A Multiobjective Variable Neighborhood Search for Solving the Motif Discovery Problem
In this work we approach the Motif Discovery Problem (MDP) by using a trajectory-based heuristic. Identifying common patterns, motifs, in deoxyribonucleic acid (DNA) sequences is a major problem in bioinformatics, and it has not yet been resolved in an efficient manner. The MDP aims to discover patterns that maximize three objectives: support, motif length, and similarity. Therefore, the use of multiobjective evolutionary techniques can be a good tool to get quality solutions. We have developed a multiobjective version of the Variable Neighborhood Search (MO-VNS) in order to handle this problem. After accurately tuning this algorithm, we also have implemented its variant Multiobjective Skewed Variable Neighborhood Search (MO-SVNS) to analyze which version achieves more complete solutions. Moreover, in this work, we incorporate the hypervolume indicator, allowing future comparisons of other authors. As we will see, our algorithm achieves very good solutions, surpassing other proposals.
KeywordsPareto Front Transcription Factor Binding Site Variable Neighborhood Variable Neighborhood Search Position Weight Matrix
Unable to display preview. Download preview PDF.
- 1.Stine, M., Dasgupta, D., Mukatira, S.: Motif discovery in upstream sequences of coordinately expressed genes. In: The 2003 Congress on Evolutionary Computation (CEC 2003), December 2003, vol. 3, pp. 1596–1603 (2003)Google Scholar
- 4.Fonseca, C.M., Paquete, L., Lopez–Ibanez, M.: An improved dimension–sweep algorithm for the hypervolume indicator. In: IEEE Congress on Evolutionary Computation (CEC 2006), July 2006, pp. 1157–1163 (2006)Google Scholar