Published online 19 October 2021 Nucleic Acids Research, 2022, Vol. 50, Database issue D1069–D1076 https://doi.org/10.1093/nar/gkab925 PopHumanVar: an interactive application for the functional characterization and prioritization of adaptive genomic variants in humans Aina Colomer-Vilaplana1,† , Jesús Murga-Moreno1,2,† , Aleix Canalda-Baltrons1 , Clara Inserte2 , Daniel Soto1 , Marta Coronado-Zamora1,2 , Antonio Barbadilla 1,2 and Sònia Casillas 1,2,* 1 Department of Genetics and Microbiology, Universitat Autònoma de Barcelona, Bellaterra, Barcelona 08193, Spain and 2 Institute of Biotechnology and Biomedicine, Universitat Autònoma de Barcelona, Bellaterra, Barcelona 08193, Spain Received August 14, 2021; Revised September 17, 2021; Editorial Decision September 24, 2021; Accepted September 28, 2021 ABSTRACT GRAPHICAL ABSTRACT Adaptive challenges that humans faced as they ex- panded across the globe left specific molecular foot- prints that can be decoded in our today’s genomes. Different sets of metrics are used to identify ge- nomic regions that have undergone selection. How- ever, there are fewer methods capable of pinpoint- ing the allele ultimately responsible for this selec- tion. Here, we present PopHumanVar, an interactive online application that is designed to facilitate the exploration and thorough analysis of candidate ge- nomic regions by integrating both functional and population genomics data currently available. Po- pHumanVar generates useful summary reports of prioritized variants that are putatively causal of re- cent selective sweeps. It compiles data and graph- ically represents different layers of information, in- cluding natural selection statistics, as well as func- INTRODUCTION tional annotations and genealogical estimations of The landscape of variation in human genomes holds the variant age, for biallelic single nucleotide variants record of our evolutionary history. Despite the numerous (SNVs) of the 1000 Genomes Project phase 3. Specif- attempts to identify selection targets in diverse populations ically, PopHumanVar amasses SNV-based informa- (1–5), or date the time of appearance of an adaptive mu- tion from GEVA, SnpEFF, GWAS Catalog, ClinVar, tation and trace its spread around the globe (6–11), how, RegulomeDB and DisGeNET databases, as well as where, and when our genomes underwent adaptation is a accurate estimations of iHS, nSL and iSAFE statis- subtle issue which is far from being resolved. tics. Notably, PopHumanVar can successfully identify One of the results of next-generation sequencing (NGS) known causal variants of frequently reported can- technologies is the 1000 Genomes Project (1000GP) (12), an didate selection regions, including EDAR in East- international research effort to generate a catalog of human genetic variation. Years after its completion, it still repre- Asians, ACKR1 (DARC) in Africans and LCT/MCM6 sents one of the largest public catalogs of human variation in Europeans. PopHumanVar is open and freely avail- and genotype data. Reporting >84 million variants, with able at https://pophumanvar.uab.cat. 2504 sequenced genomes from 26 populations, it is one of * To whom correspondence should be addressed. Tel: +34 93 581 2730; Fax: +34 93 581 2011; Email: sonia.casillas@uab.cat † The authors wish it to be known that, in their opinion, the first two authors should be regarded as joint First Authors. Present address: Marta Coronado-Zamora, Institute of Evolutionary Biology, CSIC-Universitat Pompeu Fabra, Barcelona 08003, Spain.  C The Author(s) 2021. Published by Oxford University Press on behalf of Nucleic Acids Research. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. D1070 Nucleic Acids Research, 2022, Vol. 50, Database issue the main references for population genomics in the human region close to the gene EDAR, which is associated with species. hair follicle thickness and straightness and shovel-shaped With the 1000GP, came the possibility of scanning the en- incisors in East-Asians (31–34), a region in the gene ACKR1 tire genome for signatures of natural selection, resulting in (DARC), which is associated with resistance to malaria in the piling up of genomic regions believed to have evolved Africans (35–37), as well as a region close to the genes LCT under positive selection (13–18). However, these genome- and MCM6, which is associated with lactase persistence in wide scans present two major constraints: the scarce agree- Europeans (38,39). In all three cases, PopHumanVar is able ment among studies, and the lack of in-depth characteri- to identify the causal variant reported in previous studies zation of candidate loci. In 2018, we tackled the first con- and