报告题目：Accurate identification of disease-specific non-coding risk variants based on multi-omics profiles
时 间：2016年7月15日 上午10:00
地 点：电子所楼北103生医学院会议室
联 系 人：谢建明
秦昭晖教授为我校“客座教授”。秦昭晖教授是生物信息学及生物统计学领域的国际知名专家，以首席科学家的身份主持了多项科研项目，包括美国国家卫生院（NIH）的R01、R21及美国国家科研基金（NSF）项目，并参与了大型国际HapMap项目。秦博士在包括Nature，Science等专业学术刊物上已发表论文90余篇，总引用次数超过一万次（Google Scholar）。开发或主持开发了十余种用于各类高通量实验数据分析的算法和软件。
报告摘要：The majority of variants identified by Genome-wide association studies (GWASs) fall outside of the protein-coding regions. Understanding the cryptic link between non-coding sequence variants and pathophysiology of complex diseases is a fundamental challenge. To overcome the lack of annotation in the intergenic regions, various recent computational tools have been developed to identify non-coding risk variants using genome-wide genomics and epigenomics profiling data. A common feature of these methods is that they do not distinguish risk variants associated with different diseases. Since different biological mechanisms are believed to contribute to the etiology of different diseases, it is desirable to characterize the impact of a non-coding variant in a disease-specific manner. In this work, we describe DIVAN, a data-driven, machine learning approach that aims to identify disease-specific risk variants. Using 1,806 epigenomic profiles across cell types and factors, along with other static genomic features, we adopt a novel feature selection based ensemble-learning framework to achieve this goal. Our results suggest that DIVAN has the potential to annotate novel variants in a disease-specific manner which will be important to help us understand disease etiology.
欢迎您和学生参加，谢谢！