CMStatistics 2018: Start Registration
View Submission - CMStatistics
Title: Estimation of selection-bias adjusted FDR for the overall assessment in follow-up studies Authors:  Donghwan Lee - Ewha Womans University (Korea, South) [presenting]
Boram Jeong - Ewha womans university (Korea, South)
Yudi Pawitan - Karolinska Institutet (Sweden)
Woojoo Lee - Inha University (Korea, South)
Abstract: To verify the generalizability and reproducibility of scientific findings, follow-up studies are often conducted where only a few significant discoveries in the previous experiment are re-evaluated. We consider multiple testing problems with the selection-bias and propose a way to estimate the false discovery rate (FDR) for adjusting such bias. We show how to use the selection-bias adjust FDR estimation to assess the overall significance of the discoveries based on the integrated data in follow-up studies as well as the preliminary study. Simulation studies and two metabolomics datasets are considered to illustrate the application of the proposed method in high-throughput data analysis.