Predicting Mouse Liver Microsomal Stability with "Pruned" Machine Learning Models and Public Data

Enhancing Multi-species Liver Microsomal Stability Prediction through Artificial Intelligence

Liver microsomal stability, a crucial aspect of metabolic stability, significantly impacts practical drug discovery. However, current models for predicting liver microsomal stability are based on limited molecular information from a single species. To address this limitation, we constructed the largest public database of compounds from three common species: human, rat, and mouse. Subsequently, we developed a series of classification models using both traditional descriptor-based and classic graph-based machine learning (ML) algorithms. Remarkably, the best-performing models for the three species achieved Matthews correlation coefficients (MCCs) of 0.616, 0.603, and 0.574, respectively, on the test set. Furthermore, through the construction of consensus models based on these individual models, we have demonstrated their superior predictive performance in comparison with the existing models of the same type. To explore the similarities and differences in the properties of liver microsomal stability among multispecies molecules, we conducted preliminary interpretative explorations using the Shapley additive explanations (SHAP) and atom heatmap approaches for the models and misclassified molecules. Additionally, we further investigated representative structural modifications and substructures that decrease the liver microsomal stability in different species using the matched molecule pair analysis (MMPA) method and substructure extraction techniques. The established prediction models, along with insightful interpretation information regarding liver microsomal stability, will significantly contribute to enhancing the efficiency of exploring practical drugs for development.

MS-BACL: enhancing metabolic stability prediction through bond graph augmentation and contrastive learning

MOTIVATION

Accurately predicting molecular metabolic stability is of great significance to drug research and development, ensuring drug safety and effectiveness. Existing deep learning methods, especially graph neural networks, can reveal the molecular structure of drugs and thus efficiently predict the metabolic stability of molecules. However, most of these methods focus on the message passing between adjacent atoms in the molecular graph, ignoring the relationship between bonds. This makes it difficult for these methods to estimate accurate molecular representations, thereby being limited in molecular metabolic stability prediction tasks.

RESULTS

We propose the MS-BACL model based on bond graph augmentation technology and contrastive learning strategy, which can efficiently and reliably predict the metabolic stability of molecules. To our knowledge, this is the first time that bond-to-bond relationships in molecular graph structures have been considered in the task of metabolic stability prediction. We build a bond graph based on 'atom-bond-atom', and the model can simultaneously capture the information of atoms and bonds during the message propagation process. This enhances the model's ability to reveal the internal structure of the molecule, thereby improving the structural representation of the molecule. Furthermore, we perform contrastive learning training based on the molecular graph and its bond graph to learn the final molecular representation. Multiple sets of experimental results on public datasets show that the proposed MS-BACL model outperforms the state-of-the-art model.

AVAILABILITY AND IMPLEMENTATION

The code and data are publicly available at https://github.com/taowang11/MS.

Extended study on atomic featurization in graph neural networks for molecular property prediction

Graph neural networks have recently become a standard method for analyzing chemical compounds. In the field of molecular property prediction, the emphasis is now on designing new model architectures, and the importance of atom featurization is oftentimes belittled. When contrasting two graph neural networks, the use of different representations possibly leads to incorrect attribution of the results solely to the network architecture. To better understand this issue, we compare multiple atom representations by evaluating them on the prediction of free energy, solubility, and metabolic stability using graph convolutional networks. We discover that the choice of atom representation has a significant impact on model performance and that the optimal subset of features is task-specific. Additional experiments involving more sophisticated architectures, including graph transformers, support these findings. Moreover, we demonstrate that some commonly used atom features, such as the number of neighbors or the number of hydrogens, can be easily predicted using only information about bonds and atom type, yet their explicit inclusion in the representation has a positive impact on model performance. Finally, we explain the predictions of the best-performing models to better understand how they utilize the available atomic features.

CMMS-GCL: cross-modality metabolic stability prediction with graph contrastive learning

MOTIVATION

Metabolic stability plays a crucial role in the early stages of drug discovery and development. Accurately modeling and predicting molecular metabolic stability has great potential for the efficient screening of drug candidates as well as the optimization of lead compounds. Considering wet-lab experiment is time-consuming, laborious, and expensive, in silico prediction of metabolic stability is an alternative choice. However, few computational methods have been developed to address this task. In addition, it remains a significant challenge to explain key functional groups determining metabolic stability.

RESULTS

To address these issues, we develop a novel cross-modality graph contrastive learning model named CMMS-GCL for predicting the metabolic stability of drug candidates. In our framework, we design deep learning methods to extract features for molecules from two modality data, i.e. SMILES sequence and molecule graph. In particular, for the sequence data, we design a multihead attention BiGRU-based encoder to preserve the context of symbols to learn sequence representations of molecules. For the graph data, we propose a graph contrastive learning-based encoder to learn structure representations by effectively capturing the consistencies between local and global structures. We further exploit fully connected neural networks to combine the sequence and structure representations for model training. Extensive experimental results on two datasets demonstrate that our CMMS-GCL consistently outperforms seven state-of-the-art methods. Furthermore, a collection of case studies on sequence data and statistical analyses of the graph structure module strengthens the validation of the interpretability of crucial functional groups recognized by CMMS-GCL. Overall, CMMS-GCL can serve as an effective and interpretable tool for predicting metabolic stability, identifying critical functional groups, and thus facilitating the drug discovery process and lead compound optimization.

