An improved genetic algorithm for statistical potential function design and protein structure prediction

Protein structure prediction is an important but far from being well-resolved problem in computational biology. It is generally regarded that the native structures of proteins correspond to minimumenergy states. Potential functions are useful in protein structure prediction. To obtain the optimal parameters of protein potential functions, we introduced several strategies to improve the basic Genetic Algorithm (GA). The improved GA was employed in statistical potential function design and protein structure prediction, and experimental results validate the effectiveness and efficiency of the proposed algorithm.