accurately estimate the variant age. These promising re- straint by presenting PopHumanScan (19), a genome-wide sults illustrate the exploratory potential of PopHumanVar catalog that brings together 2859 candidate regions under to push out into yet unfamiliar human adaptation signa- selection resulting from the combination of several metrics tures, including those compiled in PopHumanScan or the that capture selection in a wide range of time scales and ones that can be visually extracted from PopHuman, but selective regimes. Even though PopHumanScan compiled also any other genomic region of interest. an exhaustive list of candidate regions and cross-referenced them to 268 previous publications, it did not provide tools CONTENTS OF POPHUMANVAR to facilitate their validation nor to perform thorough anal- yses at the single nucleotide variant (SNV) level. PopHumanVar collects evolutionary data, functional anno- Integrating the numerous currently available information tations, and age information altogether. Evolutionary and layers on functional and population genetics metrics can age information have been computed on the 26 populations help portray the genomic landscape of a putatively selected of the phase 3 of the 1000GP (12), while functional annota- region and aid the prioritization of causal genetic variants. tions have been retrieved from publicly available databases These sources range from functional annotations (e.g. as- (see below). In total, 81.70 M SNVs of the 1000GP have sociations with phenotypes and diseases, implication in the information for one or more of the collected data sources. regulation of gene expression, or predicted functional ef- fects), to selection statistics based on the analysis of ge- Selection statistics and favored mutation rank nomics data, to genealogical estimations of variant age. As far as we know, even though several SNV-oriented public All selection statistics were computed on the 26 popula- online databases exist that cover one of the previous aspects tions of the phase 3 of the 1000GP, including non-inbred (see e.g. snpXplorer (20)), none of them bring both func- individuals as specified by Gazal et al. (40). We considered tional and evolutionary information all together with the autosomal biallelic SNVs that are accessible to sequenc- main focus of identifying causal variants of selective sweeps. ing techniques according to the 1000GP pilot accessibility Here, we present PopHumanVar, an interactive online ap- mask (12). We advise taking results for the four admixed- plication that is designed to facilitate the exploration and American populations with caution, as these populations thorough analysis of candidate genomic regions under se- have complex recent demographic histories that may mimic lection, generating useful summary reports of prioritized some patterns of genetic diversity that PopHumanVar uses variants that are putatively causal of recent selective sweeps. to infer selection, and thus results from these populations It compiles and graphically represents selection statistics may be difficult to interpret. based on linkage disequilibrium, a comprehensive set of functional annotations, and recent genealogical estimations Integrated Haplotype Score (iHS). Defined by Voight et of variant age for SNVs of the 26 populations of the phase 3 al. (21), it tracks the decay of haplotype homozygosity for of the 1000GP. Specifically, PopHumanVar gathers data ei- both ancestral and derived haplotypes. It has good power ther computed or compiled from the following data sources: to detect selective sweeps at a moderate frequency (50– the Integrated Haplotype Score (iHS) (21), the Number of 80%) (16,21,41). iHS was computed with selscan v1.2.0a Segregating sites by Length (nSL ) (22), the Integrated Selec- and norm v1.2.1a (42). We only considered those SNVs hav- tion of Allele Favored by Evolution (iSAFE) (23), SnpEFF ing a Minor Allele Frequency (MAF) higher than 0.05 and (24), RegulomeDB (25), ClinVar (26), GWAS Catalog (27), a maximum gap of 20 kb between consecutive SNPs when DisGeNET (28), and the Genealogical Estimation of Vari- assembling haplotypes. The recombination maps used to in- ant Age (GEVA) as obtained from the Human Genome terpolate genetic positions (necessary to compute iHS) were Dating database (or Atlas of Variant Age) (29). As such, the sex-averaged ones from Bhérer et al. (43). We obtained PopHumanVar is complementary to our previous genome estimates for a total of 12.14 M SNVs (Figure 1). Signifi- browser -PopHuman (30)- and database of candidate se- cance was assessed from the empirical distribution of iHS lection regions -PopHumanScan (19)-, allowing