Genome Research

Mutation spectra vary among populations, and a comprehensive view of this variation requires broad sampling and fine-scale population resolution. Here, we profile mutation spectra from single-nucleotide polymorphisms and small insertions and deletions in a uniformly processed collection of 3,034 globally sampled Saccharomyces cerevisiae genomes, and an extended dataset including the previously sampled global 1,011-genome collection. Major mutation spectrum patterns reported from the 1,011 genomes remain evident in the extended dataset, including the pronounced mutation-spectrum differentiation of African Beer strains. At finer population resolution, we identify additional variation among wine-associated populations, including a Georgian Wine-associated signal that is strongest among rare variants and is therefore consistent with a comparatively recent change in mutation processes. We further apply a sequence-context-aware classification to characterize small-indel profiles across the 3,034 natural yeast genomes. Together, our analyses show that expanded sampling and refined population assignments provide a more comprehensive view of mutation-spectrum variation within S. cerevisiae.

Topoisomerase cleavage complexes (TOPcc) are covalent protein-DNA intermediates formed transiently during topoisomerase catalysis and constitute a major class of DNA-protein crosslinks in human cells, yet their endogenous genomic distributions remain difficult to resolve. Here, we developed protein-DNA adduct immunoprecipitation followed by sequencing (pDAIP-seq), a highly sensitive method for detecting endogenous TOPcc that enables their genome-wide profiling under normal physiological conditions without stabilization by topoisomerase poisons. Using pDAIP-seq, we found that camptothecin treatment, required by existing TOP1cc profiling methods, not only increased TOP1cc abundance but also substantially altered its genomic distribution. Endogenous TOP1cc were globally reduced upon transcriptional suppression and increased upon TOP2A depletion, supporting TOP1cc regulation by the maintenance of genome-wide DNA topological homeostasis. Interestingly, both the regional distribution and nucleotide-scale positioning of TOP1cc, including the positions of TOP1 catalytic cleavage sites, were largely specified by local DNA sequence context. By extending pDAIP-seq to capture transient covalent protein-DNA intermediates involved in DNA repair, including TDP1-, PARP1-, and KU70-DNA adducts, we further identified endogenous TOP1cc as a potential major source of spontaneous DNA breaks in human cells. Finally, pDAIP-seq also enabled genome-wide mapping of TOP2cc, TOP3cc, and SPO11-DNA covalent intermediates in mouse testis undergoing programmed meiotic DNA cleavage. Together, these findings establish pDAIP-seq as a broadly applicable method for the mapping of covalent protein-DNA intermediates and reveal endogenous TOP1cc as a key intermediate between DNA topological regulation and human genome instability.

Genomic foundation models (GFMs) are pretrained on large genomic datasets, but their ability to predict rare disease-associated events after fine-tuning remains unclear. We tested this using TDP-43 cryptic exons, well characterized splicing events linked to neurodegenerative disease that are poorly represented in standard annotations. We compared the state-of-the-art GFM AlphaGenome with OpenSpliceAI, a specialized splice site prediction model. Fine-tuned OpenSpliceAI detected held-out cryptic exons in human and mouse and captured allele-specific splicing effects at the disease-associated UNC13A locus. In contrast, a splicing prediction head trained on AlphaGenome's frozen pretrained backbone failed to predict cryptic exons. Unfreezing and fine-tuning all model weights, however, enabled AlphaGenome to generalize to held-out data. To investigate cryptic splicing in cell types that are difficult to study experimentally, we applied fine-tuned OpenSpliceAI across the genome to study cell type-specific genes. This led to the identification of new cryptic splicing events in CLDN11 and ARHGAP23, which were subsequently validated in FTLD-TDP brain tissue. Notably, the cryptic exon in CLDN11 is predicted to generate a neoepitope that could serve as an oligodendrocyte-specific biomarker of TDP-43 pathology. Our results show that even a state-of-the-art GFM can fail to generalize to rare splicing events, even with substantial pretraining. However, both foundation and specialized models that learn to generalize to rare biological events can have broad applications in biological discovery and medicine.

