Holder： Shiyang Ma（Shanghai Jiao Tong University）
Time：2026-04-20 13:30-14:30
Location：Room 818 of the New Public Health Building of Peking University Medical School
Abstract:
In genome-wide survival association analysis, time-to-event traits offer the advantage of capturing both diagnosis status and timing, facilitating the identification of genetic variants associated with age of onset, disease progression, and lifespan. However, existing methods primarily focus on quantitative and binary traits, which do not leverage censored time information and have limitations in detecting rare variants associated with disease progression. To address these challenges, we propose a set of powerful statistical frameworks for genome-wide survival analysis, including SurvSTAAR and Cox-MK. SurvSTAAR provides a scalable pipeline for rare variant analysis of TTE traits, accounting for sample relatedness, population structure, and functional annotations. Cox-MK integrates model-X knockoff inference with saddlepoint approximation to achieve SNP-level false discovery rate (FDR) control. Extensive simulations and applications to UK Biobank data demonstrate that our methods improve statistical power while maintaining well-calibrated FDR control compared to existing approaches. Notably, our frameworks identify novel associations for Alzheimer’s disease, as well as additional candidate genes for asthma and ischemic heart disease.
About the Speaker:
马诗洋，上海交通大学医学院/数学科学学院副研究员，博士生导师，上海交通大学医学院临床医学研究院副院长。2023年入选国家海外青年高层次人才计划、上海市海外高层次人才计划。2014年本科毕业于四川大学数学学院，2019年获美国罗切斯特大学统计学博士学位，随后在哥伦比亚大学生物统计系著名遗传统计学家Iuliana Ionita-Laza教授的指导下从事博士后研究，2022年底回国加入上海交通大学。主要研究方向为遗传统计和生物医学统计。文章发表在PNAS、JASA、Nature Communications、Genome Biology、American Journal of Human Genetics、Statistics in Medicine等高水平学术期刊上，多篇论文被领域内广泛引用，单篇最高引用逾百次。主持上海市启明星项目、上海市教委人工智能促进科研范式改革赋能学科跃升计划一般项目、上海市卫健委卫生行业临床研究专项、上海交通大学 “交大之星”计划医工交叉研究基金，参与科技部重点研发项目。担任中国数学会医学数学专委会副秘书长、中国遗传学会大人群健康与常见病遗传分会委员。教学方面获上海交通大学教学成果奖一等奖，并参编教育部“101计划”统计学核心教材《生物统计》。
Your participation is warmly welcomed!
欢迎扫码关注北大统计科学中心公众号，了解更多讲座信息!
