Biology subjects

Supek, F.

Publications and source records attributed to Supek, F..

At least 19 recordsLinked to original sources

bioRxiv2026

Single genome analysis of normal cells identifies the early steps of kidney cancer.

Whole genome analyses have shown that every cell of the body accumulates thousands of genetic changes during a lifetime, with implications for cancer initiation. We have analyzed normal kidney cells from both healthy individuals and patients with the Von Hippel Lindau (VHL) syndrome, which predisposes to kidney cancer. Using a clonal-expansion method that enables the collection of high confidence genetic and gene expression data from the same cell, we define a new mutational signature that marks a specific subset of damage-activated proximal tubule cells, as well as 96% of clear cell renal cell carcinomas. The underlying mutational process is specifically active in highly transcribed DNA and impacts the occurrence of cancer driver variants in the VHL gene. Our data from normal cells, supported by in vitro experiments, point to perturbations of oxygen availability as a causative factor. These findings link hypoxia, a frequent endogenous stress in the normal renal tissue, to excessive mutation and cancer initiation.

cancer biology↗

bioRxiv2026

The genetic architecture of human programmed stop codon readthrough

Programmed translational readthrough produces C-terminally extended protein isoforms via decoding of stop codons by near-cognate tRNAs. Human genes experimentally validated as readthrough targets share a CUAG motif downstream of a UGA stop codon. However, the full sequence determinants of readthrough efficiency, how they combine, and how generalisable they are across genes remain largely unexplored. Here we use deep mutational scanning to quantify [~]1,400 sequence variants for each of the three examples of human readthrough in the genes AQP4, MAPK10 and OPRK1. In addition to the core CUAG motif, mutations that modulate readthrough elements extend up to +27 nucleotides downstream of the stop codon and across six codons (18 nucleotides) upstream. For the downstream sequence, an additive model with a sigmoidal global epistasis function captures most of the within-gene readthrough variance for double mutants (R{superscript 2}=0.84-0.96), with additional contributions from a small number of strong pairwise interactions. Mutational effects nonetheless generalise poorly between genes: only the immediate -3 to +4 nucleotide window shows consistent behaviour, while mutations in more distal positions have context-dependent effects. Combinatorial assembly of sequence blocks from different genes into chimeras reveals strong interactions (epistasis) between sequences upstream and downstream of the stop codon. This study provides comprehensive quantitative maps of the sequence determinants of human programmed readthrough and suggests that three examples of programmed readthrough are located on distinct local fitness peaks, each defined by different upstream and downstream architectures built around a shared CUAG core motif.

genomics↗

bioRxiv2026

Transformer models of mutation risk at base-pair resolution identify non-coding hotspot cancer driver mutations

Recurrent somatic mutations reveal cancer drivers, but in whole genomes many non-coding hotspots are passengers generated by localized mutational processes. We developed MutFormer, a transformer/convolutional neural net model that predicts base-pair-resolution somatic mutation risk from DNA sequence alone, separately for COSMIC signatures. Trained on >90 million high-confidence mutational signature-assigned SNVs from cancer genomes, MutFormer learns extended sequence determinants beyond trinucleotide context, often spanning up to ~20 nucleotides, and recovers APOBEC, UV, POLE and SBS17 sequence preferences as well as various additional mutation risk-prone motifs. We integrated MutFormer predictions with mutation burden, signature exposures and epigenetic covariates to model neutral recurrence of individual hotspots in >18,000 tumor whole genomes. Coding-region analyses calibrated the framework against known driver genes and AlphaMissense scores, supporting conservative false-discovery estimates. In non-coding regions, most recurrent hotspots were explained by passenger mutability, whereas selected outliers were enriched near cancer genes and supported by SpliceAI, PromoterAI, AlphaGenome and expression data. Prioritized candidates include splice-region or deep-intronic hotspots in BCL6, PTEN, TCF7L2, PBRM1, PTPRT and VHL, and promoter hotspots in SHKBP1, PRSS3 and BCL2.

