Virtual screening of Abl inhibitors from large compound libraries by support vector machines

A multiple-step screening protocol to identify norepinephrine and dopamine reuptake inhibitors for depression

Depression severely impairs the health of people all over the world. Cognitive dysfunction due to depression has resulted in a severe economic burden to family and society induced by the reduction of social functioning of patients. Norepinephrine-dopamine reuptake inhibitors (NDRIs) targeted with the human norepinephrine transporter (hNET) and distributed with the human dopamine transporter (hDAT) simultaneously treat depression and improve cognitive function, and they effectively prevent sexual dysfunction and other side effects. Because many patients continue to poorly respond to NDRIs, it is urgent to discover novel NDRI antidepressants that do not interfere with cognitive function. The aim of this work was to selectively identify novel NDRI candidates acting against hNET and hDAT from large compound libraries by a comprehensive strategy integrating support vector machine (SVM) models, ADMET, molecular docking, in vitro binding assays, molecular dynamics simulation, and binding energy calculation. First, 6522 compounds that do not inhibit the human serotonin transporter (hSERT) were obtained by SVM models of hNET, hDAT, and non-target hSERT with similarity analyses from compound libraries. ADMET and molecular docking were then used to identify compounds that could potently bind to the hNET and hDAT with satisfactory ADMET, and 4 compounds were successfully identified. According to their docking scores and ADMET information, 3719810 was advanced for profiling by in vitro assays as a novel NDRI lead compound due to its strongest druggability and balancing activities. Encouragingly, 3719810 performed comparative activities on two targets, with Ki values of 7.32 μM for hNET and 5.23 μM for hDAT. To obtain candidates with additional activities and balance the activities of 2 targets, 5 analogs were optimized, and 2 novel scaffold compounds were successively designed. By assessment of molecular docking, molecular dynamics simulations, and binding energy calculations, 5 compounds were validated as NDRI candidates with high activities, and 4 of them performed acceptable balancing activities acting on hNET and hDAT. This work supplied promising novel NDRIs for treatment of depression with cognitive dysfunction or other related neurodegenerative disorders, and also provided a strategy for highly efficient and cost-effective identification of inhibitors for dual targets with homologous non-targets.

A novel strategy for designing the magic shotguns for distantly related target pairs

Due to its promising capacity in improving drug efficacy, polypharmacology has emerged to be a new theme in the drug discovery of complex disease. In the process of novel multi-target drugs (MTDs) discovery, in silico strategies come to be quite essential for the advantage of high throughput and low cost. However, current researchers mostly aim at typical closely related target pairs. Because of the intricate pathogenesis networks of complex diseases, many distantly related targets are found to play crucial role in synergistic treatment. Therefore, an innovational method to develop drugs which could simultaneously target distantly related target pairs is of utmost importance. At the same time, reducing the false discovery rate in the design of MTDs remains to be the daunting technological difficulty. In this research, effective small molecule clustering in the positive dataset, together with a putative negative dataset generation strategy, was adopted in the process of model constructions. Through comprehensive assessment on 10 target pairs with hierarchical similarity-levels, the proposed strategy turned out to reduce the false discovery rate successfully. Constructed model types with much smaller numbers of inhibitor molecules gained considerable yields and showed better false-hit controllability than before. To further evaluate the generalization ability, an in-depth assessment of high-throughput virtual screening on ChEMBL database was conducted. As a result, this novel strategy could hierarchically improve the enrichment factors for each target pair (especially for those distantly related/unrelated target pairs), corresponding to target pair similarity-levels.