researchers values in each population separately. to focus on particular selective sweeps, pinpoint the corre- sponding causal variants, and estimate variant age. For pop- Number of Segregating sites by Length (nSL ). It is also a ulations and/or samples not included in the online applica- haplotype-based statistic. It combines information on the tion, PopHumanVar allows uploading and analyzing a VCF distribution of fragment lengths, defined by pairwise dif- file with custom data. ferences, with the distribution of the number of segregating The utility of PopHumanVar has been tested on fre- sites between all pairs of chromosomes (22). It is better than quently reported candidate genomic regions in genome- iHS at capturing soft sweeps. nSL was also computed with wide scans for positive selection in humans, including a selscan v1.2.0a and norm v1.2.1a. We only considered those Nucleic Acids Research, 2022, Vol. 50, Database issue D1071 Figure 1. Summary of the contents of PopHumanVar. Bars represent the number and percentage of single nucleotide variants (SNVs) with information for each dataset. SNVs having a MAF higher than 0.05 and a maximum gap Functional annotations of 20 kbp between consecutive SNPs when assembling hap- SnpEFF. It predicts and annotates the functional effects lotypes. We obtained estimates for a total of 12.66 M SNVs of genetic variants (24) (e.g. stop gain, splice donor vari- (Figure 1). As for iHS, significance was assessed from the ant, missense variant, intergenic region. . . ), which are clas- empirical distribution of nSL values in each population sep- sified into four different categories based on their impact arately. (i.e. high, moderate, low or modifier). We ran SnpEFF v.5.0 with default parameters and obtained annotations for 79.56 M SNVs (Figure 1). Affected genes, if any, were also Integrated Selection of Allele Favored by Evolution recorded. (iSAFE). It aims to identify the specific variant ul- timately responsible for a selective sweep (23). iSAFE exploits coalescent-based signals in the surroundings of RegulomeDB. It predicts and annotates the regulatory po- a candidate region to rank mutations according to their tential of intergenic variants (25). Evidence is compiled likelihood of having caused the selective sweep. In order from GEO (44), ENCODE (45), and the published litera- to compute iSAFE genome-wide, we analyzed overlapping ture, and it includes known, as well as predicted, regulatory sliding windows of 3 Mbp, with a 1 Mbp overlap, all DNA elements, such as regions of DNase hypersensitivity along the autosomal chromosomes (values suggested sites, transcription factor binding sites, and promoter re- by the iSAFE authors in their GitHub repository at gions that have been biochemically characterized to regulate https://github.com/alek0991/iSAFE). From each window, transcription. RegulomeDB scores the regulatory potential we kept values for the 1 Mbp middle chunk and discarded of intergenic SNVs based on overlapping supporting infor- values in the shoulders. In order to facilitate the genome- mation (i.e. 15 different scores, from 1a to 7). We retrieved wide approach, we ran iSAFE with default parameters, but RegulomeDB v.2.0.3 scores for 12.76 M SNVs (Figure 1). ignoring the gaps and increasing the maximum rank pa- rameter up to the window size (MaxRank = window = 300) ClinVar. It is one of the largest catalogs of genetic vari- in order to retrieve values for all SNVs in the window. ants that are clinically associated with diseases, together We obtained iSAFE values for a total of 42.14 M SNVs with supporting evidence (26). It rates variant-disease asso- (Figure 1). Significance was assessed from the empirical ciations into different categories (e.g. pathogenic, risk fac- distribution of iSAFE values in each population separately. tor, presenting drug response, protective, benign. . . ). We re- D1072 Nucleic Acids Research, 2022, Vol. 50, Database issue trieved ClinVar (updated on 2021/03/04) annotations for (GRCh37/hg19) or by searching a variant rsID, gene sym- 630 k SNVs (Figure 1). bol or Ensembl identifier; (ii) activating one or more popu- lations for which to display selection statistics and favored GWAS Catalog. It is a quality-controlled, manually- mutation ranks; and (iii) setting filters and parameters spe- curated, literature-derived collection of all published cific to the different visualization tabs. Changes in the FIL- genome-wide association studies (GWAS) assaying at least TERS MENU will only be applied after clicking the ‘Up- 100 000 genetic variants (27). We retrieved the number of date’ button at the bottom of the panel. associations in the GWAS Catalog