Protein Remote Homology Detection Using Order Profiles

Protein remote homology detection is a central problem in bioinformatics. In this study, we present a novel building block of proteins called order profiles (OP) to use the evolutionary information of the protein sequence frequency profiles and apply this novel building block to remote homology detection. Order profiles contain the evolutionary information extracted from the protein sequence frequency profiles outputted by PSI-BLAST. The performance of our method is further improved by applying an efficient feature extraction algorithm from natural language processing, namely, latent semantic analysis (LSA). When tested on SCOP 1.53 benchmark, our method outperforms the other building-block-based methods and related methods.