Classification of Membrane Protein Using a Multiple Feature Extraction and Support Vector MachineReportar como inadecuado




Classification of Membrane Protein Using a Multiple Feature Extraction and Support Vector Machine - Descarga este documento en PDF. Documentación en PDF para descargar gratis. Disponible también para leer online.

Membrane proteins are generally classified into the following five types: 1, type I single-pass transmembrane proteins; 2, type II single-pass transmembrane proteins; 3, multi-pass Tran membrane pro- teins; 4, lipid chain-anchored membrane proteins; and 5, GPI-anchored membrane proteins. In this article, based on the usage of the Support Vector Machine algorithm SVM, a novel type of membrane proteins se- quences, which combined the basic amino acid composition and 6 different types of physical and chemical properties of amino acid residues and long distance correlations between them, is developed for defining a protein, and getting a higher prediction accuracy in predicting the membrane protein types. Further mining the information of structures and functions in the membrane proteins showed in this paper actually yield high success rates in both the self-consistency and jackknife tests, as well as in an independent dataset test. A set of correlation coefficients defined by amino acid indexes 25 tiers in this paper: the results show that the lar- ger of the number of the tiers doesn’t mean the better of the accuracy is calculated to represent the various correlations that reflect the sequence order effect along the chain. The comparative results imply that the combination of the whole 25 tiers correlation factors did not contribute more to the classification than the most three dominant elements among them due to the interference. In this sake, the dimensionality of a given membrane protein can be decreased, which effectively reduces the calculated amount.

KEYWORDS

Membrane Protein Classification; Feature Extraction; Correlation Coefficient; Amino Acid Resi- due Index; Dimension Optimization; SVM

Cite this paper







Autor: Cong Zeng, Zhenghua Wang, Xiping He

Fuente: http://www.scirp.org/



DESCARGAR PDF




Documentos relacionados