Predicting building Schemas of protein secondary structure
By applying Genetics Algorithms
FIND,ECRC,Institute for Information Industry
Assessing accurate secondary structures of a protein involves in preparation a crystal of protein, x-ray scanning and computing. These cost a lot. Researchers have developed methods to predict secondary structures of a protein since 1960s. Recently, methods predicting protein secondary structure through the use of new algorithms such as HMM (3), neural networks (2), new evolutionary databases (2) etc.
These algorithms do help to predict protein secondary structure. However, some algorithms are like “Black Boxes”. Researchers don’t the meanings of understand enormous parameters or how the results of prediction come out but only accept them. This study intends to predict protein secondary structure schemas by genetics algorithms.
In this research, a genetic algorithm has been applied to predict building schemas of protein secondary structure. The results of this GAPS (Genetics Algorithm for Protein Secondary Structure) achieved an average Q3 score of 55%~ 65 %. Although the highest Q3 of this research is not the highest score among researches, some fundamental and useful building schemas of protein secondary structure information have been found.
Previous researches (e.g. focused on global free energy minimum of protein secondary structure) could not give us a complete understanding of the driving forces behind protein folding. Why?
Previous researches take every amino acid in the sequence into consideration. However, from the results of this study, not all the residues in a schemas effect the conformation of protein secondary structure. Only few amino acids actually effect the conformation of protein secondary structure folding.