Investigating motif selection methods for support vector machine classification of protein sequences to functional families

摘要

In this work protein sequences are assigned to functional families using support vector machine classification. The classification is performed using binary feature vectors denoting the presence or absence in the protein of highly conserved sequences of amino-acids called motifs. Since the input vectors of the classifier consist of a great number of motifs, feature selection algorithms are applied in order to select the most discriminative ones. The performance of three selection algorithms, embedded within the support vector machine architecture, has been investigated. The embedded algorithms present computational efficiency and can result to the ranking of the selected features. The experimental evaluation demonstrated the usefulness of the aforementioned approach, whereas the individual ranking for the three selection algorithms presented significant agreement.