Comparing and Validating Machine Learning Models for Mycobacterium tuberculosis Drug Discovery

Predicting the Hallucinogenic Potential of Molecules Using Artificial Intelligence

The development of new drugs addressing serious mental health and other disorders should avoid the psychedelic experience. Analogs of psychedelic drugs can have clinical utility and are termed "psychoplastogens". These represent promising candidates for treating opioid use disorder to reduce drug dependence, with rarely reported serious adverse effects. This drug abuse cessation is linked to the induction of neuritogenesis and increased neuroplasticity, a hallmark of psychedelic molecules, such as lysergic acid diethylamine. Some, but not all psychoplastogens may act through the G-protein coupled receptor (GPCR) 5HT2A whereas others may display very different polypharmacology making prediction of hallucinogenic potential challenging. In the process of developing tools to help design new psychoplastogens, we have used artificial intelligence in the form of machine learning classification models for predicting psychedelic effects using a published in vitro data set from PsychLight (support vector classification (SVC), area under the curve (AUC) 0.74) and in vivo human data derived from books from Shulgin and Shulgin (SVC, AUC, 0.72) with nested five-fold cross validation. We have also explored conformal predictors with ECFP6 and electrostatic descriptors in an effort to optimize them. These models have been used to predict known 5HT2A agonists to assess their potential to act as psychedelics and induce hallucinations for PsychLight (SVC, AUC 0.97) and Shulgin and Shulgin (random forest, AUC 0.71). We have tested these models with head twitch data from the mouse. This predictive capability is desirable to reliably design new psychoplastogens that lack in vivo hallucinogenic potential and help assess existing and future molecules for this potential. These efforts also provide useful insights into understanding the psychedelic structure activity relationship.

Near-Term Quantum Classification Algorithms Applied to Antimalarial Drug Discovery

Computational approaches are widely applied in drug discovery to explore properties related to bioactivity, physiochemistry, and toxicology. Over at least the last 20 years, the exploitation of machine learning on molecular data sets has been used to understand the structure-activity relationships that exist between biomolecules and druggable targets. More recently, these methods have also seen application for phenotypic screening data for neglected diseases such as tuberculosis and malaria. Herein, we apply machine learning to build quantum Quantitative Structure Activity Relationship models from antimalarial data sets. There is a continual need for new antimalarials to address drug resistance, and the readily available in vitro data sets could be utilized with newer machine learning approaches as these develop. Furthermore, quantum machine learning is a relatively new method that uses a quantum computer to perform the calculations. First, we present a classical-quantum hybrid computational approach by building a Latent Bernoulli Autoencoder machine learning model for compressing bit-vector descriptors to a size that can be adapted to quantum computers for classification tasks with limited loss of embedded information. Second, we apply our method for feature map compression to quantum classification algorithms, including a completely novel machine learning algorithm with no analogy in classical computers: the Quantum Fourier Transform Classifier. We apply both these approaches to build quantum machine learning models for small-molecule antimalarials with quantum simulation software and then benchmark these quantum models against classical machine learning approaches. While there are many challenges currently facing the development of reliable quantum computers, our results demonstrate that there is potential for the use of this technology in the field of drug discovery.

Computational drug repositioning identifies niclosamide and tribromsalan as inhibitors of Mycobacterium tuberculosis and Mycobacterium abscessus

Tuberculosis (TB) is still a major global health challenge, killing over 1.5 million people each year, and hence, there is a need to identify and develop novel treatments for Mycobacterium tuberculosis (M. tuberculosis). The prevalence of infections caused by nontuberculous mycobacteria (NTM) is also increasing and has overtaken TB cases in the United States and much of the developed world. Mycobacterium abscessus (M. abscessus) is one of the most frequently encountered NTM and is difficult to treat. We describe the use of drug-disease association using a semantic knowledge graph approach combined with machine learning models that has enabled the identification of several molecules for testing anti-mycobacterial activity. We established that niclosamide (M. tuberculosis IC90 2.95 μM; M. abscessus IC90 59.1 μM) and tribromsalan (M. tuberculosis IC90 76.92 μM; M. abscessus IC90 147.4 μM) inhibit M. tuberculosis and M. abscessus in vitro. To investigate the mode of action, we determined the transcriptional response of M. tuberculosis and M. abscessus to both compounds in axenic log phase, demonstrating a broad effect on gene expression that differed from known M. tuberculosis inhibitors. Both compounds elicited transcriptional responses indicative of respiratory pathway stress and the dysregulation of fatty acid metabolism.

