Multiobjectivizing the HP model for protein structure prediction

  • Authors:
  • Mario Garza-Fabre;Eduardo Rodriguez-Tello;Gregorio Toscano-Pulido

  • Affiliations:
  • Information Technology Laboratory, CINVESTAV-Tamaulipas, Cd. Victoria, Tamaulipas, México;Information Technology Laboratory, CINVESTAV-Tamaulipas, Cd. Victoria, Tamaulipas, México;Information Technology Laboratory, CINVESTAV-Tamaulipas, Cd. Victoria, Tamaulipas, México

  • Venue:
  • EvoCOP'12 Proceedings of the 12th European conference on Evolutionary Computation in Combinatorial Optimization
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

The hydrophobic-polar (HP) model for protein structure prediction abstracts the fact that hydrophobic interactions are a dominant force in the protein folding process. This model represents a hard combinatorial optimization problem, which has been widely addressed using evolutionary algorithms and other metaheuristics. In this paper, the multiobjectivization of the HP model is proposed. This originally single-objective problem is restated as a multiobjective one by decomposing the conventional objective function into two independent objectives. By using different evolutionary algorithms and a large set of test cases, the new alternative formulation was compared against the conventional single-objective problem formulation. As a result, the proposed formulation increased the search performance of the implemented algorithms in most of the cases. Both two- and three-dimensional lattices are considered. To the best of authors' knowledge, this is the first study where multiobjective optimization methods are used for solving the HP model.