Related Experiment Videos
Upward bias in estimation of genetic effects
1Department of Statistics, Stanford University, Stanford, CA 94305, USA. dos@stat.stanford.edu
American Journal of Human Genetics
|October 19, 2002
Abstract:
Because of the large number of tests for linkage that are performed in genome scans, the naive estimator of the size of a genetic effect in cases of borderline significance can be inflated and lead to unrealistic expectations for successful replication. As a remedy, this report proposes lower confidence limits that account for the multiple comparisons of the genome scan.