Abstract
Machine learning is a well-known approach for virtual screening. Recently, deep learning, a machine learning algorithm in artificial neural networks, has been applied to the advancement of precision medicine and drug discovery. In this study, we performed comparative studies between deep neural networks (DNN) and other ligand-based virtual screening (LBVS) methods to demonstrate that DNN and random forest (RF) were superior in hit prediction efficiency. By using DNN, several triple-negative breast cancer (TNBC) inhibitors were identified as potent hits from a screening of an in-house database of 165,000 compounds. In broadening the application of this method, we harnessed the predictive properties of trained model in the discovery of G protein-coupled receptor (GPCR) agonist, by which computational structure-based design of molecules could be greatly hindered by lack of structural information. Notably, a potent (~ 500 nM) mu-opioid receptor (MOR) agonist was identified as a hit from a small-size training set of 63 compounds. Our results show that DNN could be an efficient module in hit prediction and provide experimental evidence that machine learning could identify potent hits in silico from a limited training set.
Highlights
Abbreviations deep neural networks (DNN) Deep neural networks ligand-based virtual screening (LBVS) Ligand-based virtual screening random forest (RF) Random forest triple-negative breast cancer (TNBC) Triple-negative breast cancer G protein-coupled receptor (GPCR) G-protein-couple receptors artificial intelligence (AI) Artificial intelligence quantitative structure–activity relationship (QSAR) Quantitative structure–activity relationship support vector machine (SVM) Support vector machine ADME Absorption, distribution, metabolism, and excretion decision tree (DT) Decision tree k-Nearest neighbors (K-NN) K-nearest neighbors artificial neural networks (ANNs) Artificial neural networks VS Virtual screening mu-opioid receptor (MOR) Mu-opioid receptor structure-based virtual screening (SBVS) Structure-based virtual screening partial least squares (PLS) Partial least squares multiple linear regression (MLR) Multiple linear regression extended connectivity fingerprints (ECFPs) Extended connectivity fingerprints functional-class fingerprints (FCFPs) Functional-class fingerprints FRU Relative fluorescence units
An advancement in the virtual screening method was made to reduce the burden of the drug discovery/development processes in a cost-effective manner[26]
Partial least squares (PLS) and multiple linear regression (MLR) are methods used for large data manipulation and allow facile generation of the model unlike other 3D-QSAR methods
Summary
Abbreviations DNN Deep neural networks LBVS Ligand-based virtual screening RF Random forest TNBC Triple-negative breast cancer GPCR G-protein-couple receptors AI Artificial intelligence QSAR Quantitative structure–activity relationship SVMs Support vector machine ADME Absorption, distribution, metabolism, and excretion DT Decision tree K-NN K-nearest neighbors ANNs Artificial neural networks VS Virtual screening MOR Mu-opioid receptor SBVS Structure-based virtual screening PLS Partial least squares MLR Multiple linear regression ECFPs Extended connectivity fingerprints FCFPs Functional-class fingerprints FRU Relative fluorescence units. The incorporation of machine learning method for the progressive analysis of the active compounds and concurrent generation of the prediction model should address such limitations. Our study suggested that deep learning method could generate potent hit compounds in different disease areas for the drug discovery process.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
