On the applicability of genetic algorithms to protein folding

Discusses the protein folding problem and suggests the use of genetic algorithms for protein folding simulations. The issues of protein energy functions, search algorithms, and folding pathways are discussed. The authors review the current approaches to the protein folding problem, point out the limitations of the approaches, and present the genetic algorithm method, which is based on viewing evolution as an optimization process. The schemata theorem is proved in the context of protein structure, showing that during a genetic algorithm search more and more attention will be given to favorable local structures while unfavorable local structures will be rapidly abandoned. It is shown that genetic algorithms are a suitable tool in protein structure predictions. A version of the genetic algorithm is presented that is suitable for protein structure prediction. The behavior of the algorithm is explored in a single model of folding, and it is shown that the algorithm behaves as expected and is able to find the correct conformation.