AVAILABILITY AND IMPLEMENTATION

The code and data underlying this article are freely available at https://github.com/dubingxue/CMMS-GCL.

The Inhibitory Properties of a Novel, Selective LMTK3 Kinase Inhibitor

Recently, the oncogenic role of lemur tyrosine kinase 3 (LMTK3) has been well established in different tumor types, highlighting it as a viable therapeutic target. In the present study, using in vitro and cell-based assays coupled with biophysical analyses, we identify a highly selective small molecule LMTK3 inhibitor, namely C36. Biochemical/biophysical and cellular studies revealed that C36 displays a high in vitro selectivity profile and provides notable therapeutic effect when tested in the National Cancer Institute (NCI)-60 cancer cell line panel. We also report the binding affinity between LMTK3 and C36 as demonstrated via microscale thermophoresis (MST). In addition, C36 exhibits a mixed-type inhibition against LMTK3, consistent with the inhibitor overlapping with both the adenosine 5'-triphosphate (ATP)- and substrate-binding sites. Treatment of different breast cancer cell lines with C36 led to decreased proliferation and increased apoptosis, further reinforcing the prospective value of LMTK3 inhibitors for cancer therapy.

In silico and in vitro ADME-Tox analysis and in vivo pharmacokinetic study of representative pan-PDE inhibitors from the group of 7,8-disubstituted derivatives of 1,3-dimethyl-7H-purine-2,6-dione

Phosphodiesterase (PDE) inhibitors represent a wide class of chemically different compounds that have been extensively studied in recent years. Their anti-inflammatory and anti-fibrotic effects are particularly desirable in the treatment of chronic respiratory diseases, including asthma and chronic obstructive pulmonary disease (COPD). Due to diversified expression of individual PDEs within cells and/or tissues as well as PDE signaling compartmentalization, pan-PDE inhibitors (compounds capable of simultaneously blocking various PDE subtypes) are of particular interest. Recently, a large group of 7,8-disubstituted derivatives of 1,3-dimethyl-7H-purine-2,6-dione (theophylline) was designed and synthesized. These compounds were characterized as potent pan-PDE inhibitors and their prominent anti-inflammatory and anti-fibrotic activity in vitro has been proved. Herein, we investigated a general in vitro safety profile and pharmacokinetic characteristics of two leading compounds from this group: a representative compound with N'-benzylidenebutanehydrazide moiety (38) and a representative derivative containing N-phenylbutanamide fragment (145). Both tested pan-PDE inhibitors revealed no cytotoxic, mutagenic, and genotoxic activity in vitro, showed moderate metabolic stability in mouse and human liver microsomes, as well as fell into the low or medium permeation category. Additionally, 38 and 145 revealed a lack of interaction with adenosine receptors, including A1, A2A, and A2B. Pharmacokinetic analysis revealed that both tested 7,8-disubstituted derivatives of 1,3-dimethyl-7H-purine-2,6-dione were effectively absorbed from the peritoneal cavity. Simultaneously, they were extensively distributed to mouse lungs and after intraperitoneal (i.p.) administration were detected in bronchoalveolar lavage fluid. These findings provide evidence that investigated compounds represent a new drug candidates with a favorable in vitro safety profile and satisfactory pharmacokinetic properties after a single i.p. administration. As the next step, further pharmacokinetic studies after multiple i.p. and p.o. doses will be conducted to ensure effective 38 and 145 serum and lung concentrations for a longer period of time. In summary, 7,8-disubstituted derivatives of 1,3-dimethyl-7H-purine-2,6-dione represent a promising compounds worth testing in animal models of chronic respiratory diseases, the etiology of which involves various PDE subtypes.

Generation of new inhibitors of selected cytochrome P450 subtypes- In silico study

Physicochemical and pharmacokinetic compound profile has crucial impact on compound potency to become a future drug. Ligands with desired activity profile cannot be used for treatment if they are characterized by unfavourable physicochemical or ADMET properties. In the study, we consider metabolic stability and focus on selected subtypes of cytochrome P450 - proteins, which take part in the first phase of compound transformations in the organism. We develop a protocol for generation of new potential inhibitors of selected cytochrome isoforms. Its subsequent stages are composed of generation and assessment of new derivatives of known cytochrome inhibitors, docking and evaluation of the compound possible inhibition on the basis of the obtained ligand-protein complexes. Besides the library of new potential agents inhibiting particular cytochrome subtypes, we also prepare a graph neural network that predicts the change in activity for all modifications of the starting molecule. In addition, we perform a systematic statistical study on the influence of particular substitutions on the potential inhibition properties of generated compounds (both mono- and di-substitutions are considered), provide explanations of the inhibitory predictions and prepare an on-line visualization platform enabling manual inspection of the results. The developed methodology can greatly support the design of new cytochrome P450 inhibitors with the overarching goal of generation of new metabolically stable compounds. It enables instant evaluation of possible compound-cytochrome interactions and selection of ligands with the highest potential of possessing desired biological activity.

