Biology subjects

Gusev, A.

Publications and source records attributed to Gusev, A..

At least 37 records · Page 2Linked to original sources

bioRxiv2021

Allele-specific epigenetic activity in prostate cancer and normal prostate tissue implicates prostate cancer risk mechanisms

BackgroundGenome-wide association studies of prostate cancer have identified >250 significant risk loci, but the causal variants and mechanisms for these loci remain largely unknown. Here, we sought to identify and characterize risk harboring regulatory elements by integrating epigenomes from primary prostate tumor and normal tissues of 27 patients across the H3K27ac, H3K4me3, and H3K4me2 histone marks and FOXA1 and HOXB13 transcription factors. ResultsWe identified 7,371 peaks with significant allele-specificity (asQTL peaks). Showcasing their relevance to prostate cancer risk, H3K27ac T-asQTL peaks were the single annotation most enriched for prostate cancer GWAS heritability (40x), significantly higher than corresponding non-asQTL H3K27ac peaks (14x) or coding regions (14x). Surprisingly, fine-mapped GWAS risk variants were most significantly enriched for asQTL peaks observed in tumors, including asQTL peaks that were differentially imbalanced with respect to tumor-normal states. These data pinpointed putative causal regulatory elements at 20 GWAS loci, of which 11 were detected only in the tumor samples. More broadly, tumor-specific asQTLs were enriched for expression QTLs in benign tissues as well as accessible regions found in stem cells, supporting a hypothesis where some germline variants become reactivated during/after transformation and can be captured by epigenomic profiling of the tumor. ConclusionOur study demonstrates the power of allele-specificity in chromatin signals to uncover GWAS mechanisms, highlights the relevance of tumor-specific regulation in the context of cancer risk, and prioritizes multiple loci for experimental follow-up.

cancer biology↗

bioRxiv2021

Seeing beyond the target: Leveraging off-target reads in targeted clinical tumor sequencing to identify prognostic biomarkers

Clinical tumor sequencing is rapidly becoming a standard component of clinical care, providing essential information for selecting amongst treatment options and providing prognostic value. Here we develop a robust and scalable software platform (SBT: Seeing Beyond the Target) that mines discarded components of clinical sequences to produce estimates of a rich set of omics features including rDNA and mtDNA copy number, microbial species abundance, and T and B cell receptor sequences. We validate the accuracy of SBT via comparison to multimodal data from the TCGA and apply SBT to a tumor panel cohort of 2,920 lung adenocarcinomas to identify associations of clinical value. We replicated known associations of somatic events in TP53 with changes in rDNA (p=0.012); as well as diversity of BCR and TCR repertoires with the biopsy site (p=2.5x10-6, p<10-20). We observed striking differences in EGFR mutant lung cancers versus wild-type, including higher rDNA copy number and lower immune repertoire diversity. Integrating clinical outcomes, we identified significant prognostic associations with overall survival, including SBT estimates of 5S rDNA (p=1.9x10-4, hazard ratio = 1.22) and TCR diversity (p=2.7x10-3, hazard ratio=1.77). Both novel survival associations replicated in 1,302 breast carcinoma and 1,651 colorectal cancer tumors. We anticipate that feature estimates derived by SBT will yield novel biomarker hypotheses and open research opportunities in existing and emerging clinical tumor sequencing cohorts.

genomics↗

bioRxiv2021

Genetic determinants of chromatin reveal prostate cancer risk mediated by context-dependent gene regulation

Methods that link genetic variation to steady-state gene expression levels, such as expression quantitative trait loci (eQTLs), are widely used to functionally annotate trait-associated variants, but they are limited in identifying context-dependent effects on transcription. To address this challenge, we developed the cistrome-wide association study (CWAS), a framework for nominating variants that impact traits through their effects on chromatin state. CWAS associates the genetic determinants of cistromes (e.g., the genome-wide profiles of transcription factor binding sites or histone modifications) with traits using summary statistics from genome-wide association studies (GWAS). We performed CWASs of prostate cancer and androgen-related traits, using a reference panel of 307 prostate cistromes from 165 individuals. CWAS nominated susceptibility regulatory elements or androgen receptor (AR) binding sites at 52 out of 98 known prostate cancer GWAS loci and implicated an additional 17 novel loci. We functionally validated a subset of our results using CRISPRi and in vitro reporter assays. At 28 of the 52 risk loci, CWAS identified regulatory mechanisms that are not observable via eQTLs, implicating genes with complex or context-specific regulation that are overlooked by current approaches that relying on steady-state transcript measurements. CWAS genes include transcription factors that govern prostate development such as NKX3-1, HOXB13, GATA2, and KLF5. Moreover, CWAS boosts discovery power in modestly sized GWAS, identifying novel genetic associations mediated through AR binding for androgen-related phenotypes, including resistance to prostate cancer therapy. CWAS is a powerful and biologically interpretable paradigm for studying variants that influence traits by affecting context-dependent transcriptional regulation.

