Trends in OPEN ACCESS Cancer Review Single cell cancer epigenetics Marta Casado-Pelaez, 1,5,@ Alberto Bueno-Costa, 1,5,@ and Manel Esteller 1,2,3,4, ,@ * Bulk sequencing methodologies have allowed us to make great progress in cancer Highlights research. Unfortunately, these techniques lack the resolution to fully unravel the The epigenome encompasses several epigenetic mechanisms that govern tumor heterogeneity. Consequently, many mechanisms controlling gene expression that can be aberrantly regulated during novel single cell-sequencing methodologies have been developed over the past cancer development and progression. decade, allowing us to explore the epigenetic components that regulate different Tumors are highly complex and hetero- aspects of cancer heterogeneity, namely: clonal heterogeneity, tumor microen- geneous biological systems that require vironment (TME), spatial organization, intratumoral differentiation programs, the study of epigenetic alterations at a single cell resolution. metastasis, and resistance mechanisms. In this review, we explore the different sequencing techniques that enable researchers to study different aspects of Several single cell technologies devel- epigenetics (DNA methylation, chromatin accessibility, histone modifications, oped to study different layers of the epi- DNA–protein interactions, and chromatin 3D architecture) at the single cell level, genome, such as chromatin accessibility or histone modifications, have been de- their potential applications in cancer, and their current technical limitations. veloped and applied in cancer research over the past few years, improving our understanding of the mechanisms driv- Need for single cell epigenetic approaches to unravel tumor heterogeneity ing tumorigenesis. The importance of epigenetics in both basic and clinical research is indisputable. In the field of cancer, epigenetic alterations have important implications for many aspects of this disease. Although these techniques are promis- ing, most are still nascent and present Indeed, non-mutational epigenetic reprogramming was recently designated as a mechanistic limitations, such as low throughput and determinant that enables the acquisition of cancer hallmark capabilities [1]. Although it is well limited coverage. In addition, the analysis established that cancer cells may arise from genetic mutations that drive carcinogenesis, many and integration of the various single cell types of tumor lack strong genetic drivers that could explain important malignant processes, epigenomic data modalities have chal- lenges and require the development of such as tumor progression [2], resistance to therapy [3], and metastasis [4], suggesting that new computational tools. non-genetic determinants have a crucial role in cancer [5]. Thus, alterations of the epigenetic mechanisms affecting both malignant and non-malignant cells in a tumor may act as critical non-genetic determinants of cancer evolution. These epigenetic mechanisms, which regulate the expression of genes without altering the DNA sequence, fall into five main categories: 1 Josep Carreras Leukaemia Research (i) DNA methylation; (ii) chromatin accessibility; (iii) histone modifications; (iv) DNA–protein interac- Institute (IJC), Badalona, Barcelona, Catalonia, Spain tions; and (v) chromatin tridimensional architecture [6,7]. Each type of mechanism can be exper- 2 Centro de Investigacion Biomedica en imentally studied using several bulk methodologies (Box 1). Unfortunately, due to the complex Red Cancer (CIBERONC), 28029 cellular heterogeneity of many types of tumor, valuable information is lost when using these tech- Madrid, Spain 3 niques, since all the possible data that could be retrieved from a single cell point of view are Institucio Catalana de Recerca i Estudis Avançats (ICREA), Barcelona, Catalonia, masked by the bulk cell averaging. Nonetheless, with the emergence of single cell-sequencing Spain 4 technologies [5,8], many aspects of this tumoral heterogeneity that were otherwise impossible Physiological Sciences Department, School of Medicine and Health to assess are now open for exploration. Sciences, University of Barcelona (UB), Barcelona, Catalonia, Spain 5 A tumor is a highly heterogeneous entity comprising malignant and non-malignant cells, each of These authors contributed equally to which has crucial roles in cancer progression [9]. The development of single cell epigenomic the paper as first authors sequencing technologies can help to properly dissect non-genetic dependencies of malignant progression and unravel this tumor complexity. There are six important aspects of cancer biology *Correspondence: related to tumor heterogeneity in which epigenetic alterations have a key role (Figure 1): (i) clonal mesteller@carrerasresearch.org heterogeneity; (ii) TME; (iii) spatial organization and intercellular crosstalk; (iv) differentiation and (M. Esteller). @ developmental programs (phenotypic plasticity); (v) metastasis; and (vi) the appearance of Twitter: @Marta_CasadoP (M. Casado-Pelaez), @AlbertoBuenoCo new resistance mechanisms to therapy. Thus, it is necessary to develop single cell resolution (A. Bueno-Costa), and @ManelEsteller technologies that allow us to understand the epigenetic cues that are otherwise undetectable (M. Esteller). 820 Trends in Cancer, October 2022, Vol. 8, No. 10 https://doi.org/10.1016/j.trecan.2022.06.005 © 2022 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Trends in Cancer OPEN ACCESS Box 1. Bulk methodologies to analyze epigenetic mechanisms Various bulk methodologies have been used to understand epigenetic mechanisms: (i) DNA methylation, taking advantage of bisulfite chemistry, can be analyzed by whole-genome bisulfite sequencing (WGBS), reduced representation bisulfite sequencing (RRBS), or 450k/850k Illumina methylation arrays [128]; (ii) DNA accessibility is mainly studied using the assay for transposase-accessible chromatin sequencing (ATAC-seq) [129]; and (iii) histone modifications and (iv) DNA–protein interactions can be studied by chromatin immunoprecipitation sequencing (ChIP-seq) [130]; (v) chromatin 3D architecture can be explored with multiple types of chromosome conformation capture technology, such as 3C, 4C, 5C, Hi-C, promoter-capture Hi-C, and ChIA-PET [131,132]. One important common drawback among these techniques is the need for a considerable sample size, demanding thousands to millions of cells as the minimal input. Thus, these techniques are considered ‘bulk methodologies’, by which we obtain an average value from the whole-cell bulk [133]. Various epigenetic deconvolution strategies can be applied to bulk data, but with a substantial risk of retrieving artifacts or losing difficult-to-detect minor subclones [134]. Nevertheless, bulk methodologies have been indispensable tools for our current understanding of epigenetics and its relationship with cancer. For example, they allowed for the methylation-based classification of diffuse gliomas (LGm1-LGm6) [135], the potential classification of cancers of unknown primary [123], and the histone modification- based tracking of cell differentiation states [136]. using bulk methodologies. In this review, we catalog current technologies that facilitate the study of different epigenetic characteristics at the single cell level. We classify each technology based on the epigenetic mechanism under study (DNA methylation, chromatin accessibility, histone modifications and DNA–protein interactions, and chromatin 3D architecture), focusing first on mono-omic methodologies (techniques that allow the study of only one epigenetic mechanism on a single cell) and then on multi-omic methodologies (which allow the study of multiple layers of information simultaneously on a single cell). In addition, we summarize currently available single cell spatial epigenomic methodologies and their potential in cancer research. We also highlight recent discoveries and insights gained from these single cell epigenetic technologies, how they can contribute to solve many current challenges in cancer research (mostly derived from tumor heterogeneity), their current limitations, and their potential in translational/clinical scenarios. Tumor heterogeneity and its relation to epigenetic alterations Clonal heterogeneity A tumor can comprise multiple malignant subclones, each with unique genetic and epigenetic properties [10,11]. As a cancer population evolves, cells accumulate genetic and epigenetic alter- ations that contribute to the appearance of new clones that may harbor novel, selective advantages Tumor cells CSC Clonal 6 1 heterogeneity 3 1 Tumor T-cell Macrophage 2 microenvironment B-cell Fibroblast Spatial 3 organization and crosstalk Differentiation 4 programs CSC EMT 5 5 Metastasis Tumor cell Migratory tumor cell 2 4 Resistance 6 mechanisms to therapy Trends in Cancer Figure 1. The six main aspects of tumor heterogeneity that are difficult to assess using bulk methodologies. Abbreviations: CSC, cancer stem cell; EMT, epithelial–mesenchymal transition. Trends in Cancer, October 2022, Vol. 8, No. 10 821 Trends in Cancer OPEN ACCESS (e.g., enhanced proliferation, resistance to therapy, invasiveness, etc.) [5]. The detection of these clones is crucial to understand tumor progression and its impact on clinical outcome. Single cell methodologies are able to detect each one of these clones (especially minor, difficult-to-detect, subclones), thus revealing valuable prognostic information. Tumor microenvironment A tumor does not comprise solely malignant cells but harbors myriad types of non-malignant cell with distinct roles in cancer progression. The T cell content is directly associated with tumor pro- gression in many cancer types, with cytotoxic T cells (Tc) and helper T cells (Th1, Th2, and Th17) correlating with good prognosis [12]. Tumor-associated macrophages have crucial roles in cancer progression, depending on their M1/M2 differentiation state [13]. Additionally, natural killer (NK) cells, B cells, endothelial cells, fibroblasts, and other cell types participate in this complex interactome [14]. These microenvironmental interactions profoundly modulate the epigenome of both tumoral and nontumoral cells, generating an epigenetic crosstalk that directly determines cancer progression [15]. Thus, studying these epigenetic signals at the single cell level is manda- tory to decipher this complex tumoral interactome. Spatial organization and intercellular crosstalk Malignant and non-malignant cells are not randomly distributed inside a tumor, but instead occupy specific positions in the tumoral space, generating discrete cell–cell interactions that impact