A two-step target binding and selectivity support vector machines approach for virtual screening of dopamine receptor subtype-selective ligands

Artificial intelligence and machine-learning approaches in structure and ligand-based discovery of drugs affecting central nervous system

CNS disorders are indications with a very high unmet medical needs, relatively smaller number of available drugs, and a subpar satisfaction level among patients and caregiver. Discovery of CNS drugs is extremely expensive affair with its own unique challenges leading to extremely high attrition rates and low efficiency. With explosion of data in information age, there is hardly any aspect of life that has not been touched by data driven technologies such as artificial intelligence (AI) and machine learning (ML). Drug discovery is no exception, emergence of big data via genomic, proteomic, biological, and chemical technologies has driven pharmaceutical giants to collaborate with AI oriented companies to revolutionise drug discovery, with the goal of increasing the efficiency of the process. In recent years many examples of innovative applications of AI and ML techniques in CNS drug discovery has been reported. Research on therapeutics for diseases such as schizophrenia, Alzheimer's and Parkinsonism has been provided with a new direction and thrust from these developments. AI and ML has been applied to both ligand-based and structure-based drug discovery and design of CNS therapeutics. In this review, we have summarised the general aspects of AI and ML from the perspective of drug discovery followed by a comprehensive coverage of the recent developments in the applications of AI/ML techniques in CNS drug discovery.

Prediction of activity and selectivity profiles of human Carbonic Anhydrase inhibitors using machine learning classification models

The development of selective inhibitors of the clinically relevant human Carbonic Anhydrase (hCA) isoforms IX and XII has become a major topic in drug research, due to their deregulation in several types of cancer. Indeed, the selective inhibition of these two isoforms, especially with respect to the homeostatic isoform II, holds great promise to develop anticancer drugs with limited side effects. Therefore, the development of in silico models able to predict the activity and selectivity against the desired isoform(s) is of central interest. In this work, we have developed a series of machine learning classification models, trained on high confidence data extracted from ChEMBL, able to predict the activity and selectivity profiles of ligands for human Carbonic Anhydrase isoforms II, IX and XII. The training datasets were built with a procedure that made use of flexible bioactivity thresholds to obtain well-balanced active and inactive classes. We used multiple algorithms and sampling sizes to finally select activity models able to classify active or inactive molecules with excellent performances. Remarkably, the results herein reported turned out to be better than those obtained by models built with the classic approach of selecting an a priori activity threshold. The sequential application of such validated models enables virtual screening to be performed in a fast and more reliable way to predict the activity and selectivity profiles against the investigated isoforms.

Benchmarking Data Sets from PubChem BioAssay Data: Current Scenario and Room for Improvement

Developing realistic data sets for evaluating virtual screening methods is a task that has been tackled by the cheminformatics community for many years. Numerous artificially constructed data collections were developed, such as DUD, DUD-E, or DEKOIS. However, they all suffer from multiple drawbacks, one of which is the absence of experimental results confirming the impotence of presumably inactive molecules, leading to possible false negatives in the ligand sets. In light of this problem, the PubChem BioAssay database, an open-access repository providing the bioactivity information of compounds that were already tested on a biological target, is now a recommended source for data set construction. Nevertheless, there exist several issues with the use of such data that need to be properly addressed. In this article, an overview of benchmarking data collections built upon experimental PubChem BioAssay input is provided, along with a thorough discussion of noteworthy issues that one must consider during the design of new ligand sets from this database. The points raised in this review are expected to guide future developments in this regard, in hopes of offering better evaluation tools for novel in silico screening procedures.

Applications of machine-learning methods for the discovery of NDM-1 inhibitors

The emergence of New Delhi metal beta-lactamase (NDM-1)-producing bacteria and their worldwide spread pose great challenges for the treatment of drug-resistant bacterial infections. These bacteria can hydrolyze most β-lactam antibacterials. Unfortunately, there are no clinically useful NDM-1 inhibitors. In the current work, we manually collected NDM-1 inhibitors reported in the past decade and established the first NDM-1 inhibitor database. Four machine-learning models were constructed using the structural and property characteristics of the collected compounds as input training set to discover potential NDM-1 inhibitors. In order to distinguish between high active inhibitors and putative positive drugs, a three-classification strategy was introduced in our study. In detail, the commonly used positive and negative divisions are converted into strongly active, weakly active, and inactive. The accuracy of the best prediction model designed based on this strategy reached 90.5%, compared with 69.14% achieved by the traditional docking-based virtual screening method. Consequently, the best model was used to virtually screen a natural product library. The safety of the selected compounds was analyzed by the ADMET prediction model based on machine learning. Seven novel NDM-1 inhibitors were identified, which will provide valuable clues for the discovery of NDM-1 inhibitors.