genetics↗

bioRxiv2021

Germline testing data validate inferences of mutational status for variants detected from tumor-only sequencing

Structured AbstractO_ST_ABSBackgroundC_ST_ABSPathogenic germline variants (PGV) in cancer susceptibility genes are usually identified in cancer patients through germline testing of DNA from blood or saliva: their detection can impact patient treatment options and potential risk reduction strategies for relatives. PGV can also be identified, in tumor sequencing assays, often performed without matched normal specimens. It is then critical to determine whether detected variants are somatic or germline. Here, we evaluate the clinical utility of computational inference of mutational status in tumor-only sequencing compared to germline testing results. Patients and MethodsTumor-only sequencing data from 1,608 patients were retrospectively analyzed to infer germline-versus-somatic status of variants using an information-theoretic, gene-independent approach. Loss of heterozygosity (LOH) was also determined. The predicted mutational models were compared to clinical germline testing results. Statistical measures were computed to evaluate performance. ResultsTumor-only sequencing detected 3,988 variants across 70 cancer susceptibility genes for which germline testing data were available. Our analysis imputed germline-versus-somatic status for >75% of all detected variants, with a sensitivity of 65%, specificity of 88%, and overall accuracy of 86% for pathogenic variants. False omission rate was 3%, signifying minimal error in misclassifying true PGV. A higher portion of PGV in known hereditary tumor suppressors were found to be retained with LOH in the tumor specimens (72%) compared to variants of uncertain significance (58%). ConclusionsTumor-only sequencing provides sufficient power to distinguish germline and somatic variants and infer LOH. Although accurate detection of PGV from tumor-only data is possible, analyzing sequencing data in the context of specimens tumor cell content allows systematic exclusion of somatic variants, and suggests a balance between type 1 and 2 errors for identification of patients with candidate PGV for standard germline testing. Our approach, implemented in a user-friendly bioinformatics application, facilities objective analysis of tumor-only data in clinical settings. HighlightsO_LIMost pathogenic germline variants in cancer predisposition genes can be identified by analyzing tumor-only sequencing data. C_LIO_LIInformation-theoretic gene-independent analysis of common sequencing data accurately infers germline vs. somatic status. C_LIO_LIA reasonable statistical balance can be established between sensitivity and specificity demonstrating clinical utility. C_LIO_LIPathogenic germline variants are more often detected with loss of heterozygosity vs. germline variants of uncertain significance. C_LI

bioinformatics↗

bioRxiv2021

Somatic mutational profiles and germline polygenic risk scores in human cancer

The mutational profile of a cancer reflects the activity of the mutagenic processes which have been operative throughout the lineage of the cancer cell. These processes leave characteristic profiles of somatic mutations called mutational signatures. Mutational signatures, including single-based substitution (SBS) signatures, may reflect the effects of exogenous or endogenous exposures. Here, we used polygenic risk score (PRS) as proxies for exposures and examined the association between somatic mutational profiles and germline PRS in 12 cancer types from The Cancer Genome Atlas project. We found 17 statistically significant associations after Bonferroni correction (p < 3.15x10-5), including positive associations between germline inflammatory bowel disease PRS and number of somatic mutations of signature SBS1 in prostate cancer and APOBEC-related signatures in breast cancer. The age at menarche PRS was inversely associated with mutation counts of SBS1 in prostate cancer. Our analysis suggests that there are robust associations between tumor somatic mutational profiles and germline PRS. These may reflect mechanisms through hormone regulation and immunological responses that contribute to cancer etiology and drive cancer progression.

genomics↗

bioRxiv2020

H3k27ac-HiChIP in prostate cell lines identifies risk genes for prostate cancer susceptibility