disease progression [16]. Knowing the spatial distribution of each cell has been crucial for assessing the ‘heat’ of certain types of tumor (e.g., melanoma), in which the relative quantity and position of cytotoxic T cells are key determinants of cancer progression [17]. In addition, DNA methylation heterogeneity is dependent on the spatial organization of colorectal cancer (CRC) cells in patients with locoregional cancer, correlating with relapse-free and overall survival [18]. Microscopy techniques, such as immunohistochemistry. have enabled great advances in this aspect. Nevertheless, these techniques lack the resolution and specificity to unveil the different epigenetic characteristics for each cell. Thus, single cell-sequencing technologies (including cutting-edge spatial epigenomics) will enable researchers to infer how spatial cues correlate with epigenetic changes inside a tumor, which is of strong clinical value. Differentiation and developmental programs (phenotypic plasticity) The cancer stem cell hypothesis states that tumor growth depends, at least in part, on the asym- metrical divisions of malignant stem cells that differentiate to specific types of committed cancer cell [19]. In addition, depending on their epigenetic background, each malignant cell can follow a specific developmental program that will impact the progression of cancer. For example, in glio- blastoma, there are at least four types of malignant cell state program, some related to higher stemness (neural progenitor-like and oligodendrocyte progenitor-like cells), and others related to a more differentiated state (astrocyte-like and mesenchymal-like cells) [20]. The identity of each cell is maintained by epigenetic memory mechanisms (e.g., DNA methylation) that ensure full commitment to specific transcriptional programs [21]. Thus, detecting alterations in this epigenetic machinery at the single cell level may provide valuable information on potential malignant differentiation trajectories, predicting how the tumor may progress and deciding which type of treatment should be applied. Metastasis Some cancer cells acquire the ability to leave their primary site and colonize distant tissues, which is the cause of most cancer-related deaths [22]. From its transformation until its settlement on a new tissue, the metastatic cancer cell experiences drastic changes, such as acquiring a higher motility program (epithelial–to-mesenchymal transition), avoiding immune cell surveillance, and 822 Trends in Cancer, October 2022, Vol. 8, No. 10 Trends in Cancer OPEN ACCESS adapting to the new secondary site [22]. No genetic driver mutations specific to metastasis have yet been identified, suggesting that dynamic epigenetic mechanisms are involved in key steps of metastasis [4,23]. Single cell technologies will be useful to confidently detect in primary and secondary sites those cancer cells that have a metastatic-prone epigenetic background. Resistance mechanisms to therapy Certain malignant subclones that are undetectable by bulk methodologies due to their low abun- dance may harbor key mutations and epimutations that render them resistant to treatments that otherwise affect other more abundant subclones [24]. These resistant subclones are most likely to become the predominant ones after the first line of treatment, representing the most common cause of relapse. Alterations in epigenetic mechanisms have been strongly linked with antitumoral drug resistance [25]. For example, during bortezomib treatment in multiple myeloma, certain cancer subclones enter a slow-cycling, drug-tolerant reversible state, as a consequence of epi- genetic plasticity rather than of genetic determinants. Another case of non-genetically determined resistance to therapy are alterations in histone H3 lysine 4 demethylases, such as KDM5, which contribute to transcriptomic heterogeneity in breast cancer, leading to a decreased sensitivity to antiestrogens [26]. In taxane-resistant triple-negative breast cancer (TNBC), global DNA hypome- thylation and relocation of histone H3K27 trimethylation enable an epigenetic state that enables cancer cells to become resistant to paclitaxel, thus creating a new therapeutic vulnerability by using epigenetic inhibitors [27]. There are many more well-documented cases in which non- genetic determinants are the main drivers of the appearance of new resistant subclones [28]. Thus, detecting these resistant subclones early during diagnosis, using single cell epigenetic technologies, would significantly help clinicians to select the best treatment combinations. Additionally, the ability to detect minimal residual disease after treatment is fundamental, because it constitutes a prognostic biomarker that can predict relapse in some cancers [29]. Single cell mono-omic technologies to explore cancer epigenomics Single cell techniques encompass a breakthrough methodology that has revolutionized the way in which complex biological systems can be characterized by looking at one cell at a time. Single cell approaches are essential to properly examine the underlying complexity of tumors and explore cellular heterogeneity at several levels. With the advent of single cell RNA-sequencing (scRNA-seq), the transcriptome has become the molecular level most exploited by single cell technologies. It has accelerated our understanding of cancer biology, enabling the characteriza- tion of the intratumoral heterogeneity and cellular architecture of several cancer types at unprec- edented resolution [20,30–33]. Additionally, there are emerging single cell DNA-sequencing technologies that allow us to profile, in an amplicon-based and targeted manner, recurrently mutated genes, providing the genotype of every cell by detecting single nucleotide variants (SNPs) and copy number variants (CNVs) [34]. Nevertheless, transcriptional cell state diversity among malignant cells in a tumor is often independent of genetic clonal heterogeneity, highlighting the importance of developing epigenetic single cell analysis tools to assess this heterogeneity [35]. Although single cell techniques aimed at studying the epigenome have not evolved as rapidly compared with those for the transcriptome, new approaches are being developed to explore the different epigenetic mechanisms of gene regulation. DNA methylation 5-Methylcytosine (5mC) is the most well-known DNA modification. In mammals, this methylation mostly occurs in cytosines that are followed by a guanine, forming a 5′-to-3′ CpG pair. Approxi- mately 70% of all human gene promoters are enriched with multiple clustered CpG pairs, forming ‘CpG islands’ that are prone to 5mC methylation [36]. In these regions, methylation acts as a repressive switch, restricting gene expression. Additionally, 5mC can be found in other genomic Trends in Cancer, October 2022, Vol. 8, No. 10 823 Trends in Cancer OPEN ACCESS regions, such as gene bodies and distant regulatory regions (enhancers and CTCF sites), regulating their function in cis. In most types of cancer, DNA methylation is significantly deregulated. Promoter hyper/hypomethylation in tumor suppressors/oncogenes is a well-established driver of tumoral progression [37]. In addition, deregulation in enhancer methylation and other distant regulatory regions may have crucial implications in cancer by fostering tumoral epigenetic heterogeneity [38]. Bulk methodologies helped revolutionize our understanding in this area. Most of these method- ologies are based on the conversion of unmethylated cytosines to uracil after bisulfite treatment of the DNA. This allows the detection of methylated cytosines using sequencing or array-based methods [18]. However, as previously stated, bulk methodologies lack the resolution to unravel the epigenetic variability of a tumoral population. Therefore, in 2013, the first bisulfite-based sequencing methodology was developed to detect DNA methylation at the single cell level, single cell-reduced representation bisulfite sequencing (scRRBS) [39,40], which uses enzymatic cleav- age to generate DNA fragments with CpG-rich ends. Since then, multiple bisulfite-based single cell-sequencing technologies have been developed (Table 1): post-bisulfite adaptor tagging (PBAT) technology increases CpG coverage from 3–5% to an average of 18% unique CpGs (up to 48% at saturating sequencing depths) [41], enabling the establishment of new single cell DNA methylation-sequencing technologies, such as scBS-Seq [41], scWGBS [42], and scPBAT [43], with the expense of capturing less-enriched CpG regions and generating more PCR dupli- cates, especially when applying high-sequencing depths. Whether to use scRRBS-based or whole genome-based approaches depends on the user’s preference: scRRBS enables the capture of CpG-island regions (which are usually found in gene promoters), while single cell whole-genome methods capture other, less-enriched, CpG regions, although with higher CpG coverage. Interestingly, there are loci-specific, bisulfite-based alternatives, such as multiplexed- scAEBS, which enable the analysis by PCR of the methylation status of specific CpGs directly from single cells [44]. There are many other bisulfite-based mono-omic methods available, which are beyond the scope of the current review. Nevertheless, the reader can find a compre- hensive catalog of all these methods in Table 1. Unfortunately, bisulfite-based single-cell DNA methylation sequencing currently has significantly low coverage, among many other important limitations, summarized in Box 2. By contrast, unstable DNA modifications, such as 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC), and 5-carboxylcytosine (5caC), are intermediates that serve as useful fingerprints for detecting DNA regions that are being actively demethylated. Single cell detection methods for these intermediates, such as single-cell restriction endonuclease AbaSI sequencing (scAba-Seq, for detecting 5hmC) [45], chemical labeling-enabled C-to-T conversion sequencing (CLEVER-Seq, for 5fC) [46], and single cell methylase-assisted bisulfite sequencing (scMAB-Seq, for both 5fC and 5caC) [47], are promising but have not yet been applied in cancer research. In addition to bisulfite-based methodologies, novel nonbisulfite-based approaches have been developed, mostly based on methylation-sensitive restriction enzymes (MSREs). The first one developed was the restriction enzyme-based single-cell methylation assay (RSMA), which inter- rogates the CpG methylation status of specific genomic loci [48]. Some years later, genome- wide MSRE-based single cell technologies were developed, such as single-cell CpG-island sequencing (scCGI-seq) and epigenomics and genomics of single cells analyzed by restriction (epi-gSCAR), achieving up to 18.8% genome coverage [49,50]. Additionally, a novel targeted bisulfite-free method, named single cell-targeted analysis of the methylome (scTAM-seq), directly profiles 650 specific CpG sites in up to 10 000 cells using a commercial microfluidic platform [51]. By avoiding the use of bisulfite, we can ensure better DNA integrity, although nonbisulfite-based technologies still exhibit similar coverage to those that are bisulfite based. 