Identification of New Modulators and Inhibitors of Palmitoyl-Protein Thioesterase 1 for CLN1 Batten Disease and Cancer

Palmitoyl-protein thioesterase 1 (PPT1) is an understudied enzyme that is gaining attention due to its role in the depalmitoylation of several proteins involved in neurodegenerative diseases and cancer. PPT1 is overexpressed in several cancers, specifically cholangiocarcinoma and esophageal cancers. Inhibitors of PPT1 lead to cell death and have been shown to enhance the killing of tumor cells alongside known chemotherapeutics. PPT1 is hence a viable target for anticancer drug development. Furthermore, mutations in PPT1 cause a lysosomal storage disorder called infantile neuronal ceroid lipofuscinosis (CLN1 disease). Molecules that can inhibit, stabilize, or modulate the activity of this target are needed to address these diseases. We used PPT1 enzymatic assays to identify molecules that were subsequently tested by using differential scanning fluorimetry and microscale thermophoresis. Selected compounds were also tested in neuroblastoma cell lines. The resulting PPT1 screening data was used for building machine learning models to help select additional compounds for testing. We discovered two of the most potent PPT1 inhibitors reported to date, orlistat (IC50 178.8 nM) and palmostatin B (IC50 11.8 nM). When tested in HepG2 cells, it was found that these molecules had decreased activity, indicating that they were likely not penetrating the cells. The combination of in vitro enzymatic and biophysical assays enabled the identification of several molecules that can bind or inhibit PPT1 and may aid in the discovery of modulators or chaperones. The molecules identified could be used as a starting point for further optimization as treatments for other potential therapeutic applications outside CLN1 disease, such as cancer and neurological diseases.

T6496 targeting EGFR mediated by T790M or C797S mutant: machine learning, virtual screening and bioactivity evaluation study

Acquired resistance to EGFR is a major impediment in lung cancer treatment, highlighting the urgent need to discover novel compounds to overcome EGFR drug resistance. In this study, we utilized in silico methods and bioactivity evaluation for drug discovery to identify novel active anticancer agents targeting EGFRT790M/L858R and EGFRT790M/C797S/L858R. Firstly, we employed ROC-guided machine learning to retrieve nearly 7,765 compounds from a collection of three libraries (comprising over 220,000 compounds). Next, virtual screening, cluster analysis, and binding model analysis were employed to identify six potential compounds. Additionally, the kinase assay revealed that these six compounds demonstrated higher sensitivity to EGFR than c-Met. Among these compounds, T6496 inhibited both EGFRT790M/L858R and EGFRT790M/C797S/L858R kinases, with an IC50 of 3.30 and 8.72 μM. Furthermore, we evaluated the antitumor effects of the six selected compounds, and compound T6496 exhibited the strongest anticancer activity against H1975 cell lines, with an IC50 value of 2.7 μM. These results suggest that T6496 may mitigate EGFR resistance caused by T790M or C797S mutations. Moreover, the AO staining assay, JC-1 staining, ROS experiment and hemolytic toxicity evaluation revealed that T6496 could induce apoptosis in H1975 cell lines in a time-dependent and concentration-dependent manner, and is a potential compound for further structural optimization.Communicated by Ramaswamy H. Sarma.

