Predicting Protein Subcellular Localization by Fusing Binary Tree and Error-Correcting Output Coding
In this paper, a new method was applied to predict the protein subcellular localization. The features used in the paper were the Distance frequency (DF), the Physical and chemical composition (PCC) and the Pseudo Amino Acid composition (PseAA). The classifier was integrated by Binary tree and Error-Correcting Output Coding (ECOC) based six Artifical neural networks (ANN). The prediction ability was evaluated by 5-jackknife cross-validation. By comparing its results with other methods, such as Lei-SVM and ESVM, the experimental result demonstrated that our method outperformed their predictions, and indicate the new approach is feasible and effective.
Keywordssubcellular localization feature extraction Binary tree ECOC ANN ensemble classifier
Unable to display preview. Download preview PDF.
- 4.Pánek, J., Eidhammer, I., Aasland, R.: A New Method for Identification of Protein (Sub) Families in a Set of Proteins Based on Hydropathy Distribution in Proteins. Proteins: Struct. Funct. Bioinformatics 558, 923–934 (2005)Google Scholar
- 9.Zhang, S., Huang, B., Xia, X., et al.: Bioinformatics Research in Subcellular Localization of Protein. Prog. Biochem. Biophys. 34(6), 573–579 (2007)Google Scholar
- 11.Dietterich, T.G., Bakiri, G.: Solving Multiclass Learning Problems via Error-Correcting Output Codes. Artificial Intelligence Research (2), 263–286 (1995)Google Scholar