824 Trends in Cancer, October 2022, Vol. 8, No. 10 Trends in Cancer OPEN ACCESS Table 1. Currently available mono-omic methodologies for epigenetic single cell sequencing and spatial epigenomics Epigenetic Technique Designed Current applications in cancer mechanism by DNA methylation scRRBS [39,40] Chronic lymphocytic leukemia [52], glioma [53] scXRBS [145] Acute myeloid leukemia cell lines [145] scBS-seq [41] Circulating tumor cells [146], CRC [54] scWGBS [42] Chronic myeloid leukemia cell lines [42], acute promyelocytic leukemia cell line [42], circulating metastatic tumor cells [56] scPBAT [43] N/Aa PBAL [147] Acute myeloid leukemia [148] snmC-seq [149] CRC cell line [150] snmC-seq2 [151] N/A sci-MET [152] N/A MID-RRBS [153] N/A BRIF-seq [154] N/A scAba-seq [45] N/A (5hmC) CLEVER-seq [46] N/A (5fC) scMAB-seq [47] N/A (5fC/5caC) RSMA [48] CRC cell line [48] scCGI-seq [49] Chronic myeloid leukemia cell line [49] epi-gSCAR [50] Acute myeloid leukemia cell lines [50] scAEBS [44] Circulating tumor cells [44] scTEM-seq [155] Acute myeloid leukemia cell line [155] scTAM-seq [51] N/A Chromatin scATAC-seq [58,59] Basal cell carcinoma [59], multiple myeloma [61], prostate Accessibility (microfluidics cancer [62], triple-negative breast cancer [63,64], based) drug-resistant lung cancer [65], gynecological cancers [66], glioblastoma [156], renal cancer [157], breast metastasis [158], chronic lymphocytic leukemia [71], CRISPR perturbation [68–70], Barret metaplasia [159], bladder cancer [160], drug-resistant leukemic T cells [161], lung adenocarcinoma metastasis [162] scATAC-seq [60] N/A (plate based) sciATAC-seq [163] Mouse lung adenocarcinoma [164] SNuBar-ATAC [165] Lung cancer [165] HyDrop-ATAC [166] N/A scMNase-seq [73,74] N/A scGET-seq [167] Cervical cancer cell line [167] sciMAP-ATAC [109] N/A Spatial [110] N/A ATAC-seq Histone scChIP-seq [75] Breast cancer [77,78] modifications scChIC-seq [76] N/A and DNA-protein interactions iscChIC-seq [168] N/A (continued on next page) Trends in Cancer, October 2022, Vol. 8, No. 10 825 Trends in Cancer OPEN ACCESS Table 1. (continued) Epigenetic Technique Designed Current applications in cancer mechanism by scChIL-seq [169] Breast cancer cell line [169], chronic myeloid leukemia cell line [169] ACT-seq [170] N/A CoBATCH [171] N/A uliCUT&RUN [172] N/A scCUT&Tag [173,174] Glioblastoma [175], various cancer cell lines [176] AutoCUT&Tag [79] Mixed-lineage leukemia [79], chronic myeloid leukemia cell line [79] sciTIP-seq [177] CRC cell line [177] REpi-seq [178] N/A Spatial-CUT&Tag [107] N/A Epigenomic [108] N/A MERFISH Chromatin 3D scHiC [82,83] Cervical cancer cell line [83], chronic myeloid leukemia cell interactions line [83] scDAM-ID [84] Chronic myeloid leukemia cell line [84] a N/A indicates methodologies that have not yet been applied in cancer research. Although single cell DNA methylation-sequencing technologies are still nascent, they have already provided new crucial insights into important questions about tumor heterogeneity, such as intratumoral epigenetic diversity, TME, and metastasis. Using scRRBS on patient-derived chronic lymphocytic leukemia (CLL) [52] and glioblastoma samples [53], we can observe the significant intratumoral epigenetic diversity of cancer cells, showing that important epigenetic routes signif- icantly contribute to tumor heterogeneity. scBS-seq in CRC tumors showed that some partial methylation domains (PMDs) exhibit extensive epigenetic heterogeneity between cells of the same tumor, which can be explained by TME cues [54]. scBS-seq has even been applied to the study of patients with common variable immune deficiency (CVID), in which B cell differentiation is impaired and increases the risk of malignancy by up to 12-fold [55]. Lastly, scWGBS showed that circulating tumor cells (CTCs), which form clusters to increase their metastatic potential, exhibit Box 2. Current limitations of single cell epigenetic-sequencing technologies Enormous efforts have been made to develop epigenomic tools capable of profiling, individually and simultaneously, the different epigenetic layers at a single cell resolution. However, the application of these technologies is still nascent and not well established. This is partly due to the current challenges these approaches pose, such as low throughput, limited coverage per cell, and elevated costs [137]. Many experimental issues need to be tackled, such as amplification bias, differences of library sizes across samples, and DNA damage upon processing (e.g., bisulfite treatment), among others. Additionally, the numerous techniques and methodologies used for the same purpose (e.g., there are more than ten epigenetic techniques for only studying single cell DNA methylation, see Table 1 in the main text) emphasizes the lack of standardization and the need to perform benchmarking across these technologies. Conversely, computational analysis represents a major challenge due to the nature of single cell data, characterized by large-scale dimensions and high intrinsic noise and sparsity. Since the resulting data matrixes are often sparse (due to low coverage), most approaches cannot be loci specific (e.g., in single cell DNA methylation), but instead clusters are generated (t-SNEs/UMAPs) after high-dimensional data reduction. To compensate for missing values, imputation (e.g., KNN), pseudo-bulking, and genomic binning are often used. It is also necessary to evaluate the integration of multi-omics data, including paired multimodal data. New multi-omic integrative pipelines were recently designed to facilitate the analysis of multiple layers of epigenetic data simultaneously, such as MOFA+ [138,139], scMVP [140], Babel [141], Maestro [142], or EpiScanpy [116]. In addition, new user-friendly interfaces, such as ShinyArchR for scATAC-seq [143] or ChromSCape for scHi-C [144], will make analysis more accessible for nonbioinformaticians. 826 Trends in Cancer, October 2022, Vol. 8, No. 10 Trends in Cancer OPEN ACCESS significant promoter hypomethylation of key stemness-/proliferation-associated genes, mimicking embryonic stem cell biology; treatment with Na+/K+ ATPase inhibitors leads to DNA methylation of these regions, resulting in CTC dissociation and metastasis suppression [56]. Chromatin accessibility Genome-wide DNA accessibility assays facilitate the study of epigenomic alterations by identifying DNA regulatory elements (REs), such as gene promoters and enhancers, involved in the repression or activation of gene expression. These noncoding regions are relevant to understand cancer biology and identify pathways controlling tumorigenesis [57]. The most used technique to investigate this at the single cell resolution is single-cell ATAC-seq (scATAC-seq), or an alternative modality based on combinatorial cellular indexing (sciATAC-seq). Both methods are based on inserting sequencing adapters into accessible genome areas to facilitate Tn5 transposition as a measure of open chromatin [58]. There are two main protocols: microfluidics techniques encompassing either droplet-based [59] or plate-based techniques [60]. Several studies have used scATAC-seq, often in combination with scRNA-seq, to dissect the intrinsic properties of different cancer types and associate chromatin accessibility variation with gene expression. Following this approach, a study investigated the mechanisms behind cellular plasticity in multiple myeloma, highlighting the acquisition of aberrant transcriptional states within cancer cells pro- moted by epigenetic reprogramming affecting cell chromatin landscape and enhancer usage. Owing to this unprecedented resolution, it was observed that these different transcriptional states coexist within individual myeloma cells, conferring them a plasticity that leads to dedifferentiation and that promotes the expression of surface markers that are not specific to the lineage and that could represent new immunotherapeutic targets. Moreover, this study showed that myeloma cells dedifferentiate independently of their genotype, revealing the possibility to target epigenomic rather than genomic states. These insights show the potential of scATAC-seq to unveil tumoral epigenomic heterogeneity and to discover new actionable targets [61]. With the same methodology, a study conducted in prostate cancer identified pre-existing and treatment-persistent cell subsets with regenerative properties after exposure to therapy [62]. Additionally, scATAC-seq profiling of the TME in basal cell carcinoma facilitated the characterization of regulatory networks in immune, stroma, and malignant cells. This technique enabled the identification of regulatory mechanisms related to T cell exhaustion in tumor-infiltrating lymphocytes [59]. scATAC-seq is also useful to understand variations in the response to cancer treatment. For example, profiling of chromatin accessibility and the transcriptome of immune cells in patients with TNBC treated with chemotherapy alone or in combination with checkpoint inhibitors revealed that high levels of baseline CXCL13+ T cells had a pivotal role in effective responses to the combination treatment [63]. scATAC-seq also revealed synthetic lethalities and epigenetic resistance mechanisms after treatment in TNBC and lung cancer [64,65]. Additionally, the de novo acquisition of RE, analyzed by scATAC-seq, has been reported to drive oncogenic pathways in gynecological malignancies [66]. Chromatin accessibility is also relevant to differentiate between mature and stem-like populations, being the key determinants of the latter in many cancers. A recent study in primary glioblastoma using scATAC-seq revealed that tumor-initiating cells are heterogeneous, exhibiting three differ- ent states associated with survival, which are controlled by unique transcription factors (TFs) [67]. Single cell chromatin accessibility profiling can be used in combination with CRISPR/Cas9 technology to detect changes in the epigenome preceded by CRISPR perturbation and explore TF-binding dynamics in cancer [68–70]. More recently, a new protocol involving the modification Trends in Cancer, October 2022, Vol. 8, No. 10 827 Trends in Cancer OPEN ACCESS of scATAC-seq to enrich