PredMS: a random forest model for predicting metabolic stability of drug candidates in human liver microsomes

MOTIVATION

Poor metabolic stability leads to drug development failure. Therefore, it is essential to evaluate the metabolic stability of small compounds for successful drug discovery and development. However, evaluating metabolic stability in vitro and in vivo is expensive, time-consuming and laborious. In addition, only a few free software programs are available for metabolic stability data and prediction. Therefore, in this study, we aimed to develop a prediction model that predicts the metabolic stability of small compounds.

RESULTS

We developed a computational model, PredMS, which predicts the metabolic stability of small compounds as stable or unstable in human liver microsomes. PredMS is based on a random forest model using an in-house database of metabolic stability data of 1917 compounds. To validate the prediction performance of PredMS, we generated external test data of 61 compounds. PredMS achieved an accuracy of 0.74, Matthew's correlation coefficient of 0.48, sensitivity of 0.70, specificity of 0.86, positive predictive value of 0.94 and negative predictive value of 0.46 on the external test dataset. PredMS will be a useful tool to predict the metabolic stability of small compounds in the early stages of drug discovery and development.

AVAILABILITY AND IMPLEMENTATION

The source code for PredMS is available at https://bitbucket.org/krictai/predms, and the PredMS web server is available at https://predms.netlify.app.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Bayesian Modeling and Intrabacterial Drug Metabolism Applied to Drug-Resistant Staphylococcus aureus

We present the application of Bayesian modeling to identify chemical tools and/or drug discovery entities pertinent to drug-resistant Staphylococcus aureus infections. The quinoline JSF-3151 was predicted by modeling and then empirically demonstrated to be active against in vitro cultured clinical methicillin- and vancomycin-resistant strains while also exhibiting efficacy in a mouse peritonitis model of methicillin-resistant S. aureus infection. We highlight the utility of an intrabacterial drug metabolism (IBDM) approach to probe the mechanism by which JSF-3151 is transformed within the bacteria. We also identify and then validate two mechanisms of resistance in S. aureus: one mechanism involves increased expression of a lipocalin protein, and the other arises from the loss of function of an azoreductase. The computational and experimental approaches, discovery of an antibacterial agent, and elucidated resistance mechanisms collectively hold promise to advance our understanding of therapeutic regimens for drug-resistant S. aureus.

Development of Machine Learning Models and the Discovery of a New Antiviral Compound against Yellow Fever Virus

Yellow fever (YF) is an acute viral hemorrhagic disease transmitted by infected mosquitoes. Large epidemics of YF occur when the virus is introduced into heavily populated areas with high mosquito density and low vaccination coverage. The lack of a specific small molecule drug treatment against YF as well as for homologous infections, such as zika and dengue, highlights the importance of these flaviviruses as a public health concern. With the advancement in computer hardware and bioactivity data availability, new tools based on machine learning methods have been introduced into drug discovery, as a means to utilize the growing high throughput screening (HTS) data generated to reduce costs and increase the speed of drug development. The use of predictive machine learning models using previously published data from HTS campaigns or data available in public databases, can enable the selection of compounds with desirable bioactivity and absorption, distribution, metabolism, and excretion profiles. In this study, we have collated cell-based assay data for yellow fever virus from the literature and public databases. The data were used to build predictive models with several machine learning methods that could prioritize compounds for in vitro testing. Five molecules were prioritized and tested in vitro from which we have identified a new pyrazolesulfonamide derivative with EC₅₀ 3.2 μM and CC₅₀ 24 μM, which represents a new scaffold suitable for hit-to-lead optimization that can expand the available drug discovery candidates for YF.

[Predict of metabolic stability of xenobiotics by the PASS and GUSAR programs]

Metabolic stability refers to the susceptibility of compounds to the biotransformation; it is characterized by such pharmacokinetic parameters as half-life (T1/2) and clearance (CL). Generally, these parameters are estimated by in vitro assays, which are based on cells or subcellular fractions (mainly liver microsomal enzymes) and serve as models of the processes occurring in living organisms. Data obtained from the experiments are used to build QSAR (Quantitative Structure-Activity Relationship) models. More than 8000 compounds with known CL and/or T1/2 values obtained in vitro using human liver microsomes were selected from the freely available ChEMBL v.27 database. GUSAR (General Unrestricted Structure-Activity Relationships) and PASS (Prediction of Activity Spectra for Substances) softwares were used to make quantitative and classification models. The quality of the models was evaluated using 5-fold cross-validation. Compounds were subdivided into "stable" and "unstable" by means of the following threshold parameters: T1/2 = 30 minutes, CL = 20 ml/min/kg. The accuracy of the models ranged from 0.5 (calculated in 5-fold CV on the test set for the half-life prediction quantitative model) to 0.96 (calculated in 5-fold CV on the test set for the clearance prediction classification model).

