Fuzzy pattern trees for predicting the protein binding sites

Abstract

Protein molecules are very important in the living organisms, since they are involved in many processes in the organisms. The knowledge of their functions is crucial for designing new drugs. There are various experimental methods for determining their functions, but they are very complex, so the number of known protein structures with undetermined functions is growing too fast. Therefore, one of the main research directions in bioinformatics community is investigating new computational methods for determining the protein functions. In this research paper, we present a twostep fuzzy pattern tree based method for predicting the binding sites of the proteins. Further, this method could be incorporated in a framework for protein function annotation. The binding sites of the proteins are the amino acid residues where interactions between protein structures occur, while their features determine the functions that the proteins have in these interactions. In the first step of our method, we extract the most important features of the amino acids of the protein molecules. In the second step, using the amino acids’ features we induce fuzzy pattern trees that would be used to classify the amino acids as binding or non-binding sites. We present some experimental results of the evaluation of the fuzzy pattern trees based method.