transposable-accessible mitochondrial DNA and to infer tumoral clonal dynamics has been developed [71,72]. In addition to these approaches, another technique, called single cell micrococcal nuclease sequencing (scMNase-seq), allows the measurement of chromatin compaction as well as nucle- osome positioning in single cells [73,74]. To do so, the linker DNA between the nucleosome cores is digested by MNase, enabling subsequent sequencing of the protected DNA regions. Histone modifications and DNA–protein interactions Another layer of epigenetic regulation is orchestrated by a range of post-translational modifica- tions in histones, which lead to transcriptionally permissive or repressive chromatin states. The characterization of these histone marks at a single cell resolution can help to properly dissect regulatory heterogeneity within cancerous cells, enabling the detection of rare chromatin states that might be associated with drug resistance and relapse. This is achievable owing to the development of several techniques that facilitate the profiling of histone modifications in single cells. One such technique is single-cell ChIP-seq (scChIP-seq), which can map histone marks, TFs, and other DNA-interacting proteins within the genome [75]. Similarly, single-cell chromatin immunocleavage sequencing (scChIC-seq) is a method to analyze histone modifications at a single cell resolution. By the binding of MNase to a specific antibody that targets a histone modification of interest, the surrounding DNA is cleaved into small fractions followed by sequencing [76]. Other novel techniques are detailed in Table 1. These methodologies can facilitate the segmentation of tumor cells based on their histone modification profiles, fostering the identification of rare populations with distinct properties. By applying scCHIP-seq in patient-derived xenograft models of TNBC, a repressive H3K27me3 mark was found to be depleted in several genes associated with chemotherapy resistance in a subset of resistant cells. Interestingly, these chromatin features were also found in a fraction of cells from the untreated drug-sensitive tumor, suggesting that cells with epigenetic characteris- tics of drug resistance already exist before therapy [77]. Another recent study in TNBC, also using scCHIP-seq, showed that cancer cells exhibit bivalent chromatin states with permissive H3K4me3 and repressive H3K27me3 modifications, indicating that those cells that manage to deplete H3K27me3 marks are prone to becoming chemotolerant. The authors also showed that treating cells with a combination of a H3K27me3 demethylase inhibitor and chemotherapy reduced the number of persistent cells, suggesting an enhanced therapeutic approach for TNBC [78]. Finally, a similar study was performed in samples from patients with mixed-lineage leukemia using an automatized variant of scCUT&Tag (AutoCUT&Tag) for studying the leukemo- genic histone methyltransferase KMT2A, uncovering the epigenetic heterogeneity of patient samples and predicting sensitivity to DOT1L pharmacological inhibition [79]. These promising discoveries reinforce the potential of the aforementioned methodologies to study tumor cell epigenomes from a single cell perspective. Chromatin 3D architecture The 3D structure of the genome is governed by chromosome conformation and folding within the nucleus and has been reported to have a significant role in the regulation of gene expression by, for instance, promoting the interaction of enhancers and promoters that regulates the expression of specific genes [80]. The nuclear architecture comprises chromosomal compartments that facilitate the emergence of trans-regulatory elements, long-range loops, topologically associating domains (TADs), and lamina-associated domains (LADs). In some cancers, this genome architec- ture is disrupted due to genomic rearrangement or structural variations, affecting the regulatory landscape of the cancer cell [81]. 828 Trends in Cancer, October 2022, Vol. 8, No. 10 Trends in Cancer OPEN ACCESS Single cell Hi-C (scHi-C) enables in-depth interaction profiling by relying on proximity ligation followed by massively parallel sequencing. It helps to decipher millions of loci concomitantly and assess cell–cell variability in terms of chromosome structure within the nucleus [82,83]. Although there are many studies that have used Hi-C to explore the genome architecture of different types of cancer, they have been performed at the bulk level. Additionally, a technology called single cell DNA adenine methyltransferase identification (scDamID), enables genome- wide mapping of LADs at a single cell resolution [84]. Currently, neither single cell 3D approaches have been applied in cancer research. Single cell epigenetic multi-omic approaches Complex biological systems, such as cancer, cannot be fully understood by either exploration at the bulk level or by inspection of individual layers of information. Therefore, it is crucial to integrate different ‘omics at a single cell resolution to decipher the mechanisms and regulatory heterogeneity that drive cancer development and progression (Figure 2) [85]. The recent development of tools has permitted more comprehensive epigenomic analysis by simultaneously exploring several layers of the epigenome in individual cells. Single-cell Nucleosome Trends in Cancer Figure 2. Single cell-sequencing and spatial technologies to study the different epigenetic mechanisms in cancer. Abbreviations: ACT-seq, antibody-guided chromatin tagmentation sequencing [170]; CLEVER-seq, chemical- labeling-enabled C-to-T conversion sequencing [46]; CoBATCH, combinatorial barcoding and targeted chromatin release [171]; epigenomic MERFISH, epigenomic multiplexed error robust fluorescence in situ hybridization [108], scABA-seq, single cell restriction endonuclease AbaSI sequencing [45]; scATAC-seq, single cell sequencing assay for transposase-accessible chromatin [59]; scBS-seq, single cell bisulfite sequencing [41]; scChIC-seq, single cell chromatin immunocleavage sequencing [76]; scChIL-seq, single cell chromatin integration labeling [169]; scChIP-seq, single cell chromatin immunoprecipitation followed by sequencing [75]; scCUT&Tag, single cell cleavage under targets and tagmentation [174]; scDamID, single cell DNA adenine methyltransferase identification [84]; scHi-C, single cell Hi-C [82,83]; sciMAP-ATAC, single cell combinatorial indexing on microbiopsies assigned to positions for the assay for transposase accessible chromatin [109]; sci-MET, single cell combinatorial indexing for methylation analysis [152]; scMAB-seq, single cell methylase-assisted bisulfite sequencing [47]; scMNase-seq, single cell micrococcal nuclease sequencing [74]; scRRBS-seq, single cell reduced representation bisulfite sequencing [39,40]; scTAM-seq, single cell targeted analysis of the methylome [51]; scTEM-seq, single cell transposable element methylation sequencing [155]; spatial ATAC-seq, spatially resolved chromatin accessibility profiling [110]; spatial CUT&Tag, spatial histone modification profiling with cleavage under targets and tagmentation [107]; uliCUT&RUN, ultra-low-input cleavage under targets and release using nuclease [172]. Trends in Cancer, October 2022, Vol. 8, No. 10 829 Trends in Cancer OPEN ACCESS Occupancy and Methylome sequencing (NOME)-seq enables the profiling of chromatin accessibil- ity and DNA methylation status of each cell [86]. Furthermore, it is also possible to simultaneously analyze the chromatin 3D structure and the methylome in individual cells with tools such as scMethyl-HiC and sn-m3C-seq [87,88]. Additionally, single-cell Chromatin Overall Omic-scale Landscape Sequencing (scCOOL-seq) can provide a more in-depth characterization of the epige- nome, by exploring the nucleosome positioning, chromatin accessibility, DNA methylation, copy number alterations (CNAs), and ploidy in each single cell [89]. A variation of scCOOL-seq that also combines scRNA-seq (scCOOL-seq + scRNA-seq) allowed the identification of two novel prognosis pancreatic ductal adenocarcinoma biomarkers (ZNF667 and ZNF667-AS1) [90]. As a side note, in a similar fashion to some bulk DNA methylation array-based methods [91], CNAs can also be inferred from mono-omic single cell DNA methylation-sequencing technologies, by splitting the genome in windows/bins of equal size, calculating the number of CpGs detected per window, and normalizing by the total number of detected CpGs per cell, using non-malignant cells as a baseline reference [35]. It is relevant to combine epigenomic approaches with the simultaneous interrogation of the transcriptome to understand how the epigenome landscape of a cell is reflected in its transcrip- tional state. In most cases, expression heterogeneity is not fully explained by genetic alterations, but rather by cell–cell epigenetic cues. An example of this is the joint multiplexed-scRRBS/ scRNA-seq dual-protocol (Smart-RRBS), optimized by the Landau lab, which allows the simulta- neous interrogation of DNA methylation, transcriptomics, and CNAs in a single cell and has been successfully used to unravel the epigenetic clonal heterogeneity found in CLL and glioblastoma cells [35,92]; both of these studies show how important the intratumoral epigenetic heterogeneity in cancer progression is: in the first study, Gaiti et al. demonstrated that epigenetic information enables the mapping of CLL lineage history and the prediction of its evolution after therapy [92]. In the second study, Chaligne et al. demonstrated the epigenetic heritability in glioma cells, with key differences in cell plasticity states between IDH-mutant cells and IDH-wild-type glioblastoma [35]. Another multi-omic protocol is scTrio-seq, which also interrogates single cell DNA methylome, transcriptome, and CNAs, and has enabled study of tumoral heterogeneity by identifying epigenet- ically diverse cancer subpopulations in samples from patients with either CRC or hepatocellular cancer that may reflect the different lineages in the tumor [93,94]. There are many additional tools that combine single cell transcriptomics with DNA methylation (scM&T-seq [95], scTrio-seq [94], and snmCT-seq [96]), chromatin accessibility (scCAT-seq [97] and joint scATAC+RNA-seq [98]), DNA–protein interactions (scDam&T-seq [99]) and multiple epigenomic layers (scNMT-seq [100] and scNOMeRe-seq [101]). Every additional piece of information that we could gain in a single cell will allow a better under- standing of the co-occurrence of certain genetic/transcriptomic/epigenomic states in a tumor. Currently, scCOOL-seq coupled with scRNA-seq [90] is the one technique that simultaneously analyzes the most