Genome-wide association studies (GWAS) have identified more than 140 prostate cancer (PrCa) risk regions which provide potential insights into causal mechanisms. Multiple lines of evidence show that a significant proportion of PrCa risk can be explained by germline causal variants that dysregulate nearby target genes in prostate-relevant tissues thus altering disease risk. The traditional approach to explore this hypothesis has been correlating GWAS variants with steady-state transcript levels, referred to as expression quantitative trait loci (eQTLs). In this work, we assess the utility of chromosome conformation capture (3C) coupled with immunoprecipitation (HiChIP) to identify target genes for PrCa GWAS risk loci. We find that interactome data confirms previously reported PrCa target genes identified through GWAS/eQTL overlap (e.g., MLPH). Interestingly, HiChIP identified links between PrCa GWAS variants and genes well-known to play a role in prostate cancer biology (e.g., AR) that are not detected by eQTL-based methods. We validate these findings through CRISPR interference (CRISPRi) perturbation of the variant-containing regulatory elements for NKX3-1 and AR in the LNCaP cell line. Our results demonstrate that looping data harbor additional information beyond eQTLs and expand the number of PrCa GWAS loci that can be linked to candidate susceptibility genes.

genomics↗

bioRxiv2020

Leveraging expression from multiple tissues using sparse canonical correlation analysis (sCCA) and aggregate tests improves the power of transcriptome-wide association studies (TWAS)

Transcriptome-wide association studies (TWAS) test the association between traits and genetically predicted gene expression levels. The power of a TWAS depends in part on the strength of the correlation between a genetic predictor of gene expression and the causally relevant gene expression values. Consequently, TWAS power can be low when expression quantitative trait locus (eQTL) data used to train the genetic predictors have small sample sizes, or when data from causally relevant tissues are not available. Here, we propose to address these issues by integrating multiple tissues in the TWAS using sparse canonical correlation analysis (sCCA). We show that sCCA-TWAS combined with single-tissue TWAS using an aggregate Cauchy association test (ACAT) outperforms traditional single-tissue TWAS. In empirically motivated simulations, the sCCA+ACAT approach yielded the highest power to detect a gene associated with phenotype, even when expression in the causal tissue was not directly measured, while controlling the Type I error when there is no association between gene expression and phenotype. For example, when gene expression explains 2% of the variability in outcome, and the GWAS sample size is 20,000, the average power difference between the ACAT combined test of sCCA features and single-tissue, versus single-tissue combined with Generalized Berk-Jones (GBJ) method, single-tissue combined with S-MultiXcan or summarizing cross-tissue expression patterns using Principal Component Analysis (PCA) approaches was 5%, 8%, and 38%, respectively. The gain in power is likely due to sCCA cross-tissue features being more likely to be detectably heritable. When applied to publicly available summary statistics from 10 complex traits, the sCCA+ACAT test was able to increase the number of testable genes and identify on average an additional 400 additional gene-trait associations that single-trait TWAS missed. Our results suggest that aggregating eQTL data across multiple tissues using sCCA can improve the sensitivity of TWAS while controlling for the false positive rate.Author summary Transcriptome-wide association studies (TWAS) can improve the statistical power of genetic association studies by leveraging the relationship between genetically predicted transcript expression levels and an outcome. We propose a new TWAS pipeline that integrates data on the genetic regulation of expression levels across multiple tissues. We generate cross-tissue expression features using sparse canonical correlation analysis and then combine evidence for expression-outcome association across cross- and single-tissue features using the aggregate Cauchy association test. We show that this approach has substantially higher power than traditional single-tissue TWAS methods. Application of these methods to publicly available summary statistics for ten complex traits also identifies associations missed by single-tissue methods.View Full Text

genetics↗

bioRxiv2020

Identity-by-descent detection across 487,409 British samples reveals fine-scale population structure, evolutionary history, and trait associations