Detecting somatic mutations is a cornerstone of cancer genomics and clinical genotyping, traditionally relying on whole exome (WES) or whole genome sequencing (WGS) data. While RNA sequencing (RNA-seq) is primarily used to assess gene expression, it can also be utilized for somatic variant detection. RNA-seq inherently captures expressed variants, thereby reducing the identification of passenger mutations and mitigating the annotation biases associated with WES. We performed somatic variant calling and cancer driver detection from tumor RNA-seq data from 1,359 RNA-seq samples with matched WGS from the Pan-Cancer Analysis of Whole Genomes (PCAWG) Consortium of the International Cancer Genome Consortium (ICGC) and The Cancer Genome Atlas (TCGA). Our approach using RNA-seq data, alone, identified at least one putative driver gene in 15 of 16 cancer types, consistent with driver predictions from WGS. We further observed that the variant allele frequency (VAF) of RNA-seq-derived variants reflects their predicted functional impact, providing information specific to transcriptomic data. Because RNA-seq data can be generated and analyzed at lower computational cost and time than WGS or WES, it represents a resource-efficient alternative for cancer driver detection. This study provides an extensive pan-cancer resource of RNA-seq-derived variants from the PCAWG consortium and supports the utility of RNA-seq as a complementary approach for characterizing expressed somatic variation in cancer.

Infected cells rapidly adapt their transcriptional profile to upregulate genes that limit pathogen replication and promote inflammation. Viruses also induce widespread changes in host gene expression, and both host and virus further remodel the transcriptome through targeted mRNA degradation, modulation of alternative splicing, and alterations at transcript 5 and 3 prime ends. The complexity of the response to infection makes it difficult to understand which transcriptional changes are driven by the virus and which reflect the host antiviral program. Short-read RNA sequencing has been the primary tool for studying these changes, but its inability to resolve full-length transcript isoforms limits isoform switching-based analyses, narrowing our understanding of how transcript usage shapes immunity. Here we profiled the host transcriptome in response to interferon-{beta} treatment or influenza A virus infection using both short- and long-read RNA sequencing analyzed with the SQANTI3 pipeline. Because isoform switching and transcript abundance change are distinct regulatory events, we focused our analysis on differential transcript usage to uncover a layer of transcriptome remodeling that standard expression analysis cannot detect. Pathway analysis revealed distinct regulons for genes undergoing transcriptional upregulation compared with those undergoing isoform switching. These data provide a rich resource for querying the isoform changes underlying shared and distinct transcriptional programs triggered by viral infection or innate immune stimulation. To identify putative mechanisms coordinating these stimulation-dependent isoform changes, we integrated 232 eCLIP datasets and found a small set of RNA-binding proteins localized at these isoform switches, including DDX3X. Comparing the responses to interferon and to virus highlighted RNA-binding proteins associated with infection-dependent immune remodeling. Altogether, these results demonstrate the value of hybrid sequencing for deciphering the post-transcriptional regulatory code active during the antiviral response.

Signal-to-noise ratios in copy-number estimates across genomic loci improve with the use of large panels of germline reference data. However, in the case of sex chromosomes, germline samples differ between males and females (males have one X chromosome (chrX) and one Y chromosome (chrY), while females have two chrXs), limiting the size of the reference sets to those that match the sex of the index sample. Frequent mosaic loss of chrY in normal male tissue also leads to biases when not accounted for. Here we report TangentXY, a copy-number normalization method that uses both male and female reference samples, correcting for sex differences in chrX and accounting for chrY mosaicism, to assess sex chromosome copy numbers for samples of either sex. We show that TangentXY maintains sex-specific differences in chrX and chrY while minimizing noise. We also used TangentXY to analyze associations between copy-number alterations in sex chromosomes and cancer characteristics, and found that losses in the chrX but not chrY are correlated with TP53 mutation, that gain of chrX can predict patient prognosis in certain tumor types, and that the vast majority of gains of chrX in males remain active, whereas gains of chrX in females are divided approximately equally between active and inactive chrXs. We also found that loss of chrY correlates with inactivating driver mutations in escape genes in chrX, suggesting a mechanism by which its frequent loss promotes cancers in males.

