Biocomputing: Computational Structural Biology
Introduction
We specialize in computational design of biomolecules, which includes rational drug design and protein engineering. Our computational work is strongly driven by the underlying biological phenomena, thus it is interdisciplinary in nature. Specifically, three-dimensional atomic structure of biomolecules, especially proteins, is the conceptual vehicle for all our investigations. The initial task is always to visualize the 3D structural basis of biochemical or genetically observed phenomena.
Our primary assets in achieving this visualization are 1) the growing database of experimental 3D biomolecular structures (the Protein Data Bank); 2) powerful protein structure prediction and bioinformatics computational tools. The protein structure prediction tools require two major components: First, we should calculate all essential free-energy terms of a trial protein conformation with an accuracy sufficient to ensure the uniqueness of the native conformation. Second, we need a procedure to locate the global minimum of this energy function in the giant space of conformational possibilities using a limited number of function evaluations.
We utilize efficient algorithms to align sequences, calculate electrostatic polarization energy, surface energy and side-chain entropy and incorporate them into a global optimization procedure. This procedure allows us to predict, in full atomic detail, the solution structure of small molecule:protein complexes, peptides, side chains and loops directly from the sequence and without additional assumptions. This capability allows us to bring 3D structural information from the database to bear on a wide range of biological problems both for protein design purposes and for small molecule drug design.
The range of problems currently under consideration and the trajectories of the investigations are described in the "Current Projects/Lab Member " section.