Detection of Identical-By-Descent (IBD) segments provides a fundamental measure of genetic relatedness and plays a key role in a wide range of genomic analyses. We developed a new method, called FastSMC, that enables accurate biobank-scale detection of IBD segments transmitted by common ancestors living up to several hundreds of generations in the past. FastSMC combines a fast heuristic search for IBD segments with accurate coalescent-based likelihood calculations and enables estimating the age of common ancestors transmitting IBD regions. We applied FastSMC to 487,409 phased samples from the UK Biobank and detected the presence of [~]214 billion IBD segments transmitted by shared ancestors within the past 1,500 years. We quantified time-dependent shared ancestry within and across 120 postcodes, obtaining a fine-grained picture of genetic relatedness within the past two millennia in the UK. Sharing of common ancestors strongly correlates with geographic distance, enabling the localization of a samples birth coordinates from genomic data. We sought evidence of recent positive selection by identifying loci with unusually strong shared ancestry within recent millennia and we detected 12 genome-wide significant signals, including 7 novel loci. We found IBD sharing to be highly predictive of the sharing of ultra-rare variants in exome sequencing samples from the UK Biobank. Focusing on loss-of-function variation discovered using exome sequencing, we devised an IBD-based association test and detected 29 associations with 7 blood-related traits, 20 of which were not detected in the exome sequencing study. These results underscore the importance of modelling distant relatedness to reveal subtle population structure, recent evolutionary history, and rare pathogenic variation.

genetics↗

bioRxiv2020

Mechanism of monoallelic expression and allelic rheostat role of DNA methylation

In diploid cells, maternal and paternal copies of genes usually have similar transcriptional activity. Mammalian allele-specific epigenetic mechanisms such as X-chromosome inactivation (XCI) and imprinting were historically viewed as rare exceptions to this rule. The discovery of mitotically stable monoallelic autosomal expression (MAE) a decade ago revealed an additional allele-specific mode regulating thousands of mammalian genes. However, despite its prevalence, the mechanistic basis of MAE remains unknown. To uncover the mechanism of MAE maintenance, we devised a small-molecule screen for reactivation of silenced alleles across multiple loci using targeted RNA sequencing. Contrary to previous reports, we identified DNA methylation as a key mechanism of MAE mitotic maintenance. In contrast with the binary choice of the active allele in XCI, stringent transcriptome-wide analysis revealed MAE as a regulatory mode with tunable control of allele-specific expression, dependent on the extent of DNA methylation. In a subset of MAE genes, allelic imbalance was insensitive to changes in DNA methylation, implicating additional mechanisms in MAE maintenance in these loci. Our findings identify a key mechanism of MAE maintenance, reveal tunability of this mode of gene regulation, and provide the essential platform for probing the biological role of MAE in development and disease.

genomics↗

bioRxiv2019

Predicting master transcription factors from pan-cancer expression data

The function of critical developmental regulators can be subverted by cancer cells to control expression of oncogenic transcriptional programs. These "master transcription factors" (MTFs) are often essential for cancer cell survival and represent vulnerabilities that can be exploited therapeutically. The current approaches to identify candidate MTFs examine super-enhancer associated transcription factor-encoding genes with high connectivity in network models. This relies on chromatin immunoprecipitation-sequencing (ChIP-seq) data, which is technically challenging to obtain from primary tumors, and is currently unavailable for many cancer types and clinically relevant subtypes. In contrast, gene expression data are more widely available, especially for rare tumors and subtypes where MTFs have yet to be discovered. We have developed a predictive algorithm called CaCTS (Cancer Core Transcription factor Specificity) to identify candidate MTFs using pan-cancer RNA-sequencing data from The Cancer Genome Atlas. The algorithm identified 273 candidate MTFs across 34 tumor types and recovered known tumor MTFs. We also made novel predictions, including for cancer types and subtypes for which MTFs have not yet been characterized. Clustering based on MTF predictions reproduced anatomic groupings of tumors that share 1-2 lineage-specific candidates, but also dictated functional groupings, such as a squamous group that comprised five tumor subtypes sharing 3 common MTFs. PAX8, SOX17, and MECOM were candidate factors in high-grade serous ovarian cancer (HGSOC), an aggressive tumor type where the core regulatory circuit is currently uncharacterized. PAX8, SOX17, and MECOM are required for cell viability and lie proximal to super-enhancers in HGSOC cells. ChIP-seq revealed that these factors co-occupy HGSOC regulatory elements globally and co-bind at critical gene loci including MUC16 (CA-125). Addiction to these factors was confirmed in studies using THZ1 to inhibit transcription in HGSOC cells, suggesting early down-regulation of these genes may be responsible for cytotoxic effects of THZ1 on HGSOC models. Identification of MTFs across 34 tumor types and 140 subtypes, especially for those with limited understanding of transcriptional drivers paves the way to therapeutic targeting of MTFs in a broad spectrum of cancers.

