|
Abstract:
|
Motif discovery is a general and important problem in bioinformatics, as motifs often are used to infer biologically important sites in bio-molecular sequences. Many problems in bioinformatics are naturally cast in terms of sequences, and distance measures for sequences derived from edit distance is fundamental in bioinformatics. Geometric Crossover is a representation-independent definition of crossover based on a distance on the solution space. Using a distance measure that is tailored to the problem at hand allows the design of crossovers that embed problem knowledge in the search. In this paper we apply this theoretically motivated operator to motif discovery in protein sequences and report encouraging experimental results.
|
