Gamiv: A Genetic Algorithm For Identifying Variable-Length Motifs In Noncoding Dna

GAMI uses a genetic algorithm to identify putatively conserved inotifs of a pre-selected length in noncoding DNA froin diverse species. In this work, I present an extension to the systei, GAMIV, that identifies putatively conserved inotifs of variable length. The systein begins with an initial set of very short inotifs and allows thein to grow through a pair of custoin operators. A fitness function that rewards both inotif conservation and inotif length is used to evolve a population of conserved inotifs of variable length. This paper describes the inotivation for GAMIV, discusses the design of the systei, and presents initial results for the systeln. Based on these initial results, GAMIV is a pronising tool for the inference of variable-length inotifs in noncoding DNA.