cancer biology↗

bioRxiv2019

Quantifying genetic effects on disease mediated by assayed gene expression levels

Disease variants identified by genome-wide association studies (GWAS) tend to overlap with expression quantitative trait loci (eQTLs). However, it remains unclear whether this overlap is driven by mediation of genetic effects on disease by expression levels, or whether it primarily reflects pleiotropic relationships instead. Here we introduce a new method, mediated expression score regression (MESC), to estimate disease heritability mediated by the cis-genetic component of assayed steady-state gene expression levels, using summary association statistics from GWAS and eQTL studies. We show that MESC produces robust estimates of expression-mediated heritability across a wide range of simulations. We applied MESC to GWAS summary statistics for 42 diseases and complex traits (average N = 323K) and cis-eQTL data across 48 tissues from the GTEx consortium. We determined that a statistically significant but low proportion of disease heritability (mean estimate 11% with S.E. 2%) is mediated by the cis-genetic component of assayed gene expression levels, with substantial variation across diseases (point estimates from 0% to 38%). We further partitioned expression-mediated heritability across various gene sets. We observed an inverse relationship between cis-heritability of expression and disease heritability mediated by expression, suggesting that genes with weaker eQTLs have larger causal effects on disease. Moreover, we observed broad patterns of expression-mediated heritability enrichment across functional gene sets that implicate specific gene sets in disease, including loss-of-function intolerant genes and FDA-approved drug targets. Our results demonstrate that eQTLs estimated from steady-state expression levels in bulk tissues are informative of regulatory disease mechanisms, but that such eQTLs are insufficient to explain the majority of disease heritability. Instead, additional assays are necessary to more fully capture the regulatory effects of GWAS variants.

genetics↗

bioRxiv2019

Allelic imbalance reveals widespread germline-somatic regulatory differences and prioritizes risk loci in Renal Cell Carcinoma

Determining the function of non-coding regulatory variants in cancer is a key challenge transcriptional biology. We investigated genetic (germline and somatic) determinants of regulatory mechanisms in renal cell carcinoma (RCC) using H3K27ac ChIP-seq data in 10 matched tumor/normal samples and RNA-seq data from 496/66 tumor/normal samples from The Cancer Genome Atlas (TCGA). Unsupervised clustering of H3K27ac activity cleanly separated tumor from normal individuals, highlighting extensive epigenetic reprogramming during transformation. We developed a novel method to test each chromatin feature for evidence of an allele-specific quantitative trait locus (asQTL) and evaluate tumor/normal differences in allele-specificity (d-asQTLs) while modelling local structural variation and read overdispersion. At an FDR of 5%, we identified 1,356 unique asQTL chromatin peaks in normal tissues; 2,868 in tumors; and 1,054 d-asQTLs (primarily imbalanced in tumor). The d-asQTL peaks were significantly enriched for RCC genome-wide association study (GWAS) heritability (32x, P=1.8x10-3), more so than any other functional feature including all H3K27ac peaks (12x), super-enhancers (5x), and asQTL genes (4x). Intersection of asQTLs with RCC GWAS loci identified putative functional features for 6/17 known loci including tumor-specific activity at SCARB1, a cholesterol metabolism mediator, which has recently been implicated in RCC progression. We validated the asQTL variant through CRISPR interference (CRISPRi) and demonstrated a concomitant allelic effect on the overlapping enhancer and on downstream SCARB1 expression. Knockdowns of master transcription factors (TFs) involved in the hypoxia pathway altered the expression of SCARB1 in a kidney cancer cell line, consistent with a variant-TF interaction. Genome-wide, d-asQTLs were significantly enriched for tumor-specific binding of hypoxic transcription factors, implicating a more general mechanism for polygenic germline-somatic interaction.