v1.0.2, as well as the spe- cific traits reported, over 149 k SNVs (Figure 1). Selection (iHS & nSL ). iHS and nSL value distributions within and across the genomic region of interest are dis- DisGeNET. It is one of the largest publicly available col- played for each of the selected populations. Significant val- lections of genes and variants associated with human dis- ues are indicated by the golden horizontal lines and in the eases (28). It integrates data from expert-curated reposito- interactive hover panel, and can be filtered from the FIL- ries, homogeneously annotated with controlled vocabular- TERS MENU (default empirical P-value ≤ 0.005, cus- ies and community-driven ontologies. It provides original tomizable by the user). metrics to assist the prioritization of genotype-phenotype relationships, such as disease specificity, evidence index, or number of Pubmed identifiers. We retrieved DisGeNET Favored mutation (iSAFE). As above, iSAFE value dis- v.7.0 annotations for 189 k SNVs (Figure 1). tributions within and across the genomic region of interest are displayed for each of the selected populations. For sim- plicity, only scores higher than 0.05 are represented. Signifi- Age estimation cant values are indicated by the golden horizontal lines and Human Genome Dating (or Atlas of Variant Age). It gath- in the interactive hover panel, and can be filtered from the ers age estimation results for more than 45 M variants in FILTERS MENU (default empirical P-value ≤ 0.0001 fol- the human genome, computed using the Genealogical Es- lowing Akbari et al. (23), customizable by the user). timation of Variant Age (GEVA) (29). GEVA is a method that exploits coalescent modeling to infer the time to the Functional description. It includes different representa- most recent common ancestor (TMRCA) between individ- tions of the annotations retrieved or computed from ual genomes based on three different clock models, consid- SnpEFF, RegulomeDB, ClinVar, GWAS Catalog and Dis- ering: (i) mutation events that occur independently in each GeNET. Several filters can be applied from the FILTERS lineage and pile up as the ancestral haplotype is passed on MENU. over the generations (i.e. mutational clock); (ii) recombina- tion events that shorten the length of the ancestral haplo- type, independently in each lineage and across generations Age information. Genealogical estimations of variant age (i.e. recombination clock); or (iii) both (i.e. joint clock). We are represented for each SNV across the genomic region retrieved age estimates, as well as the corresponding quality of interest. Several filters and parameters can be applied scores, for all three clock models, from the Atlas of Vari- from the FILTERS MENU, including the clock model and ant Age database (downloaded on 2021/06/26) for a total the units of age in generations or years. In the graph, size of 43.23 M SNVs (Figure 1). and color represent the quality score of the estimations. OVERVIEW OF THE POPHUMANVAR INTERFACE Summary report. This section aims to summarize all the evolutionary and functional information gathered for the The PopHumanVar interface is divided into four main sec- genomic region of interest through a JBrowse implementa- tions: (i) Stats Visualization represents the main navigation tion showing gene annotations overlapping the region, di- interface and provides several interactive graphs to aid the rect links to PopHuman (30) and PopHumanScan (19), an exploration and prioritization of genomic variants in the eloquent summary graph including selection statistics and region of interest (Figure 2); (ii) Download provides tools functional data, a list of top-20 automatically-prioritized to customize batch downloads from the database; (iii) Up- putatively causal variants, and representations of EHH and load Data allows uploading and analyzing a VCF file with haplotype furcations around any SNV of the top-20 pri- custom data; and (iv) Tutorial describes the database and oritized variants (by default the one having the most ex- presents a step-by-step usage example. treme iSAFE value; computed with rehh (46)). In the sum- mary graph, iSAFE scores are represented for all SNVs across the genomic region of interest. Color represents the Stats visualization strongest SnpEFF functional effect of each variant, and size Stats visualization unfolds five subitems, each pointing to represents its combined iHS + nSL value. All the informa- a visualization tab with one or more interactive graphs tion displayed in this section refers to a specific population, and tables. Note that while in the Stats Visualization, an which can be chosen from the right-side menu –DISPLAY additional