage of events supported by N long reads hot source L1 elements per donor. Data are mean ± s.d. number of elements per is represented (N ranges from 0–1 to more than 5). MEIs supported by at least donor. e, The novel Plinian source element on 7p12.3 mediates 72 transductions two Nanopore reads were considered to be true positives (blue palette) and among only 6 cancer samples. This generates a transduction that induces the were classified as false positives (red) otherwise. The total number of germline deletion of the tumour-suppressor gene CDKN2A. f, Violin plots show the MEIs per allele frequency bin is shown on the right. j, Correlation between estimated number of distinct germline MEI alleles per PCAWG donor. The box predicted MEI lengths from Illumina and Nanopore data. Two-sided hypothesis denotes the interquartile range, with the median marked as a white point. The testing was performed using Spearman’s rank correlations without correction whiskers extend as far as the range or 1.5× the interquartile range, whichever is for multiple testing. less. Donors are grouped according to their genetic ancestry: AFR, African; Article Extended Data Fig. 13 | See next page for caption. Extended Data Fig. 13 | Different mechanisms of telomere lengthening in mutations that occurred before duplications involving the relevant cancer. a, Scatter plot showing the four clusters of tumour-specific telomere chromosome (including whole-genome duplications); clonal [late], clonal patterns identified across PCAWG samples, together with the clusters of mutations that occurred after such duplications; and clonal [NA], mutations matched normal samples, generated by t-distributed stochastic neighbour that occurred when no duplication was observed. d, Relationship between the embedding. Circles represent tumour samples and triangles represent estimated number of stem cell divisions per year and rate of telomere matched normal samples. Points are coloured by tissue of origin. Data are maintenance abnormalities across tumour types. The analysis uses data on based on n = 2,518 tumour samples and their matched normal samples. estimated rates of stem cell division per year across n = 19 tissue types b, Patterns of comutation of the relevant driver mutations across individual previously collated from the literature82. Tumour types are coloured according patients. Columns in plot represent individual patients, coloured by type of to the scheme shown in Extended Data Fig. 3. Two-sided hypothesis testing was abnormality observed. c, Distribution of clonality of driver mutations in genes performed using likelihood ratio tests on Poisson regression models with no relevant to telomere maintenance across clusters. Clonal [early], clonal correction for multiple tests. Article Extended Data Table 1 | Overview of the tumour types included in PCAWG project Adeno., adenocarcinoma; Ca., carcinoma; Comb., combined; F, female; HCC, hepatocellular carcinoma; M, male; Med, median; 10–90th, 10–90th centiles; SCC, squamous cell carcinoma. Extended Data Table 2 | Ethical considerations of genomic cloud computing Article Extended Data Table 3 | Scientific output using PCAWG data, in bite-size chunks Key findings are described further in associated papers4–18. nature research | reporting summary Peter Campbell, Joshua Stuart, Gad Getz, Jan Corresponding author(s): Korbel, Lincoln Stein Last updated by author(s): Oct 23, 2019 Reporting Summary Nature Research wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency in reporting. For further information on Nature Research policies, see Authors & Referees and the Editorial Policy Checklist. Statistics For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section. n/a Confirmed The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly The statistical test(s) used AND whether they are one- or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section. A description of all covariates tested A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals) For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted Give P values as exact values whenever suitable. For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes Estimates of effect sizes (e.g. Cohen's d, Pearson's r), indicating how they were calculated Our web collection on statistics for biologists contains articles on many of the points above. Software and code Policy information about availability of computer code Data collection Data and metadata were collected from International Cancer Genome Consortium (ICGC) consortium members using custom software packages designed by the ICGC Data Coordinating Centre. The general-purpose core libraries and utilities underlying this software have been released under the GPLv3 open source license as the "Overture" package and are available at https://www.overture.bio. Other data collection software used in this effort, such as ICGC-specific portal user interfaces, are available upon request to contact@overture.bio. Data analysis The workflows executing core WGS alignment, QC and variant-calling software are packaged as executable Dockstore images and available at: https://dockstore.org/search?labels.value.keyword=pcawg&searchMode=files. Individual software components are as follows: BWA-MEM v0.78.8-r455; DELLY v0.6.6; ACEseq v1.0.189; DKFZ somatic SNV workflow v1.0.132-1; Platypus v0.7.4; ascatNgs v1.5.2; BRASS v4.012; grass v1.1.6; CaVEMan v1.50; Pindel v1.5.7; ABSOLUTE/JaBbA v1.5; SvABA 2015-05-20; dRanger 2016-03-13; BreakPointer 2015-12-22; MuTect v1.1.4; MuSE v1.0rc; SMuFIN 2014-10-26; OxoG 2016-4-28; VAGrENT v2.1.2; ANNOVAR v2014Nov12; VariantBAM v2017Dec12; SNV-Merge v2017May26; SV-MERGE v2017Dec12; DKFZ v2016Dec15; Chromothripsis Explorer v1.0. For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors/reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Research guidelines for submitting code & software for further information. Data October 2018 Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: - Accession codes, unique identifiers, or web links for publicly available datasets - A list of figures that