Logistic Regression of Ligands of Chemotaxis Receptors Offers Clues about Their Recognition by Bacteria

Because of relative simplicity of signal transduction pathway, bacterial chemotaxis sensory systems have been expected to be applied to biosensor. Tar and Tsr receptors mediate chemotaxis of Escherichia coli and have been studied extensively as models of chemoreception by bacterial two-transmembrane receptors. Such studies are typically conducted using two canonical ligands: l-aspartate for Tar and l-serine for Tsr. However, Tar and Tsr also recognize various analogs of aspartate and serine; it remains unknown whether the mechanism by which the canonical ligands are recognized is also common to the analogs. Moreover, in terms of engineering, it is important to know a single species of receptor can recognize various ligands to utilize bacterial receptor as the sensor for wide range of substances. To answer these questions, we tried to extract the features that are common to the recognition of the different analogs by constructing classification models based on machine-learning. We computed 20 physicochemical parameters for each of 38 well-known attractants that act as chemoreception ligands, and 15 known non-attractants. The classification models were generated by utilizing one or more of the seven physicochemical properties as descriptors. From the classification models, we identified the most effective physicochemical parameter for classification: the minimum electron potential. This descriptor that occurred repeatedly in classification models with the highest accuracies, This descriptor used alone could accurately classify 42/53 of compounds. Among the 11 misclassified compounds, eight contained two carboxyl groups, which is analogous to the structure of characteristic of aspartate analog. When considered separately, 16 of the 17 aspartate analogs could be classified accurately based on the distance between their two carboxyl groups. As shown in these results, we succeed to predict the ligands for bacterial chemoreceptors using only a few descriptors; single descriptor for single receptor. This result might be due to the relatively simple topology of bacterial two-transmembrane receptors compared to the G-protein-coupled receptors of seven-transmembrane receptors. Moreover, this distance between carboxyl groups correlated with the receptor binding affinity of the aspartate analogs. In view of this correlation, we propose a common mechanism underlying ligand recognition by Tar of compounds with two carboxyl groups.

Discovery of molecular mechanism of a clinical herbal formula upregulating serum HDL-c levels in treatment of metabolic syndrome by in vivo and computational studies

Decreased HDL cholesterol (HDL-c) is considered as an independent risk factor of cardiovascular disease in metabolic syndrome (Mets). Wendan decoction (WDD), a famous clinical traditional Chinese medicine formula in Mets in China, which can obviously up-regulate serum HDL-c levels in Mets. However, till now, the molecular mechanism of up-regulation still remained unclear. In this study, an integrated approach that combined serum ABCA1 in vivo assay, QSAR modeling and molecular docking was developed to explore the molecular mechanism and chemical substance basis of WDD upregulating HDL-c levels. Compared with Mets model group, serum ABCA1 and HDL-c levels intervened by two different doses of WDD for two weeks were significantly up-regulated. Then, kohonen and LDA were applied to develop QSAR models for ABCA1 up-regulators based flavonoids. The derived QSAR model produced the overall accuracy of 100%, a very powerful tool for screening ABCA1 up-regulators. The QSAR model prediction revealed 67 flavonoids in WDD were ABCA1 up-regulators. Finally, they were subjected to the molecular docking to understand their roles in up-regulating ABCA1 expression, which led to discovery of 23 ABCA1 up-regulators targeting LXR beta. Overall, QSAR modeling and docking studies well accounted for the observed in vivo activities of ABCA1 affected by WDD.

Pharmacological relationships and ligand discovery of G protein-coupled receptors revealed by simultaneous ligand and receptor clustering

Conventional ligand and receptor similarity methods have been extensively used for exposing pharmacological relationships and drug lead discovery. They may in some cases neglect minor relationships useful for target hopping particularly against the remote family members. To complement the conventional methods for capturing these minor relationships, we developed a new method that uses a SLARC (Simultaneous Ligand And Receptor Clustering) 2D map to simultaneously characterize the ligand structural and receptor binding-site sequence relationships of a receptor family. The SLARC maps of the rhodopsin-like GPCR family comprehensively revealed scaffold hopping, target hopping, and multi-target relationships for the ligands of both homologous and remote family members. Their usefulness in new ligand discovery was validated by guiding the prospective discovery of novel indole piperazinylpyrimidine dual-targeting adenosine A2A receptor antagonist and dopamine D2 agonist compounds. The SLARC approach is useful for revealing pharmacological relationships and discovering new ligands at target family levels.

