Genomic Analysis of Inherited Breast and Ovarian Cancer

The discovery of BRCA1 and BRCA2 and other genes of homologous recombination repair, and the characterization of the roles of these genes in inherited predisposition to breast and ovarian cancer, have had a tremendous impact on cancer prevention and treatment. Yet family history remains one of the most important

risk factors for these cancers, and a great deal of that risk remains unexplained. We propose two hypotheses.: (1) In some families, breast cancer is due to previously cryptic non-coding variation altering transcription and/or expression of known breast cancer genes (BRCA1, BRCA2, PALB2, ATM, CHEK2, BARD1, BRIP1, RAD51C,

RAD51D, and TP53); and (2) In some families, breast cancer is due to rare or private non-coding variants regulating expression of critical genes with no cancer-predisposing alleles in coding sequence. To test these hypotheses, we will deploy new technologies: long-read sequencing of genomic and cDNA; Fiber-

seq identification of regions of open chromatin in cell types of choice; massively parallel reporter assays (MPRA) to compare effects of variant versus reference alleles for many sites simultaneously; and CRISPR inhibition (CRISPRi) of candidate regulatory regions. We propose to integrate these approaches to discover and

characterize classes of non-coding pathogenic variation in 1253 extended families, severely affected with breast and/or ovarian cancer, but for whom no causal variants have been found by any current genomic technology. In Aim 1, we will use adaptive-sampling long-read genomic sequencing to determine the spectrum of all classes

of rare non-coding variants in the extended TADs of the known breast cancer genes. In Aim 2. we will use long- read cDNA sequencing to characterize transcriptional consequences of the complex SVs and rare deep intronic SNVs co-segregating with breast cancer (from Aim 1). In Aim 3, we will use Fiber-seq analyses (already in hand)

of fallopian tube epithelial cells and MCF10A mammary cells to identify candidate regulatory regions of breast cancer genes. Then, for rare variants in these peaks that are co-inherited with breast cancer in unsolved families, we will use MPRA to test allele-specific differences in reporter activity; and in parallel silence the regions with

CRISPRi to evaluate their effects on gene expression. In Aim 4, we will extend Aim 3 genome-wide, to identify variants near other breast cancer-relevant genes that lie in open chromatin regions, that co-segregate with breast cancer in unsolved families, and that reveal allele-specific differences in regulatory activity.

We expect to discover clinically meaningful genetic variation in non-coding regulatory regions, including in repetitive genomic neighborhoods particularly subject to complex rearrangement. We expect that this approach will be directly applicable to any complex diseases with an inherited genetic influence.