menu –FILTERS MENU– adds to the left- OPTIONS–. Changes in the DISPLAY OPTIONS menu side panel of the application. It allows: (i) choosing a ge- will be applied after clicking the ‘Refresh’ button at the bot- nomic region of interest, either by entering its coordinates tom of the panel. Nucleic Acids Research, 2022, Vol. 50, Database issue D1073 Figure 2. Simplified representation of the PopHumanVar interface. Some representative graphs of each of the five elements of the Stats Visualization section of the database are represented; from left to right, top to bottom: Summary Report, Selection, Favored Mutation, Functional Description and Age Information. Download upon binding to its ligand, induces an intracellular cascade leading to the activation of the transcription factor NF-␬B. Download unfolds two subitems: Current Region and Batch EDAR is a well-studied gene involved in the development Download. Both are used to download PopHumanVar data of hair follicles, teeth, and sweat glands (31–34). It has fre- in tabular files. Current Region is the most customizable op- quently been reported in genome-wide scans for positive se- tion and allows specifying how filters should be applied and lection in humans (19,47–49) and is one of the candidate re- which data should be included in the downloaded files. The gions cataloged in PopHumanScan (Figure 3). PopHuman- right-side menu is used to set these parameters. Batch Down- Scan reports signatures of selection for haplotype-based load is used to make bulk downloads of the whole database statistics (i.e. iHS and XP-EHH) in East-Asian populations, contents. Data for a maximum of 50 Mbp (which may be especially in the Southern Han Chinese (CHS) population. split into several regions) can be retrieved at a time. Alter- In addition, the region shows extreme values (i.e. more than natively, data for whole chromosomes can be downloaded two standard deviations away from the mean value) both for in compressed tabular files. the haplotype-based statistics iHS and XP-EHH, and the Site Frequency Spectrum (SFS)-based statistics Tajima’s D Upload data and Fay and Wu’s H, as displayed in PopHuman. Although both PopHuman and PopHumanScan bring our attention This section allows uploading a VCF file with custom data, to this region, none of them allows us to shift to the SNV which may cover up to 2 Mbp of genomic sequence for one level and determine which variant was selected and when. single population. The data is processed automatically by Instead of reporting summary statistics for a genomic re- the PopHumanVar pipeline and results are sent by email as gion of interest, as PopHuman and PopHumanScan do, the a dynamic Shiny markdown file. PopHumanVar application presented here reports informa- tion at the SNV level and helps prioritize causal variants Tutorial of selective sweeps. In the EDAR gene region, apart from gathering abundant functional annotations and evolution- This section documents the data used and the procedures ary statistics, PopHumanVar prioritizes the protein-coding implemented in PopHumanVar, and includes a complete tu- missense variant rs3827760 (A > G) as the top causal vari- torial introducing the usage of the database through a step- ant (Figure 3), which happens to be the known causal vari- by-step worked example. ant of a well-studied selective sweep in East-Asians (34,49). The derived G allele (Val370Ala substitution) is found at high frequency in East-Asian populations (87%), as well as POPHUMANVAR WITH AN EXAMPLE: SELECTION Native American populations (39%) (31). It was driven to AT THE EDAR LOCUS high frequency in East-Asia by positive selection prior to 10 Here, we illustrate the usage of PopHumanVar with an ex- 000 years ago (6,31). In the GWAS catalog, it is reported to ample. This section summarizes the main findings, while the be associated with ear, eyebrow and chin morphology, and Tutorial section of the application contains a step-by-step male-pattern baldness. The prioritized variant rs3827760 guide of the same example. reports the highest iSAFE––as well as iHS and nSL ––values For this case study, we will focus on a genomic region of in the region. 