genetics↗

bioRxiv2019

Germline features associated with immune infiltration in solid tumors

Given the clinical success of immune checkpoint blockade (ICB) across a diverse set of solid tumors, and the emerging role for different immune infiltrates in contributing to response to ICB, a comprehensive assessment of the properties that dictate immune infiltrations may reveal new biological insights and inform the development of new effective therapies. Multiple studies have examined somatic and functional immune properties associated with different tumor infiltrates; however, germline features that associate with specific immune infiltrates in cancers have been incompletely characterized. Here, we analyzed over 7 million autosomal germline variants in the TCGA cohort (5788 European-ancestry samples across 30 cancer types) and tested for pan-cancer association with established immune-related phenotypes that describe the tumor immune microenvironment. We identified: one SNP associated with the fraction of follicular helper T cells in bulk tumor; 77 unique candidate genes, some of which are involved in cytokine-mediated signaling (e.g. CNTF and TRIM34) and cancer pathogenesis (e.g. ATR and AKAP9); and subnetworks with genes that are part of DNA repair (RAD51 and XPC) and transcription elongation (CCNT2) pathways. We found a positive association between polygenic risk for rheumatoid arthritis and absolute fraction of infiltrating CD8 T cells. Overall, we identified multiple germline genetic features associated with specific tumor-immune phenotypes across cancer, and developed a framework for probing inherited features that contribute to variation in immune infiltration.

genetics↗

bioRxiv2019

Non-coding Somatic Mutations Converge on the PAX8 Pathway in Epithelial Ovarian Cancer

Transcriptional regulation is highly disease and cell-type specific. We performed H3K27ac chromatin immunoprecipitation and transcriptomic sequencing in primary tumors for the four different subtypes of invasive epithelial ovarian cancer (OC). Histotype-specific regulatory elements (REs) were enriched in enhancers (P<0.001). In silico prediction of putative target genes for histotype-specific REs identified genes (WFDC2, P=5.5x10-5) and pathways (PI3K-Akt signaling, P<0.002) known to be involved in OC development. Some genes (e.g. PAX8 and CA125) are associated with super-enhancers (SEs) in all OCs, while others are histotype-specific, including PPP1R3B which is associated with SEs specific to clear cell OC. Integrated analysis of active chromatin landscapes with somatic single nucleotide variants (SNVs) from whole genome sequencing (WGS) of 232 primary OCs identified frequently mutated REs, including the KLF6 promoter (P=8.2x10-8) and a putative enhancer at chromosome 6p22.1 (P<0.05). In high-grade serous OCs, somatic SNVs clustered in binding sites for the PAX8 binding partner TEAD4 (P=6x10-11), while the collection of cis regulatory elements associated with PAX8 was the most frequently mutated set of enhancers in OC (P=0.003). Functional analyses supported our findings: Knockdown of PPP1R3B in clear cell OC cells significantly reduced intracellular glycogen content, a signature feature of this histotype; and stable knockout of a 635 bp region in the 6p22.1 enhancer induced downregulation of two predicted target genes, ZSCAN16 and ZSCAN12 (P=6.6 x 10-4 and P=0.02). In summary, we have characterized histotype-specific epigenomic and transcriptomic landscapes in OC and defined likely functional REs based on somatic mutation analysis of ovarian tumors.

cancer biology↗

bioRxiv2019

Tumor mutational landscape is a record of the pre-malignant state

Chromatin structure has a major influence on the cell-specific density of somatic mutations along the cancer genome. Here, we present a pan-cancer study in which we searched for the putative cancer cell-of-origin of 2,550 whole genomes, representing 32 cancer types by matching their mutational landscape to the regional patterns of chromatin modifications ascertained in 104 normal tissue types. We found that, in almost all cancer types, the cell-of-origin can be predicted solely from their DNA sequences. Our analysis validated the hypothesis that high-grade serous ovarian cancer originates in the fallopian tube and identified distinct origins of breast cancer subtypes. We also demonstrated that the technique is equally capable of identifying the cell-of-origin for a series of 2,044 metastatic samples from 22 of the tumor types available as primaries. Moreover, cancer drivers, whether inherited or acquired, reside in active chromatin regions in the respective cell-of-origin. Taken together, our findings highlight that many somatic mutations accumulate while the chromatin structure of the cell-of-origin is maintained and that this historical record, captured in the DNA, can be used to identify the often elusive cancer cell-of-origin.

genomics↗

bioRxiv2018

Annotations capturing cell-type-specific TF binding explain a large fraction of disease heritability