layers. There are more multi-omic approaches available, catalogued in Table 2. Spatial epigenomics Although all these technologies provide relevant information on the epigenome of each individual cell, its spatial localization within the tumor is lost upon dissociation. Spatial technologies based on RNA, DNA, multiplexed fluorescence, and isotope labeling now enable exploration of the cell type composition of a tissue preserving the spatial information [102–104]. Incorporation of the spatial context has helped investigation of the architecture of several tumors, as well as understanding how both malignant and non-malignant cells, such as stromal and immune 830 Trends in Cancer, October 2022, Vol. 8, No. 10 Trends in Cancer OPEN ACCESS Table 2. Currently available multi-omic methodologies for epigenetic single cell sequencing Technique Basis Designed by Current applications in cancer scM&T-seq Methylome + transcriptome + [95] N/Aa genotype scMT-seq Methylome + transcriptome [179] N/A Smart-RRBS Methylome + transcriptome + [92,180] Chronic lymphocytic leukemia CNA [91], Glioblastoma [35] scmCT-seq Methylome + transcriptome [96] N/A scGEM Methylome + transcriptome + [181] Lung cancer [181] genotype scTrio-seq Methylome + transcriptome + [94] CRC [93], hepatocellular CNA carcinoma [94] scNMT-seq Methylome + transcriptome + [100] N/A accessibility scNOME-seq Methylome + accessibility [86] Chronic myeloid leukemia cell line [86] scNOMeRe-seq Methylome + transcriptome + [101] N/A accessibility scCOOL-seq Methylome + accessibility + [89] N/A CNA + ploidy iscCOOL-seq Methylome + accessibility + [182] N/A CNA + ploidy scCOOL-seq + scRNA-seq Methylome + accessibility + [90] Pancreatic ductal carcinoma CNA + transcriptome [90] scCAT-seq Accessibility + transcriptome [97] Lung cancer [97], various cancer cell lines [97] Joint Accessibility + transcriptome [98] Chronic myeloid leukemia cell scATAC-Seq/scRNA-seq line [98] scDAM&T-seq Protein-DNA interactome + [99] Chronic myeloid leukemia cell transcriptome line [99] T-ATAC-seq Accessibility + TCR [183] Acute T cell leukemia cell line sequencing [183] scMethyl-HiC Methylome + chromatin [86] N/A conformation sn-m3C-seq Methylome + chromatin [87] N/A conformation SNARE-seq Accessibility + transcriptome [184] Chronic myeloid leukemia cell line [184] Paired-seq Accessibility + transcriptome [185] Hepatocellular carcinoma cell line [185] Paired-Tag Histone modifications + [186] N/A transcriptome scChaRM-seq Methylome + transcriptome + [187] N/A accessibility NEAT-seq Accessibility + transcriptome + [188] Chronic myeloid leukemia cell nuclear proteomics line [188] ASAP-seq Accessibility + proteomics + [189] N/A mitochondrial DNA capture Spear-ATAC Accessibility + short guide [70] N/A RNA integration detection Pi-ATAC Accessibility + proteomics [190] Chronic myeloid leukemia cell line [190] (continued on next page) Trends in Cancer, October 2022, Vol. 8, No. 10 831 Trends in Cancer OPEN ACCESS Table 2. (continued) Technique Basis Designed by Current applications in cancer scCUT&TAG-pro DNA–protein interactions + [191] N/A surface proteins SHARE-seq Accessibility + transcriptome [192] N/A Multi-scale multiplexed FISH Transcriptome + chromatin [193] N/A imaging technology for conformation simultaneous imaging of genomic loci alone or with nascent RNA transcripts Multimodal approach Histone modifications + [194] N/A (DNA seqFISH+, multiplexed transcriptome + chromatin immunofluorescence + RNA conformation seqFISH) a N/A indicates methodologies that have not yet been applied in cancer research. cells, are displayed within the TME [105,106]. The opportunity to leverage this methodology by also spatially profiling epigenomic states within the tissue would provide exceptional resolution of the mechanisms driving cancer evolution. Recently, a study presented a spatial histone modification-profiling approach based on a combination of in-tissue deterministic barcoding and cleavage under targets and tagmentation (CUT&Tag) chemistry, called spatial-CUT&Tag. By genome-wide mapping of different histone modifications across a tissue section of mouse embryos, the authors were able to recapitulate the spatial position of the different cell types and characterize how epigenetic regulation is displayed throughout the tissue [107]. Epigenomic multiplexed error robust fluorescence in situ hybridization (Epigenomic MERFISH) is another recent tool for spatially histone modification profiling, which involves in situ tagmentation and tran- scription followed by highly multiplexed imaging. This high-resolution targeted approach allows mapping of single cell epigenomes in a tissue region [108]. In addition, there are spatially resolved epigenomics approaches that rely on measuring chromatin accessibility. Single cell combinatorial indexing on Microbiopsies Assigned to Positions for the Assay for Transposase Accessible Chromatin (sciMAP-ATAC) is a method that profiles the chromatin states of individual cells in a cubic region defined by a microbiopsy punch [109]. Finally, spatial ATAC-seq is another recent method for chromatin accessibility mapping that can reveal the epigenetic landscape at cellular level over a tissue section [110]. Although these developing methodologies are recent and have not yet been applied in cancer research, they hold the potential to explore how regulatory programs occur throughout the tumor tissue. By incorporating the spatial context, we will be able to visualize how cancer subclones with distinct epigenetic properties, such as stem-like subclones, are localized and investigate whether their position compromises their future clonal evolution and expansion. Apart from spatial tumoral heterogeneity, they will also contribute to understanding how cancer cells interact with each other and with the TME. This crucial information can shed light on epigenetic mechanisms behind the variability in patient outcomes and resistance to treatments, such as the aberrant regulatory pro- grams that facilitate cancer cell evasion of the immune system. Another potential benefit of these technological breakthroughs is the possibility to improve or design new therapeutic approaches by leveraging the localization of the subclones with distinct regulatory mechanisms. Nevertheless, these spatial epigenomics technologies work only on fresh tissue sections. Currently, it is possible to analyze formalin-fixed paraffin-embedded (FFPE) samples for spatial transcriptomics [111]. Given that FFPE is the standard format to archive and preserve clinical 832 Trends in Cancer, October 2022, Vol. 8, No. 10 Trends in Cancer OPEN ACCESS samples, FFPE-based spatial epigenomics techniques would represent a breakthrough in this Outstanding questions field. [112]. How can we overcome and improve experimental limitations, such as limited throughput, low coverage, and high Concluding remarks and future perspectives costs, of single cell epigenomics The emergence of single cell-sequencing technologies over the past few years has opened a new technologies? dimension for how to study cancer biology. Genetic and transcriptomic high-throughput single cell-sequencing technologies have already proven to be extraordinary tools to ultimately dissect Will robust computational tools capable of analyzing and integrating all single cell tumors cell by cell, ‘atomizing’ the complex tumoral heterogeneity that otherwise would be epigenomic layers of information, also impossible to decipher [8]. Nevertheless, cancer is, by nature, an epigenetic disease, the progres- combined with other single cell ‘omics sion of which is largely dependent on non-genetic determinants [5]. Indeed, epigenetics is a data, be developed in the near future? crucial mechanistic layer without which essential information about tumor heterogeneity would Will single cell epigenomics approaches be lost. Scientists across the globe have developed dozens of new technologies that allow the provide new insights that promote and analysis epigenetic mechanisms at the single cell level, including the development of multi-omic accelerate the development of novel approaches to simultaneously unravel multiple layers of information (Tables 1 and 2). Unfortu- therapies in cancer? nately, due to their novelty and complexity, most of these technologies have not yet been applied How is epigenomic heterogeneity in cancer research (see Outstanding questions). Nonetheless, there are extraordinary examples displayed across the cellular distribution showing how useful epigenetic single cell technologies are at unraveling tumoral heterogeneity, of a tumor and how is this related to the encouraging the development of new, more fine-tuned methodologies that may overcome cellular crosstalk between cancerous cells and with the TME? current limitations (Box 2). The creation of automatized, high-throughput platforms for single- cell epigenetics, similar to those already applied for microfluidics-based scATAC-seq [59], would help to standardize this process, making it more ‘user-friendly’ and more available to any type of research laboratory. Another possible alternative would be the use of novel long-read, third-generation sequencing technologies, such as Oxford Nanopore, especially at detecting single cell DNA methylation, which would not require any bisulfite conversion or PCR amplification steps (using PCR-free based barcoding), thus avoiding DNA degradation and elongation biases [113]. This technology is already being used to detect DNA 5mC methylation at the bulk level (PromethIONTM Nanopore platform) [114]. In addition, a new method for sequencing the tran- scriptome at the single cell level, named scCOLOR-seq, was recently developed using Nanopore technology [115]. Despite this progress, no epigenetic long-read single cell sequencing has yet been developed. Furthermore, there is a need for benchmarking and standardization of bioinfor- matic pipelines when analyzing single cell epigenetic-sequencing data, making this process more comprehensible for all bioinformaticians [116]. Despite these limitations, single cell epigenetic technologies have a bright future ahead, especially when we consider all the potential benefits that they could offer from a clinical viewpoint [117]. At present, several bulk epigenetic-based tests are used as diagnostic/prognostic tools in cancer. For example, MGMT and MLH1 promoter methylation analysis by methylation-sensitive PCR (MSP) or pyrosequencing is commonly used to prognosticate and diagnose patients with glio- blastoma and sporadic CRC, respectively [118,119]. Continuing with glioblastoma, nearly all IDH-mutant gliomas manifest a positive CpG island methylator phenotype (G-CIMP+), linking this DNA methylation status to a favorable prognosis [120]. In addition, and similar to the well- known ColoPrint and MammaPrint gene expression-based CRC and breast cancer classifiers [121,122], several epigenetic DNA methylation-based signatures show promising clinical results, as in the classification of cancers of unknown primary using bulk-based DNA methylation arrays [123] or in the early detection of blood-circulating tumor cells from patients with CRC, breast or prostate cancer [44,124,125]. Even though these bulk-based epigenetic clinical approaches are useful, their major drawback lies in the unsolved tumor heterogeneity. Under the bulk average may lie important low-abundant, undetected subclones that are key for the prognosis of certain cancers, such as glioblastoma, one of the most heterogeneous cancers [126]. Additionally, the different types of non-malignant cell in the TME also have a significant clinical impact on patient Trends in Cancer, October 2022, Vol. 8, No. 10 833 Trends in Cancer OPEN ACCESS outcome, such as in melanoma, where determining whether the tumor is ‘hot’ or ‘cold’ (tumors with high or low content of immune cells, respectively) is crucial for predicting responses to immu- notherapy [127]. Thus, applying single cell epigenetic technologies will enable the discovery of new predictive/ diagnostic biomarkers to detect these subclones, to develop highly specific and personalized therapeutic strategies that will avoid potential resistance to therapy, and even detecting early metastatic events. In addition, since FFPE-archived samples are the most common way to pre- serve patient biopsies, it is necessary to develop suitable tools to study FFPE samples at the single cell level in a retrospective manner. Of course, many obstacles lie along this long road to clinics: first, these technologies are economically expensive and, thus, unaffordable by hospital services; second, the lack of commercially standardized and automatized protocols; and lastly, the enormous complexity of the generated data calls for simplified, targeted-based analyses. It is difficult to say whether this technology will be applied in clinics in the immediate future, but what is clear is that all the knowledge that we will gain from these single cell epigenetic approaches over the next few years will be groundbreaking, in terms of both basic and clinical scenarios. Acknowledgments We thank CERCA Programme/Generalitat de Catalunya for institutional support. A.B.C. is a fellow of the Spanish Ministry of Education and Vocational Training, under FPU contract no. FPU17/02423. M.E. is an ICREA Research Professor. This work was supported by the Health Department PERIS-project no. SLT/002/16/00374 and AGAUR-projects no. 2017SGR1080 of the Catalan Government (Generalitat de Catalunya); Ministerio de Ciencia e Innovación (MCI), Agencia Estatal de Investigación (AEI) and European Regional Development Fund (ERDF) project no. RTI2018-094049-B-I00; the Cellex Foundation; and ‘la Caixa’ Banking Foundation (LCF/PR/GN18/51140001). Declaration of interests M.E. is a consultant for Ferrer International and Quimatryx. The other authors declare no conflict of interest. References 1. Hanahan, D. (2022) Hallmarks of cancer: new dimensions. Cancer 13. Baghban, R. et al. (2020) Tumor microenvironment complexity Discov. 12, 31–46 and therapeutic implications at a glance. Cell Commun. Signal 2. Turajilic, S. et al. (2019) Resolving genetic heterogeneity in can- 18, 59 cer. Nat. Rev. Genet. 20, 404–416 14. Anderson, N.M. and Simon, M.C. (2020) The tumor microenvi- 3. Shaffer, S.M. et al. (2017) Rare cell variability and drug-induced ronment. Curr. Biol. 30, R921–R925 reprogramming as a mode of cancer drug resistance. Nature 15. Marks, D.L. (2016) Epigenetic control of the tumor microenvi- 546, 431–435 ronment. Epigenomics 8, 1671–1687 4. Chen, J.F. and Yan, Q. (2021) The roles of epigenetics in cancer 16. Noble, R. et al. (2022) Spatial structure governs the mode of progression and metastasis. Biochem. J. 478, 3373–3393 tumour evolution. Nat. Ecol. Evol. 6, 207–217 5. Nam, A.S. et al. (2021) Integrating genetic and non-genetic 17. Trujillo, J.A. et al. (2018) T cell-inflamed versus non-T cell- determinants of cancer evolution by single-cell multi-omics. inflamed tumors: a conceptual framework for cancer immuno- Nat. Rev. Genet. 22, 3–18 therapy drug development and combination therapy selection. 6. Allis, C.D. and Jenuwein, T. (2016) The molecular hallmarks of Cancer Immunol. Res. 6, 990–1000 epigenetic control. Nat. Rev. Genet. 17, 487–500 18. Martínez-Cardús, A. et al. (2016) Epigenetic homogeneity within 7. Darwiche, N. (2020) Epigenetic mechanisms and the hallmarks colorectal tumors predicts shorter relapse-free and overall survival of cancer: an intimate affair. Am. J. Cancer Res. 10, 1954–1978 times for patients with locoregional cancer. Gastroenterology 151, 8. Yalan, L. et al. (2021) Applications of single-cell sequencing in 961–972 cancer research: progress and perspectives. J. Hematol. 19. Lim, J.R. et al. (2021) Cancer stem cell characteristics and their Oncol. 14, 91 potential as therapeutic targets. Med. Oncol. 38, 76 9. Dagogo-Jack, I. and Shaw, A.T. (2018) Tumour heterogeneity 20. Neftel, C. et al. (2019) An integrative model of cellular and resistance to cancer therapies. Nat. Rev. Clin. Oncol. 15, states, plasticity, and genetics for glioblastoma. Cell 178, 81–94 835–849 10. McGranahan, N. and Swanton, C. (2017) Clonal heterogeneity and 21. Lee, H.J. (2014) Reprogramming the methylome: erasing memory tumor evolution: past, present, and the future. Cell 168, 613–628 and creating diversity. Cell Stem Cell 14, 710–719 11. Oakes, C.C. et al. (2014) Evolution of DNA methylation is linked 22. Fares, J. et al. (2020) Molecular principles of metastasis: a hall- to genetic aberrations in chronic lymphocytic leukemia. Cancer mark of cancer revisited. Signal Transduct. Target. Ther. 5, 28 Discov. 4, 348–361 23. Patel, S.A. and Vanharanta, S. (2017) Epigenetic determinants 12. Tay, R.E. et al. (2021) Revisiting the role of CD4 + T cells in of metastasis. Mol. Oncol. 11, 79–96 cancer immunotherapy-new insights into old paradigms. 24. Wang, X. et al. (2019) Drug resistance and combating drug Cancer Gene Ther. 28, 5–17 resistance in cancer. Cancer Drug Resist 2, 141–160 834 Trends in Cancer, October 2022, Vol. 8, No. 10 Trends in Cancer OPEN ACCESS 25. Hayashi, T. and Konishi, I. (2021) Correlation of anti-tumour Published online April 12. 2022. https://doi.org/10.1101/ drug resistance with epigenetic regulation. Br. J. Cancer 124, 2022.04.11.487648 681–682 52. Pastore, A. et al. (2019) Corrupted coordination of epigenetic 26. Hinohara, K. et al. (2018) KDM5 histone demethylase activity links modifications leads to diverging chromatin states and transcrip- cellular transcriptomic heterogeneity to therapeutic resistance. tional heterogeneity in CLL. Nat. Commun. 10, 1874 Cancer Cell 34, 939–953 53. Johnson, K.C. et al. (2021) Single-cell multimodal glioma 27. Deblois, G. et al. (2020) Epigenetic switch-induced viral mimicry analyses identify epigenetic regulators of cellular plasticity and evasion in chemotherapy-resistant breast cancer. Cancer environmental stress response. Nat. Genet. 53, 1456–1468 Discov. 10, 1312–1329 54. Huang, Y. et al. (2021) Comprehensive analysis of partial 28. Marine, J.-C. et al. (2020) Non-genetic mechanisms of thera- methylation domains in colorectal cancer based on single-cell peutic resistance in cancer. Nat. Rev. Cancer 20, 743–756 methylation profiles. Brief. Bioinform. 22, bbab267 29. Schuurhuis, G.J. et al. (2018) Minimal/measurable residual 55. Rodríguez-Ubreva, J. et al. (2022) Single-cell atlas of common disease in AML: a consensus document from the European variable immunodeficiency shows germinal center-associated Leukemia Net MRD Working Party. Blood 131, 1275–1291 epigenetic dysregulation in B-cell responses. Nat. Commun. 30. van Galen, P. et al. (2019) Single-cell RNA-seq reveals AML 13, 1779 hierarchies relevant to disease progression and immunity. Cell 56. Gkountela, S. et al. (2019) Circulating tumor cell clustering 176, 1265–1281 shapes DNA methylation to enable metastasis seeding. Cell. 31. Costa, A. et al. (2018) Fibroblast heterogeneity and immuno- 176, 98–112 suppressive environment in human breast cancer. Cancer Cell 57. Sur, I. and Taipale, J. (2016) The role of enhancers in cancer. 33, 463–479 Nat. Rev. Cancer 16, 483–493 32. Aoki, T. et al. (2020) Single-cell transcriptome analysis reveals 58. Buenrostro, J.D. et al. (2015) Single-cell chromatin accessibility disease-defining t-cell subsets in the tumor microenvironment reveals principles of regulatory variation. Nature 523, 486–490 of classic hodgkin lymphoma. Cancer Discov. 10, 406–421 59. Satpathy, A.T. et al. (2019) Massively parallel single-cell chro- 33. Campillo-Marcos, I. et al. (2021) Single-cell technologies and matin landscapes of human immune cell development and analyses in hematopoiesis and hematological malignancies. intratumoral T cell exhaustion. Nat. Biotechnol. 37, 925–936 Exp. Hematol. 98, 1–13 60. Chen, X. et al. (2018) A rapid and robust method for single cell 34. Miles, L.A. et al. (2020) Single-cell mutation analysis of clonal chromatin accessibility profiling. Nat. Commun. 9, 5345 evolution in myeloid malignancies. Nature 587, 477–482 61. Frede, J. et al. (2021) Dynamic transcriptional reprogramming 35. Chaligne, R. et al. (2021) Epigenetic encoding, heritability and leads to immunotherapeutic vulnerabilities in myeloma. Nat. plasticity of glioma transcriptional cell states. Nat. Genet. 53, Cell Biol. 23, 1199–1211 1469–1479 62. Taavitsainen, S. et al. (2021) Single-cell ATAC and RNA 36. Deaton, A.M. and Bird, A. (2011) CpG islands and the regulation sequencing reveal pre-existing and persistent cells associated of transcription. Genes Dev. 25, 1010–1022 with prostate cancer relapse. Nat. Commun. 12, 5307 37. Berdasco, M. and Esteller, M. (2019) Clinical epigenetics: seizing 63. Zhang, Y. et al. (2021) Single-cell analyses reveal key immune opportunities for translation. Nat. Rev. Genet. 20, 109–127 cell subsets associated with response to PD-L1 blockade in 38. Bell, R.E. et al. (2016) Enhancer methylation dynamics contrib- triple-negative breast cancer. Cancer Cell 39, 1578–1593.e8 ute to cancer plasticity and patient mortality. Genome Res. 26, 64. Shu, S. et al. (2020) Synthetic lethal and resistance interactions 601–611 with BET bromodomain inhibitors in triple-negative breast cancer. 