genomics↗

bioRxiv2026

Multidimensional mutational phenotypes of MMR deficiency in human cancer cells

Mismatch repair deficiency (MMRd) is a central driver of cancer genome evolution, yet it is commonly reduced to a binary MSI-high versus microsatellite-stable call. Here we combine multiplex CRISPR/Cas12a editing, long-term mutation accumulation experiments in human cells, and whole-genome sequencing to map the mutational consequences of single and pairwise perturbations across canonical and accessory MMR genes, together with selected base excision repair and direct-repair factors. Across 72 isogenic lineages representing 58 genotypes, we recover reproducible substitution, indel and short-tandem-repeat mutational states that distinguish MSH2, MSH6, MLH1, PMS2, MSH3 and MLH3 deficiencies. A dedicated STR spectrum reveals genome-wide evidence for MSH3-driven EMAST and enables design of compact loci panels for mechanism-aware MSI classification. Experimental spectra generalize across independent cell systems and to cancer whole genome sequences, where Elastic-net classifiers accurately identify causal MMR genes in Lynch tumors and uncover common occurrences of intermediate MMRd states. Pairwise perturbations further reveal repair crosstalk reflected in mutational spectra, including enhancement of MUTYH-associated oxidative mutagenesis by MMR loss. Together, these findings recast mismatch repair deficiency as a multidimensional genomic phenotype and provide a framework for interpreting DNA repair deficiencies by bridging mutational phenotypes from experimental systems and human cancers.

genomics↗

bioRxiv2026

Quantitative prediction of nonsense-mediated mRNA decay across human genes by genomic language model and large-scale mutational scanning

The molecular consequences of protein truncating variants depend strongly on whether their transcripts are eliminated by nonsense-mediated mRNA decay (NMD), yet NMD is still predicted largely from a small set of binary positional rules. How individual premature termination codons (PTCs) engage NMD across genes and transcript contexts therefore remains incompletely resolved. Here we integrated endogenous allele-specific PTC expression from large-scale genomic data, mRNA language-model prediction and high-throughput mutational scanning to revisit the rules that govern mammalian NMD. Using allele-specific expression measurements from large human cohorts, we trained NMDetective-AI on [~]14,000 somatic PTCs, and after testing it on [~]1,800 germline PTCs, found that it improves on previous models, with its accuracy approaching the reproducibility of the underlying measurements. We then generated experimental maps of NMD with deep mutational scanning, including [~]450 PTCs nearby the 50-nt penultimate exon boundary, [~]950 engineered PTCs across 9 exon lengths to resolve long-exon escape, and [~]11k PTCs across 139 genes to quantify and refine start-proximal evasion. Our results support the positional logic of mammalian NMD yet show that this logic is implemented quantitatively: the classical rules resolve into graded, gene-dependent response curves whose boundaries are shaped by transcript architecture and modulated by local sequence context. Applying this framework to population, disease and cancer datasets further identifies genes in which NMD is predicted to aggravate or ameliorate the effects of truncating variants, providing a basis for variant interpretation and for prioritizing NMD-directed therapies.

genomics↗

bioRxiv2025

Intrinsic DNA sequence determinants and tissue-specific regulation of human replication origins

The accurate duplication of the genome relies on the spatiotemporal control of DNA replication initiation, yet the determinants specifying mammalian origin locations remain elusive. By developing ORIFormer, a transformer-based neural network, we decode a complex, conserved DNA sequence grammar that accurately predicts initiation sites. This approach uncovers novel determinants, which we validate by demonstrating selection and molecular effects of motif-altering genetic variants. To characterize tissue-specific usage, we developed MuSAS, a statistical genomic method leveraging widespread mutational strand asymmetries, such as from APOBEC and mismatch repair deficiency, to map initiation zones across diverse somatic tissues. Integrating these modalities reveals that while intrinsic DNA sequence features establish a high-potential landscape of constitutive origins, tissue-specific usage is governed by local chromatin accessibility acting as a permissive switch. We suggest that human replication initiation is driven by a deterministic genetic code modulated by the epigenetic landscape, providing a unified framework for understanding genome copying.

genomics↗

bioRxiv2025

Dual inhibition of the nonsense mediated mRNA decay enhances tumour immunogenicity, drives immunoediting, and potentiates checkpoint blockade