Target-Based Small Molecule Drug Discovery for Colorectal Cancer: A Review of Molecular Pathways and In Silico Studies

Colorectal cancer is one of the most prevalent cancer types. Although there have been breakthroughs in its treatments, a better understanding of the molecular mechanisms and genetic involvement in colorectal cancer will have a substantial role in producing novel and targeted treatments with better safety profiles. In this review, the main molecular pathways and driver genes that are responsible for initiating and propagating the cascade of signaling molecules reaching carcinoma and the aggressive metastatic stages of colorectal cancer were presented. Protein kinases involved in colorectal cancer, as much as other cancers, have seen much focus and committed efforts due to their crucial role in subsidizing, inhibiting, or changing the disease course. Moreover, notable improvements in colorectal cancer treatments with in silico studies and the enhanced selectivity on specific macromolecular targets were discussed. Besides, the selective multi-target agents have been made easier by employing in silico methods in molecular de novo synthesis or target identification and drug repurposing.

Synthesis, single crystal (XRD), Hirshfeld surface analysis, computational study (DFT) and molecular docking studies of (E)-4-((2-hydroxy-3,5-diiodobenzylidene)amino)-N-(pyrimidine)-2-yl) benzenesulfonamide

The Schiff base (E)-4-((2-hydroxy-3,5-diiodobenzylidene)amino)-N-(pyrimidine)-2-yl) benzene sulfonamide (DIDA) compound was synthesis with condensation of 3,5-diiodosalicylaldehyde and sulfadiazine. The compound characterized with FTIR, X-ray crystallography and electronic spectra. The titled compound associated with experimental and theoretical method, DFT used for the theoretical method. The IR was calculated from DFT mode with B3LYP/GENSEP basic set. The electronic spectra computed from TD-DFT method with CAM-B3LYP functional, with IEFPCM solvation model and DMSO used as the solvent. Wave function based properties like localized orbital locator, electron localization function and non-covalent interactions have been studied extensively. The ADMET properties of the compound DIDA indicated that the compound has excellent drug likeness properties and PASS studies showed that it has anti-infective properties, which is confirmed by a docking score of -7.4 kcal/mol.

A network pharmacology study on the triterpene saponins from Medicago sativa L. for the treatment of Neurodegenerative diseases

Neurodegenerative diseases (NDDs) are characterized by progressive and irreversible, is a kind of complex illnesses, and the long-term therapy which is frequently associated with adverse side effects. Medicago sativa L., widely consumed as a vegetable, has the effects of improving memory and relieving central nervous system diseases. However, there are less studies on its specific mechanism for NDDs. In this investigation, we applied a method of network pharmacology, which combined molecular docking and network analysis to decipher the mechanisms of M. sativa in NDDs. The pharmacological system generated 55 triterpene saponins from M. sativa, and predicted 27 potential targets with 100 pathways in the treatment of NDDs. As a result, 13 compounds, 10 target proteins, and 6 signaling pathways were found to play important roles in the treatment of NDDs. In addition, in vitro experiments of isolates confirmed activities for NDDs, which were consistent with the results of network pharmacology prediction. PRACTICAL APPLICATIONS: Medicago sativa L. has been widely consumed as a vegetable, which possesses many nutritional components. As a functional food stuff, M. sativa can improve human health, such as memory improving activities, relieving central nervous system diseases, immunomodulatory, antioxidant, anticancer, and anti-inflammatory. In this article, the mechanism of triterpene saponins from M. sativa against NDDs was successfully predicted by network pharmacology method. The results will serve as a reference of M. sativa against NDDs.

A Structure-Based Drug Discovery Paradigm

Structure-based drug design is becoming an essential tool for faster and more cost-efficient lead discovery relative to the traditional method. Genomic, proteomic, and structural studies have provided hundreds of new targets and opportunities for future drug discovery. This situation poses a major problem: the necessity to handle the “big data” generated by combinatorial chemistry. Artificial intelligence (AI) and deep learning play a pivotal role in the analysis and systemization of larger data sets by statistical machine learning methods. Advanced AI-based sophisticated machine learning tools have a significant impact on the drug discovery process including medicinal chemistry. In this review, we focus on the currently available methods and algorithms for structure-based drug design including virtual screening and de novo drug design, with a special emphasis on AI- and deep-learning-based methods used for drug discovery.

