A study on recognition of classically and non-classically secreted proteins from cancer cells based on support vector machine

Abstract:

Based on support vector machine (SVM) algorithm, a fast and accurate method is proposed to distinguish the classically and non-classically secreted proteins from cancer cells. By a strict feature selection, the optimal feature set is obtained which consists of amino acid composition (AAC), position specificity score matrix (PSSM) and signal peptide (SP). The test results show that our method has strong ability to distinguish the non-classically secreted proteins (NCSPs) from the classically secreted proteins (CSPs) of cancer cells, which may provide theoretical reference for finding common biomarkers among different kinds of cancers.

Key words:Support vector machine; Cancer; Non-classically secreted protein; Position specific scoring matrix; Signal peptide