HECNet: a hierarchical approach to enzyme function classification using a Siamese Triplet Network

Abstract

Motivation Understanding an enzyme’s function is one of the most crucial problem domains in computational biology. Enzymes are a key component in all organisms and many industrial processes as they help in fighting diseases and speed up essential chemical reactions. They have wide applications and therefore, the discovery of new enzymatic proteins can accelerate biological research and commercial productivity. Biological experiments, to determine an enzyme’s function, are time-consuming and resource expensive. Results In this study, we propose a novel computational approach to predict an enzyme’s function up to the fourth level of the Enzyme Commission (EC) Number. Many studies have attempted to predict an enzyme’s function. Yet, no approach has properly tackled the fourth and final level of the EC number. The fourth level holds great significance as it gives us the most specific information of how an enzyme performs its function. Our method uses innovative deep learning approaches along with an efficient hierarchical classification scheme to predict an enzyme’s precise function. On a dataset of 11,353 enzymes and 402 classes, we achieved a hierarchical accuracy and Macro-F1 score of 91.2% and 81.9%, respectively, on the 4th level. Moreover, our method can be used to predict the function of enzyme isoforms with considerable success. This methodology is broadly applicable for genome-wide prediction that can subsequently lead to automated annotation of enzyme databases and the identification of better/cheaper enzymes for commercial activities. Availability The web-server can be freely accessed at http://hecnet.cbrlab.org/. Supplementary information Supplementary data are available at Bioinformatics online.

A hierarchical deep learning based approach for multi‐functional enzyme classification

“…The model used in mlHECNet is similar to the one used in our study on mono‐functional enzymes 2 . There are two types of features: length dependent and length independent features.…”

“…The data gathered from PSSMs underwent normalization as performed by HECNet. 2 The PSSMs were vertically stacked together and the mean and variance, in depth, was calculated for each position. Hence, we acquired position specific means and standard deviations for the PSSM.…”

“…Therefore, to address the complex nature of this problem, several studies have attempted to predict the function of an enzyme using computational approaches. HECNet, 2 ECPred, 3 DEEPre, 4 and EzyPred 5 are some of the major enzyme function prediction tools. Most of these methods use machine/deep learning to solve this problem.…”

“…Therefore, we aim to achieve multi‐functional enzyme classification till the bottom of the hierarchy using nonredundant sequences. For this, we propose a deep learning based M ulti‐ L abel H ierarchical approach to E nzyme function C lassification (mlHECNet) which is inspired by our previous study, HECNet (a hierarchical approach to Enzyme Function Classification using a Siamese Triplet Network) 2 . Moreover, we have proposed an appropriate loss function, inspired by Zhang et al, 15 and an assisted learning scheme which resulted in a deep neural network that gave us highly accurate predictions up to the fourth level of the EC number for multi‐functional enzymes.…”

TripletProt: Deep Representation Learning of Proteins Based On Siamese Networks

“…Language modeling approaches on proteins [8], [17], [18] utilize protein sequences as an analogy to sentences in NLP. HECNet [19] employs the siamese network for enzyme function classification. In this classification framework, a domain specific feature extraction is used, while in the present work we produce generic embeddings for proteins.…”