Drug-Like Small Molecule HSP27 Functional Inhibitor Sensitizes Lung Cancer Cells to Gefitinib or Cisplatin by Inducing Altered Cross-Linked Hsp27 Dimers

Relationships between heat shock protein 27 (HSP27) and cancer aggressiveness, metastasis, drug resistance, and poor patient outcomes in various cancer types including non-small cell lung cancer (NSCLC) were reported, and inhibition of HSP27 expression is suggested to be a possible strategy for cancer therapy. Unlike HSP90 or HSP70, HSP27 does not have an ATP-binding pocket, and no effective HSP27 inhibitors have been identified. Previously, NSCLC cancer cells were sensitized to radiation and chemotherapy when co-treated with small molecule HSP27 functional inhibitors such as zerumbone (ZER), SW15, and J2 that can induce abnormal cross-linked HSP27 dimer. In this study, cancer inhibition effects of NA49, a chromenone compound with better solubility, longer circulation time, and less toxicity than J2, were examined in combination with anticancer drugs such as cisplatin and gefitinib in NSCLC cell lines. When the cytotoxic drug cisplatin was treated in combination with NA49 in epidermal growth factor receptors (EGFRs) WT cell lines, sensitization was induced in an HSP27 expression-dependent manner. With gefitinib treatment, NA49 showed increased combination effects in both EGFR WT and Mut cell lines, also with HSP27 expression-dependent patterns. Moreover, NA49 induced sensitization in EGFR Mut cells with a secondary mutation of T790M when combined with gefitinib. Augmented tumor growth inhibition was shown with the combination of cisplatin or gefitinib and NA49 in nude mouse xenograft models. These results suggest the combination of HSP27 inhibitor NA49 and anticancer agents as a candidate for overcoming HSP27-mediated drug resistance in NSCLC patients.

A Machine Learning Strategy for Drug Discovery Identifies Anti-Schistosomal Small Molecules

Schistosomiasis is a chronic and painful disease of poverty caused by the flatworm parasite Schistosoma. Drug discovery for antischistosomal compounds predominantly employs in vitro whole organism (phenotypic) screens against two developmental stages of Schistosoma mansoni, post-infective larvae (somules) and adults. We generated two rule books and associated scoring systems to normalize 3898 phenotypic data points to enable machine learning. The data were used to generate eight Bayesian machine learning models with the Assay Central software according to parasite’s developmental stage and experimental time point (≤24, 48, 72, and >72 h). The models helped predict 56 active and nonactive compounds from commercial compound libraries for testing. When these were screened against S. mansoni in vitro, the prediction accuracy for active and inactives was 61% and 56% for somules and adults, respectively; also, hit rates were 48% and 34%, respectively, far exceeding the typical 1–2% hit rate for traditional high throughput screens.

Random Forest Model Prediction of Compound Oral Exposure in the Mouse

An early hurdle in the optimization of small-molecule chemical probes and drug discovery entities is the attainment of sufficient exposure in the mouse via oral administration of the compound. While computational approaches have attempted to predict molecular properties related to the mouse pharmacokinetic (PK) profile, we present herein a machine learning approach to specifically predict the oral exposure of a compound as measured in the mouse snapshot PK assay. A random forest workflow was found to produce the best cross-validation and external test set statistics after processing of the input data set and optimization of model features. The modeling approach should be useful to the chemical biology and drug discovery communities to predict this key molecular property and afford chemical entities of translational significance.

Current development of integrated web servers for preclinical safety and pharmacokinetics assessments in drug development

In drug development, preclinical safety and pharmacokinetics assessments of candidate drugs to ensure the safety profile are a must. While in vivo and in vitro tests are traditionally used, experimental determinations have disadvantages, as they are usually time-consuming and costly. In silico predictions of these preclinical endpoints have each been developed in the past decades. However, only a few web-based tools have integrated different models to provide a simple one-step platform to help researchers thoroughly evaluate potential drug candidates. To efficiently achieve this approach, a platform for preclinical evaluation must not only predict key ADMET (absorption, distribution, metabolism, excretion and toxicity) properties but also provide some guidance on structural modifications to improve the undesired properties. In this review, we organized and compared several existing integrated web servers that can be adopted in preclinical drug development projects to evaluate the subject of interest. We also introduced our new web server, Virtual Rat, as an alternative choice to profile the properties of drug candidates. In Virtual Rat, we provide not only predictions of important ADMET properties but also possible reasons as to why the model made those structural predictions. Multiple models were implemented into Virtual Rat, including models for predicting human ether-a-go-go-related gene (hERG) inhibition, cytochrome P450 (CYP) inhibition, mutagenicity (Ames test), blood-brain barrier penetration, cytotoxicity and Caco-2 permeability. Virtual Rat is free and has been made publicly available at https://virtualrat.cmdm.tw/.

