时间地点：2016.6.10 10:00-11:00AM  N602

报告人：Guolian Kang

摘要: In many case-control designs of genome-wide association studies (GWAS) or next generation sequencing studies (NGS), extensive data about correlated secondary binary and continuous traits that may share the common genetic variants with the primary disease are available. Investigation of these secondary traits may provide critical insights about the disease etiology or pathology and enhance the primary GWAS or NGS results. Logistic regression (LG)-based method using retrospective likelihood conditional on disease status were developed in this regard. However, due to inappropriate modeling of the relationship between primary disease status, secondary trait, genetic variations and its instable algorithm, they yield severe inflated type I errors when both traits are correlated and the tested SNP is associated with the primary disease, especially at a stringent significance level for rare variants. To solve this issue, we propose a novel set-valued (SV) model that efficiently models the dichotomizing process of underlying continuous variables to generate binary traits and models their relationship. Extensive simulation studies and 7 secondary traits analyses in a GWAS of benign ethnic neutropenia show that our SV method not only controls type I error rate well at a significance level of 10^-5, but has larger power than LG-based method especially for the secondary continuous trait and rare variants. By contrast, the LG-based method has severe inflated type I error in most of cases. Because of the striking advantage of the SV method, we strongly recommend the SV method be employed instead of the LG-based method for secondary binary and continuous traits analyses in a case-control sequencing study.

个人简介：康国莲博士，一直从事统计遗传和系统生物学研究。已发表SCI学术论文50余篇，其中包括Nature Genetics, Journal of Clinical Oncology等，被独立引用300次以上。其多个研究成果被世界著名媒体报道（美国： NewsRX.com 和ScienceDaily， 委内瑞拉: International Adaptogens，英国Nature Review Genetics））并被收录于书《Transgénicos》 （古巴）。 主持1项由美国NHLBI资助的高水平科研项目的Data Coordinating Center; 6项由美国NCI, NIH资助的科研项目的leader biostatistician. 参与评审NIH项目，是多个国际期刊的编委。