SVMDLF: A novel R-based Web application for prediction of dipeptidyl peptidase 4 inhibitors

Dipeptidyl peptidase 4 (DPP4) is a well-known target for the antidiabetic drugs. However, currently available DPP4 inhibitor screening assays are costly and labor-intensive. It is important to create a robust in silico method to predict the activity of DPP4 inhibitor for the new lead finding. Here, we introduce an R-based Web application SVMDLF (SVM-based DPP4 Lead Finder) to predict the inhibitor of DPP4, based on support vector machine (SVM) model, predictions of which are confirmed by in vitro biological evaluation. The best model generated by MACCS structure fingerprint gave the Matthews correlation coefficient of 0.87 for the test set and 0.883 for the external test set. We screened Maybridge database consisting approximately 53,000 compounds. For further bioactivity assay, six compounds were shortlisted, and of six hits, three compounds showed significant DPP4 inhibitory activities with IC₅₀ values ranging from 8.01 to 10.73 μm. This application is an OpenCPU server app which is a novel single-page R-based Web application for the DPP4 inhibitor prediction. The SVMDLF is freely available and open to all users at http://svmdlf.net/ocpu/library/dlfsvm/www/ and http://www.cdri.res.in/svmdlf/.

Multiple machine learning based descriptive and predictive workflow for the identification of potential PTP1B inhibitors

In insulin and leptin signaling pathway, Protein-Tyrosine Phosphatase 1B (PTP1B) plays a crucial controlling role as a negative regulator, which makes it an attractive therapeutic target for both Type-2 Diabetes (T2D) and obesity. In this work, we have generated classification models by using the inhibition data set of known PTP1B inhibitors to identify new inhibitors of PTP1B utilizing multiple machine learning techniques like naïve Bayesian, random forest, support vector machine and k-nearest neighbors, along with structural fingerprints and selected molecular descriptors. Several models from each algorithm have been constructed and optimized, with the different combination of molecular descriptors and structural fingerprints. For the training and test sets, most of the predictive models showed more than 90% of overall prediction accuracies. The best model was obtained with support vector machine approach and has Matthews Correlation Coefficient of 0.82 for the external test set, which was further employed for the virtual screening of Maybridge small compound database. Five compounds were subsequently selected for experimental assay. Out of these two compounds were found to inhibit PTP1B with significant inhibitory activity in in-vitro inhibition assay. The structural fragments which are important for PTP1B inhibition were identified by naïve Bayesian method and can be further exploited to design new molecules around the identified scaffolds. The descriptive and predictive modeling strategy applied in this study is capable of identifying PTP1B inhibitors from the large compound libraries.

Towards cheminformatics-based estimation of drug therapeutic index: Predicting the protective index of anticonvulsants using a new quantitative structure-index relationship approach

The overall efficacy and safety profile of a new drug is partially evaluated by the therapeutic index in clinical studies and by the protective index (PI) in preclinical studies. In-silico predictive methods may facilitate the assessment of these indicators. Although QSAR and QSTR models can be used for predicting PI, their predictive capability has not been evaluated. To test this capability, we developed QSAR and QSTR models for predicting the activity and toxicity of anticonvulsants at accuracy levels above the literature-reported threshold (LT) of good QSAR models as tested by both the internal 5-fold cross validation and external validation method. These models showed significantly compromised PI predictive capability due to the cumulative errors of the QSAR and QSTR models. Therefore, in this investigation a new quantitative structure-index relationship (QSIR) model was devised and it showed improved PI predictive capability that superseded the LT of good QSAR models. The QSAR, QSTR and QSIR models were developed using support vector regression (SVR) method with the parameters optimized by using the greedy search method. The molecular descriptors relevant to the prediction of anticonvulsant activities, toxicities and PIs were analyzed by a recursive feature elimination method. The selected molecular descriptors are primarily associated with the drug-like, pharmacological and toxicological features and those used in the published anticonvulsant QSAR and QSTR models. This study suggested that QSIR is useful for estimating the therapeutic index of drug candidates.

