Author(s): Ratan Mani Prasad and Rajeev Kumar Pathak

Abstract: Multiple Sequence Alignment (MSA) is a foundational task in bioinformatics, essential for understanding evolutionary relationships, predicting protein structure, and discovering conserved functional regions among sequences. Traditional alignment methods (such as progressive or consistency-based algorithms) often struggle when sequence divergence is high or when there are many gaps, and heuristic optimization techniques, while useful, may suffer from premature convergence or slow exploration of the search space. Motivated by these limitations, this work proposes an improved Particle Swarm Optimization (PSO) algorithm with adaptive parameter control for MSA, aiming to enhance alignment quality, accelerate convergence, and balance exploration versus exploitation more effectively. In the proposed method, the key PSO parameters—namely inertia weight, cognitive coefficient, and social coefficient are dynamically adjusted based on the current swarm performance. Specifically, inertia weight decays from a high value to a low value as diversity among particles decreases; cognitive and social coefficients are modulated in response to stagnation detection and alignment fitness improvement rates. Each particle encodes a candidate alignment using gap insertion, deletion, and residue matching operations, evaluated using a fitness function combining Sum-of-Pairs (SP) score and a gap penalty scheme that is both position-sensitive and structure aware. The algorithm is benchmarked on widely used MSA datasets (e.g. BAliBASE, SABmark), comparing to standard PSO, ClustalW, MUSCLE, and other heuristic evolutionary approaches.

DOI: 10.33545/26633582.2025.v7.i2c.218

Pages: 226-234 | Views: 137 | Downloads: 81

Download Full Article: Click Here