Immune checkpoint blockade has transformed cancer therapy, but current biomarkers such as tumour mutation burden often fail to reliably predict clinical benefit. One proposed reason for this discrepancy is the activity of nonsense-mediated mRNA decay (NMD), a cellular quality-control pathway that degrades mutant transcripts, potentially reducing the presentation of neoantigens that would otherwise stimulate anti-tumour immunity. To address this, we identified publicly available small-molecule inhibitors targeting the NMD factors SMG5 and SMG7 (NMDi), which we have prioritise as the NMD components most strongly associated with NMD efficiency across TCGA tumours, and evaluated their therapeutic potential in vivo. In a syngeneic, DNA repair deficient mouse model of lung adenocarcinoma, NMDi treatment alone substantially reduced tumour burden, and its combination with anti-PD-1 therapy led to additive benefit either treatment alone. These effects were immune-dependent and specifically required CD8{square} T cells. Transcriptomic and single-cell analyses revealed that NMDi reprograms the tumour immune microenvironment, enriching for clonally expanded, cytotoxic CD8{square} T cells and altering macrophage states toward those associated with tumour regression. Whole-genome sequencing of tumours revealed that NMDi also promotes immunoediting, leading to negative selection against immunogenic mutations and coding indels, with selective pressure comparable to or greater than that induced by anti-PD-1 treatment. Guided by these results, we generated a new protein language model trained on tumour mutations that can identify neopeptides with immunogenic properties, revealing that the NMDi tumour landscape provides a rich genomic and immunopeptidomic setting for exposing neoantigens that evade immunoediting. Together, these pre-clinical and integrative genomic-transcriptomic insights position NMD inhibition as a promising immunostimulatory strategy that can potentiate immune checkpoint blockade efficacy in tumours.

cancer biology↗

bioRxiv2025

Characterization of Somatic Mutations in Human tRNA Genes Reveals Tumor-Specific Mutational Loads, and the Generation of tRNA Variants that Alter the Genetic Code

Transfer RNA genes (tDNAs) are essential genomic elements that safeguard translational fidelity. Using the T2T version of the human genome we have mapped the position of human tDNAs and analyzed their individual transcriptional activities. Then we have characterized, at single base resolution, the impact of somatic mutations in human tDNAs and its relationship to the transcriptional status of each gene. We confirm that tDNAs are hotspots for somatic mutagenesis, and show that they display mutational loads that are directly proportional to their transcription rates. Highly transcribed tDNAs in tumors or healthy tissues accumulate mutations at rates up to nine-fold higher than highly transcribed protein-coding genes. Mutational loads at tDNAs are tumor-specific, and increase with patient age. Mutations at structurally conserved tRNA positions appear to be under negative selection. Strikingly, anticodon nucleotides crucial for decoding frequently acquire somatic mutations, readily generating chimeric tRNAs species capable of systematically introducing amino acid substitutions across the proteome. Our results reveal a previously unrecognized source of somatic heterogeneity in human cancer and aging tissues that may directly impact upon translation efficiency and fidelity, and cause cell-specific proteostasis degeneration.

genomics↗

bioRxiv2024

Allele-specific expression is selected in tumorigenesis, results from epigenetic changes and has prognostic relevance

Allelic imbalance (AI) in levels of mRNAs in cancer is widely appreciated to result from somatic copy number alterations (CNA) affecting one allele. Apart from CNA, other mechanisms could lead to imbalanced mRNA expression of the alleles, and similarly drive cancer evolution by epistatic interactions with the somatic mutations. By integrating genomic and transcriptomic pan-cancer data, we show that mRNA allelic imbalance favoring the mutant allele in driver genes is subject to positive selection, generating second-hit events often independently of somatic CNA. In some cases, the somatic coding mutations could induce allele-specific expression directly, e.g. with splicing-altering exonic mutations, which can be selected in various cancer genes. However, in the majority of cases, these and related somatic mutation effects (which might in principle alter transcription output via impacting promoters or intragenic enhancers) do not explain the CNA-independent mRNA-level AI, suggesting prevalent epigenetic alterations affecting alleles differently in tumors. Importantly, the mRNA AI events associate with worse overall survival across all cancer types, outperforming various other predictive markers. Our study suggests that mRNA allelic imbalances can occur independently of CNA but similarly function as second-hit events to somatic mutations, driving tumorigenesis, and so represent valuable prognostic biomarkers for cancer patient stratification.

cancer biology↗

bioRxiv2024

The DiffInvex evolutionary model for conditional somatic selection identifies chemotherapy resistance genes in 10,000 cancer genomes

