Feature Extraction in Transformed Domain for Protein Superfamily Identification from Amino Acid Sequences

摘要

This paper deals with the task of identification of protein superfamily, which is a group of proteins that share similarity in structure and function, by looking at its amino acid sequence. The novelty of the attempt lies in analyzing the amino acid sequences in the transformed domain, where the features are extracted using Fourier and wavelet transforms. Comparative study in the two domains reveals that the features extracted using wavelet transform are more robust in sequence classification than those extracted using Fourier transform. Sequences belonging to two superfamilies, Globin and Ras, are considered for the experiments which are obtained from the Protein Information Resource (PIR) database.