Pruned Machine Learning Models to Predict Aqueous Solubility

Solubility is a key metric for therapeutic compounds. Conversely, insoluble compounds cloud the accuracy of assays at all stages of chemical biology and drug discovery. Herein, we disclose naïve Bayesian classifier models to predict aqueous solubility. Publicly accessible aqueous solubility data were used to create two full, or nonpruned, training sets. These two sets were also combined to create a full fused set, and a training set comprised of a literature collation of solubility data was also considered as a reference. We tested different extents of data pruning on the training sets and constructed machine learning models that were evaluated with two independent, external test sets that contained compounds that were different from the training sets. The best pruned and fused model was significantly more accurate, in comparison to either the full model or the full fused model, with the prediction of these external test sets. By carefully removing data from the training set, less information can be used to create more accurate machine learning models for aqueous solubility. This knowledge and the curated training sets should prove useful to future machine learning approaches.

Machine Learning Platform to Discover Novel Growth Inhibitors of Neisseria gonorrhoeae

PURPOSE

To advance fundamental biological and translational research with the bacterium Neisseria gonorrhoeae through the prediction of novel small molecule growth inhibitors via naïve Bayesian modeling methodology.

METHODS

Inspection and curation of data from the publicly available ChEMBL web site for small molecule growth inhibition data of the bacterium Neisseria gonorrhoeae resulted in a training set for the construction of machine learning models. A naïve Bayesian model for bacterial growth inhibition was utilized in a workflow to predict novel antibacterial agents against this bacterium of global health relevance from a commercial library of >10⁵ drug-like small molecules. Follow-up efforts involved empirical assessment of the predictions and validation of the hits.

RESULTS

Specifically, two small molecules were found that exhibited promising activity profiles and represent novel chemotypes for agents against N. gonorrrhoeae.

CONCLUSIONS

This represents, to the best of our knowledge, the first machine learning approach to successfully predict novel growth inhibitors of this bacterium. To assist the chemical tool and drug discovery fields, we have made our curated training set available as part of the Supplementary Material and the Bayesian model is accessible via the web. Graphical Abstract.

Entering the era of computationally driven drug development

Historically, failure rates in drug development are high; increased sophistication and investment throughout the process has shifted the reasons for attrition, but the overall success rates have remained stubbornly and consistently low. Only 8% of new entities entering clinical testing gain regulatory approval, indicating that significant obstacles still exist for efficient therapeutic development. The continued high failure rate can be partially attributed to the inability to link drug exposure with the magnitude of observed safety and efficacy-related pharmacodynamic (PD) responses; frequently, this is a result of nonclinical models exhibiting poor prediction of human outcomes across a wide range of disease conditions, resulting in faulty evaluation of drug toxicology and efficacy. However, the increasing quality and standardization of experimental methods in preclinical stages of testing has created valuable data sets within companies that can be leveraged to further improve the efficiency and accuracy of preclinical prediction for both pharmacokinetics (PK) and PD. Models of Quantitative structure-activity relationships (QSAR), physiologically based pharmacokinetics (PBPK), and PK/PD relationships have also improved efficiency. Founded on a core understanding of biochemistry and physiological interactions of xenobiotics, these in silico methods have the potential to increase the probability of compound success in clinical trials. Integration of traditional computational methods with machine-learning approaches and existing internal pharma databases stands to make a fundamental impact on the speed and accuracy of predictions during the process of drug development and approval.

Metabolic stability and its role in the discovery of new chemical entities

Determination of metabolic profiles of new chemical entities is a key step in the process of drug discovery, since it influences pharmacokinetic characteristics of therapeutic compounds. One of the main challenges of medicinal chemistry is not only to design compounds demonstrating beneficial activity, but also molecules exhibiting favourable pharmacokinetic parameters. Chemical compounds can be divided into those which are metabolized relatively fast and those which undergo slow biotransformation. Rapid biotransformation reduces exposure to the maternal compound and may lead to the generation of active, non-active or toxic metabolites. In contrast, high metabolic stability may promote interactions between drugs and lead to parent compound toxicity. In the present paper, issues of compound metabolic stability will be discussed, with special emphasis on its significance, in vitro metabolic stability testing, dilemmas regarding in vitro-in vivo extrapolation of the results and some aspects relating to different preclinical species used in in vitro metabolic stability assessment of compounds.