Predicting Subtype Selectivity for Adenosine Receptor Ligands with Three-Dimensional Biologically Relevant Spectrum (BRS-3D)

Adenosine receptors (ARs) are potential therapeutic targets for Parkinson’s disease, diabetes, pain, stroke and cancers. Prediction of subtype selectivity is therefore important from both therapeutic and mechanistic perspectives. In this paper, we introduced a shape similarity profile as molecular descriptor, namely three-dimensional biologically relevant spectrum (BRS-3D), for AR selectivity prediction. Pairwise regression and discrimination models were built with the support vector machine methods. The average determination coefficient (r²) of the regression models was 0.664 (for test sets). The 2B-3 (A_2Bvs A₃) model performed best with q² = 0.769 for training sets (10-fold cross-validation), and r² = 0.766, RMSE = 0.828 for test sets. The models’ robustness and stability were validated with 100 times resampling and 500 times Y-randomization. We compared the performance of BRS-3D with 3D descriptors calculated by MOE. BRS-3D performed as good as, or better than, MOE 3D descriptors. The performances of the discrimination models were also encouraging, with average accuracy (ACC) 0.912 and MCC 0.792 (test set). The 2A-3 (A_2Avs A₃) selectivity discrimination model (ACC = 0.882 and MCC = 0.715 for test set) outperformed an earlier reported one (ACC = 0.784). These results demonstrated that, through multiple conformation encoding, BRS-3D can be used as an effective molecular descriptor for AR subtype selectivity prediction.

Getting the most out of PubChem for virtual screening

INTRODUCTION

With the emergence of the 'big data' era, the biomedical research community has great interest in exploiting publicly available chemical information for drug discovery. PubChem is an example of public databases that provide a large amount of chemical information free of charge.

AREAS COVERED

This article provides an overview of how PubChem's data, tools, and services can be used for virtual screening and reviews recent publications that discuss important aspects of exploiting PubChem for drug discovery.

EXPERT OPINION

PubChem offers comprehensive chemical information useful for drug discovery. It also provides multiple programmatic access routes, which are essential to build automated virtual screening pipelines that exploit PubChem data. In addition, PubChemRDF allows users to download PubChem data and load them into a local computing facility, facilitating data integration between PubChem and other resources. PubChem resources have been used in many studies for developing bioactivity and toxicity prediction models, discovering polypharmacologic (multi-target) ligands, and identifying new macromolecule targets of compounds (for drug-repurposing or off-target side effect prediction). These studies demonstrate the usefulness of PubChem as a key resource for computer-aided drug discovery and related area.

Predicting subtype selectivity of dopamine receptor ligands with three-dimensional biologically relevant spectrum

We applied a novel molecular descriptor, three-dimensional biologically relevant spectrum (BRS-3D), in subtype selectivity prediction of dopamine receptor (DR) ligands. BRS-3D is a shape similarity profile calculated by superimposing the objective compounds against 300 template ligands from sc-PDB. First, we constructed five subtype selectivity regression models between DR subtypes D1-D2, D1-D3, D2-D3, D2-D4, and D3-D4. The models' 10-fold cross-validation-squared correlation coefficient (Q² , for training sets) and determination coefficient (R² , for test sets) were in the range of 0.5-0.7 and 0.6-0.8, respectively. Then, four pair-wise (D1-D2, D2-D3, D2-D4, and D3-D4) and a multitype (D2, D3, and D4) classification models were developed with the prediction accuracies around or over 90% (for test sets). Lastly, we compared the performances of the models developed on BRS-3D and classical descriptors. The results showed that BRS-3D performed similarly to classical 2D descriptors and better than other 3D descriptors. Combining BRS-3D and 2D descriptors can further improve the prediction performance. These results confirmed the capacity of BRS-3D in the prediction of DR subtype-selective ligands.