Tumors often show an initial response to chemotherapy, but then develop resistance, leading to relapse and poor prognosis. We hypothesized that a genomic comparison of mutations in pre-treated versus treatment-naive tumors would serve to identify genes that confer resistance. A challenge in such an analysis is that therapy alters mutation burdens and signatures, confounding association studies and complicating identifying causal, selected mutations. We developed DiffInvex, a framework for identifying changes in selection acting on individual genes in somatic genomes. Crucially, DiffInvex draws on a mutation rate baseline that accounts for these shifts in neutral mutagenesis during cancer evolution. We applied DiffInvex to 9,953 cancer whole-genomes from 29 cancer types from 8 studies, containing both WGS of treatment-naive tumors and tumors pre-treated by various drugs, identifying genes where point mutations are under conditional positive or negative selection for a certain chemotherapeutic, suggesting resistance mechanisms occurring via point mutation. DiffInvex confirmed well-known chemoresistance-driver mutations in EGFR, ESR1, KIT and AR genes as being under conditional positive selection, with additional cancer types identified for EGFR and KIT. Additionally, DiffInvex identified 11 genes with treatment-associated selection for different classes of therapeutics. In most cases, these genes were common cancer genes including PIK3CA, APC, MAP2K4 and MAP3K1. This suggests that tumor resistance to therapy via mutation often occurs via selective advantages conferred by known driver genes, rather than via mutations in specialized resistance genes. Various gene-chemotherapy associations were further supported in tests for functional impact of mutations, again implemented in a conditional selection setting, as well as replicating in independent panel or exome sequencing data. In addition to nominating drug resistance genes that could be targeted by future therapeutics, DiffInvex can also be applied to diverse analysis in cancer evolution, such as comparing normal and tumoral tissues, or analyzing subclonal evolution, identifying changes in selection over time.

genomics↗

bioRxiv2024

Variable efficiency of nonsense-mediated mRNA decay across human tissues, tumors and individuals

Nonsense-mediated mRNA decay (NMD) is a quality-control pathway that degrades mRNA bearing premature termination codons (PTCs) resulting from mutation or mis-splicing, and that additionally participates in gene regulation of unmutated transcripts. We analyzed [~]10,000 exomes and [~]27,000 transcriptomes from human tumors and healthy tissues to quantify individual-level NMD efficiency, and assess its variability between tissues and between individuals. This was done by monitoring allele-specific expression of germline PTCs, and independently supported by mRNA levels of endogenous NMD target transcripts. Nervous system and reproductive system tissues have lower NMD efficiency than other tissues such as the digestive tract. Next, there is considerable systematic inter-individual variability in NMD efficiency, and we identify two underlying mechanisms. First, in cancers there are somatic copy number alterations that robustly associate with NMD efficiency, prominently the commonly-occurring gain at chromosome 1q that encompasses two core NMD genes: SMG5 and SMG7 and additional functionally interacting genes such as PMF1 and GON4L. Second, loss-of-function germline variants in various genes such as the KDM6B chromatin modifier can associate with higher or lower NMD efficiency in individuals, affecting different tissues thereof. Variable NMD efficiency should have clinical implications as it modulates positive selection upon somatic nonsense mutations in tumor suppressor genes, and is associated with survival of cancer patients, with relevance to predicting immunotherapy responses across cancer types.

genomics↗

bioRxiv2023

Mutation rate heterogeneity at the sub-gene scale due to local DNA hypomethylation

Local mutation rates are highly heterogeneous across the human genome. This variability was better studied at the scale of megabase-sized chromosomal domains on the one extreme, and at the scale of oligonucleotides at the other extreme. The intermediate, kilobase-scale heterogeneity in mutation risk was less studied. Here, by analyzing thousands of somatic genomes, we considered the hypothesis there are mutation risk gradients along gene bodies, representing a genomic scale spanning roughly 1 kb - 10 kb, and that different mutational mechanisms are differently distributed across gene segments. The main intragenic heterogeneity concerns several kilobases at the transcription start site and further into 5 ends of gene bodies, which are commonly hypomutated with respect to several mutational signatures, most prominently the ubiquitous mutational signature of C>T changes at CpG dinucleotides. Width and shape of this mutational coldspot at 5 gene ends is variable across genes, and corresponds to variable interval of lowered DNA methylation across genes. These hypomutated genic intervals correspond to hypomethylation that can originate from various causes, including intragenic enhancers, Polycomb-marked regions, or chromatin loop anchor points. Tissue-specific DNA hypomethylation begets tissue-specific local hypomutation. However, direction of mutation rate effect is inverted for some mutational processes, where signatures of AID/APOBEC3 cytosine deaminase activity are actually increased in hypomethylated regions. Overall, local DNA methylation determines mutation rate heterogeneity at the sub-gene level, and can generate either mutational coldspots or hotspots, depending on the mutagen exposure history of a cell.