Exploiting machine learning for end-to-end drug discovery and development

A variety of machine learning methods such as naive Bayesian, support vector machines and more recently deep neural networks are demonstrating their utility for drug discovery and development. These leverage the generally bigger datasets created from high-throughput screening data and allow prediction of bioactivities for targets and molecular properties with increased levels of accuracy. We have only just begun to exploit the potential of these techniques but they may already be fundamentally changing the research process for identifying new molecules and/or repurposing old drugs. The integrated application of such machine learning models for end-to-end (E2E) application is broadly relevant and has considerable implications for developing future therapies and their targeting.

Naïve Bayesian Models for Vero Cell Cytotoxicity

PURPOSE

To advance translational research of potential therapeutic small molecules against infectious microbes, the compounds must display a relative lack of mammalian cell cytotoxicity. Vero cell cytotoxicity (CC₅₀) is a common initial assay for this metric. We explored the development of naïve Bayesian models that can enhance the probability of identifying non-cytotoxic compounds.

METHODS

Vero cell cytotoxicity assays were identified in PubChem, reformatted, and curated to create a training set with 8741 unique small molecules. These data were used to develop Bayesian classifiers, which were assessed with internal cross-validation, external tests with a set of 193 compounds from our laboratory, and independent validation with an additional diverse set of 1609 unique compounds from PubChem.

RESULTS

Evaluation with independent, external test and validation sets indicated that cytotoxicity Bayesian models constructed with the ECFP_6 descriptor were more accurate than those that used FCFP_6 fingerprints. The best cytotoxicity Bayesian model displayed predictive power in external evaluations, according to conventional and chance-corrected statistics, as well as enrichment factors.

CONCLUSIONS

The results from external tests demonstrate that our novel cytotoxicity Bayesian model displays sufficient predictive power to help guide translational research. To assist the chemical tool and drug discovery communities, our curated training set is being distributed as part of the Supplementary Material. Graphical Abstract Naive Bayesian models have been trained with publically available data and offer a useful tool for chemical biology and drug discovery to select for small molecules with a high probability of exhibiting acceptably low Vero cell cytotoxicity.

Novel Pyrimidines as Antitubercular Agents

Mycobacterium tuberculosis infection is responsible for a global pandemic. New drugs are needed that do not show cross-resistance with the existing front-line therapeutics. A triazine antitubercular hit led to the design of a related pyrimidine family. The synthesis of a focused series of these analogs facilitated exploration of their in vitro activity, in vitro cytotoxicity, and physiochemical and absorption-distribution-metabolism-excretion properties. Select pyrimidines were then evaluated for their pharmacokinetic profiles in mice. The findings suggest a rationale for the further evolution of this promising series of antitubercular small molecules, which appear to share some similarities with the clinical compound PA-824 in terms of activation, while highlighting more general guidelines for the optimization of small-molecule antitubercular agents.

Data Mining and Computational Modeling of High-Throughput Screening Datasets

We are now seeing the benefit of investments made over the last decade in high-throughput screening (HTS) that is resulting in large structure activity datasets entering public and open databases such as ChEMBL and PubChem. The growth of academic HTS screening centers and the increasing move to academia for early stage drug discovery suggests a great need for the informatics tools and methods to mine such data and learn from it. Collaborative Drug Discovery, Inc. (CDD) has developed a number of tools for storing, mining, securely and selectively sharing, as well as learning from such HTS data. We present a new web based data mining and visualization module directly within the CDD Vault platform for high-throughput drug discovery data that makes use of a novel technology stack following modern reactive design principles. We also describe CDD Models within the CDD Vault platform that enables researchers to share models, share predictions from models, and create models from distributed, heterogeneous data. Our system is built on top of the Collaborative Drug Discovery Vault Activity and Registration data repository ecosystem which allows users to manipulate and visualize thousands of molecules in real time. This can be performed in any browser on any platform. In this chapter we present examples of its use with public datasets in CDD Vault. Such approaches can complement other cheminformatics tools, whether open source or commercial, in providing approaches for data mining and modeling of HTS data.

Comparison of Deep Learning With Multiple Machine Learning Methods and Metrics Using Diverse Drug Discovery Data Sets

