Leveraging whole-exome sequence data from diverse biobanks and cohorts to study rare coding variation in prostate cancer

摘要

Abstract There is strong evidence that prostate cancer (PCa) is a heritable phenotype. In addition to greater risk observed in men with a family history of PCa, genome-wide association studies (GWAS) have identified over 400 common independent risk variants, which explain ~40% of the familial risk. It is increasingly recognized that much of the unknown heritability for PCa may also be due to variants of low minor allele frequency (<1%). While large, multi- ancestry genome-wide reference panels (e.g., TOPMed) have been developed to facilitate studies of less common alleles (down to 0.1%), they cannot be used to enumerate and accurately study very rare alleles that can only be characterized via sequencing. Pathogenic variants in DNA repair pathway genes (e.g., BRCA2, ATM, NBN, CHEK2, PALB2), identified through candidate gene studies, provide strong support for exceedingly rare (<0.1%) protein coding variation contributing to overall PCa and aggressive disease susceptibility. Unfortunately, we remain limited in our ability to comprehensively survey and study very rare variation genome- or exome-wide due to high sequencing costs, limiting current sample sizes. Here, we propose to combine existing whole-exome (WES) and whole-genome (WGS) sequence data from multi-ancestry biobanks and cohorts to conduct the first, large-scale study of rare coding variation in PCa and to integrate tumor somatic and germline mutation data to elucidate the biology of gene-risk associations. In Aim 1, we will leverage existing WES data for >90,000 PCa cases (58,000 European ancestry, 20,000 African ancestry, 4,000 Asian ancestry and 6,700 Latino/Hispanic) and >500,000 controls within biobanks and cohorts in the US and UK and conduct exome-wide analyses of overall PCa and aggressive disease phenotypes. In Aim 2, we will examine the combined effect of rare coding variants and a polygenic risk score (PRS) on risk of overall PCa and aggressive disease and estimate absolute risks for the combined effects of rare coding variants and PRS in prospective biobanks and cohorts across populations. In Aim 3, we will integrate somatic tumor and germline variation data to inform genes and biological pathways involved in PCa and aggressive disease. For this Aim, we have assembled a somatic resource of >7,000 PCa patients with germline exome/PRS data and somatic mutation profiling from WES and WGS studies, including >2,000 with transcriptomic data. We expect this study to provide the most comprehensive and well-powered examination of rare coding variation in PCa across populations to date. Findings from this study will have immediate clinical translation by informing personalized risk prediction and the development of novel risk-based screening strategies for overall and aggressive PCa. Integrating germline and somatic data will also define biological mechanisms that may be clinically important for understanding how to treat and prevent PCa and lethal disease across populations.