报告人：董献军 （哈佛大学）

报告时间：2021年11月22日，9:00-10:00

报告地点：腾讯会议 ID：779 893 536

摘要： While lots of efforts have improved our understanding of the function of the human genome in the last decades, the human genome is still full of puzzles. For example, half of the human genome is constituted of retrotransposon elements, whose functions are still largely unknown. Most of the disease-associated risk loci are in non-coding regions. Their mutation won’t change the protein, but how they function in diseases is largely unclear. Recently development in next-generation sequencing techniques, especially the single-cell omics and spatial transcriptomics, allows us to measure the biological signals in the single-cell and temporospatial level. However, integrating these high-throughput, high-dimensional data with minimal bias from the covariances can be difficult. Moreover, other data types including clinical, imaging, daily activities, are being accessible for both patients and healthy individuals. How to intergrade the heterogenous data together and make insight from it is a challenge for us. In this talk, Dr. Dong will share few mathematics challenges that he and his team are facing in normalizing, integrating, and interpreting their biomedical data.

讲者简介：董献军 现任哈佛大学助理教授，Brigham and Women’s Hospital (BWH)神经内科系研究员 ，精准神经医学生物计算部主任，BWH生物信息学及基因组学中心主任，主要研究方向包括基因转录调控以其在神经疾病中的作用，基因组非编码区的功能研究，以及大数据和机器学习在复杂疾病中的应用。2008年博士毕业于挪威卑尔根大学生物信息及基因组学专业，后赴美攻读博士后，在麻省大学医学院作为核心数据科学家参与人类基因组注释项目，该项目在2012年被Science杂志评为2012年度十大科学突破之一。自2010年至今，在Nature, Science, Cell, Nature Neuroscience, Nature Methods, Genome Biology, Bioinformatics 等期刊共发表学术论文38篇, 其中一作或通讯作者12篇，H-index为22，累计引用超过17,000。

邀请人：张雷洪