The Agile particle swarm optimizer applied to proteomic pattern matching and discovery

Abstract

Pattern discovery in protein structures is a fundamental task in computational biology, with important applications in protein structure prediction, profiling and alignment. We propose a novel approach for pattern matching and discovery in protein structures using particle swarm-based flying windows over potentially promising regions of the search space. Using a heuristic search, based on particle swarm optimization is, however, easily trapped in local optima due to the sparse nature of the problem search space. Thus, we introduce a novel fitness-based stagnation detection technique that effectively and efficiently restarts the search process to escape potential local optima. The proposed approach predicts an imminent stagnation situation using a novel way that collectively incorporates the already-calculated fitness performances of the swarm particles relative to the objective function, instead of repeatedly calculating their pairwise distances. Our approach is first applied to protein contact maps, which are the 2D compact representation of protein structures. Then, it is generalized to work on classical and advanced (shifted/rotated) benchmark optimization functions. The experimental results show that the proposed fitness-based approach not only demonstrates efficient convergence (up to 3 times faster), but also significantly outperforms the commonly used distance-based method (using Wilcoxon rank-sum test at 95 % confidence level).