BME Seminar Series - Dr. Jason Yang, Massachusetts Institute of Technology

Abstract: 

"A White-Box Machine Learning Approach for Revealing Molecular Mechanisms"

Recent advances in high-throughput experimental technologies and data analyses now enable unprecedented observation, quantification and association of biological signals with cellular phenotypes. However, current techniques for extracting biological information from large datasets are frequently unable to provide casually mechanistic biological insights. Here, we will describe a “white-box” machine learning approach integrating prospective network modeling with biochemical screens to discover experimentally testable hypotheses. We will demonstrate how this approach enabled the novel discovery that purine biosynthesis is involved in bactericidal antibiotic lethality, through its coupling to central carbon metabolism. Additionally, we will discuss new insights gained from applying this approach to study isoniazid lethality in Mycobacterium tuberculosis. We propose that such approaches can be generalized for investigating any quantifiable phenotype using relevant biological networks.

Bio:

Dr. Jason Yang is a Research Scientist in the lab of Dr. James Collins at MIT and the Broad Institute. He received his Bachelor’s from Johns Hopkins University with a double major in Biomedical Engineering and Electrical Engineering, where he worked with Dr. Raimond Winslow on mathematical models of cardiomyocyte electrophysiology. Jason received his Ph.D. in Biomedical Engineering from the University of Virginia as a graduate trainee of Dr. Jeffrey Saucerman, studying mechanisms of β-adrenergic signaling regulation in cardiomyocytes with mechanistic modeling and live cell imaging. As a postdoctoral fellow and now research scientist with Dr. Collins, Jason has been developing systems approaches for understanding mechanisms of antibiotic lethality. He is the recipient of a K99/R00 Pathway to Independence Award from NIGMS and is interested in developing network modeling- and machine-learning based approaches for advancing precision medicine.