Washington University School of Medicine

Division of Biostatistics
Seminar Series Fall 2007

Design and Analysis for Studying Major Cancer Genes

Sining Chen
Department of Environmental Health Sciences
Johns Hopkins School of Public Health

Friday, November 16, 2007, 12:30–1:30 pm

GEMS classroom, 3rd Floor in Shriner's Building
Coffee, tea, and cookies will be provided


Abstract

Using BRCA1 and BRCA2 in familial breast cancer as an example, we discuss a number of designs and analyses for characterizing rare genetic variants in hereditary syndromes. Specifically, we present two approaches to analyze heavily ascertained family data. The retrospective likelihood approach uses the likelihood of genetic test results conditioning on the family history. It takes full advantage of the Mendelian transmission of mutated alleles and provides unbiased estimates. The ascertainment-conditional likelihood approach conditions on the event of ascertainment and models the probability of ascertainment through a family history summary variable. It has higher efficiency than the retrospective likelihood and remains unbiased under appropriate modeling of the ascertainment event. Finally we present a meta-analysis of BRCA penetrance studies.