Abstract
The X-ray or NMR methods are usually used to get the tertiary structure of protein . Both of them are time-consuming and needed a lot of resource . Therefore, the researchers wish to utilize computers to predict the conformation structure more quickly, and then to study the function of protein produced in biological organism . At present, the difficulty of this field is the total conformations are too numerous, and it is so complicated that would be out of current computation power . Meanwhile, the convergent energy is another problem .It often converges to a so- called " local minimun ", not the " global minimun " energy, and could not get the exact solution of that structure . Genetic algorithm uses the evolutionary principle to simulate our natural environment . Like species they struggle for survival among things by the law of natural selection, only the survival of the fittest can live. So does this algorithm . It is composed of three major subroutines, CROSSOVER、MUTATION、SELECTION to search for the optimal elite . The method is to select a good character from mother generation and change their bits at random to each other for new one . Sometimes the new generations may have more chances to survive . Generations after generations, we must get the best individual finally . Our research goal is try to make a topic of amino acid sequence and develop a computer program by using this algorithm . We also combine this with one software named ECEPP ( Empirical Confirmation Energy Program for Peptides ) to calculate the energy and produce the conformation . The well-known structure of Met-enkephalin which is made of five amino acid residues( Tyr-Gly-Gly-Phe-Met ) and is computed .We have obtained the exact solution successfully . In future, we may apply it to more complicated polypeptides .