QSAR Study of the Inhibitors of the Acetyl-CoA Carboxylase 1 and 2 using Bayesian Regularized Genetic Neural Networks: A Comparative Study

Abolfazl Valadkhani,Ahmad Mani-Varnosfaderani,Mohammad Asadollahi-Baboli

doi:10.5935/0103-5053.20150017

Abstract

Linear and non-linear quantitative structure-activity relationship (QSAR) models were presented for modeling and predicting anti-diabetic activities of a set of inhibitors of acetyl-CoA carboxylase 1 and 2 (ACC1 and ACC2). Different algorithms were utilized to choose the best variables among large numbers of descriptors and then these selected descriptors were used for non-linear (artificial neural network) and linear (multiple linear regression) modeling. The variable selection methods were consisted of stepwise-multiple linear regression (stepwise-MLR), successive projections algorithm (SPA), genetic algorithm-multiple linear regression (GA-MLR) and Bayesian regularized genetic neural networks (BRGNNs). The prediction abilities of the models were evaluated by Monte Carlo cross validation (MCCV) in variable selection and modeling steps. The results revealed that the best variables for describing the inhibition mechanism of ACC were among topological charge indices, radial distribution function, geometrical, and autocorrelation descriptors. The statistical parameters of R2 and root mean square error (RMSE) indicated that BRGNNs is superior for modeling the inhibitory activity of ACC modulators over the other methods. The sensitivity analysis together with the frequency of the selected molecular descriptors in this work can establish an understanding to the mechanism of ACC inhibitory activity of small molecules.

Highlights

Diabetes mellitus is a typical metabolic disorder characterized by abnormally high levels of plasma glucose or hyperglycemia.[1]
The results demonstrated that the variables which were chosen in a way in which non-linear interactions were considered between the descriptors and the activity of compounds, were totally different compared to linear based selection methods and showed even more acceptable correlation to the activities
The results revealed that for modeling both pIC50 and ligand efficiency (LE) values, Bayesian regularized genetic neural networks (BRGNNs) is a robust algorithm for the variable selection and modeling procedures, simultaneously

Summary

Introduction

Diabetes mellitus is a typical metabolic disorder characterized by abnormally high levels of plasma glucose or hyperglycemia.[1]. Many recent studies have demonstrated that the obesity is one of the most important risk factors for the prevention of type 2 diabetes and its related co-morbid conditions.[4] In obese individuals, adipose tissue releases increased amounts of non-esterified fatty acids, glycerols, hormones, proinflammatory cytokines and other factors that are involved in the development of insulin resistance.[5] A drug agent that would be expected to impact type 2 diabetes and obesity would have potential to positively affect health outcomes for diabetes and the obese. Acetyl-coenzyme A carboxylase (ACC) has crucial roles in fatty acid metabolism in most living organisms and represents an attractive target for drug discovery. This heterodimeric protein which has two known isoforms, ACC1 and ACC2, is composed of carboxyltransferase (CT), biotin carboxyl carrier protein (BCCP), and biotin carboxylase (BC) domains, whose

Objectives

Results

Conclusion