Protein structural class determination using support vector machines

Zerrin Isik, Benin Yanikoglu, Ugur Sezerman

Research output: Contribution to JournalArticleAcademicpeer-review

Abstract

Proteins can be classified into four structural classes (all-α, all-β, α/β, α+β) according to their secondary structure composition. In this paper, we predict the structural class of a protein from its Amino Acid Composition (AAC) using Support Vector Machines (SVM). A protein can be represented by a 20 dimensional vector according to its AAC. In addition to the AAC, we have used another feature set, called the Trio Amino Acid Composition (Trio AAC) which takes into account the amino acid neighborhood information. We have tried both of these features, the AAC and the Trio AAC, in each case using a SVM as the classification tool, in predicting the structural class of a protein. According to the Jackknife test results, Trio AAC feature set shows better classification performance than the AAC feature. © Springer-Verlag 2004.
Original languageEnglish
Pages (from-to)82-89
JournalLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume3280
DOIs
Publication statusPublished - 2004
Externally publishedYes

Fingerprint

Dive into the research topics of 'Protein structural class determination using support vector machines'. Together they form a unique fingerprint.

Cite this