Protein Folding Prediction

The search for an accurate and efficient protein conformation predicting method started around 1972 with a theoretical study on the ribonuclease folding. Surprisingly, physicist, chemists, and mathematicians have only been making very minute progress towards the empirical prediction algorithm of three-dimensional protein folding structures. The three most commonly accepted methods used in the past decade consist of homology prediction, folding recognition, and ab initio determination. Our research takes the ab inito approach using computational genetic algorithm (GA) search method. The GA is based on the basic concept of evolutionary natural selection for the “fittest” individual. To this technique we added an adaptation search with gene alteration. We have successfully developed a GA program that minimizes the potential energy of target sequence and generates the corresponding Cartesian coordinates for each atom. By using a three dimensional graphing software, Accelry’s, we were able to visualize the predicted conformation and compare to natural conformation.