David Quigley, PhD

The genomics of prostate cancer
I am a translational geneticist focused on improving outcomes for patients with metastatic prostate cancer. My research focuses on understanding how defects in DNA repair genes affect tumor genomes, and how to detect and overcome resistance to targeted PARP inhibitor therapy. I am also developing methods to model DNA repair defects in prostate and breast cancer in the laboratory.

During the last two years I have led analyses that identified of multi-clonal BRCA2 reversion mutations in metastatic prostate cancer patients who developed resistance to talazoparib and olaparib using liquid biopsy (Quigley et al. Cancer Discovery 2017) and deep whole genome and transcriptome sequencing of 101 prostate cancer metastases (Quigley et al, Cell, 2018). Key findings from our whole genome study include: Data and analysis from this paper are available on this website.

Genetics and Genomics of cancer susceptibility
During my graduate training with Allan Balmain, Anne-Lise Børresen-Dale, and Vessela Kristensen I identified candidate mechanisms of germline cancer susceptibility variants in mouse models of skin cancer and human breast cancer by combining bioinformatic and genetic approaches. To this end, I developed methods for expression Quantitative Trait Loci (eQTL) network analysis to identify germline variants that influence genes with a common functional role. eQTL network analysis identifies the causal effects of individual variants on organism-level phenotypes while simultaneously identifying a mechanism of action.

The first step in understanding the genetic architecture of human disease is to identify germline and acquired genetic variants associated with disease initiation or progression.

A cancer susceptibility locus may encompass hundreds of genes. To identify the relevant targets of the variant we have taken a systems genetics approach that measures the influence of that locus on intermediate phenotypes such as mRNA expression or DNA copy number changes. This influence is then synthesized through a data-driven approach guided by a biological understanding of the affected tissue. I have used genetically heterogeneous mouse models of skin cancer to identify networks of expression Quantitative Trait Loci (QTL) associated with sets of genes that are functionally related because they produce proteins that are part of a common structure or because they are steps in a gene pathway. This approach is reviewed in (Quigley et al. Nature Reviews: Cancer 2009).

Combining these networks with phenotype QTL identified plausible candidate genes under common genetic control with matched phenotypes (e.g., hemoglobin pathway genes that share common genetic control with the Red Blood Cell count phenotype). We used this method to identify the stem cell marker Lgr5 as a candidate driver of normal hair follicle activity and the vitamin D receptor as a driver of skin tumor development (Quigley et al. Nature 2009).

Tumor context interacts with germline and somatic variation
As tumors develop, germline influence on gene expression is supplanted by the effects of somatic genome alterations and changes in the tumor microenvironment. The DMBA/TPA mouse skin tumor system develops frequent trisomy on chromosome seven. I showed that genes on this chromosome have a disproportionate loss of significant eQTL and corresponding increase in expression of Cyclin D1, a driver of tumorigenesis located on chromosome seven (Quigley et al. Genome Biology 2011). I have also used network eQTL methods to identify genetic variants only relevant in the context of a tumor. We identified a candidate mechanism for the breast cancer susceptibility locus at 5p12 (Quigley et al. Molecular Oncology 2014). We showed the risk genotype was associated with increased expression and decreased promoter methylation of MRPS30, a mitochondrial ribosomal protein. This association was exclusive in ER-positive breast tumors and not found in normal breast tissue. In an analysis of somatic TP53 mutation in the 2000 tumor METABRIC cohort, we showed that lymphocyte invasion in IC10/Basal-like breast tumors is associated with wild-type TP53 (Quigley et al., Molecular Cancer Research 2014).
EmailDavid.Quigley at ucsf.edu
Address UCSF Helen Diller Comprehensive Cancer Center
1450 Third St. Room 207
Box 0128
San Francisco, CA 94143-0875

Updated November 2019