1.15 Mbp in chromosome 2 (chr2:109500927–109615828; Two additional case studies, that of genes ACKR1 GRCh37/hg19). The region contains the gene EDAR – (DARC) in Africans and LCT/MCM6 in Europeans, are Ectodysplasin A Receptor–, a cell-surface receptor that, shown in the Supplementary Data. D1074 Nucleic Acids Research, 2022, Vol. 50, Database issue Figure 3. Characterization and variant prioritization in the EDAR gene region, which is associated with hair follicle thickness and straightness and shovel- shaped incisors in East-Asians. Complementary information obtained from PopHumanScan (top) and PopHumanVar (bottom) is shown. Color labels at the right of the PopHumanVar section represent, from top to bottom: iSAFE, iHS, nSL (top 0.5%), SnpEff effect, GWAS Catalog hits, and Atlas of Variant Age variant age in generations. CONCLUSION vance our understanding of how environmental and social challenges have shaped our genomes through the action of The PopHumanVar interactive application presented here, natural selection. successfully tested by confirmatory results on the EDAR gene and two other well-known case studies, demonstrates its exploratory potential to prioritize variants in regions holding signatures of natural selection. Contrary to other IMPLEMENTATION AND AVAILABILITY SNV-oriented public online databases, the PopHumanVar PopHumanVar is based on the Shiny framework (50) for de- approach brings both functional and evolutionary informa- velopment of web-based applications using the R program- tion all together, including natural selection statistics, func- ming environment (51). Interactive plots are implemented tional annotations and genealogical estimations of variant with plotly (52). Interactive tables are generated with DT age, and goes one step forward in the task of identifying (53), an R-based interface to the JavaScript DataTables li- and dating the emergence of variants that were putatively brary. The genome browser integrated into the Summary causal of the corresponding selective sweeps. In this way, Report section is implemented using the JBrowseR package PopHumanVar eases the description and thorough anal- (54). User queries are processed by R and sent to a Mari- ysis of yet unfamiliar human adaptation signatures such aDB database. All scripts are available in the GitHub repos- as those compiled in PopHumanScan or the ones that can itory (https://github.com/ainacolovila/PopHumanVar). be visually extracted from PopHuman. Future implemen- PopHumanVar is served with Apache on a CentOS 7.2 tations to PopHumanVar will include the development of a Linux x64 server with 16 Intel Xeon 2.4 GHz processors pre-processing module that returns uniform, adequate data and 32 GB RAM. All data, tools and support resources pro- from any human variation source data, so that additional vided by the PopHumanVar database are open and freely populations not in the 1000GP, or new 1000GP samples, available at https://pophumanvar.uab.cat. PopHumanVar can be easily incorporated in PopHumanVar. All in all, we is accessible and legible on computer, phone and tablet think that the public release of PopHumanVar will help ad- screens. Nucleic Acids Research, 2022, Vol. 50, Database issue D1075 SUPPLEMENTARY DATA 12. Auton,A., Abecasis,G.R., Altshuler,D.M., Durbin,R.M., Abecasis,G.R., Bentley,D.R., Chakravarti,A., Clark,A.G., Supplementary Data are available at NAR Online. Donnelly,P., Eichler,E.E. et al. (2015) A global reference for human genetic variation. Nature, 526, 68–74. 13. Johnson,K.E. and Voight,B.F. (2018) Patterns of shared signatures of ACKNOWLEDGEMENTS recent positive selection across human populations. Nat. Ecol. Evol., 2, 713–720. The authors would like to thank the Port d’Informació 14. Sugden,L.A., Atkinson,E.G., Fischer,A.P., Rong,S., Henn,B.M. and Cientı́fica (PIC) of the UAB for providing the informatics Ramachandran,S. (2018) Localization of adaptive variants in human genomes using averaged one-dependence estimation. Nat. Commun., infrastructure in which most of the population genomics 9, 703. statistics have been computed, and help on using it. We 15. Sabeti,P.C., Schaffner,S.F., Fry,B., Lohmueller,J., Varilly,P., also thank Esteve Sanz for providing some data manage- Shamovsky,O., Palma,A., Mikkelsen,T.S., Altshuler,D. and ment utilities, Laia Carrillo for evaluating the PopHuman- Lander,E.S. (2006) Positive natural selection in the human lineage. Science, 312, 1614–1620. Var data on several case regions, and members of the Ge- 16. Sabeti,P.C., Varilly,P., Fry,B., Lohmueller,J., Hostetter,E., nomics, Bioinformatics and Evolutionary Biology group for Cotsapas,C., Xie,X., Byrne,E.H., McCarroll,S.A., Gaudet,R. et al. testing the database implementation. Finally, we thank two (2007) Genome-wide detection and characterization of positive anonymous referees for very helpful comments on the Po- selection in human populations. Nature, 449, 913–918. pHumanVar implementation and manuscript. 17. Biswas,S. and Akey,J.M. (2006) Genomic insights into positive selection. Trends Genet., 22, 437–446. 