A Promising Similarity Based Secondary Structure Prediction Method

Abstract

Assigning secondary structure to amino acid sequences is challenging task due to complexity of protein folding. Protein structures differ enormously. Here, we analyze the mapping between amino acid sequence and secondary structure in a set of 80,592 non-redundant protein chains from the PDB (Protein Data Bank). To identify local conserved regions, we restricted our attention only to the components of these structures of window sizes from 7, to 45. In this article, we examined the issue of how far secondary structure of proteins can be predicted based on the similar segments of solved structures currently deposited in PDB. It is seen that for almost all proteins, secondary structures can be predicted with a mean accuracy of 92%. This accuracy is the highest in the literature.