Finding an optimal loci combination of variable number tandem repeats using genetic algorithms

In genotyping, finding the right combination of variable number tandem repeat (VNTR) loci is a research interest in multilocus VNTR analysis (MLVA). The intention is to obtain a set of VNTR loci that has the highest discriminatory power using the least number of loci possible. In this paper, a modified genetic algorithm (GA) is proposed to find an optimum set of loci from a VNTR genotyping of Mycobacterium tuberculosis. The performance of the modified GA is compared with results of the basic GA and the exhaustive method. The effectiveness and efficiency of the modified GA are discussed. The results indicate that the modified GA is a promising method for finding an optimum set of VNTR loci from a MLVA database.