Protein Fold Recognition and Remote Homology Detection Based on Profile-Level Building Blocks

Abstract

Protein remote homology detection and fold recognition are central problems in bioinformatics. In this paper, two kinds of profile-level building blocks of protein sequences, binary profiles and N-nary profiles, are presented, which contain the evolutionary information of the protein sequence frequency profile. The two building blocks are applied for protein remote homology and fold detection tasks. The latent semantic analysis (LSA) model is adopted to further improve the performance of our methods. Experiment results show that the methods based on profile-level building blocks give better results compared to related methods.