BENZ WS: the Bologna ENZyme Web Server for four-level EC number annotation

Abstract

The Bologna ENZyme Web Server (BENZ WS) annotates four-level Enzyme Commission numbers (EC numbers) as defined by the International Union of Biochemistry and Molecular Biology (IUBMB). BENZ WS filters a target sequence with a combined system of Hidden Markov Models, modelling protein sequences annotated with the same molecular function, and Pfams, carrying along conserved protein domains. BENZ returns, when successful, for any enzyme target sequence an associated four-level EC number. Our system can annotate both monofunctional and polyfunctional enzymes, and it can be a valuable resource for sequence functional annotation.

PredictEFC: a fast and efficient multi-label classifier for predicting enzyme family classes

“…As all previous models were established on different datasets, it was difficult to fairly compare PredictEFC with previous models. In view of this, we selected the models with web-server, including ECpred [ 22 ] and BENZ WS [ 21 ] for comparing their performance on two independent datasets. We directly input the sequences of enzymes in two independent datasets into above two web-servers and captured the predicted results for counting measurements listed in Eqs.…”

“…A recent method, UDSMProt [ 20 ], was also designed for this purpose. The later methods are devoted to predicting entire EC numbers of enzymes, such as BENZ WS [ 21 ], ECPred [ 22 ], DEEPre [ 23 ], and EFICAz 2.5 [ 24 ]. Extraction of informative features from enzymes is an important step for designing most above methods.…”

“…Selection of proper classification algorithms is another important step to build the efficient machine learning based methods. Several algorithms have been adopted, such as artificial neuron network (ANN) [ 25 ], support vector machine (SVM) [ 1 , 2 , 4 , 7 , 11 , 16 ], Bayesian [ 3 ], nearest neighbor algorithm (NNA) [ 6 , 10 , 17 , 19 ], linear discriminant analysis (LDA) [ 8 , 9 ], hidden Markov model [ 21 ], ensemble learning [ 22 , 24 ], and deep learning algorithms [ 20 , 23 ]. When investigating the family classes of enzymes, all methods, except BENZ WS [ 21 ], considered six family classes of enzymes.…”

“…Several algorithms have been adopted, such as artificial neuron network (ANN) [ 25 ], support vector machine (SVM) [ 1 , 2 , 4 , 7 , 11 , 16 ], Bayesian [ 3 ], nearest neighbor algorithm (NNA) [ 6 , 10 , 17 , 19 ], linear discriminant analysis (LDA) [ 8 , 9 ], hidden Markov model [ 21 ], ensemble learning [ 22 , 24 ], and deep learning algorithms [ 20 , 23 ]. When investigating the family classes of enzymes, all methods, except BENZ WS [ 21 ], considered six family classes of enzymes. However, in 2018, IUBMB added the seventh family class (translocases), inducing the limitations on applications of these methods.…”

Enzyme Substrate Prediction from Three-Dimensional Feature Representations Using Space-Filling Curves

“…As protein function can be described in multiple ways, we focus here on the enzyme-to-substrate mapping problem as a specific instance of the larger protein functional annotation challenge. Predicting enzyme function is of practical value in the context of metabolic engineering, where it can extend the set of enzymes to support novel pathways. − Several structure-based − and/or sequence-based ,− approaches to enzyme function prediction have been reported. The large majority of these methods focus only on predicting the enzyme family, typically in the form of its Enzyme Commission (EC) number, a four-digit number that maps enzymes to a hierarchical classification scheme.…”