报告题目：Model-based approach to assess replicability for large-scale high-throughput biomedical studies

报 告 人：焦泽宇（中国科学技术大学）

报告时间：2025年12月26日  16:00

报告地点：格物楼528报告厅

报告摘要：

The development of high-throughput multi-omics platforms and recent breakthroughs in artificial intelligence (AI) have enhanced our understanding of the molecular mechanisms underlying various human behaviors and diseases. In recent years, association analyses remain fundamental in large-scale high-throughput biomedical studies, yet replicability challenges arise in these studies. Given the continuous expansion for large-scale multi-omics association studies, there is still a lack of systematic replicability investigations. Recently, our studies addressed this foundational issue and delivered three insights: (1) We developed a model-based approach to assess replicability for large-scale high-throughput MRI-based association studies; (2) We demonstrated the merits and challenges for plasma proteomics in the replicability of associations; and (3) Based on kernel density estimation (KDE) approach, we proposed a novel framework to assess the theoretical replicability and sample size requirements for large-scale multi-omics association studies.

报告人简介：

焦泽宇，中国科学技术大学数学学院博士后研究员。本科毕业于北京科技大学，2022年获得复旦大学应用数学博士学位。主要研究领域是生物医学数据的挖掘与分析，包括医学影像数据、蛋白组学、基因组学、代谢组学等多组学问题分析以及组学数据关联分析的可重复性问题研究。在Neuroimage、Translational psychiatry、Cerebral Cortex等杂志发表多篇论文。