Multistage virtual screening and identification of novel HIV-1 protease inhibitors by integrating SVM, shape, pharmacophore and docking methods

The HIV-1 protease has proven to be a crucial component of the HIV replication machinery and a reliable target for anti-HIV drug discovery. In this study, we applied an optimized hierarchical multistage virtual screening method targeting HIV-1 protease. The method sequentially applied SVM (Support Vector Machine), shape similarity, pharmacophore modeling and molecular docking. Using a validation set (270 positives, 155,996 negatives), the multistage virtual screening method showed a high hit rate and high enrichment factor of 80.47% and 465.75, respectively. Furthermore, this approach was applied to screen the National Cancer Institute database (NCI), which contains 260,000 molecules. From the final hit list, 6 molecules were selected for further testing in an in vitro HIV-1 protease inhibitory assay, and 2 molecules (NSC111887 and NSC121217) showed inhibitory potency against HIV-1 protease, with IC50 values of 62 μM and 162 μM, respectively. With further chemical development, these 2 molecules could potentially serve as HIV-1 protease inhibitors.

Recent progresses in the exploration of machine learning methods as in-silico ADME prediction tools

In-silico methods have been explored as potential tools for assessing ADME and ADME regulatory properties particularly in early drug discovery stages. Machine learning methods, with their ability in classifying diverse structures and complex mechanisms, are well suited for predicting ADME and ADME regulatory properties. Recent efforts have been directed at the broadening of application scopes and the improvement of predictive performance with particular focuses on the coverage of ADME properties, and exploration of more diversified training data, appropriate molecular features, and consensus modeling. Moreover, several online machine learning ADME prediction servers have emerged. Here we review these progresses and discuss the performances, application prospects and challenges of exploring machine learning methods as useful tools in predicting ADME and ADME regulatory properties.

In silico identification of novel kinase inhibitors targeting wild-type and T315I mutant ABL1 from FDA-approved drugs

The constitutively active fusion protein BCR-ABL1 is the major cause of chronic myeloid leukemia (CML), and selective inhibition of ABL1 is a promising approach for the treatment of CML. Reported drugs worked well in clinical practice, such as imatinib, dasatinib, nilotinib and bosutinib. However, resistance arises due to ABL1 mutation in patients, especially the T315I gate-keeper mutation. Thus, wide spectrum drugs targeting ABL1 are urgently needed. In order to screen potential drugs targeting wild-type ABL1 and T315I mutant ABL1, 1408 FDA approved small molecule drugs were subjected to molecular docking. With subsequent molecular dynamic (MD) simulation and MM/GBSA binding free energy calculation and energy decomposition, we identified chlorhexidine and sorafenib as potential "new use" drugs targeting wild-type ABL1, while nicergoline and plerixafor targeted T315I ABL1. Meanwhile, we also found that residues located in the ATP-binding site and A-loop motif played key roles in drug discovery towards ABL1. These findings may not only serve as a paradigm for the repositioning of existing approved drugs, but also instill new vitality to ABL1-targeted anti-CML therapeutics.

Predictions of BuChE inhibitors using support vector machine and naive Bayesian classification techniques in drug discovery

