EC Number Prediction

Number of genome sequence projects of various organisms have resulted in generation of a large amount of gene and protein sequence information. The focus is now on the identification and functional characterization of proteins encoded by these genomes. Enzyme Commission number (EC number) is a numerical classification scheme for enzymes, based on the chemical reactions they catalyze. EC numbers represent enzymes and enzyme genes (genomic information), but they are also utilized as identifiers of enzymatic reactions (chemical information). The scheme is a hierarchical organization of enzyme reactions into six main classes i.e. oxidoreductases, transferases, hydrolases, lyases, isomerases and ligases which are then further split at three hierarchical levels Due to the recent efforts of structural genomics initiatives a large and growing number of enzymes have no functional annotation whilst Experimental functional characterization is time-consuming and expensive. High-precision EC number assignment is of utmost importance for studies such as metabolic pathway reconstruction, understanding evolutionary relationships in pathways and metabolite prediction, etc. Hence there is a vital requisite for improved computational techniques for precise prediction and assignment of EC number. Here, we present ECpred: a predictive model to assign EC number to the enzyme with unidentified function using two supervised machine learning approaches k Nearest Neighbor (k-NN) and Probabilistic Neural Network (PNN). The final prediction is made on the basis of a consensus of the predictions made by selected algorithm and a probability is assigned to it. ECpred classifies an enzyme in one of 3349 EC numbers.