The Optimal Side-chain Positioning of Proteins via the Beta-complex

초록

Protein structure determines its function and there have been tremendous amount of studies on the relationship between structure and function in molecular biology. One of such efforts, becoming more important in these days, is computational approaches to predict protein structures (possibly with experimental validations). From structural point of view, a protein consists of two parts: backbone and side-chains. Assuming that a backbone structure is known, the side-chain positioning problem (SCP-problem) is to predict the optimal conformation of the structure of side-chains in that the potential energy of an entire protein is minimized. The SCP-problem has many important applications. For example, accepting that a protein design problem is to find the optimal sequence which is best compatible with a give backbone, the protein design problem contains the SCP-problem as one of its subproblems. Another examples is homology modeling which generates a backbone structure from a given sequence by collecting a number of seemingly good substructures from databases where the measure of “good”-ness is appropriately defined. However, it is well-known that the SCP-problem is NP-hard thus heuristic algorithms of various kinds are inevitably invited in its solution process. This paper proposes an efficient algorithm for solving the SCP-problem by using the beta-complex. It turns out that the algorithm obtains the quality of the solution very close to the optimal solution in a reasonable time.