Title:Protein structure prediction as a hard optimization problem: the genetic algorithm approach

Abstract: Protein structure prediction can be shown to be an NP-hard problem; the number of conformations grows exponentially with the number of residues. The native conformations of proteins occupy a very small subset of these, hence an exploratory, robust search algorithm, such as a genetic algorithm (GA), is required. The dynamics of GAs tend to be complicated and problem-specific. However, their empirical success warrants their further study. In this paper, guidelines for the design of genetic algorithms for protein structure prediction are determined. To accomplish this, the performance of the simplest genetic algorithm is investigated for simple lattice-based protein structure prediction models (which is extendible to real-space), using energy minimization. The study has led us to two important conclusions for `protein-structure-prediction-genetic-algorithms'. Firstly, they require high resolution building blocks attainable by multi-point crossovers and secondly they require a local dynamics operator to `fine tune' good conformations. Furthermore, we introduce a statistical mechanical approach to analyse the genetic algorithm dynamics and suggest a convergence criterion using a quantity analogous to the free energy of population.