Transcription and RNA processing are tightly coupled to chromatin organization, yet how nucleosome positioning is coordinated across the promoter and gene body to shape alternative isoform expression remains poorly understood. This question has been difficult to address because short-read chromatin profiling cannot simultaneously capture promoter and gene-body chromatin states on the same DNA molecule, and short-read RNA sequencing cannot resolve full-length transcripts without requiring computational assembly. To address this, we deleted two conserved chromatin remodelers, ISW1 and CHD1, in Saccharomyces cerevisiae and profiled chromatin fibers, nascent RNA, and full-length mature RNA using Oxford Nanopore long-read sequencing. We developed EpiFLAIR, a computational framework that assigns discrete long-range chromatin states across entire genes and integrates these states with isoform-level transcriptional outputs. Using EpiFLAIR, we show that single-molecule chromatin states together with sequence motifs are strong predictors of alternative TSS isoform expression. Loss of ISW1 and CHD1 produced divergent transcriptional changes by activating transcriptionally silenced promoters while repressing active ones. Single-molecule chromatin profiling suggested that this redistribution of transcriptional activity arose, in part, from an anticorrelation between gene-body and promoter chromatin accessibility along the same chromatin fiber, a finding uniquely enabled by long-read single-molecule resolution. Furthermore, disruption of regular nucleosome spacing impaired RNAPII processivity and co-transcriptional splicing, consequently altering mature RNA levels and splicing isoform composition. Together, our findings demonstrate that nucleosomes maintain transcriptional fidelity during initiation, elongation, and splicing to regulate isoform expression. More broadly, long-read single-molecule chromatin profiling represents a powerful approach for linking chromatin organization to isoform-level gene regulation.

Droplet-based and split-pool barcoding are the two dominant high-throughput single-cell RNA-seq paradigms, yet their systematic biases remain incompletely characterized. We compared seven matched 10X Genomics Chromium and Parse Biosciences Evercode WT sample pairs from five studies spanning multiple tissues and species focusing on gene counts and compositional differences rather than cell-type resolution. By separating Parse reads by priming strategy, we distinguished biases arising from priming chemistry from platform-level effects. These effects were largely independent, with platform-level differences substantially larger. Priming strategy introduced a GC-content bias within Parse, whereas the two platforms showed a consistent compositional shift across priming strategies, supported by differential capture of nuclear- and cytoplasmic-enriched transcripts. We interpret this shift as a relative enrichment of nuclear RNA in Parse. Finally, using contamination measures that do not assume a droplet-specific model, we found background RNA contamination in Parse to be comparable in magnitude to that in 10X, contrary to claims of substantially lower contamination. Together, these results indicate that expression measurements and cell-type composition are not directly interchangeable between the two platforms.

Nucleosomes regulate DNA accessibility and thereby influence gene expression, DNA repair, and mutagenesis. Although DNA sequence strongly determines nucleosome positioning, the higher-order sequence grammar underlying nucleosome organization and its relationship to cancer-associated mutational and epigenetic processes remain poorly understood. We developed DeepND, an interpretable deep learning framework that distinguishes nucleosomal from inter-nucleosomal DNA and identifies sequence features associated with nucleosome architecture. Our interpretable framework enabled analysis of sequence patterns and identified 210 sequence motifs of 2-16 bp length that were significantly enriched or depleted in nucleosomal DNA, substantially extending the previously characterized nucleosomal motifs. These motifs revealed distinct sequence grammars within nucleosomal footprints and linker regions, including patterns associated with inward- and outward-facing minor grooves. Forty-four of these motifs showed cancer-type-specific enrichment or depletion of somatic mutations, and six motifs were associated with differential DNA methylation levels. In lung and endometrial cancers, mutations exhibited a significant ~10-bp periodic enrichment pattern across distinct minor-groove orientations, consistent with differential sensitivity of mutational and DNA repair processes to nucleosomal rotational positioning. We further identified sequence-specific methylation patterns consistent with reduced accessibility of nucleosomal DNA to DNA methyltransferases, including differential methylation at inward- and outward-facing CpG motifs. Together, these findings provide a sequence-resolved characterization of nucleosome-level organization and its link to cancer-specific mutational processes and epigenetic landscapes.