From immunology to artificial intelligence: revolutionizing latent tuberculosis infection diagnosis with machine learning

Latent tuberculosis infection (LTBI) has become a major source of active tuberculosis (ATB). Although the tuberculin skin test and interferon-gamma release assay can be used to diagnose LTBI, these methods can only differentiate infected individuals from healthy ones but cannot discriminate between LTBI and ATB. Thus, the diagnosis of LTBI faces many challenges, such as the lack of effective biomarkers from Mycobacterium tuberculosis (MTB) for distinguishing LTBI, the low diagnostic efficacy of biomarkers derived from the human host, and the absence of a gold standard to differentiate between LTBI and ATB. Sputum culture, as the gold standard for diagnosing tuberculosis, is time-consuming and cannot distinguish between ATB and LTBI. In this article, we review the pathogenesis of MTB and the immune mechanisms of the host in LTBI, including the innate and adaptive immune responses, multiple immune evasion mechanisms of MTB, and epigenetic regulation. Based on this knowledge, we summarize the current status and challenges in diagnosing LTBI and present the application of machine learning (ML) in LTBI diagnosis, as well as the advantages and limitations of ML in this context. Finally, we discuss the future development directions of ML applied to LTBI diagnosis.

High-Throughput Phenotypic Screening and Machine Learning Methods Enabled the Selection of Broad-Spectrum Low-Toxicity Antitrypanosomatidic Agents

Broad-spectrum anti-infective chemotherapy agents with activity against Trypanosomes, Leishmania, and Mycobacterium tuberculosis species were identified from a high-throughput phenotypic screening program of the 456 compounds belonging to the Ty-Box, an in-house industry database. Compound characterization using machine learning approaches enabled the identification and synthesis of 44 compounds with broad-spectrum antiparasitic activity and minimal toxicity against Trypanosoma brucei, Leishmania Infantum, and Trypanosoma cruzi. In vitro studies confirmed the predictive models identified in compound 40 which emerged as a new lead, featured by an innovative N-(5-pyrimidinyl)benzenesulfonamide scaffold and promising low micromolar activity against two parasites and low toxicity. Given the volume and complexity of data generated by the diverse high-throughput screening assays performed on the compounds of the Ty-Box library, the chemoinformatic and machine learning tools enabled the selection of compounds eligible for further evaluation of their biological and toxicological activities and aided in the decision-making process toward the design and optimization of the identified lead.

1-Sulfonyl-3-amino-1H-1,2,4-triazoles as Yellow Fever Virus Inhibitors: Synthesis and Structure-Activity Relationship

Yellow fever virus (YFV) transmitted by infected mosquitoes causes an acute viral disease for which there are no approved small-molecule therapeutics. Our recently developed machine learning models for YFV inhibitors led to the selection of a new pyrazolesulfonamide derivative RCB16003 with acceptable in vitro activity. We report that the N-phenyl-1-(phenylsulfonyl)-1H-1,2,4-triazol-3-amine class, which was recently identified as active non-nucleoside reverse transcriptase inhibitors against HIV-1, can also be repositioned as inhibitors of yellow fever virus replication. As compared to other Flaviviridae or Togaviridae family viruses tested, both compounds RCB16003 and RCB16007 demonstrate selectivity for YFV over related viruses, with only RCB16007 showing some inhibition of the West Nile virus (EC50 7.9 μM, CC50 17 μM, SI 2.2). We also describe the absorption, distribution, metabolism, and excretion (ADME) in vitro and pharmacokinetics (PK) for RCB16007 in mice. This compound had previously been shown to not inhibit hERG, and we now describe that it has good metabolic stability in mouse and human liver microsomes, low levels of CYP inhibition, high protein binding, and no indication of efflux in Caco-2 cells. A single-dose oral PK study in mice has a T1/2 of 3.4 h and Cmax of 1190 ng/mL, suggesting good availability and stability. We now propose that the N-phenyl-1-(phenylsulfonyl)-1H-1,2,4-triazol-3-amine class may be prioritized for in vivo efficacy testing against YFV.