Structural Investigation for Optimization of Anthranilic Acid Derivatives as Partial FXR Agonists by in Silico Approaches

In this paper, a three level in silico approach was applied to investigate some important structural and physicochemical aspects of a series of anthranilic acid derivatives (AAD) newly identified as potent partial farnesoid X receptor (FXR) agonists. Initially, both two and three-dimensional quantitative structure activity relationship (2D- and 3D-QSAR) studies were performed based on such AAD by a stepwise technology combined with multiple linear regression and comparative molecular field analysis. The obtained 2D-QSAR model gave a high predictive ability (R²(train) = 0.935, R²(test) = 0.902, Q²(LOO) = 0.899). It also uncovered that number of rotatable single bonds (b_rotN), relative negative partial charges (RPC(-)), oprea's lead-like (opr_leadlike), subdivided van der Waal's surface area (SlogP_VSA2) and accessible surface area (ASA) were important features in defining activity. Additionally, the derived3D-QSAR model presented a higher predictive ability (R²(train) = 0.944, R²(test) = 0.892, Q²(LOO) = 0.802). Meanwhile, the derived contour maps from the 3D-QSAR model revealed the significant structural features (steric and electronic effects) required for improving FXR agonist activity. Finally, nine newly designed AAD with higher predicted EC50 values than the known template compound were docked into the FXR active site. The excellent molecular binding patterns of these molecules also suggested that they can be robust and potent partial FXR agonists in agreement with the QSAR results. Overall, these derived models may help to identify and design novel AAD with better FXR agonist activity.

2D and 3D QSAR models for identifying diphenylpyridylethanamine based inhibitors against cholesteryl ester transfer protein

Cholesteryl ester transfer protein (CETP) inhibitors hold promise as new agents against coronary heart disease. Molecular modeling techniques such as 2D-QSAR and 3D-QSAR analysis were applied to establish models to distinguish potent and weak CETP inhibitors. 2D and 3D QSAR models-based a series of diphenylpyridylethanamine (DPPE) derivatives (newly identified as CETP inhibitors) were then performed to elucidate structural and physicochemical requirements for higher CETP inhibitory activity. The linear and spline 2D-QSAR models were developed through multiple linear regression (MLR) and support vector machine (SVM) methods. The best 2D-QSAR model obtained by SVM gave a high predictive ability (R(2)train=0.929, R(2)test=0.826, Q(2)LOO=0.780). Also, the 2D-QSAR models uncovered that SlogP_VSA0, E_sol and Vsurf_DW23 were important features in defining activity. In addition, the best 3D-QSAR model presented higher predictive ability (R(2)train=0.958, R(2)test=0.852, Q(2)LOO=0.734) based on comparative molecular field analysis (CoMFA). Meanwhile, the derived contour maps from 3D-QSAR model revealed the significant structural features (steric and electronic effects) required for improving CETP inhibitory activity. Consequently, twelve newly designed DPPE derivatives were proposed to be robust and potent CETP inhibitors. Overall, these derived models may help to design novel DPPE derivatives with better CETP inhibitory activity.

A lazy learning-based QSAR classification study for screening potential histone deacetylase 8 (HDAC8) inhibitors

Histone deacetylases 8 (HDAC8) is an enzyme repressing the transcription of various genes including tumour suppressor gene and has already become a target of human cancer treatment. In an effort to facilitate the discovery of HDAC8 inhibitors, two quantitative structure-activity relationship (QSAR) classification models were developed using K nearest neighbours (KNN) and neighbourhood classifier (NEC). Molecular descriptors were calculated for the data set and database compounds using ADRIANA.Code of Molecular Networks. Principal components analysis (PCA) was used to select the descriptors. The developed models were validated by leave-one-out cross validation (LOO CV). The performances of the developed models were evaluated with an external test set. Highly predictive models were used for database virtual screening. Furthermore, hit compounds were subsequently subject to molecular docking. Five hits were obtained based on consensus scoring function and binding affinity as potential HDAC8 inhibitors. Finally, HDAC8 structures in complex with five hits were also subjected to 5 ns molecular dynamics (MD) simulations to evaluate the complex structure stability. To the best of our knowledge, the NEC classification model used in this study is the first application of NEC to virtual screening for drug discovery.

Predicting selective liver X receptor β agonists using multiple machine learning methods

The classification models for predicting selective LXRβ agonists were firstly established using multiple machine learning methods. The top models can predict selective LXRβ agonists with chemical structure diversity.