Machine learning methods have been applied to many data sets in pharmaceutical research for several decades. The relative ease and availability of fingerprint type molecular descriptors paired with Bayesian methods resulted in the widespread use of this approach for a diverse array of end points relevant to drug discovery. Deep learning is the latest machine learning algorithm attracting attention for many of pharmaceutical applications from docking to virtual screening. Deep learning is based on an artificial neural network with multiple hidden layers and has found considerable traction for many artificial intelligence applications. We have previously suggested the need for a comparison of different machine learning methods with deep learning across an array of varying data sets that is applicable to pharmaceutical research. End points relevant to pharmaceutical research include absorption, distribution, metabolism, excretion, and toxicity (ADME/Tox) properties, as well as activity against pathogens and drug discovery data sets. In this study, we have used data sets for solubility, probe-likeness, hERG, KCNQ1, bubonic plague, Chagas, tuberculosis, and malaria to compare different machine learning methods using FCFP6 fingerprints. These data sets represent whole cell screens, individual proteins, physicochemical properties as well as a data set with a complex end point. Our aim was to assess whether deep learning offered any improvement in testing when assessed using an array of metrics including AUC, F1 score, Cohen's kappa, Matthews correlation coefficient and others. Based on ranked normalized scores for the metrics or data sets Deep Neural Networks (DNN) ranked higher than SVM, which in turn was ranked higher than all the other machine learning methods. Visualizing these properties for training and test sets using radar type plots indicates when models are inferior or perhaps over trained. These results also suggest the need for assessing deep learning further using multiple metrics with much larger scale comparisons, prospective testing as well as assessment of different fingerprints and DNN architectures beyond those used.

Addressing the Metabolic Stability of Antituberculars through Machine Learning

We present the first prospective application of our mouse liver microsomal (MLM) stability Bayesian model. CD117, an antitubercular thienopyrimidine tool compound that suffers from metabolic instability (MLM t_1/2 < 1 min), was utilized to assess the predictive power of our new MLM stability model. The S-substituent was removed, a set of commercial reagents was utilized to construct a virtual library of 411 analogues, and our MLM stability model was applied to prioritize 13 analogues for synthesis and biological profiling. In MLM stability assays, all 13 analogues had superior metabolic stability to the parent compound, and six new analogues had acceptable MLM t_1/2 values greater than or equal to 60 min. It is noteworthy that whole-cell efficacy and lack of relative mammalian cell cytotoxicity could not be predicted simultaneously. These results support the utility of our new MLM stability model in chemical tool and drug discovery optimization efforts.

Machine learning and docking models for Mycobacterium tuberculosis topoisomerase I

There is a shortage of compounds that are directed towards new targets apart from those targeted by the FDA approved drugs used against Mycobacterium tuberculosis. Topoisomerase I (Mttopo I) is an essential mycobacterial enzyme and a promising target in this regard. However, it suffers from a shortage of known inhibitors. We have previously used computational approaches such as homology modeling and docking to propose 38 FDA approved drugs for testing and identified several active molecules. To follow on from this, we now describe the in vitro testing of a library of 639 compounds. These data were used to create machine learning models for Mttopo I which were further validated. The combined Mttopo I Bayesian model had a 5 fold cross validation receiver operator characteristic of 0.74 and sensitivity, specificity and concordance values above 0.76 and was used to select commercially available compounds for testing in vitro. The recently described crystal structure of Mttopo I was also compared with the previously described homology model and then used to dock the Mttopo I actives norclomipramine and imipramine. In summary, we describe our efforts to identify small molecule inhibitors of Mttopo I using a combination of machine learning modeling and docking studies in conjunction with screening of the selected molecules for enzyme inhibition. We demonstrate the experimental inhibition of Mttopo I by small molecule inhibitors and show that the enzyme can be readily targeted for lead molecule development.

Collaborative drug discovery for More Medicines for Tuberculosis (MM4TB)

Neglected disease drug discovery is generally poorly funded compared with major diseases and hence there is an increasing focus on collaboration and precompetitive efforts such as public-private partnerships (PPPs). The More Medicines for Tuberculosis (MM4TB) project is one such collaboration funded by the EU with the goal of discovering new drugs for tuberculosis. Collaborative Drug Discovery has provided a commercial web-based platform called CDD Vault which is a hosted collaborative solution for securely sharing diverse chemistry and biology data. Using CDD Vault alongside other commercial and free cheminformatics tools has enabled support of this and other large collaborative projects, aiding drug discovery efforts and fostering collaboration. We will describe CDD's efforts in assisting with the MM4TB project.

Learning from the past for TB drug discovery in the future

Tuberculosis drug discovery has shifted in recent years from a primarily target-based approach to one that uses phenotypic high-throughput screens. As examples of this, through our EU-funded FP7 collaborations, New Medicines for Tuberculosis was target-based and our more-recent More Medicines for Tuberculosis project predominantly used phenotypic screening. From these projects we have examples of success (DprE1) and failure (PimA) going from drug to target and from target to drug, respectively. It is clear that we still have much to learn about the drug targets and the complex effects of the drugs on Mycobacterium tuberculosis. We propose a more integrated approach that learns from earlier drug discovery efforts that could help to move drug discovery forward.