Antiviral Evaluation of Dispirotripiperazines against Hepatitis B Virus

Hepatitis B virus (HBV) is a hepatotropic DNA virus that replicates by reverse transcription. It chronically infects >296 million people worldwide, including ∼850,000 in the USA, and kills 820,000 annually worldwide. Current nucleos(t)ide analogue (NA) or pegylated interferon α therapies do not eradicate the virus and would benefit from a complementary antiviral drug. We performed a preliminary screen of 28 dispirotripiperazines against HBV, identifying 9 hits with EC50 of 0.7-25 μM. Compound 11826096 displays the most potent activity and represents a promising lead for future optimization. While the mechanism of action is unknown, preliminary assays limit possible targets to activities involved in RNA accumulation, translation, capsid assembly, and/or capsid stability. In addition, we built machine learning models to determine if they were able to predict the activity of this series of compounds. The novelty of these molecules indicated they were outside of the applicability domain of these models.

Machine Learning Models Identify New Inhibitors for Human OATP1B1

The uptake transporter OATP1B1 (SLC01B1) is largely localized to the sinusoidal membrane of hepatocytes and is a known victim of unwanted drug-drug interactions. Computational models are useful for identifying potential substrates and/or inhibitors of clinically relevant transporters. Our goal was to generate OATP1B1 in vitro inhibition data for [3H] estrone-3-sulfate (E3S) transport in CHO cells and use it to build machine learning models to facilitate a comparison of seven different classification models (Deep learning, Adaboosted decision trees, Bernoulli naïve bayes, k-nearest neighbors (knn), random forest, support vector classifier (SVC), logistic regression (lreg), and XGBoost (xgb)] using ECFP6 fingerprints to perform 5-fold, nested cross validation. In addition, we compared models using 3D pharmacophores, simple chemical descriptors alone or plus ECFP6, as well as ECFP4 and ECFP8 fingerprints. Several machine learning algorithms (SVC, lreg, xgb, and knn) had excellent nested cross validation statistics, particularly for accuracy, AUC, and specificity. An external test set containing 207 unique compounds not in the training set demonstrated that at every threshold SVC outperformed the other algorithms based on a rank normalized score. A prospective validation test set was chosen using prediction scores from the SVC models with ECFP fingerprints and were tested in vitro with 15 of 19 compounds (84% accuracy) predicted as active (≥20% inhibition) showed inhibition. Of these compounds, six (abamectin, asiaticoside, berbamine, doramectin, mobocertinib, and umbralisib) appear to be novel inhibitors of OATP1B1 not previously reported. These validated machine learning models can now be used to make predictions for drug-drug interactions for human OATP1B1 alongside other machine learning models for important drug transporters in our MegaTrans software.

A machine learning-based diagnostic model for children with autism spectrum disorders complicated with intellectual disability

Background

Early detection of children with autism spectrum disorder (ASD) and comorbid intellectual disability (ID) can help in individualized intervention. Appropriate assessment and diagnostic tools are lacking in primary care. This study aims to explore the applicability of machine learning (ML) methods in diagnosing ASD comorbid ID compared with traditional regression models.

Method

From January 2017 to December 2021, 241 children with ASD, with an average age of 6.41 ± 1.96, diagnosed in the Developmental Behavior Department of the Children's Hospital Affiliated with the Medical College of Zhejiang University were included in the analysis. This study trained the traditional diagnostic models of Logistic regression (LR), Support Vector Machine (SVM), and two ensemble learning algorithms [Random Forest (RF) and XGBoost]. Socio-demographic and behavioral observation data were used to distinguish whether autistic children had combined ID. The hyperparameters adjustment uses grid search and 10-fold validation. The Boruta method is used to select variables. The model's performance was evaluated using discrimination, calibration, and decision curve analysis (DCA).

