Binary Classification of Uncharacterized Proteins into DNA Binding - Non-DNA Binding Proteins from Sequence Derived Features Using ANN

Abstract

The problem for predicting DNA binding and non-DNA binding proteins from protein sequence information is still an open problem in bioinformatics. It is further becoming more important as the number of sequenced information grows exponentially over time. Sequence similarity matrices are a useful approach to provide functional annotation, but its use is sometime limited, prompting the development and use of machine learning methods. We implemented a novel approach for predicting the DNA binding and non-DNA binding proteins from its amino acid sequence using artificial neural network (ANN). The ANN used in this study is a feed-forward neural network with a standard back propagation training algorithm. Using 62 sequence features alone, we have been able to achieve 72.99% correct prediction of proteins into DNA binding/non-DNA binding (in the set of 1000 proteins). For the complete set of 62 parameters using 5 fold cross-validated classification, ANN model revealed a superior model (accuracy = 72.99%, Qpred = 73.952%, sensitivity = 81.53% and specificity = 72.54%).