Guest Seminar - Causal Inference, analysis of whole genome sequencing data, and Immunosequencing for cancer research

: In the talk, I will cover three different topics in statistical genetics/genomics. First, leveraging increasingly available GWAS summary statistics enables the discovery of causal relationships among complex traits and human diseases but also presents significant challenges. We proposed a novel Mendelian randomization (MR) method called MR SPI to automatically select valid genetic instruments to infer causality between complex traits and diseases. MR SPI demonstrates superior performance in our extensive real data analysis of 146 exposure-outcome pairs. Second, we developed a new method ERStruct based on modern random matrix theory to infer population structure in whole genome sequencing data which is fundamentally important in population genetics and in genetic association studies. Moreover, we proposed a powerful method MultiSTAAR to detect rare variants associated with complex traits by incorporating functional annotations. Third, I will briefly introduce T-cell receptor sequencing and its use for cancer research.