Amyotrophic lateral sclerosis (ALS) shows well-established sex differences in incidence, age at onset and progression, but their biological basis is unclear. We hypothesised that the DNA methylation changes associated with ALS differ between the sexes, which would point to sex-specific disease processes. To test this, we performed a blood-based epigenome-wide association meta-analysis of ALS cases and controls from Project MinE, an international ALS genomics consortium, and an independent Australian cohort, comprising 8,946 individuals after quality control (5,129 males and 3,817 females; 8,942 with complete covariates in the primary models). We analysed 781,648 autosomal CpGs using complementary sex-stratified case-control models and an ALS-by-sex interaction model that directly tested whether ALS-associated methylation effects differed by sex. Sex-stratified analyses identified no genome-wide significant differences between male and female ALS effects. In contrast, interaction testing identified seven robust sex-dependent ALS-associated differentially methylated positions (P < 9 x 10-8), mapping to KCNQ1, ADPRH, HLA-F, GHRLOS, IRF5, RAD51B and one intergenic region. The strongest signal was at KCNQ1 (P = 3.1 x 10-9). Six of the seven remained genome-wide significant in a four-stratum sensitivity analysis, and none was detected in previous conventional ALS or population sex-associated methylation analyses. Together, these findings suggest a distinct sex-conditional component of the ALS blood methylome, implicating immune, metabolic and ion-channel loci and providing new candidates for investigating biological heterogeneity between males and females with ALS.

Cell fate transitions involve major changes to the chromatin landscape, some of which are thought to deepen the basin around a newly adopted cell state, stabilizing gene expression programs that might otherwise be transient. Transcription factors (TFs) are central to this process, binding a subset of their genome-wide motif occurrences to create accessible regions that can act as cell type-specific enhancers. However, apart from a few well-studied cases, it is not clear which of the ~1,600 mammalian TFs, when newly expressed, can directly or indirectly initiate newly accessible sites, nor why only some motif occurrences gain accessibility. Here, we applied pooled CRISPR activation (CRISPRa) to upregulate endogenous TFs to physiological levels in mouse embryonic stem cells (mESCs). After screening guides targeting 240 developmental TFs for transcriptional response, we selected 60 and jointly profiled gene expression and chromatin accessibility in ~140,000 single cells. Some TFs (e.g. Cdx2, Sox10) appeared to initiate new accessibility on their own, predominantly at distal sites that are closed in mESCs but coincide with cell type-specific enhancers later in development. Others appeared to partner with "co-TFs", either by leveraging preexisting mESC heterogeneity (e.g. {uparrow}Grhl1 in Tfap2c+ mESCs [-&gt;] opening of sites bearing motifs for both), or by hierarchical activation (e.g. {uparrow}Sox7 [-&gt;] {uparrow}Gata6 [-&gt;] opening of sites bearing either or both motifs). Such dependencies shape which sites are used and which fates are reachable, and may explain the differential behavior of TFs with nearly identical consensus motifs. Thus, the capacity of a newly expressed TF to open chromatin strongly depends on the cell type, cell state and gene regulatory network it lands in. Charting these dependencies will inform how cell fate transitions are initiated in development, homeostasis and disease, and how to engineer them in silico, in vitro and in vivo.

High-accuracy, high-resolution benchmark data are fundamental to improving the accuracy and generalizability of sequencing technologies and associated algorithms. Most existing benchmarks for DNA sequencing rely on mapping to a single linear reference, omitting inaccessible and highly repetitive complex genomic regions. Here, we present the Genome in a Bottle (GIAB) Consortium's HG002 v5.0q benchmark set, a more comprehensive variant benchmark derived from a fully phased, highly curated, and accurate diploid assembly of HG002. Our approach leverages recent advances in long-read sequencing and assembly to provide confident small- and structural-variant calls across previously inaccessible regions, including segmental duplications, homopolymers, tandem repeats, and complex immune loci. The v5.0 benchmark significantly expands the callable genome relative to v4.2.1, adding x200 Mbp of sequence and increasing variant detection by 18 % for small variants and 3-fold for structural variants compared to previous benchmarks. The benchmark regions include more variants in repetitive regions, including 1.2 x segmental duplications, 1.8 x tandem repeats, and 2.0 x homopolymers, compared to v4.2.1. Separate and combined small- and structural-variant benchmarks are available for three versions of the human reference genome (GRCh37, GRCh38, and CHM13). We validate the reliability of this benchmark set by comparing multiple sequencing technologies and variant calling pipelines and curating differences. The v5.0q benchmark provides a critical resource for the development and validation of next-generation sequencing technologies and analysis methods.

