Bivariate estimation of distribution algorithms for protein structure prediction.

Abstract

A real-valued bivariate ‘Estimation of Distribution Algorithm’ specific for the ab initio and full-atom Protein Structure Prediction problem is proposed. It is known that this is a multidimensional and multimodal problem. In order to deal with the multimodality and the correlation of dihedral angles φ and ψ, we developed approaches based on Kernel Density Estimation and Finite Gaussian Mixtures. Simulation results have shown that both techniques are promising when applied to that problem.