genomics↗

bioRxiv2023

Genome-scale quantification and prediction of pathogenic stop codon readthrough by small molecules

Premature termination codons (PTCs) cause [~]10-20% of Mendelian diseases and are the major mechanism of tumor suppressor gene inactivation in cancer. A general strategy to alleviate the effects of PTCs would be to promote translational readthrough. Nonsense suppression by small molecules has proven effective in diverse disease models, but translation into the clinic is hampered by ineffective readthrough of many PTCs. Here we directly tackle the challenge of defining drug efficacy by quantifying readthrough of [~]5,800 human pathogenic stop codons by 8 drugs. We find that different drugs promote readthrough of complementary subsets of PTCs defined by local sequence context. This allows us to build interpretable models that accurately predict drug-induced readthrough genome-wide. Accurate readthrough quantification and prediction will empower clinical trial design and the development of personalized nonsense suppression therapies.

genomics↗

bioRxiv2023

Copy number losses of oncogenes and gains of tumor suppressor genes generate common driver events of human cancer

Cancer driver genes can be under positive selection for various types of genetic alterations, including gain-of-function or loss-of-function point mutations (single-nucleotide variants, SNV), small indels, copy number alterations (CNA) and other structural variants. We studied the landscape of interactions between these different types of alterations affecting the same gene by a statistical method, MutMatch, which can test for significant differences in selection, while accounting for various causes of mutation risk heterogeneity. Analyzing [~]18,000 cancer exomes and genomes, we found that known oncogenes simultaneously exhibit signatures of positive selection and also negative selection, where the latter can mask the former. Consistently, focussing on known positively selected regions identifies additional tumor types where an oncogene is relevant. Next, we characterized the landscape of CNA-dependent selection effects, revealing a general trend of increased positive selection on oncogene mutations not only upon CNA gains but also upon CNA deletions. Conversely, we observe a positive interaction between mutations and CNA gains in tumor suppressor genes. Thus, two-hit events involving point mutations and CNA are universally observed on driver genes regardless of the type of CNA, and may signal new therapeutic opportunities that have been overlooked. An explicit focus on the somatic CNA two-hit events can identify additional driver genes relevant to a tumor type. By a global analysis of CNA-selection effects across many driver genes and tissues, we identified at least four independently varying signatures, and thus generated a comprehensive, data-driven classification of cancer genes by mechanisms of (in)activation by genetic alterations.

genomics↗

bioRxiv2022

Prevalence, causes and impact of TP53-loss phenocopying events in human tumors

TP53 is a master tumor suppressor gene, mutated in approximately half of all human cancers. Given the many regulatory roles of the corresponding p53 protein, it is possible to infer loss of p53 activity -- which may occur from trans-acting alterations -- from gene expression patterns. We apply this approach to transcriptomes of ~8,000 tumors and ~1,000 cell lines, estimating that 12% and 8% of tumors and cancer cell lines phenocopy TP53 loss: they are likely deficient in the activity of the p53 pathway, while not bearing obvious TP53 inactivating mutations. While some of these are explained by amplifications in the known phenocopying genes MDM2, MDM4 and PPM1D, others are not. An analysis of cancer genomic scores jointly with CRISPR/RNAi genetic screening data identified an additional TP53-loss phenocopying gene, USP28. Deletions in USP28 are associated with a TP53 functional impairment in 2.9-7.6% of breast, bladder, lung, liver and stomach tumors, and are comparable to MDM4 amplifications in terms of effect size. Additionally, in the known CNA segments harboring MDM2, we identify an additional co-amplified gene (CNOT2) that may cooperatively boost the TP53 functional inactivation effect. An analysis using the phenocopy scores suggests that TP53 (in)activity commonly modulates associations between anticancer drug effects and relevant genetic markers, such as PIK3CA and PTEN mutations, and should thus be considered as a relevant interacting factor in personalized medicine studies. As a resource, we provide the drug-marker associations that differ depending on TP53 functional status.

