光華講壇——社會名流與企業家論壇第 5199 期
主題：Computationally scalable method for association testing in whole-genome sequencing
主辦單位：数据科学与商业智能联合实验室 统计学院 科研处
Yaowu Liu currently is a postdoctoral fellow at the Harvard School of Public Health. He obtained his Ph.D. in statistics at Purdue University. His research interests include scalable methods for large-scale statistical inference, high-dimensional data analysis, and statistical genetics.
With the advent of massive DNA sequencing data, there is a strong need for methods that are both statistically efficient and computationally scalable. For example, ongoing large-scale whole genome sequencing (WGS) studies are currently producing hundreds of terabytes of WGS data from tens of thousands of individuals with a wide spectrum of phenotypes. In this talk, I present a novel p-value combination method based on the Cauchy distribution. This method is very general and enjoys many good properties, such as simple, computationally efficient, resistant to correlation, and powerful against sparse alternatives. The effectiveness of the proposed methods is demonstrated by a WGS analysis of blood-related phenotypes.