18. Akey,J.M. (2009) Constructing genomic maps of positive selection in humans: where do we go from here? Genome Res., 19, 711–722. FUNDING 19. Murga-Moreno,J., Coronado-Zamora,M., Bodelón,A., Barbadilla,A. and Casillas,S. (2019) PopHumanScan: the online catalog of human Ministerio de Economı́a y Competitividad (Spain); ERDF genome adaptation. Nucleic Acids Res., 47, D1080–D1089. funds [CGL2017-89160P to M.S., A.B.]; AGAUR (Gener- 20. Tesi,N., van der Lee,S., Hulsman,M., Holstege,H. and alitat de Catalunya) [2017SGR-1379 to A.R.]; Secretaria Reinders,M.J.T. (2021) snpXplorer: a web application to explore d’Universitats i Recerca de la Generalitat de Catalunya and human SNP-associations and annotate SNP-sets. Nucleic Acids Res., 49, W603–W612. the European Social Fund [2020FI B-01045 to A.C.-V.]; 21. Voight,B.F., Kudaravalli,S., Wen,X. and Pritchard,J.K. (2006) A map Departament de Genètica i de Microbiologia (UAB) [PIF of recent positive selection in the human genome. PLOS Biol., 4, e72. to J.M.-M.]. Funding for open access charge: Ministerio de 22. Ferrer-Admetlla,A., Liang,M., Korneliussen,T. and Nielsen,R. (2014) Economı́a y Competitividad (Spain). On detecting incomplete soft or hard selective sweeps using haplotype Conflict of interest statement. None declared. structure. Mol. Biol. Evol., 31, 1275–1291. 23. Akbari,A., Vitti,J.J., Iranmehr,A., Bakhtiari,M., Sabeti,P.C., Mirarab,S. and Bafna,V. (2018) Identifying the favored mutation in a positive selective sweep. Nat. Methods, 15, 279–282. REFERENCES 24. Cingolani,P., Platts,A., Wang,L.L., Coon,M., Nguyen,T., Wang,L., 1. Nielsen,R., Akey,J.M., Jakobsson,M., Pritchard,J.K., Tishkoff,S. and Land,S.J., Lu,X. and Ruden,D.M. (2012) A program for annotating Willerslev,E. (2017) Tracing the peopling of the world through and predicting the effects of single nucleotide polymorphisms, genomics. Nature, 541, 302–310. SnpEff. Fly (Austin), 6, 80–92. 2. Fan,S., Hansen,M.E.B., Lo,Y. and Tishkoff,S.A. (2016) Going global 25. Boyle,A.P., Hong,E.L., Hariharan,M., Cheng,Y., Schaub,M.A., by adapting local: a review of recent human adaptation. Science, 354, Kasowski,M., Karczewski,K.J., Park,J., Hitz,B.C., Weng,S. et al. 54–59. (2012) Annotation of functional variation in personal genomes using 3. Hinds,D.A., Stuve,L.L., Nilsen,G.B., Halperin,E., Eskin,E., RegulomeDB. Genome Res., 22, 1790–1797. Ballinger,D.G., Frazer,K.A. and Cox,D.R. (2005) Whole-genome 26. Landrum,M.J., Lee,J.M., Riley,G.R., Jang,W., Rubinstein,W.S., patterns of common DNA variation in three human populations. Church,D.M. and Maglott,D.R. (2014) ClinVar: public archive of Science, 307, 1072–1079. relationships among sequence variation and human phenotype. 4. Altshuler,D., Donnelly,P. and The International HapMap Nucleic Acids Res., 42, D980–D985. Consortium (2005) A haplotype map of the human genome. Nature, 27. MacArthur,J., Bowler,E., Cerezo,M., Gil,L., Hall,P., Hastings,E., 437, 1299–1320. Junkins,H., McMahon,A., Milano,A., Morales,J. et al. (2017) The 5. Altshuler,D.M., Gibbs,R.A., Peltonen,L., Altshuler,D.M., new NHGRI-EBI Catalog of published genome-wide association Gibbs,R.A., Peltonen,L., Dermitzakis,E., Schaffner,S.F., Yu,F., studies (GWAS Catalog). Nucleic Acids Res., 45, D896–D901. Peltonen,L. et al. (2010) Integrating common and rare genetic 28. Piñero,J., Ramı́rez-Anguita,J.M., Saüch-Pitarch,J., Ronzano,F., variation in diverse human populations. Nature, 467, 52–58. Centeno,E., Sanz,F. and Furlong,L.I. (2020) The DisGeNET 6. Smith,J., Coop,G., Stephens,M. and Novembre,J. (2018) Estimating knowledge platform for disease genomics: 2019 update. Nucleic Acids time to the common ancestor for a beneficial allele. Mol. Biol. Evol., Res., 48, D845–D855. 35, 1003–1017. 29. Albers,P.K. and McVean,G. (2020) Dating genomic variants and 7. Speidel,L., Forest,M., Shi,S. and Myers,S.R. (2019) A method for shared ancestry in population-scale sequencing data. PLOS Biol., 18, genome-wide genealogy estimation for thousands of samples. Nat. e3000586. Genet., 51, 1321–1329. 30. Casillas,S., Mulet,R., Villegas-Mirón,P., Hervas,S., Sanz,E., 8. Speidel,L., Cassidy,L., Davies,R.W., Hellenthal,G., Skoglund,P. and Velasco,D., Bertranpetit,J., Laayouni,H. and Barbadilla,A. (2018) Myers,S.R. (2021) Inferring population histories for ancient genomes PopHuman: the human population genomics browser. Nucleic Acids using genome-wide genealogies. Mol. Biol. Evol., 38, 3497–3511. Res., 46, D1003–D1010. 9. Bergström,A., Stringer,C., Hajdinjak,M., Scerri,E.M.L. and 31. Bryk,J., Hardouin,E., Pugach,I., Hughes,D., Strotmann,R., Skoglund,P. (2021) Origins of modern human ancestry. Nature, 590, Stoneking,M. and Myles,S. (2008) Positive selection in fast Asians for 229–237. an EDAR allele that enhances NF-␬B activation. PLoS One, 3, e2209. 