Whole Genome Sequencing Defines the Genetic Heterogeneity of Familial Pancreatic Cancer

Abstract

Pancreatic cancer is projected to become the second leading cause of cancer-related death in the United States by 2020. A familial aggregation of pancreatic cancer has been established, but the cause of this aggregation in most families is unknown. To determine the genetic basis of susceptibility in these families, we sequenced the germline genomes of 638 patients with familial pancreatic cancer and the tumor exomes of 39 familial pancreatic adenocarcinomas. Our analyses support the role of previously identified familial pancreatic cancer susceptibility genes such as BRCA2, CDKN2A, and ATM, and identify novel candidate genes harboring rare, deleterious germline variants for further characterization. We also show how somatic point mutations that occur during hematopoiesis can affect the interpretation of genome-wide studies of hereditary traits. Our observations have important implications for the etiology of pancreatic cancer and for the identification of susceptibility genes in other common cancer types.

Significance: The genetic basis of disease susceptibility in the majority of patients with familial pancreatic cancer is unknown. We whole genome sequenced 638 patients with familial pancreatic cancer and demonstrate that the genetic underpinning of inherited pancreatic cancer is highly heterogeneous. This has significant implications for the management of patients with familial pancreatic cancer.

Figure 1

A. Overview of filter-based strategy to identify novel candidate familial pancreatic cancer susceptibility genes. 638 familial pancreatic cancer (FPC) patients from 593 kindreds. FPC patients were selected from 10 high-risk family registries in North America with diverse ascertainment screens including: internet recruitments, medical genetics clinics, and tertiary care facilities. Demographic and sample data for the 638 FPC patients is shown in Table 1. FPC patient samples were whole genome sequenced, aligned to the human genome build hg19 before variant calling and annotation. One FPC patient from each kindred was arbitrarily selected for filter-based analyses. All germline variants in selected FPC patients were identified and, depending on analysis, filtered by: 1) functional consequence of variant, 2) frequency of the variant in FPC patients and variant databases (the 1000 Genomes Project and EVS), 3) zygosity of variant, and 4) variant quality. The following analyses were then performed on filtered variants: 1) analysis of premature truncating variants (PTVs), 2) in-depth analysis of selected genes, and 3) analysis of variant segregation in affected members of a kindred. B. Distribution of private heterozygous PTVs in FPC patients selected for filter-based analyses. Blue bars indicate number of genes within each PTV category. Grey bars indictate cumulative number of genes.