Associate Professor, Department of Computer Science
George Mason University
All processes that maintain and replicate a living cell involve moving biomolecules. The energy landscape underscores the inherent nature of biomolecules as dynamic systems interconverting between structures with varying energies. Biomolecular structural transitions regulate diverse processes, such as allosteric signaling and catalysis, and occur on 0.1−10Å length scales and nanosecond-seconds time scales. To understand the dynamic interplay across such disparate spatio-temporal scales, link it to the atomic-scale physicochemical basis of dynamical behavior of single molecules and their interactions, and, ultimately, relate it to cellular function, we need efficient and reliable methods to expose biological and disease-related structures and transitions in biomolecules. In this talk I will summarize some of our algorithmic efforts in this direction. Inspiration for our work comes from a combination of biology and other science and engineering fields that model dynamic systems. Specifically, I will describe stochastic optimization algorithms we have designed that are proving powerful at extracting detailed, sample-based representations of energy landscapes. When further enriched with connectivity information, such representations readily yield structural transitions and summary statistics on dynamics. I will highlight findings that showcase the ability of our in-silico research to provide, for the first time, detailed maps of energy landscapes of diseased protein variants and uncover alterations to dynamics that help formulate hypotheses on how pathogenic mutations percolate to dysfunction.