genomics↗

bioRxiv2022

Redistribution of mutation rates across chromosomal domains in human cancer genomes

Somatic mutations in human cells have a highly heterogeneous genomic distribution, with increased burden in late-replication time (RT), heterochromatic domains of chromosomes. This regional mutation density (RMD) landscape is known to vary between cancer types, in association with tissue-specific RT or chromatin organization. Here, we hypothesized that regional mutation rates additionally vary between individual tumors in a manner independent of cell type, and that recurrent alterations in DNA replication programs and/or chromatin organization may underlie this. Here, we identified various RMD signatures that describe a global genome-wide mutation redistribution across many megabase-sized domains in >4000 tumors. We identified two novel global RMD signatures of somatic mutation landscapes that were universally observed across various cancer types. First, we identified a mutation rate redistribution preferentially affecting facultative heterochromatin, Polycomb-marked domains, and enriched in subtelomeric regions. This RMD signature strongly reflects regional plasticity in DNA replication time and in heterochromatin domains observed across tumors and cultured cells, which was linked with a stem-like phenotype and a higher expression of cell cycle genes. Consistently, occurrence of this global mutation pattern in cancers is associated with altered cell cycle control via loss of activity of the RB1 tumor suppressor gene. Second, we identified another independant global RMD signature associated with loss-of-function of the TP53 pathway, mainly affecting the redistribution of mutation rates away from late RT regions. The local mutation supply towards 26%-75% cancer driver genes is altered in the tumors affected by the global RMD signatures detected herein, including additionally a known pattern of a general loss of mutation rate heterogeneity due to DNA repair failures that we quantify. Our study highlights that somatic mutation rates at the domain scale are variable across tumors in a manner associated with loss of cell cycle control via RB1 or TP53, which may trigger the local remodeling of chromatin state and the RT program in cancers.

genomics↗

bioRxiv2022

Proton and alpha radiation-induced mutational profiles in human cells

Ionizing radiation (IR) is known to be DNA damaging and mutagenic, however less is known about which mutational footprints result from exposures of human cells to different types of IR. We were interested in the mutagenic effects of particle radiation exposures on genomes of various human cell types, in order to gauge the genotoxic risks of galactic cosmic radiation, and of certain types of tumor radiotherapy. To this end, we exposed cultured cell lines from the human blood, breast and lung to intermittent proton and alpha particle (helium nuclei) beams at doses sufficient to considerably affect cell viability. Whole-genome sequencing revealed that mutation rates were not overall markedly increased upon proton and alpha exposures. However, there were changes in mutation spectra and distributions, such as the increases in clustered mutations and of certain types of indels and structural variants. The spectrum of mutagenic effects of particle beams may be cell-type and/or genetic background specific. Overall, the mutational effects of recurrent exposures to proton and alpha radiation on human cells in culture appear subtle, however further work is warranted to understand effects of long-term exposures on various human tissues.

genetics↗

bioRxiv2021

The impact of rare germline variants on human somatic mutation processes

Somatic mutations are an inevitable component of ageing and the most important cause of cancer. The rates and types of somatic mutation vary across individuals, but relatively few inherited influences on mutation processes are known. We performed a comprehensive gene-based rare variant association study with diverse mutational processes, using human cancer genomes from over 11,000 individuals of European ancestry. By combining burden and variance tests, we identify 207 associations involving 15 somatic mutational phenotypes and 42 genes that replicated in an independent data set at a FDR of 1%. We associated rare inherited deleterious variants in novel genes such as MSH3, EXO1, SETD2, and MTOR with two different forms of DNA mismatch repair deficiency, and variants in genes such as EXO1, PAXIP1, and WRN with deficiency in homologous recombination repair. In addition, we identified associations with other mutational processes, such as APEX1 with APOBEC-signature mutagenesis. Many of the novel genes interact with each other and with known mutator genes within cellular sub-networks. Considered collectively, damaging variants in the newly-identified genes are prevalent in the population. We suggest that rare germline variation in diverse genes commonly impacts mutational processes in somatic cells.

genomics↗