Department of Statistics Seminar
North Carolina State University

presents

Dr. Kathryn Roeder

Carnegie Mellon University

"False discovery or missed discovery? Using linkage to improve power of association tests."

ABSTRACT

In genetic epidemiology tens of thousands of genomic regions may be tested in a genetic association study to locate alleles that increase the risk for complex diseases. Testing such a large number of hypotheses exacerbates the trade-off between power and Type I error control, making it more difficult to detect small but important signals in the data. Multiple testing problems of this nature are well suited for application of the false discovery rate (FDR) principle which can improve power somewhat. To further enhance the power we consider a new approach which involves weighting the hypotheses based on prior knowledge. Typically some of the hypotheses under investigation can be considered more likely to be non-null than others. For example, linkage studies can provide useful guidance for choosing weights in an association study. We present a method for multiple hypothesis testing that maintains control of FDR while incorporating prior information about the hypotheses. The prior information takes the form of p-value weights. If the assignment of weights is positively associated with the truth, the procedure improves power considerably. Remarkably, the loss in power is small even when the weights are poorly chosen. This desirable property arises because of an asymmetry in the distribution of p-values under the null and alternative hypotheses.

Friday, April, 15, 2005

3:35 - 4:35 pm

206 Cox Hall

Refreshments will be served on the second floor of Dabney Hall (left of Room 222) at 3:00 pm.