A similarity matrix-based hybrid algorithm for the contact map overlaps problem

This paper proposes a similarity matrix-based hybrid algorithm for the contact map overlap (CMO) problem in protein structure alignment. In this algorithm, Genetic Algorithm (GA) is used as a framework, in which the initial solutions are constructed with similarity matrix heuristic, and Extremal Optimization (EO) is embedded as a mutated operator. In this process, EO quickly approaches near-optimal solutions and GA generates improved global approximations. Five similarity measurements including ratio, inner product, cosine function, Jaccard index and Dice coefficient have been exploited to compute the similarity matrix between two contact maps. The simulations demonstrate that our algorithm is significantly faster and gets better results for most of the test sets.