Biological Network Alignment Using Hybrid Genetic Algorithm and Simulated Annealing

Abstract

This research demonstrates how we can improve the efficiency of protein-protein interaction (PPI) network alignment using soft computing. In Bioinformatics, biological network alignment is particularly important for its use in identifying cellular pathways, discovering new drugs, and detecting disease progression. Also, network alignment is used in social networks, ontology matching, pattern recognition, and natural language processing. In this regard, the main challenge is that the problem of finding the alignments in two graphs is NP-hard, therefore, accurate algorithms can only be used for very small instances. For real and relatively large cases, typically (meta)heuristic methods, which can find approximate solutions in reasonable time, are used. In this regard, we propose a new hybrid metaheuristic algorithm, called SAGA. The SAGA proposed method is applied the simulated annealing in the crossover operation of genetic algorithm. Concerning the integrated network alignment, SAGA first finds the local alignments and then it discovers the existing global network alignments. We implement the SAGA network aligner on python 3.6 and obtained experimental results on five eukaryotic species of the Biogrid dataset. The experimental results show that SAGA network aligner can achieve a better mapping than some of the state-of-the-art algorithms. Based on the experimental results, the proposed integrated network aligner can balance the functional quality and topological quality criteria that are significant in Bioinformatics.