An Improved Adaptive Genetic Algorithm For Multiple Sequence Alignment Applications

Bioinformatics is a new crossing discipline which utilizes biology, applied mathematics, computer science, it is one of the great frontiers of life sciences and natural sciences.In bioinformatics, biological sequence alignment is the most foundation problem, and multiple sequence alignment is the most fundamental task of biologica l sequences alignment. How to obtain better alignment of sequences efficiently is the important theme in it.The problem of multiple sequence alignment is a NP-Complete problem,and dynamic programming algorithm is the basic algorithm , it can find a highest score (or least costly) sequences topology, however, increasing number of sequences, the index calculation will show growth sharply. In actual operation, the algorithm will get the best rate and obtain better alignment of sequences than the strike the right balance. Genetic algorithms are an effective way to use this approximation algorithm,but genetic algorithms has such disadvantages as premature convergence, low robustness.Therefore, this paper presents an improved adaptive genetic algorithm, using the fitness of the calibration and diversity groups in a process of evolution,at the same time,using the most commonly sigmoid function of construction neuronal activation function, optimized crossover and mutation rate and improved the capacity and stability of algorithm.Bit climbing is a local search algorithm, it have strong capability of local refinement . Based on binary code for the composition and function optimization, use of this method is appropriate. Because genetic algorithm's macro search capability is greater then its local refinement,so this paper use bit climbing strategies to improve the precision of fitness in the patter of the adaptive genetic algorithm.This paper design and development a multiple sequence analysis procedures based on improved adaptive genetic algorithm, then applied to more specific nucleic acid and protein multiple sequences, to the experimental data, compared to the standard genetic algorithm and the traditional adaptive genetic algorithm to achieve the sequence of procedures,it shows new algorithm has faster quality and stability of multiple sequence alignment.