39. Guo, H. et al. (2013) Single-cell methylome landscapes of Mol. Cell 78, 1096–1113 mouse embryonic stem cells and early embryos analyzed 65. Kashima, Y. et al. (2021) Single-cell analyses reveal diverse using reduced representation bisulfite sequencing. Genome mechanisms of resistance to EGFR tyrosine kinase inhibitors Res. 23, 2126–2135 in lung cancer. Cancer Res. 81, 4835–4848 40. Guo, H. et al. (2015) Profiling DNA methylome landscapes of 66. Regner, M.J. et al. (2021) A multi-omic single-cell landscape of mammalian cells with single-cell reduced-representation human gynecologic malignancies. Mol. Cell 81, 4924–4941 bisulfite sequencing. Nat. Protoc. 10, 645–659 67. Guilhamon, P. et al. (2021) Single-cell chromatin accessibility 41. Smallwood, S.A. et al. (2014) Single-cell genome-wide bisulfite profiling of glioblastoma identifies an invasive cancer stem cell sequencing for assessing epigenetic heterogeneity. Nat. Methods population associated with lower survival. eLife 10, 1–20 11, 817–820 68. Liscovitch-Brauer, N. et al. (2021) Profiling the genetic determi- 42. Farlik, M. et al. (2015) Single-cell DNA methylome sequencing nants of chromatin accessibility with scalable single-cell and bioinformatic inference of epigenomic cell-state dynamics. CRISPR screens. Nat. Biotechnol. 39, 1270–1277 Cell Rep. 10, 1386–1397 69. Rubin, A.J. et al. (2019) Coupled single-cell CRISPR screening 43. Kobayashi, H. et al. (2016) Repetitive DNA methylome analysis and epigenomic profiling reveals causal gene regulatory networks. by small-scale and single-cell shotgun bisulfite sequencing. Cell 176, 361–376 Genes Cells 21, 1209–1222 70. Pierce, S.E. et al. (2021) High-throughput single-cell chromatin 44. Pixberg, C.F. (2017) Analysis of DNA methylation in single circu- accessibility CRISPR screens enable unbiased identification of lating tumor cells. Oncogene 36, 3223–3231 regulatory networks in cancer. Nat. Commun. 12, 2969 45. Mooijman, D. et al. (2016) Single-cell 5hmC sequencing reveals 71. Lareau, C.A. et al. (2021) Massively parallel single-cell mito- chromosome-wide cell-to-cell variability and enables lineage chondrial DNA genotyping and chromatin profiling. Nat. reconstruction. Nat. Biotechnol. 34, 852–856 Biotechnol. 39, 451–461 46. Zhu, C. et al. (2017) Single-cell 5-formylcytosine landscapes of 72. Penter, L. et al. (2021) Longitudinal single-cell dynamics of mammalian early embryos and ESCs at single-base resolution. chromatin accessibility and mitochondrial mutations in chronic Cell Stem Cell 20, 720–731 lymphocytic leukemia mirror disease history. Cancer Discov. 47. Wu, X. et al. (2017) Simultaneous mapping of active DNA 11, 3048–3063 demethylation and sister chromatid exchange in single cells. 73. Lai, B. et al. (2018) Principles of nucleosome organization re- Genes Dev. 31, 511–523 vealed by single-cell micrococcal nuclease sequencing. Nature 48. Kantlehner, M. et al. (2011) A high-throughput DNA methylation 562, 281–285 analysis of a single cell. Nucleic Acids Res. 39, e44 74. Gao, W. et al. (2020) Genome-wide profiling of nucleosome 49. Han, L. et al. (2017) Bisulfite-independent analysis of CpG position and chromatin accessibility in single cells using island methylation enables genome-scale stratification of single scMNase-seq. Nat. Protoc. 15, 68–85 cells. Nucleic Acids Res. 45, e77 75. Rotem, A. et al. (2015) Single-cell ChIP-seq reveals cell 50. Niemöller, C. et al. (2021) Bisulfite-free epigenomics and geno- subpopulations defined by chromatin state. Nat. Biotechnol. mics of single cells through methylation-sensitive restriction. 33, 1165–1172 Commun. Biol. 4, 153 76. Ku, W.L. et al. (2019) Single-cell chromatin immunocleavage 51. Bianchi, A. et al. (2022) scTAM-seq enables targeted high- sequencing (scChIC-seq) to profile histone modification. Nat. confidence analysis of DNA methylation in single cells. bioRxiv Methods 16, 323–325 Trends in Cancer, October 2022, Vol. 8, No. 10 835 Trends in Cancer OPEN ACCESS 77. Grosselin, K. et al. (2019) High-throughput single-cell ChIP-seq 104. Lewis, S.M. et al. (2021) Spatial omics and multiplexed imaging identifies heterogeneity of chromatin states in breast cancer. to explore cancer biology. Nat. Methods 18, 997–1012 Nat. Genet. 51, 1060–1066 105. Wu, Y. et al. (2022) Spatiotemporal immune landscape of colo- 78. Marsolier, J. et al. (2022) H3K27me3 conditions chemo- rectal cancer liver metastasis at single-cell level. Cancer Discov. tolerance in triple-negative breast cancer. Nat. Genet. 54, 12, 134–153 459–468 106. Nieto, P. et al. (2021) A single-cell tumor immune atlas for 79. Janssens, D.H. et al. (2021) Automated CUT&Tag profiling of precision oncology. Genome Res. 31, 1913–1926 chromatin heterogeneity in mixed-lineage leukemia. Nat. 107. Deng, Y. (2022) Spatial-CUT&Tag: spatially resolved chromatin Genet. 53, 1586–1596 modification profiling at the cellular level. Science 375, 681–686 80. Azagra, A. et al. (2020) From loops to looks: transcription factors 108. Lu, T. et al. (2022) Spatially resolved epigenomic profiling of single and chromatin organization shaping terminal B cell differentiation. cells in complex tissues. bioRxiv Published online February 19, Trends Immunol. 41, 46–60 2022. https://doi.org/10.1101/2022.02.17.480825 81. Akdemir, K.C. et al. (2020) Disruption of chromatin folding 109. Thornton, C.A. (2021) Spatially mapped single-cell chromatin domains by somatic genomic rearrangements in human cancer. accessibility. Nat. Commun. 12, 1274 Nat. Genet. 52, 294–305 110. Deng, Y. et al. (2021) Spatial-ATAC-seq: spatially resolved 82. Nagano, T. et al. (2013) Single-cell Hi-C reveals cell-to-cell chromatin accessibility profiling of tissues at genome scale variability in chromosome structure. Nature 502, 59–64 and cellular level. bioRxiv Published online June 7, 2021. 83. Ramani, V. et al. (2017) Massively multiplex single-cell Hi-C. https://doi.org/10.1101/2021.06.06.447244 Nat. Methods 14, 263–266 111. Gracia Villacampa, E. et al. (2021) Genome-wide spatial expres- 84. Kind, J. et al. (2015) Genome-wide maps of nuclear lamina in- sion profiling in formalin-fixed tissues. Cell Genom. 1, 100065 teractions in single human cells. Cell 163, 134–147 112. Zhang, H. et al. (2022) Profiling chromatin accessibility in 85. Zhu, C. et al. (2020) Single-cell multimodal omics: the power of formalin-fixed paraffin-embedded samples. Genome Res. 32, many. Nat. Methods 17, 11–14 150–161 86. Pott, S. (2017) Simultaneous measurement of chromatin acces- 113. Batista, F.M. et al. (2020) Whole genome sequencing of hepatitis sibility, DNA methylation, and nucleosome phasing in single A virus using a PCR-free single-molecule nanopore sequencing cells. eLife 6, e23203 approach. Front. Microbiol. 11, 874 87. Li, G. et al. (2019) Joint profiling of DNA methylation and chro- 114. Liu, Y. et al. (2021) DNA methylation-calling tools for Oxford matin architecture in single cells. Nat. Methods 16, 991–993 Nanopore sequencing: a survey and human epigenome-wide 88. Lee, D.S. et al. (2019) Simultaneous profiling of 3D genome evaluation. Genome Biol. 22, 295 structure and DNA methylation in single human cells. Nat. 115. Philpott, N. (2021) Nanopore sequencing of single-cell Methods 16, 999–1006 transcriptomes with scCOLOR-seq. Nat. Biotechnol. 39, 89. Guo, F. et al. (2017) Single-cell multi-omics sequencing of 1517–1520 mouse early embryos and embryonic stem cells. Cell Res. 27, 116. Danese, A. et al. (2021) EpiScanpy: integrated single-cell 967–988 epigenomic analysis. Nat. Commun. 12, 5228 90. Fan, X. et al. (2022) Integrated single-cell multiomics analysis 117. Bond, D.R. et al. (2020) Single-cell epigenomics in cancer: reveals novel candidate markers for prognosis in human charting a course to clinical impact. Epigenomics 12, pancreatic ductal adenocarcinoma. Cell Discov. 8, 13 1139–1151 91. Blecua, P. et al. (2022) Refinement of computational identifica- 118. Johannessen, L.E. et al. (2018) MGMT gene promoter methyla- tion of somatic copy number alterations using DNA methylation tion status - assessment of two pyrosequencing kits and three microarrays illustrated in cancers of unknown primary. Brief. methylation-specific PCR methods for their predictive capacity Bioinform. 6, bbac161 in glioblastomas. Cancer Genomics Proteomics 15, 437–446 92. Gaiti, F. et al. (2019) Epigenetic evolution and lineage histories 119. Locke, W.J. et al. (2019) DNA methylation cancer biomarkers: of chronic lymphocytic leukaemia. Nature 569, 576–580 translation to the clinic. Front. Genet. 10, 1150 93. Bian, S. (2018) Single-cell multiomics sequencing and analyses 120. Malta, T.M. et al. (2018) Glioma CpG island methylator phenotype of human colorectal cancer. Science 362, 1060–1063 (G-CIMP): biological and clinical implications. Neuro-Oncology 94. Hou, Y. et al. (2016) Single-cell triple omics sequencing reveals 20, 608–620 genetic, epigenetic, and transcriptomic heterogeneity in hepa- 121. Kopetz, S. et al. (2015) Genomic classifier ColoPrint predicts tocellular carcinomas. Cell Res. 26, 304–319 recurrence in stage II colorectal cancer patients more accurately 95. Angermueller, C. et al. (2016) Parallel single-cell sequencing links than clinical factors. Oncologist 20, 127–133 transcriptional and epigenetic heterogeneity. Nat. Methods 13, 122. Brandão, M. et al. (2019) Mammaprint™: a comprehensive 229–232 review. Future Oncol. 15, 207–224 96. Luo, C. et al. (2018) Multi-omic profiling of transcriptome and 123. Moran, S. et al. (2016) Epigenetic profiling to classify cancer of DNA methylome in single nuclei with molecular partitioning. unknown primary: a multicentre, retrospective analysis. Lancet bioRxiv Published online October 4, 2018. https://doi.org/10. Oncol. 17, 1386–1395 1101/434845 124. Nassar, F.J. et al. (2021) Methylated circulating tumor DNA as 97. Liu, L. et al. (2019) Deconvolution of single-cell multi-omics a biomarker for colorectal cancer diagnosis, prognosis, and layers reveals regulatory heterogeneity. Nat. Commun. 10, 470 prediction. Clin. Epigenetics 13, 111 98. Reyes, M. et al. (2019) Simultaneous profiling of gene expres- 125. Pixberg, C.F. et al. (2015) Characterization of DNA methylation sion and chromatin accessibility in single cells. Adv. Biosyst. in circulating tumor cells. Genes (Basel) 6, 1053–1075 3, 1900065 126. Friedmann-Morvinski, D. (2014) Glioblastoma heterogeneity 99. Rooijers, K. et al. (2019) Simultaneous quantification of protein– and cancer cell plasticity. Crit. Rev. Oncol. 19, 327–336 DNA contacts and transcriptomes in single cells. Nat. Biotechnol. 127. Olbryt, M. et al. (2020) Biological factors behind melanoma 37, 766–772 response to immune checkpoint inhibitors. Int. J. Mol. Sci. 21, 100. Clark, S.J. et al. (2018) ScNMT-seq enables joint profiling of 4071 chromatin accessibility DNA methylation and transcription in 128. Ortiz-Barahona, V. et al. (2020) Use of DNA methylation profiling single cells. Nat. Commun. 