Butyrylcholinesterase (BuChE, EC 3.1.1.8) is an important pharmacological target for Alzheimer's disease (AD) treatment. However, the currently available BuChE inhibitor screening assays are expensive, labor-intensive, and compound-dependent. It is necessary to develop robust in silico methods to predict the activities of BuChE inhibitors for the lead identification. In this investigation, support vector machine (SVM) models and naive Bayesian models were built to discriminate BuChE inhibitors (BuChEIs) from the noninhibitors. Each molecule was initially represented in 1870 structural descriptors (1235 from ADRIANA.Code, 334 from MOE, and 301 from Discovery studio). Correlation analysis and stepwise variable selection method were applied to figure out activity-related descriptors for prediction models. Additionally, structural fingerprint descriptors were added to improve the predictive ability of models, which were measured by cross-validation, a test set validation with 1001 compounds and an external test set validation with 317 diverse chemicals. The best two models gave Matthews correlation coefficient of 0.9551 and 0.9550 for the test set and 0.9132 and 0.9221 for the external test set. To demonstrate the practical applicability of the models in virtual screening, we screened an in-house data set with 3601 compounds, and 30 compounds were selected for further bioactivity assay. The assay results showed that 10 out of 30 compounds exerted significant BuChE inhibitory activities with IC50 values ranging from 0.32 to 22.22 μM, at which three new scaffolds as BuChE inhibitors were identified for the first time. To our best knowledge, this is the first report on BuChE inhibitors using machine learning approaches. The models generated from SVM and naive Bayesian approaches successfully predicted BuChE inhibitors. The study proved the feasibility of a new method for predicting bioactivities of ligands and discovering novel lead compounds.

The influence of the inactives subset generation on the performance of machine learning methods

BACKGROUND

A growing popularity of machine learning methods application in virtual screening, in both classification and regression tasks, can be observed in the past few years. However, their effectiveness is strongly dependent on many different factors.

RESULTS

In this study, the influence of the way of forming the set of inactives on the classification process was examined: random and diverse selection from the ZINC database, MDDR database and libraries generated according to the DUD methodology. All learning methods were tested in two modes: using one test set, the same for each method of inactive molecules generation and using test sets with inactives prepared in an analogous way as for training. The experiments were carried out for 5 different protein targets, 3 fingerprints for molecules representation and 7 classification algorithms with varying parameters. It appeared that the process of inactive set formation had a substantial impact on the machine learning methods performance.

CONCLUSIONS

The level of chemical space limitation determined the ability of tested classifiers to select potentially active molecules in virtual screening tasks, as for example DUDs (widely applied in docking experiments) did not provide proper selection of active molecules from databases with diverse structures. The study clearly showed that inactive compounds forming training set should be representative to the highest possible extent for libraries that undergo screening.

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

Target selective drugs, such as dopamine receptor (DR) subtype selective ligands, are developed for enhanced therapeutics and reduced side effects. In silico methods have been explored for searching DR selective ligands, but encountered difficulties associated with high subtype similarity and ligand structural diversity. Machine learning methods have shown promising potential in searching target selective compounds. Their target selective capability can be further enhanced. In this work, we introduced a new two-step support vector machines target-binding and selectivity screening method for searching DR subtype-selective ligands, which was tested together with three previously-used machine learning methods for searching D1, D2, D3 and D4 selective ligands. It correctly identified 50.6%–88.0% of the 21–408 subtype selective and 71.7%–81.0% of the 39–147 multi-subtype ligands. Its subtype selective ligand identification rates are significantly better than, and its multi-subtype ligand identification rates are comparable to the best rates of the previously used methods. Our method produced low false-hit rates in screening 13.56 M PubChem, 168,016 MDDR and 657,736 ChEMBLdb compounds. Molecular features important for subtype selectivity were extracted by using the recursive feature elimination feature selection method. These features are consistent with literature-reported features. Our method showed similar performance in searching estrogen receptor subtype selective ligands. Our study demonstrated the usefulness of the two-step target binding and selectivity screening method in searching subtype selective ligands from large compound libraries.

