Computational chemistry study of 3D‐structure‐function relationships for enzymes based on Markov models for protein electrostatic, HINT, and van der Waals potentials

Abstract

In a significant work, Dobson and Doig (J Mol Biol 2003, 330, 771) illustrated protein prediction as enzymatic or not from spatial structure without resorting to alignments. They used 52 protein features and a nonlinear support vector machine model to classify more than 1000 proteins collected from the PDB with a 77% overall accuracy. The most useful features were: the secondary-structure content, the amino acid frequencies, the number of disulphide bonds, and the largest cleft size. Working on the same dataset used by D&D, in this article we reported a good and simple model, based on the Markov chain models (MCM), to classify protein 3D structures as enzymatic or not, taking into consideration the spatial structure without resorting to alignments. Here we define, for the first time, a general MCM to calculate the electrostatic potential, molecular vibrations, van der Waals (vdw) interactions, and hydrophobic interactions (HINT) and use them in comparative studies of potential fields and/or protein function prediction. The dataset is composed of 1371 proteins divided into 689 enzymes and 682 nonenzymes, all proteins were collected from the PDB. The best model we found was a linear model carried out with the linear discriminant analysis; it was able to classify 74.18% of the proteins using only two electrostatic potentials. In the work described here, we define 3D-HINT potentials (mu(k)) and use them for the first time to derive a classifier for protein enzymes. We analyzed ROC curves, domain of applicability, parametric assumptions, desirability maps, and also tested other nonlinear artificial neural network models which did not improve the linear model. In closing, this MCM allows a fast calculation and comparison of different potentials deriving into accurate protein 3D structure-function relationships, notably simpler than the previous.

Prediction of ketoacyl synthase family using reduced amino acid alphabets

“…The functions of proteins correlate with their threedimensional (3D) structures. Based on the information of the 3D structure of proteins, González-Díaz and colleagues developed some models and web servers to discriminate between enzymes and nonenzymes [14,15,39], predict enzyme classes [13], and recognize protein kinases [26,27]. They also developed some quantitative structureactivity relationship (QSAR)-based methods [16,24,25] to classify polygalacturonases and nonpolygalacturonases [1], discriminate dyneins from nondyneins [17], and predict RNase scores [23], achieving encouraging results.…”

Using entropy of drug and protein graphs to predict FDA drug-target network: Theoretic-experimental study of MAO inhibitors and hemoglobin peptides from Fasciola hepatica

“…For the calculation, the MARCH-INSIDE software always uses the full matrix, never a submatrix, but the last summation term may run either for all aminoacids or only for some specific protein regions (R) denoted as: c for core, i for inner, m for middle, and s for surface regions, respectively). Consequently, we can calculate different T q k (R) for the aminoacids contained in the regions (c, i, m, s, or t) and placed at a topological distance k each other within this orbit (k is the order) [32,38,39,46,47]. In this work, we have calculated altogether 5 (types of regions) Â 6(orders considered) ¼ 30 T q k (R) indices for each protein.…”

Computational Analysis of Amino Acid Mutation: A Proteome Wide Perspective

“…For instance, Gonzalez-Díaz and co-workers developed method called MARCH-INSIDE that may be used to classify proteins according to their thermal stability [34], predict protein function [35][36][37][38][39], or predict drug-protein target interactions [40]. The method use structural network parameters derived with Markov chains theory as molecular descriptors [3,41,42].…”