10. Kelleher,J., Wong,Y., Wohns,A.W., Fadil,C., Albers,P.K. and 32. Enattah,N.S., Sahi,T., Savilahti,E., Terwilliger,J.D., Peltonen,L. and McVean,G. (2019) Inferring whole-genome histories in large Järvelä,I. (2002) Identification of a variant associated with adult-type population datasets. Nat. Genet., 51, 1330–1338. hypolactasia. Nat. Genet., 30, 233–237. 11. Rasmussen,M.D., Hubisz,M.J., Gronau,I. and Siepel,A. (2014) 33. Kamberov,Y.G., Wang,S., Tan,J., Gerbault,P., Wark,A., Tan,L., Genome-wide inference of ancestral recombination graphs. PLoS Yang,Y., Li,S., Tang,K., Chen,H. et al. (2013) Modeling recent Genet., 10, e1004342. D1076 Nucleic Acids Research, 2022, Vol. 50, Database issue human evolution in mice by expression of a selected EDAR variant. 44. Davis,S. and Meltzer,P.S. (2007) GEOquery: a bridge between the Cell, 152, 691–702. Gene Expression Omnibus (GEO) and BioConductor. 34. Park,J.-H., Yamaguchi,T., Watanabe,C., Kawaguchi,A., Haneji,K., Bioinformatics, 23, 1846–1847. Takeda,M., Kim,Y.-I., Tomoyasu,Y., Watanabe,M., Oota,H. et al. 45. Consortium,T.E.P. (2004) The ENCODE (ENCyclopedia Of DNA (2012) Effects of an Asian-specific nonsynonymous EDAR variant on Elements) Project. Science, 306, 636–640. multiple dental traits. J. Hum. Genet., 57, 508–514. 46. Klassmann,A. and Gautier,M (2020) Detecting selection using 35. Yin,Q., Srivastava,K., Gebremedhin,A., Makuria,A.T. and Extended Haplotype Homozygosity-based statistics on unphased or Flegel,W.A. (2018) Long-range haplotype analysis of the malaria unpolarized data. Authorea doi: parasite receptor gene ACKR1 in an East-African population. Hum. https://doi.org/10.22541/au.160405572.29972398/v1, 30 October Genome Var., 5, 26. 2020, preprint: not peer reviewed. 36. Schmid,P., Ravenell,K.R., Sheldon,S.L. and Flegel,W.A. (2012) 47. Savolainen,O., Lascoux,M. and Merilä,J. (2013) Ecological genomics DARC alleles and Duffy phenotypes in African Americans. of local adaptation. Nat. Rev. Genet., 14, 807–820. Transfusion (Paris), 52, 1260–1267. 48. Racimo,F., Gokhman,D., Fumagalli,M., Ko,A., Hansen,T., 37. McManus,K.F., Taravella,A.M., Henn,B.M., Bustamante,C.D., Moltke,I., Albrechtsen,A., Carmel,L., Huerta-Sánchez,E. and Sikora,M. and Cornejo,O.E. (2017) Population genetic analysis of the Nielsen,R. (2017) Archaic adaptive introgression in TBX15/WARS2. DARC locus (Duffy) reveals adaptation from standing variation Mol. Biol. Evol., 34, 509–524. associated with malaria resistance in humans. PLOS Genet., 13, 49. Fujimoto,A., Kimura,R., Ohashi,J., Omi,K., Yuliwulandari,R., e1006560. Batubara,L., Mustofa,M.S., Samakkarn,U., Settheetham-Ishida,W., 38. Tishkoff,S.A., Reed,F.A., Ranciaro,A., Voight,B.F., Babbitt,C.C., Ishida,T. et al. (2008) A scan for genetic determinants of human hair Silverman,J.S., Powell,K., Mortensen,H.M., Hirbo,J.B., Osman,M. morphology: EDAR is associated with Asian hair thickness. Hum. et al. (2007) Convergent adaptation of human lactase persistence in Mol. Genet., 17, 835–843. Africa and Europe. Nat. Genet., 39, 31–40. 50. Chang,W., Cheng,J., Allaire,J.J., Sievert,C., Schloerke,B., Xie,Y., 39. Ingram,C.J.E., Mulcare,C.A., Itan,Y., Thomas,M.G. and Allen,J., McPherson,J., Dipert,A. and Borges,B. (2021) shiny: Web Swallow,D.M. (2009) Lactose digestion and the evolutionary genetics Application Framework for R. of lactase persistence. Hum. Genet., 124, 579–591. 51. Core Team,R (2020) In: R: A Language and Environment for 40. Gazal,S., Sahbatou,M., Babron,M.-C., Génin,E. and Statistical Computing R Foundation for Statistical Computing. Leutenegger,A.-L. (2015) High level of inbreeding in final phase of Vienna, Austria. 1000 Genomes Project. Sci. Rep., 5, 17453. 52. Sievert,C. (2020) In: Interactive Web-Based Data Visualization with R, 41. Pickrell,J.K., Coop,G., Novembre,J., Kudaravalli,S., Li,J.Z., plotly, and shiny Chapman and Hall/CRC. Absher,D., Srinivasan,B.S., Barsh,G.S., Myers,R.M., Feldman,M.W. 53. Xie,Y., Cheng,J. and Tan,X. (2021) In: DT: A Wrapper of the et al. (2009) Signals of recent positive selection in a worldwide sample JavaScript Library ‘DataTables’. of human populations. Genome Res., 19, 826–837. 54. Hershberg,E.A., Stevens,G., Diesh,C., Xie,P., De,Jesus, Martinez,T., 42. Szpiech,Z.A. and Hernandez,R.D. (2014) selscan: an efficient Buels,R., Stein,L. and Holmes,I. (2021) JBrowseR: an R interface to multithreaded program to perform EHH-based scans for positive the JBrowse 2 genome browser. Bioinformatics, selection. Mol. Biol. Evol., 31, 2824–2827. https://doi.org/10.1093/bioinformatics/btab459. 43. Bhérer,C., Campbell,C.L. and Auton,A. (2017) Refined genetic maps reveal sexual dimorphism in human meiotic recombination at multiple scales. Nat. Commun., 8, 14994.