Copy number variants (CNVs) are a significant source of genomic variation in humans. They can span large genomic regions that encompass whole genes and their flanking regions, having the potential to alter the number of functional copies in an individual. Typically, CNVs are analysed individually or combined into regions without regard to the genes they span. Here, we employ a gene-centric approach to link changes in whole-gene copy number to phenotype. We first examined the relationship between CNVs spanning whole genes and gene expression using data from GTEx and found that predicted effects on gene expression are often realised. Applying this method, we identified 34 gene deletions associated with cancer incidence, including several novel candidate associations. Most notably, we discovered that gene deletions in the 13q14 region are linked to increased cancer risk, a novel extension of a known somatic association to the germline, which we did not find using region-based methods. Statistical associations were found between whole-gene loss of tumour suppressor genes and overall cancer risk, as hypothesized. It was additionally found that the total number of whole-gene losses was strongly associated with increased cancer risk, even when known cancer-related genes are not included in the count. Finally, we extended these methods to non-cancer phenotypes, uncovering significant associations with common diseases, including asthma and heart attack.

Diploid genome assemblies are now routinely available, but most read aligners were designed for haploid references, which have long been the gold standard. When reads are aligned to a diploid assembly, the aligner sees two nearly identical alignments to either haplotype, thus reducing the mapping quality (MapQ) score to reflect this ambiguity. This can cause downstream tools to discard reads from easily mappable regions. Here, we present HipHap (HIgh- Performance HAPlotype assigner) to resolve this issue by aligning reads to each haplotype assembly separately and assigning each read to its best-supported haplotype. HipHap increased the fraction of reads mapping to the diploid genome with high MapQ, outperforming alignment to either parental haplotype. Additionally, we introduce a haplotype assignment quality score (HapQ) in HipHap to quantify confidence in the haplotype of origin of a read. HipHap is implemented in Rust, supports SAM, BAM, CRAM, and PAF formats, and is freely available at: https://github.com/jheinz27/hiphap.

Noncoding regulatory elements in the human genome can control gene expression through distinct mechanisms, including acting as enhancers, silencers, or topological elements. However, we lack tools to identify all such classes systematically and quantify their effects on gene expression in an endogenous genomic context. Here we develop Swap-seq, a method that combines pooled twin prime editing with flow-sorting to delete hundreds of endogenous regulatory sequences and measure their quantitative effects on the expression of a nearby gene. We apply Swap-seq to dissect the regulatory elements that control PPIF expression in THP-1 monocytes. We quantify the effects of distal and intronic enhancers, characterize CTCF insulators, and discover a silencer that regulates distal genes through histone deacetylase (HDAC)-mediated repression of neighboring enhancers. Hundreds of other genomic elements share its chromatin signatures and transcription factor binding sites, suggesting that this silencer represents a broader class of repressive elements. We leverage Swap-seq data to benchmark sequence-to-function models and show that they often capture the local activity of regulatory elements but fail to link them to distal target genes. Swap-seq thus provides a generalizable tool to discover regulatory elements across mechanistic classes, characterize their effects on gene expression, and generate large-scale datasets for evaluating sequence-to-function models.

For the purpose of differential expression (DE) analysis in single-cell RNA-sequencing (scRNA-Seq), phenotype differences between samples are often tested within specific cell types. Cell types are regularly imputed by clustering the same gene expression data which is later used for phenotype testing. This creates the potential for a "double-dipping" or post-selection inference problem resulting in inflated rates of false discoveries. While this selection bias is known to inflate significance in cell-type marker identification, its effect on sample-level phenotype testing, e.g. in patient cohorts, has never been explored despite the growing preponderance of this type of analysis. To address this, we perform an extensive simulation study and demonstrate that naive clustering on uncorrected embeddings can severely inflate the False Discovery Rate (FDR) in the presence of strong phenotypic differences. However, we further show that applying batch-correction methods to remove phenotypic effects prior to clustering resolves the FDR inflation with no obvious loss of power. Finally, we provide measures of phenotypic imbalance that can be applied to real datasets which closely track the false discovery proportion and thus can be used to as part of data exploration to gauge the risk of post-selection inflation of p-values in a particular dataset.

