Apoptosis Protein Subcellular Location Prediction Based on Position-Specific Scoring Matrix

作者：Yao Yu Hua^*; Shi Zhuo Xing; Dai Qi

来源：Journal of Computational and Theoretical Nanoscience, 2014, 11(10): 2073-2078.

DOI：10.1166/jctn.2014.3607

• Apoptosis Proteins
• Subcellular Location
• PSI-BLAST
• Support Vector Machine
• Position-Specific Scoring Matrix

Based on Position-Specific Scoring Matrix (PSSM), average mutation probability from one particular amino acid to 20 types of residues and average mutation rate of 20 types of amino acids within query sequences during evolution are extracted, and the new method which combines these evolutionary information is proposed for apoptosis protein subcellular location prediction. Principal component analysis is employed to extract useful features. The proposed method is tested by the support vector machine classifier, and the prediction accuracy in dataset ZD98 and CL317 reaches 92.9% and 90.5%, respectively. The experiment results obtained by jackknife test can almost reach the highest level through comparison with other methods. In addition, it's worth to pointing out that the proposed method is better at small set predicting than others methods. All of the results confirm that the proposed novel sequence information obtained from Position-Specific Scoring Matrix is remarkable. It heralds that the proposed method might serve as an efficient prediction model for apoptosis protein subcellular location prediction.

• 出版日期2014-10
• 单位浙江理工大学