Academy of Mathematics and Systems Science, CAS Colloquia & Seminars | Speaker: | Qin Zhaohui,Department of Biostatistics and Bioinformatics | Inviter: | 张世华 研究员 | Title: | Harnessing machine learning and omics big data to discover novel risk loci of complex human diseases | Time & Venue: | 2019.7.11 10:00 N218 | Abstract: | Over the past fifteen years, Genome-Wide Association Studies (GWASs) have identified tens of thousands of disease-associated variants. Despite that, more causal variants remain at large given the majority of the genetic component of complex diseases is still unaccounted for. An added layer of complexity is that the most of the variant identified by GWAS is non-coding, which makes functional interpretation extremely challenging. Instead of pursuing even larger and more expensive GWASs or WGS studies in which new findings are expected to have diminishing effect sizes, we propose a complementary, supervised machine learning-based computational strategy to identifying new genetic loci. This new method builds off the tremendous investment in functional genomics and epigenomics resources from large consortia such as ENCODE and REMC. We intend to achieve two goals. First, identify genomic and epigenomic hallmarks shown at known loci linking to complex diseases. Second, using the profiles identified to derive genome-wide risk scores and use the scores to discover novel genomic loci linked to complex diseases. Application to real human disease data has the potential to lead to new hypothesis into the mechanisms and treatment targets and strategies. | | | |