The Next Era: Deep Learning in Pharmaceutical Research

Over the past decade we have witnessed the increasing sophistication of machine learning algorithms applied in daily use from internet searches, voice recognition, social network software to machine vision software in cameras, phones, robots and self-driving cars. Pharmaceutical research has also seen its fair share of machine learning developments. For example, applying such methods to mine the growing datasets that are created in drug discovery not only enables us to learn from the past but to predict a molecule's properties and behavior in future. The latest machine learning algorithm garnering significant attention is deep learning, which is an artificial neural network with multiple hidden layers. Publications over the last 3 years suggest that this algorithm may have advantages over previous machine learning methods and offer a slight but discernable edge in predictive performance. The time has come for a balanced review of this technique but also to apply machine learning methods such as deep learning across a wider array of endpoints relevant to pharmaceutical research for which the datasets are growing such as physicochemical property prediction, formulation prediction, absorption, distribution, metabolism, excretion and toxicity (ADME/Tox), target prediction and skin permeation, etc. We also show that there are many potential applications of deep learning beyond cheminformatics. It will be important to perform prospective testing (which has been carried out rarely to date) in order to convince skeptics that there will be benefits from investing in this technique.

Machine Learning Model Analysis and Data Visualization with Small Molecules Tested in a Mouse Model of Mycobacterium tuberculosis Infection (2014-2015)

The renewed urgency to develop new treatments for Mycobacterium tuberculosis (Mtb) infection has resulted in large-scale phenotypic screening and thousands of new active compounds in vitro. The next challenge is to identify candidates to pursue in a mouse in vivo efficacy model as a step to predicting clinical efficacy. We previously analyzed over 70 years of this mouse in vivo efficacy data, which we used to generate and validate machine learning models. Curation of 60 additional small molecules with in vivo data published in 2014 and 2015 was undertaken to further test these models. This represents a much larger test set than for the previous models. Several computational approaches have now been applied to analyze these molecules and compare their molecular properties beyond those attempted previously. Our previous machine learning models have been updated, and a novel aspect has been added in the form of mouse liver microsomal half-life (MLM t_1/2) and in vitro-based Mtb models incorporating cytotoxicity data that were used to predict in vivo activity for comparison. Our best Mtbin vivo models possess fivefold ROC values > 0.7, sensitivity > 80%, and concordance > 60%, while the best specificity value is >40%. Use of an MLM t_1/2 Bayesian model affords comparable results for scoring the 60 compounds tested. Combining MLM stability and in vitroMtb models in a novel consensus workflow in the best cases has a positive predicted value (hit rate) > 77%. Our results indicate that Bayesian models constructed with literature in vivoMtb data generated by different laboratories in various mouse models can have predictive value and may be used alongside MLM t_1/2 and in vitro-based Mtb models to assist in selecting antitubercular compounds with desirable in vivo efficacy. We demonstrate for the first time that consensus models of any kind can be used to predict in vivo activity for Mtb. In addition, we describe a new clustering method for data visualization and apply this to the in vivo training and test data, ultimately making the method accessible in a mobile app.

Open drug discovery for the Zika virus

The Zika virus (ZIKV) outbreak in the Americas has caused global concern that we may be on the brink of a healthcare crisis. The lack of research on ZIKV in the over 60 years that we have known about it has left us with little in the way of starting points for drug discovery. Our response can build on previous efforts with virus outbreaks and lean heavily on work done on other flaviviruses such as dengue virus. We provide some suggestions of what might be possible and propose an open drug discovery effort that mobilizes global science efforts and provides leadership, which thus far has been lacking. We also provide a listing of potential resources and molecules that could be prioritized for testing as in vitro assays for ZIKV are developed. We propose also that in order to incentivize drug discovery, a neglected disease priority review voucher should be available to those who successfully develop an FDA approved treatment. Learning from the response to the ZIKV, the approaches to drug discovery used and the success and failures will be critical for future infectious disease outbreaks.

Open Source Bayesian Models. 3. Composite Models for Prediction of Binned Responses

Bayesian models constructed from structure-derived fingerprints have been a popular and useful method for drug discovery research when applied to bioactivity measurements that can be effectively classified as active or inactive. The results can be used to rank candidate structures according to their probability of activity, and this ranking benefits from the high degree of interpretability when structure-based fingerprints are used, making the results chemically intuitive. Besides selecting an activity threshold, building a Bayesian model is fast and requires few or no parameters or user intervention. The method also does not suffer from such acute overtraining problems as quantitative structure–activity relationships or quantitative structure–property relationships (QSAR/QSPR). This makes it an approach highly suitable for automated workflows that are independent of user expertise or prior knowledge of the training data. We now describe a new method for creating a composite group of Bayesian models to extend the method to work with multiple states, rather than just binary. Incoming activities are divided into bins, each covering a mutually exclusive range of activities. For each of these bins, a Bayesian model is created to model whether or not the compound belongs in the bin. Analyzing putative molecules using the composite model involves making a prediction for each bin and examining the relative likelihood for each assignment, for example, highest value wins. The method has been evaluated on a collection of hundreds of data sets extracted from ChEMBL v20 and validated data sets for ADME/Tox and bioactivity.