Approximately 8% of the human genome is derived from ancient retroviral infections. The most recently integrated of these endogenous retroviruses is the HERV-K(HML-2) clade, whose expression has been associated with cancer, amyotrophic lateral sclerosis, and embryogenesis. Studies of HERV expression, particularly HML-2, have relied predominantly on short-read sequencing. However, the high similarity among HML-2 proviruses prevents many short reads from being assigned uniquely to individual loci. We therefore compared haplotype-resolved long-read genome assemblies from 292 donors to resolve variation in proviral structure and coding capacity. Several loci previously thought to be fixed were structurally polymorphic. Tandem arrays occurred at 13 loci and contained up to six proviral copies in a single array. At 8q11.23, we identified a previously undescribed full-length provirus in one haplotype. All 583 other haplotypes carried a solo-LTR. We found that standard reference genomes failed to represent the coding capacity retained in many individuals, whose proviruses contained intact open reading frames despite disruptive mutations in the reference sequences. Short-read genotypes left 32.5% of the tested donor-variant pairs unresolved at sites associated with viral reading frames. These findings show why HML-2 expression must be interpreted in the context of the structural and coding alleles each individual carries.

Recurrent pregnancy loss (RPL) affects 5% of couples worldwide. Over half of cases remain unexplained (uRPL), in part because paternal factors are rarely evaluated beyond paternal karyotype. Sperm DNA methylation is a putative paternal contributor to RPL. Using ONT direct whole methylome sequencing, we generated the first and largest cohort of locus-resolved, single-molecule sperm methylome maps from men with uRPL to date (n=39). We identified 294 differentially methylated regions (DMRs), of which 158 (53.7%) overlapped AluY repetitive elements, representing significant enrichment (z=40.9); most were hypomethylated in uRPL sperm and spanned the full length of individual AluY insertions. Because of single-molecule resolution, we show that the methylation differences found in uRPL sperm occur in a subpopulation of sperm rather than a uniform DNA methylation change across all sperm. Affected loci were underrepresented for imprinting-associated chromatin marks and instead overrepresented for H2A.Z. Although DMR-associated genes were enriched for neuronal GO terms, many converge on reported roles in implantation, placentation, and early embryonic development. AluY epimutation represents a previously unrecognized, predominant signature of paternal epigenetic perturbation in uRPL, establishing repetitive elements as a new axis of sperm epigenetic risk.

Aging is characterized by a progressive loss of proteostasis. Transfer RNAs (tRNAs) are essential regulators of translation, yet their dynamics during aging remain poorly understood due to challenges in sequencing highly modified RNAs. Here we present a benchmarked Nanopore direct RNA sequencing (RNA004 chemistry) resource that profiles the Saccharomyces cerevisiae tRNAome during replicative aging at single-molecule resolution. Using in vitro transcribed tRNA controls, we establish modification detection thresholds and validate key findings with orthogonal Illumina sequencing. While overall tRNA abundance remains largely stable, our resource reveals age-associated terminal A cleavage at the 3' CCA tail of mature tRNAs, targeted T-loop and anticodon modification changes, and single-molecule evidence of modification co-occurrence. This dataset provides a resource for exploring tRNA regulation, translation fidelity, and longevity.

Pooled Perturb-seq screens are often interpreted through a linear steady-state model in which each targeted gene supplies an additive perturbation input. Although this formulation identifies an operator on the observed response subspace in closed form, identifiability alone does not establish that inferred operators generalize to new perturbation targets. We evaluated fixed program-space additive inputs in three Perturb-seq screens -- Replogle K562 essential, Replogle RPE1 essential, and Jost 2020 -- using target-grouped nested cross-validation and matched signal-to-noise linear-truth controls. For inverse operator prediction, real held-out error remained at the predict-zero baseline (0.96, 1.00, and 1.00), whereas matched linear controls achieved 0.18, 0.53, and 0.72. This discrepancy persisted across random K562 target panels, an overdetermined Jost dimension sweep, direction-only fits, ridge and truncated-SVD estimators, and five linear ground-truth ensembles. Forward target-held-out prediction was more nuanced: fixed, footprint, and learned linear encodings modestly beat a predict-training-mean baseline in K562 but not in RPE1 or Jost, and the K562 gains recovered only about 7% of the gap to the matched linear control. On the tested screens, fitted fixed program-space operators should not be interpreted as validated target-generalizing regulatory operators. We release anchor-op, a reproducible toolkit for target-grouped evaluation, identifiability checks, and matched-geometry positive controls for perturbation-response models.