קולוקוויום בביה"ס למדעי המחשב - Integrative multiscale modeling of dynamic biological systems
Living cells, tissues, and entire organisms can all be modeled as dynamic systems of moving parts. From the interactions between these parts emerges the collective behavior of these systems, which ultimately yields Life. Experimental or theoretical methods for characterizing biological systems are typically restricted to only subsets of relevant spatial and temporal scales, limiting their utility for characterizing these systems holistically. Integrative multiscale modeling is a general approach for bridging this gap by systematically combining data from different methods at different scales. Using this approach, we modeled the nucleocytoplasmic transport system, which regulates the molecular traffic in and out of the cell’s nucleus. This system comprises many millions of atoms that interact over a wide range of timescales and plays a key role in both health and disease. We modeled the entire transport system based on data from a large number of experimental and theoretical sources, specified at multiple scales and varying degrees of data uncertainty. The resulting spatiotemporal model recapitulates independent measurements of transport rates and elucidates fundamental properties of the transport system, including the mechanism by which it filters by size and how it functions both rapidly and selectively at the same time. As we now strive to expand the scope and depth of our modeling, I will also discuss our ongoing efforts to develop new and more efficient algorithms for integrative modeling of dynamic biological systems in high-dimensional configuration spaces, by efficiently propagating local computations in a hierarchy of spatial and temporal scales of representation.
Barak completed his BSc studies in the Computer Science/Life Science program at the Hebrew University, where he conducted his very first research project under the guidance of Prof. Nati Linial. He went on to receive the Dean’s prize for Master’s students in Computer Science at the Weizmann Institute for his thesis work on applying machine learning and computer vision algorithms to characterize protein structures, under the supervision of Prof. Ronen Basri and Prof. Gideon Schreiber. He conducted his doctoral work under the supervision of Prof. Ora Schueler-Furman at the Hebrew University and Prof. Dan Halperin from the School of Computer Science at Tel-Aviv University, developing and applying optimization and robotic motion planning algorithms to study peptide-protein interactions and protein allostery. He conducted his postdoctoral studies with bioinformatics pioneer Prof. Andrej Sali at the University of California, San Francisco, focusing on integrative multiscale modeling of the nucleocytoplasmic transport system, a complex and dynamic biological system whose modeling requires the integration of data from a large number of different sources and the computations of the system's dynamics in high-dimensional configuration spaces over long timescales.