Identifying the functional characteristic in new annotated proteins is a challenging problem. With the existing sequence similarity search method like BLAST, scope is limited and accuracy is less. Rather than using sequence information alone, we have explored the usage of several composition, hybrid methods, and machine learning to improve the prediction of lipid-transfer proteins. In this paper, we have discussed an approach for genome wide prediction of LTP proteins in rice genome based on amino acid composition using support vector machine (SVM) algorithm. A predictive accuracy of 100 % was obtained for the module implemented with SVM using polynomial kernel. This approach was compared with an All-plant method comprising of six different plants (wheat, maize, barley, arabidopsis, tomato and soybean) which gave an accuracy of only 70 % for SVM.
CITATION STYLE
Hemalatha, N., & Narayanan, N. K. (2016). A support vector machine approach for LTP using amino acid composition. In Lecture Notes in Electrical Engineering (Vol. 396, pp. 13–23). Springer Verlag. https://doi.org/10.1007/978-81-322-3589-7_2
Mendeley helps you to discover research relevant for your work.