Discovery of novel Pim-1 kinase inhibitors by a hierarchical multistage virtual screening approach based on SVM model, pharmacophore, and molecular docking

In this investigation, we describe the discovery of novel potent Pim-1 inhibitors by employing a proposed hierarchical multistage virtual screening (VS) approach, which is based on support vector machine-based (SVM-based VS or SB-VS), pharmacophore-based VS (PB-VS), and docking-based VS (DB-VS) methods. In this approach, the three VS methods are applied in an increasing order of complexity so that the first filter (SB-VS) is fast and simple, while successive ones (PB-VS and DB-VS) are more time-consuming but are applied only to a small subset of the entire database. Evaluation of this approach indicates that it can be used to screen a large chemical library rapidly with a high hit rate and a high enrichment factor. This approach was then applied to screen several large chemical libraries, including PubChem, Specs, and Enamine as well as an in-house database. From the final hits, 47 compounds were selected for further in vitro Pim-1 inhibitory assay, and 15 compounds show nanomolar level or low micromolar inhibition potency against Pim-1. In particular, four of them were found to have new scaffolds which have potential for the chemical development of Pim-1 inhibitors.

Discovery of novel mGluR1 antagonists: a multistep virtual screening approach based on an SVM model and a pharmacophore hypothesis significantly increases the hit rate and enrichment factor

Development of glutamate non-competitive antagonists of mGluR1 (Metabotropic glutamate receptor subtype 1) has increasingly attracted much attention in recent years due to their potential therapeutic application for various nervous disorders. Since there is no crystal structure reported for mGluR1, ligand-based virtual screening (VS) methods, typically pharmacophore-based VS (PB-VS), are often used for the discovery of mGluR1 antagonists. Nevertheless, PB-VS usually suffers a lower hit rate and enrichment factor. In this investigation, we established a multistep ligand-based VS approach that is based on a support vector machine (SVM) classification model and a pharmacophore model. Performance evaluation of these methods in virtual screening against a large independent test set, M-MDDR, show that the multistep VS approach significantly increases the hit rate and enrichment factor compared with the individual SB-VS and PB-VS methods. The multistep VS approach was then used to screen several large chemical libraries including PubChem, Specs, and Enamine. Finally a total of 20 compounds were selected from the top ranking compounds, and shifted to the subsequent in vitro and in vivo studies, which results will be reported in the near future.

Exploiting PubChem for Virtual Screening

IMPORTANCE OF THE FIELD: PubChem is a public molecular information repository, a scientific showcase of the NIH Roadmap Initiative. The PubChem database holds over 27 million records of unique chemical structures of compounds (CID) derived from nearly 70 million substance depositions (SID), and contains more than 449,000 bioassay records with over thousands of in vitro biochemical and cell-based screening bioassays established, with targeting more than 7000 proteins and genes linking to over 1.8 million of substances. AREAS COVERED IN THIS REVIEW: This review builds on recent PubChem-related computational chemistry research reported by other authors while providing readers with an overview of the PubChem database, focusing on its increasing role in cheminformatics, virtual screening and toxicity prediction modeling. WHAT THE READER WILL GAIN: These publicly available datasets in PubChem provide great opportunities for scientists to perform cheminformatics and virtual screening research for computer-aided drug design. However, the high volume and complexity of the datasets, in particular the bioassay-associated false positives/negatives and highly imbalanced datasets in PubChem, also creates major challenges. Several approaches regarding the modeling of PubChem datasets and development of virtual screening models for bioactivity and toxicity predictions are also reviewed. TAKE HOME MESSAGE: Novel data-mining cheminformatics tools and virtual screening algorithms are being developed and used to retrieve, annotate and analyze the large-scale and highly complex PubChem biological screening data for drug design.

In-silico approaches to multi-target drug discovery : computer aided multi-target drug design, multi-target virtual screening

Multi-target drugs against selective multiple targets improve therapeutic efficacy, safety and resistance profiles by collective regulations of a primary therapeutic target together with compensatory elements and resistance activities. Efforts have been made to employ in-silico methods for facilitating the search and design of selective multi-target agents. These methods have shown promising potential in facilitating drug discovery directed at selective multiple targets.