Journal of Biotechnology & Biomaterials
Open Access

Abstract

Classification of enzymes based on their stability and specificity is one of the most important tasks for their application. Various classifiers such as HMM (Hidden Markov Model) and Neural Network Classifiers previously used are found to be less sensitive and accurate as compared to Support Vector Machines (SVM?s). Continuous increase in genomic/proteomic data new nitrilases are being discovered whose annotation and functional assignment of classes to these through various wet lab techniques involve lots of time consuming and laborious experiments. Machine learning techniques such as SVM?s can be effectively used to complement them saving time, money and also applicable to various other proteins. In view of this the present investigations aims to describe a novel approach for predicting the two sub classes of the nitrilase on basis of their amino acid composition, a powerful method developed from statistical learning with wide applicability in the field of proteomics. The application of SVM tool have clearly differentiated two of nitrilases with accuracy of 88.46%, specificity 80.77%, accuracy 84.62% and MCC (Mathews Correlation Coefficient) of about 0.69 for the amino acid composition, whereas dipeptide and split amino acid composition showed accuracy of 88.46 % & 84.62% and MCC of 0.73 & 0.69 respectively for aliphatic and aromatic nitrilases.

Biography

Nikhil Sharma is presently working as a research scholar in the Department of Biotechnology and is working in the area of microbial enzyme technology both in vitro and in silico. He has contributed research article & chapter to renowned journals with other many in pipeline.