Understanding Epigenomic Drivers of Human Diseases

New genomic technologies generate unprecedented data on the inner workings of

cells. To harness this data to understand and diagnose human diseases, we need to

tackle several challenges (opportunities) in machine learning.

In this talk, I will describe my research combining computational tools with high-
throughput experiments to probe epigenomic regulations of cellular function. I

will discuss a new method to identify changes in the epigenome across human

populations that are associated with disease. We validated the method on

experimental data and applied it to rheumatoid arthritis, breast and colon cancer

data to identify novel epigenetic biomarkers. In the second part of the talk, I will

discuss more broadly new directions in unsupervised learning and Bayesian

optimization with applications in epigenomics and synthetic biology.
James Zou is a Ph.D. candidate at Harvard and the Broad Institute, and is also a Simons Research Fellow at U.C. Berkeley. He is interested in developing new probabilistic models and machine learning algorithms, and in applying them to tackle challenges in computational biology and synthetic biology. He is also interested in multi-agent systems and algorithms inspired by biology. His research has been supported by a NSF Graduate Fellowship and a Gates Cambridge Scholarship.