Machine Learning For Ubiquitous Genomics And Precision Medicine
What can we learn about our societies and ourselves by decoding the genomes of hundreds of thousands of individuals? I will describe two projects that address complementary aspects of this question. In the first part of the talk, I will discuss integrating genomics with computational social sciences to characterize human mating patterns over the last few generations. The second part of the talk will investigate how we can estimate harmful mutations that are present in the general population but have not been observed yet. Patterns that arise from both analyses have significant implications for understanding the dynamics of human diseases and provide a roadmap for precision medicine. I'll highlight the machine learning methods that we've developed to tackle both projects. This talk will be accessible to a general CS audience.
James Zou is a postdoc at MSRNE. He received his Ph.D. from Harvard in 2014 and was recently a Simons fellow at U.C. Berkeley. He is broadly interested in machine learning and applications in understanding human evolution and diseases.