have associated raw data - A description of any restrictions on data availability WGS somatic and germline variant calls, mutational signatures, subclonal reconstructions, transcript abundance, splice calls and other core data generated by the ICGC/TCGA Pan-cancer Analysis of Whole Genomes Consortium are available for download at https://dcc.icgc.org/releases/PCAWG. Additional information on accessing the data, including raw read files, can be found at https://docs.icgc.org/pcawg/data/. In accordance with the data access policies of the ICGC and TCGA 1 projects, most molecular, clinical and specimen data are in an open tier which does not require access approval. To access potentially identification information, such as germline alleles and underlying sequencing data, researchers will need to apply to the TCGA Data Access Committee (DAC) via dbGaP (https:// nature research | reporting summary dbgap.ncbi.nlm.nih.gov/aa/wga.cgi?page=login) for access to the TCGA portion of the dataset, and to the ICGC Data Access Compliance Office (DACO; http:// icgc.org/daco) for the ICGC portion. In addition, to access somatic single nucleotide variants derived from TCGA donors, researchers will also need to obtain dbGaP authorization Field-specific reporting Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection. Life sciences Behavioural & social sciences Ecological, evolutionary & environmental sciences For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf Life sciences study design All studies must disclose on these points even when the disclosure is negative. Sample size We compiled an inventory of matched tumour/normal whole cancer genomes in the ICGC Data Coordinating Centre. Most samples came from treatment-naïve, primary cancers, but there were a small number of donors with multiple samples of primary, metastatic and/or recurrent tumours. Our inclusion criteria were: (i) matched tumour and normal specimen pair; (ii) a minimal set of clinical fields; and (iii) characterisation of tumour and normal whole genomes using Illumina HiSeq paired-end sequencing reads. We collected genome data from 2,834 donors, representing all ICGC and TCGA donors that met these criteria at the time of the final data freeze in autumn 2014. No statistical methods were used to predetermine sample size. Data exclusions After quality assurance, data from 176 donors were excluded as unusable. Reasons for data exclusions included inadequate coverage, extreme bias in coverage across the genome, evidence for contamination in samples and excessive sequencing errors (for example, through 8- oxoguanine). Exclusion criteria were pre-determined. Replication In order to evaluate the performance of each of the mutation-calling pipelines and determine an integration strategy, we performed a large- scale deep sequencing validation experiment. We selected a pilot set of 63 representative tumour/normal pairs, on which we ran the three core pipelines, together with a set of 10 additional somatic variant-calling pipelines contributed by members of the SNV Calling Working Group. Overall, the sensitivity and precision of the consensus somatic variant calls were 95% (CI90%: 88-98%) and 95% (CI90%: 71-99%) respectively for SNVs. For somatic indels, sensitivity and precision were 60% (34-72%) and 91% (73-96%) respectively. Regarding SVs, we estimate the sensitivity of the merging algorithm to be 90% for true calls generated by any one caller; precision was estimated as 97.5% - that is, 97.5% of SVs in the merged SV call-set have an associated copy number change or balanced partner rearrangement. Randomization No randomisation was performed - this was a descriptive study, not an experimental study. Blinding No blinding was undertaken - this was a descriptive study, not an experimental study. Reporting for specific materials, systems and methods We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material, system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response. Materials & experimental systems Methods n/a Involved in the study n/a Involved in the study Antibodies ChIP-seq Eukaryotic cell lines Flow cytometry Palaeontology MRI-based neuroimaging Animals and other organisms Human research participants Clinical data Human research participants October 2018 Policy information about studies involving human research participants Population characteristics Patient-by-patient clinical data are provided in the marker paper for the PCAWG consortium (Extended Data Table 1 of that manuscript). Demographically, the cohort included 1,469 males (55%) and 1,189 females (45%), with a mean age of 56 years (range, 1-90 years). Using population ancestry-differentiated single nucleotide polymorphisms (SNPs), the ancestry distribution was heavily weighted towards donors of European descent (77% of total) followed by East Asians (16%), as expected for large contributions from European, North American and Australian projects. We consolidated histopathology descriptions of the tumour samples, using the ICD-0-3 tumour site controlled vocabulary. Overall, the PCAWG data set comprises 38 distinct tumour 2 types. While the most common tumour types are included in the dataset, their distribution does not match the relative population incidences, largely due to differences among contributing ICGC/TCGA groups in numbers sequenced. nature research | reporting summary Recruitment Patients were recruited by the participating centres following local protocols. Samples obtained had to meet criteria on amount of tumour DNA available, meaning that the cohort is potentially somewhat biased towards larger tumours. Otherwise, we anticipate no major recruitment biases. Ethics oversight The Ethics oversight for the PCAWG protocol was undertaken by the TCGA Program Office and the Ethics and Governance Committee of the ICGC. Each individual ICGC and TCGA project that contributed data to PCAWG had their own local arrangements for ethics oversight and regulatory alignment. Note that full information on the approval of the study protocol must also be provided in the manuscript. October 2018 3