DeepEC which is a learning-powered computational framework that allows high quality and throughput prediction of enzyme commission number which is essential for the better and accurate understanding of enzyme function.
A team of Dr. Jae Yong Ryu and distinguished Professor Sang Yup Lee reported that this computational framework predicts enzyme commission number with very high precision and in a high-throughput manner.
DeepEC needs protein sequence as an input so it can show enzyme commission (EC) number as a result. Enzymes are generally proteins that boost up the biochemical reaction which consists of four levels. Thus the identification of EC is very essential for understanding enzyme functions and metabolism.
EC number is needed encoding enzyme during special procedure called genome annotation procedure. As EC number is very important, several prediction tools have been developed but they need improvement.
DeepEC uses three convolutional neural networks as major engine predicts the EC number and also implements homologu analysis for EC number, if the CNNs do not give clear results. DeepEC was developed by standard dataset covering 4,669 EC numbers and 1,388,606 protein.
DeepEC compared with five other EC number predictor tools and DeepEC made most precise and fastest prediction. It required smaller disk space so it is ideal third party component.
The study was published in Proceedings of the National Academy of Sciences on 20 june.
Source
Deep learning enables high-quality and high-throughput prediction of enzyme commission numbers
Deep Learning-Powered ‘DeepEC’ Helps Accurately Understand Enzyme Functions