报告题目:
Demystifying Associational and Causal Interpretationsfrom High-dimensional Data: Applications to Gut Microbiome-Brain Axis

报告人:
刘金苑 美国范德堡大学生物统计系助理教授，博士生导师
报告时间: 6月26日（周三）上午10:00-11:30

报告地点:一综B316会议室

Abstract:
  Technological breakthroughs such as high-throughput sequencing are generating flourishing high-dimensional data that provoke statistical innovations. Since directly modeling such data is oftentimes C hallenged by multiple testing and weak signals, an emerging alternative is to first apply feature aggregation to the high-dimensional feature at the outset. One example is to compare two subjects' features using dissimilarity metrics, yielding metrics with“between-subject attributes.”In the first half of this talk, I will extend the classical generalized linear models (GLM) to establish a distance-based regression paradigm for between-subject attributes,which aggregate individual signals for scientific-relevant associational insights.Despite exciting evidence revealed about many complex diseases， the scientific frontier is shifting towards discovering the underlying causalmechanisms, where confounding is the primary hurdle. In the second half, wegeneralize the classical average causal effect (ACE) to accommodate high-dimensional potential outcomes by deploying the feature aggregation usingwithin- and between-subject metrics. Further, we derive the efficient influencefunction (EI F) for feature-aggregated metrics, which yields the efficient“doublyrobust (DR)”causality estimators to improve robustness and efficiency. Finaly,we illustrate the proposed approaches to decipher the timely issue of the“gut-brain axis” among the humanmicrobiome，cognitive, and psychosocialmeasurements.

报告人简介:
  刘金苑，美国范德堡大学生物统计系助理教授，博士生导师。2022年获美国加州大学圣地亚哥分校生物统计学博士学位，主要研究方向为高维度数据分析、半参统计推断，因果推断等，应用领域包含复杂高维人体肠道微生物、可穿戴设备等。在统计学期刊Biometrics， Scandinavian Journal of Statistics,Statistics in Medicine，以及应用期刊包含Nature,JAMA Psychiatry等期刊发表SCI研究论文40余篇。