Abstract
This paper presents a Fragmented protein sequence alignment using two-layer PSO (FTLPSO) method to overcome the drawbacks of particle swarm optimization (PSO) and improve its performance in solving multiple sequence alignment (MSA) problem. The standard PSO suffers from the trapping in local optima, and its disability to do better alignment for longer sequences. To overcome these problems, a fragmentation technique is first introduced to divide the longer datasets to a number of fragments. Then a two-layer PSO algorithm is applied to align each fragment, which has ability to deal with unconstrained optimization problems and increase diversity of particles. The proposed method is tested on some Balibase benchmarks of different lengths. The numerical results are compared with CLUSTAL Omega, CLUSTAL W2, TCOFFEE, KALIGN, and DIALIGN-PFAM. It has been shown that better alignment scores have been achieved using the proposed technique FTLPSO. Further, studies on PSO update equation’s parameters and the parameters of the used scoring functions are presented and discussed.