9, 781 in translational oncology. Semin. Cancer Biol. 83, 523–535 101. Wang, Y. et al. (2021) Single-cell multiomics sequencing reveals 129. Marinov, G.K. and Shipony, Z. (2021) Interrogating the accessible the functional regulatory landscape of early embryos. Nat. chromatin landscape of eukaryote genomes using ATAC-seq. Commun. 12, 1247 Methods Mol. Biol. 2243, 183–226 102. Ståhl, P.L. et al. (2016) Visualization and analysis of gene 130. Nakato, R. and Sakata, T. (2021) Methods for ChIP-seq analysis: expression in tissue sections by spatial transcriptomics. Science a practical workflow and advanced applications. Methods 187, 353, 78–82 44–53 103. Rodriques, S.G. et al. (2019) Slide-seq: a scalable technology for 131. Sati, S. and Cavalli, G. (2016) Chromosome conformation cap- measuring genome-wide expression at high spatial resolution. ture technologies and their impact in understanding genome Science 363, 1463–1467 function. Chromosoma 126, 33–44 836 Trends in Cancer, October 2022, Vol. 8, No. 10 Trends in Cancer OPEN ACCESS 132. Javierre, B.M. et al. (2016) Lineage-specific genome architec- 159. Singh, H. et al. (2021) Hybrid stomach-intestinal chromatin ture links enhancers and non-coding disease variants to target states underlie human Barrett’s metaplasia. Gastroenterology gene promoters. Cell 167, 1369–1384 161, 924–939.e11 133. Carter, B. and Zhao, K. (2021) The epigenetic basis of cellular 160. Wang, H. et al. (2021) Single-cell analyses reveal mechanisms heterogeneity. Nat. Rev. Genet. 22, 235–250 of cancer stem cell maintenance and epithelial-mesenchymal 134. Chakravarthy, A. et al. (2018) Pan-cancer deconvolution of transition in recurrent bladder cancer. Clin. Cancer Res. 27, tumour composition using DNA methylation. Nat. Commun. 6265–6278 9, 3220 161. Schwartz, G.W. et al. (2021) TooManyPeaks identifies drug- 135. Ceccarelli, M. et al. (2016) Molecular profiling reveals biologically resistant-specific regulatory elements from single-cell leukemic discrete subsets and pathways of progression in diffuse glioma. epigenomes. Cell Rep. 36, 109575 Cell 164, 550–563 162. Pierce, S.E. et al. (2021) LKB1 inactivation modulates chromatin 136. Völker-Albert, M. et al. (2020) Histone modifications in stem cell accessibility to drive metastatic progression. Nat. Cell Biol. 23, development and their clinical implications. Stem Cell Rep. 15, 915–924 1196–1205 163. Cusanovich, D.A. et al. (2015) Multiplex single cell profiling of 137. Mehrmohamadi, M. et al. (2021) A comparative overview of chromatin accessibility by combinatorial cellular indexing. Science epigenomic profiling methods. Front. Cell Dev. Biol. 9, 714687 348, 910–914 138. Argelaguet, R. et al. (2018) Multi-Omics Factor Analysis—a 164. Del Priore, I. et al. (2021) Protocol for single-cell ATAC sequencing framework for unsupervised integration of multi-omics data using combinatorial indexing in mouse lung adenocarcinoma. sets. Mol. Syst. Biol. 14, e8124 STAR Protoc. 2, 100583 139. Argelaguet, R. et al. (2020) MOFA+: a statistical framework for 165. Wang, K. et al. (2021) Simple oligonucleotide-based multi- comprehensive integration of multi-modal single-cell data. Ge- plexing of single-cell chromatin accessibility. Mol. Cell 81, nome Biol. 21, 111 4319–4332 140. Li, G. et al. (2022) A deep generative model for multi-view pro- 166. De Rop, F.V. et al. (2022) Hydrop enables droplet-based single- filing of single-cell RNA-seq and ATAC-seq data. Genome Biol. cell ATAC-seq and single-cell RNA-seq using dissolvable 23, 20 hydrogel beads. Elife 11, e73971 141. Wu, K.E. et al. (2021) BABEL enables cross-modality translation 167. Tedesco, M. et al. (2022) Chromatin Velocity reveals epigenetic between multiomic profiles at single-cell resolution. Proc. Natl. dynamics by single-cell profiling of heterochromatin and Acad. Sci. U. S. A. 118, e2023070118 euchromatin. Nat. Biotechnol. 40, 235–244 142. Wang, C. et al. (2020) Integrative analyses of single-cell 168. Ku, W.L. et al. (2021) Profiling single-cell histone modifications transcriptome and regulome using MAESTRO. Genome Biol. using indexing chromatin immunocleavage sequencing. Genome 21, 198 Res. 31, 1831–1842 143. Sharma, A. et al. (2021) ShinyArchR.UiO: user-friendly, integrative 169. Harada, A. et al. (2019) A chromatin integration labelling method and open-source tool for visualisation of single-cell ATAC-seq enables epigenomic profiling with lower input. Nat. Cell Biol. 21, data using ArchR. Bioinformatics 38, 834–836 287–296 144. Prompsy, P. et al. (2020) Interactive analysis of single-cell 170. Carter, B. et al. (2019) Mapping histone modifications in low cell epigenomic landscapes with ChromSCape. Nat. Commun. number and single cells using antibody-guided chromatin 11, 5702 tagmentation (ACT-seq). Nat. Commun. 10, 3747 145. Shareef, S.J. et al. (2021) Extended-representation bisulfite 171. Wang, Q. et al. (2019) CoBATCH for high-throughput single-cell sequencing of gene regulatory elements in multiplexed samples epigenomic profiling. Mol. Cell 76, 206–216 and single cells. Nat. Biotechnol. 39, 1086–1094 172. Hainer, S.J. et al. (2019) Profiling of pluripotency factors in single 146. Chen, H. et al. (2021) Single-cell DNA methylome analysis of cells and early embryos. Cell 177, 1319–1329 circulating tumor cells. Chin. J. Cancer Res. 33, 391–404 173. Kaya-Okur, H.S. et al. (2019) CUT&Tag for efficient epigenomic 147. Hui, T. et al. (2018) High-resolution single-cell DNA methylation profiling of small samples and single cells. Nat. Commun. 10, 1930 measurements reveal epigenetically distinct hematopoietic 174. Bartosovic, M. et al. (2021) Single-cell CUT&Tag profiles histone stem cell subpopulations. Stem Cell Rep. 11, 578–592 modifications and transcription factors in complex tissues. Nat. 148. Grants, J.M. (2020) Altered microRNA expression links IL6 and Biotechnol. 39, 825–835 TNF-induced inflammaging with myeloid malignancy in humans 175. Wu, S.J. (2021) Single-cell CUT&Tag analysis of chromatin and mice. Blood 135, 2235–2251 modifications in differentiation and tumor progression. Nat. 149. Luo, C. et al. (2017) Single-cell methylomes identify neuronal Biotechnol. 39, 819–824 subtypes and regulatory elements in mammalian cortex. Science 176. Hui, W.W.I. et al. (2021) Single-cell mapping of DNA 357, 600–604 G-quadruplex structures in human cancer cells. Sci. Rep. 150. Meir, Z. (2020) Single-cell analysis of clonal maintenance of 11, 23641 transcriptional and epigenetic states in cancer cells. Nat. 177. Bartlett, D.A. et al. (2021) High-throughput single-cell epigenomic Genet. 52, 709–718 profiling by targeted insertion of promoters (TIP-seq). J. Cell Biol. 151. Luo, C. et al. (2018) Robust single-cell DNA methylome profiling 220, e202103078 with snmC-seq2. Nat. Commun. 9, 3824 178. Ohnuki, H. et al. (2021) Iterative epigenomic analyses in the 152. Mulqueen, R.M. et al. (2018) Highly scalable generation of DNA same single cell. Genome Res. 31, 1819–1830 methylation profiles in single cells. Nat. Biotechnol. 36, 428–431 179. Hu, Y. et al. (2019) Simultaneous profiling of mRNA tran- 153. Ma, S. et al. (2018) Cell-type-specific brain methylomes profiled scriptome and DNA methylome from a single cell. Methods via ultralow-input microfluidics. Nat. Biomed. Eng. 2, 183–194 Mol. Biol. 1979, 363–377 154. Li, X. et al. (2019) BRIF-seq: bisulfite-converted randomly inte- 180. Gu, H. et al. (2021) Smart-RRBS for single-cell methylome and grated fragments sequencing at the single-cell level. Mol. transcriptome analysis. Nat. Protoc. 16, 4004–4030 Plant 12, 438–446 181. Cheow, L.F. et al. (2016) Single-cell multimodal profiling reveals 155. Hunt, K.V. et al. (2022) scTEM-seq: single-cell analysis of trans- cellular epigenetic heterogeneity. Nat. Methods 13, 833–836 posable element methylation to link global epigenetic heteroge- 182. Gu, C. et al. (2019) Integrative single-cell analysis of tran- neity with transcriptional programs. Sci. Rep. 12, 5776 scriptome, DNA methylome and chromatin accessibility in 156. Wang, L. et al. (2019) The phenotypes of proliferating glioblas- mouse oocytes. Cell Res. 29, 110–123 toma cells reside on a single axis of variation. Cancer Discov. 183. Satpathy, A.T. et al. (2018) Transcript-indexed ATAC-seq for 9, 1708–1719 precision immune profiling. Nat. Med. 24, 580–590 157. Wang, Q. et al. (2022) Single-cell chromatin accessibility land- 184. Chen, S. et al. (2019) High-throughput sequencing of the tran- scape in kidney identifies additional cell-of-origin in heteroge- scriptome and chromatin accessibility in the same cell. Nat. nous papillary renal cell carcinoma. Nat. Commun. 13, 31 Biotechnol. 37, 1452–1457 158. Xu, K. et al. (2021) Integrative analyses of scRNA-seq and 185. Zhu, C. et al. (2019) An ultra high-throughput method for single- scATAC-seq reveal CXCL14 as a key regulator of lymph node cell joint analysis of open chromatin and transcriptome. Nat. metastasis in breast cancer. Hum. Mol. Genet. 30, 370–380 Struct. Mol. Biol. 26, 1063–1070 Trends in Cancer, October 2022, Vol. 8, No. 10 837 Trends in Cancer OPEN ACCESS 186. Zhu, C. et al. (2021) Joint profiling of histone modifications and 190. Chen, X. et al. (2018) Joint single-cell DNA accessibility and transcriptome in single cells from mouse brain. Nat. Methods protein epitope profiling reveals environmental regulation of 18, 283–292 epigenomic heterogeneity. Nat. Commun. 9, 4590 187. Yan, R. et al. (2021) Decoding dynamic epigenetic landscapes 191. Zhang, B. et al. (2022) Characterizing cellular heterogeneity in chro- in human oocytes using single-cell multi-omics sequencing. matin state with scCUT&Tag-pro. Nat. Biotechnol. Published online Cell Stem Cell 28, 1641–1656 March 24, 2022. https://doi.org/10.1038/s41587-022-01250-0 188. Chen, A.F. et al. (2022) NEAT-seq: simultaneous profiling of 192. Ma, S. et al. (2020) Chromatin potential identified by shared intra-nuclear proteins, chromatin accessibility, and gene single-cell profiling of RNA and chromatin. Cell 183, 1103–1116 expression in single cells. Nat. Methods 19, 547–553 193. Su, J.-H. et al. (2020) Genome-scale imaging of the 3D organization 189. Mimitou, E.P. et al. (2021) Scalable, multimodal profiling of and transcriptional activity of chromatin. Cell 182, 1641–1659 chromatin accessibility, gene expression and protein levels in 194. Takei, Y. et al. (2021) Integrated spatial genomics reveals global single cells. Nat. Biotechnol. 39, 1246–1258 architecture of single nuclei. Nature 590, 344–350 838 Trends in Cancer, October 2022, Vol. 8, No. 10