An Inspection of the Multiple Alignment Method with Use of a Genetic Algorithm

We proposed a method for amino acid sequence alignment problems by using a genetic algorithm[1][2]. The resulted alignments obtained by applying our method to the data sets of rather small number of comparatively short sequences with higher similarity were satisfactory ones, even better than those of CLUSTALW[3] with respect to e.g., the alignment score, the numbers of inserted gaps and matches. For the data sets including large number of sequences with less similarity, however, rather poor alignment results were obtained. In order to nd out major impediments in the method and to improve it further, it seems to be necessary to investigate the performance of the method by applying it to various data sets of di erent characteristics such as the number of sequences, sequence length and similarity, etc..