Speaker: Liming Liang, PhD
          Associate Professor of Statistical Genetics
          Deparment of Epidemiology
          Department of Biostatistics，Harvard T.H. Chan School of Public Health 
          Web: https://www.hsph.harvard.edu/liming-liang

Time : 10:00-12:00 am , Dec. 11（Monday）
Venue: Room 300, SIBS Main Building, Yueyang Road 320
Host： Prof. Guoqing Zhang
        CAS-MPG Partner Institute for Computational Biology

Title：Power and method to detect cell-type-specific associations in EWAS using cell mixture samples

Abstract：
        Epigenome-wide association study (EWAS) of DNA methylation is a useful tool to identify specific CpG loci whose alterations in methylation level were implicated in disease etiology or trait development. Due to the limitation of available isolated cell types and the purity of tissue dissection, most large-scale EWAS measured methylation in a mixture of heterogeneous cell types, such as whole blood. It is often of concern that whether the cell composition would confound the methylation-disease/trait association, and more importantly whether true associations with methylation within specific causally related cell types could be detected with sufficient power, especially when such causally related cell types were not common in the cell mixture. Through two large scale EWAS for total serum IgE level and coronary heart disease, we will discuss key factors, novel statistical method and useful strategy to detect cell specific signals while appropriately remove confondouning effect due to cell heterogenity.

        All are welcome！