It is widely known that regulatory variation plays a major role in complex disease and that cell-type-specific binding of transcription factors (TF) is critical to gene regulation, but genomic annotations from directly measured TF binding information are not currently available for most cell-type-TF pairs. Here, we construct cell-type-specific TF binding annotations by intersecting sequence-based TF binding predictions with cell-type-specific chromatin data; this strategy addresses both the limitation that identical sequences may be bound or unbound depending on surrounding chromatin context, and the limitation that sequence-based predictions are generally not cell-type-specific. We evaluated different combinations of sequence-based TF predictions and chromatin data by partitioning the heritability of 49 diseases and complex traits (average N=320K) using stratified LD score regression with the baseline-LD model (which is not cell-type-specific). We determined that 100bp windows around MotifMap sequenced-based TF binding predictions intersected with a union of six cell-type-specific chromatin marks (imputed using ChromImpute) performed best, with an 58% increase in heritability enrichment compared to the chromatin marks alone (11.6x vs 7.3x; P = 9 x 10-14 for difference) and a 12% increase in cell-type-specific signal conditional on annotations from the baseline-LD model (P = 8 x 10-11 for difference). Our results show that intersecting sequence-based TF predictions with cell-type-specific chromatin information can help refine genome-wide association signals.

genetics↗

bioRxiv2018

Large-scale transcriptome-wide association study identifies new prostate cancer risk regions

Although genome-wide association studies (GWAS) for prostate cancer (PrCa) have identified more than 100 risk regions, most of the risk genes at these regions remain largely unknown. Here, we integrate the largest PrCa GWAS (N=142,392) with gene expression measured in 45 tissues (N=4,458), including normal and tumor prostate, to perform a multi-tissue transcriptomewide association study (TWAS) for PrCa. We identify 235 genes at 87 independent 1Mb regions associated with PrCa risk, 9 of which are regions with no genome-wide significant SNP within 2Mb. 24 genes are significant in TWAS only for alternative splicing models in prostate tumor thus supporting the hypothesis of splicing driving risk for continued oncogenesis. Finally, we use a Bayesian probabilistic approach to estimate credible sets of genes containing the causal gene at pre-defined level; this reduced the list of 235 associations to 120 genes in the 90% credible set. Overall, our findings highlight the power of integrating expression with PrCa GWAS to identify novel risk loci and prioritize putative causal genes at known risk loci.

genetics↗

bioRxiv2018

Multi-Tissue Transcriptome-Wide Association Studies Identify 21 Novel Candidate Susceptibility Genes for High Grade Serous Epithelial Ovarian Cancer

Genome-wide association studies (GWASs) have identified about 30 different susceptibility loci associated with high grade serous ovarian cancer (HGSOC) risk. We sought to identify potential susceptibility genes by integrating the risk variants at these regions with genetic variants impacting gene expression and splicing of nearby genes. We compiled gene expression and genotyping data from 2,169 samples for 6 different HGSOC-relevant tissue types. We integrated these data with GWAS data from 13,037 HGSOC cases and 40,941 controls, and performed a transcriptome-wide association study (TWAS) across >70,000 significantly heritable gene/exon features. We identified 24 transcriptome-wide significant associations for 14 unique genes, plus 90 significant exon-level associations in 20 unique genes. We implicated multiple novel genes at risk loci, e.g. LRRC46 at 19q21.32 (TWAS P=1x10-9) and a PRC1 splicing event (TWAS P=9x10-8) which was splice-variant specific and exhibited no eQTL signal. Functional analyses in HGSOC cell lines found evidence of essentiality for GOSR2, INTS1, KANSL1 and PRC1; with the latter gene showing levels of essentiality comparable to that of MYC. Overall, gene expression and splicing events explained 41% of SNP-heritability for HGSOC (s.e. 11%, P=2.5x10-4), implicated at least one target gene for 6/13 distinct genome-wide significant regions and revealed 2 known and 26 novel candidate susceptibility genes for HGSOC.\n\nSTATEMENT OF SIGNIFICANCEFor many ovarian cancer risk regions, the target genes regulated by germline genetic variants are unknown. Using expression data from >2,100 individuals, this study identified novel associations of genes and splicing variants with ovarian cancer risk; with transcriptional variation now explaining over one-third of the SNP-heritability for this disease.

genomics↗