The focus of my research is genetic epidemiology and other problems at the interface between statistics and genetics. This interface is fertile ground for statistics: research problems frequently involve complex high dimensional data types structured by both characterized and as yet uncharacterized biological processes. Elucidating these biological relationships holds promise for identifying individuals at high risk of cancer, modifiable risk factors or clinically distinct cancer subtypes.

I am interested in Bayesian statistical methods for cancer risk assessment focusing on assessment of evidence for gene by environment and gene by gene interactions, models for SNP data and statistical inference regarding cancer risk using high-risk family data. Examples include (1) the problem of identifying gene-gene and gene-environment interactions, (2) Bayesian approaches to genome-wide association studies, and (3) BRCA1/2 gene characterization and risk modeling. I am also interested in methods for classifying unknown variants of disease genes and for analyzing high throughput genetic and genomic data such as those measuring gene expression and SNP genotypes using modern array technologies.

My research necessarily collaborative in nature. My primary collaborators include epidemiologists, oncologists, molecular cancer biologists, geneticists computer scientists and other statisticians here at Duke and elsewhere. Duke's Program in Computational Biology and Bioinformatics (CBB) plays an role in fostering a number of these interactions.