Alignment Of Protein Structures With A Memetic Evolutionary Algorithm

CATEGORY: Real-World Applications Structural comparison of proteins is a core problem in modern biomedical research. Identifying structural similarities is essential for the assessment of the relationship between structure and function in proteins, and structural comparison techniques play a key role in applications like rational drug design. In this paper we consider a technique for protein structure comparison known as the maximum contact map overlap problem. In this problem, the similarity between two protein structures is computed by aligning the proteins to maximize the number of shared contacts in their corresponding contact maps. We present a new approach to this problem that uses a Multimeme evolutionary algorithm. The best solution found by our algorithm provides a lower bound on the value of the optimal structural alignment between the proteins. We have evaluated the Multimeme algorithm on a range of benchmark problems and compared with previous heuristics. We apply a linear programming method, which provides an upper bound, to assess the accuracy of our solutions. Our experiments show that the Multimeme evolutionary algorithm represents a significant improvement on the current state of the art in metaheuristics for this problem.