Result

Among 241 autistic children, 98 (40.66%) were ASD comorbid ID. The four diagnostic models can better distinguish whether autistic children are complicated with ID, and the accuracy of SVM is the highest (0.836); SVM and XGBoost have better accuracy (0.800, 0.838); LR has the best sensitivity (0.939), followed by SVM (0.952). Regarding specificity, SVM, RF, and XGBoost performed significantly higher than LR (0.355). The AUC of ML (SVM, 0.835 [95% CI: 0.747-0.944]; RF, 0.829 [95% CI: 0.738-0.920]; XGBoost, 0.845 [95% CI: 0.734-0.937]) is not different from traditional LR (0.858 [95% CI: 0.770-0.944]). Only SVM observed a good calibration degree. Regarding DCA, LR, and SVM have higher benefits in a wider threshold range.

Conclusion

Compared to the traditional regression model, ML model based on socio-demographic and behavioral observation data, especially SVM, has a better ability to distinguish whether autistic children are combined with ID.

Machine Learning for Discovery of New ADORA Modulators

Adenosine (ADO) is an extracellular signaling molecule generated locally under conditions that produce ischemia, hypoxia, or inflammation. It is involved in modulating a range of physiological functions throughout the brain and periphery through the membrane-bound G protein-coupled receptors, called adenosine receptors (ARs) A₁AR, A_2AAR, A_2BAR, and A₃AR. These are therefore important targets for neurological, cardiovascular, inflammatory, and autoimmune diseases and are the subject of drug development directed toward the cyclic adenosine monophosphate and other signaling pathways. Initially using public data for A₁AR agonists we generated and validated a Bayesian machine learning model (Receiver Operator Characteristic of 0.87) that we used to identify molecules for testing. Three selected molecules, crisaborole, febuxostat and paroxetine, showed initial activity in vitro using the HEK293 A₁AR Nomad cell line. However, radioligand binding, β-arrestin assay and calcium influx assay did not confirm this A₁AR activity. Nevertheless, several other AR activities were identified. Febuxostat and paroxetine both inhibited orthosteric radioligand binding in the µM range for A_2AAR and A₃AR. In HEK293 cells expressing the human A_2AAR, stimulation of cAMP was observed for crisaborole (EC₅₀ 2.8 µM) and paroxetine (EC₅₀ 14 µM), but not for febuxostat. Crisaborole also increased cAMP accumulation in A_2BAR-expressing HEK293 cells, but it was weaker than at the A_2AAR. At the human A₃AR, paroxetine did not show any agonist activity at 100 µM, although it displayed binding with a K_i value of 14.5 µM, suggesting antagonist activity. We have now identified novel modulators of A_2AAR, A_2BAR and A₃AR subtypes that are clinically used for other therapeutic indications, and which are structurally distinct from previously reported tool compounds or drugs.

Chalcones from Angelica keiskei (ashitaba) inhibit key Zika virus replication proteins

Zika virus (ZIKV) is a dangerous human pathogen and no antiviral drugs have been approved to date. The chalcones are a group of small molecules that are found in a number of different plants, including Angelica keiskei Koidzumi, also known as ashitaba. To examine chalcone anti-ZIKV activity, three chalcones, 4-hydroxyderricin (4HD), xanthoangelol (XA), and xanthoangelol-E (XA-E), were purified from a methanol-ethyl acetate extract from A. keiskei. Molecular and ensemble docking predicted that these chalcones would establish multiple interactions with residues in the catalytic and allosteric sites of ZIKV NS2B-NS3 protease, and in the allosteric site of the NS5 RNA-dependent RNA-polymerase (RdRp). Machine learning models also predicted 4HD, XA and XA-E as potential anti-ZIKV inhibitors. Enzymatic and kinetic assays confirmed chalcone inhibition of the ZIKV NS2B-NS3 protease allosteric site with IC₅₀s from 18 to 50 µM. Activity assays also revealed that XA, but not 4HD or XA-E, inhibited the allosteric site of the RdRp, with an IC₅₀ of 6.9 µM. Finally, we tested these chalcones for their anti-viral activity in vitro with Vero cells. 4HD and XA-E displayed anti-ZIKV activity with EC₅₀ values of 6.6 and 22.0 µM, respectively, while XA displayed relatively weak anti-ZIKV activity with whole cells. With their simple structures and relative ease of modification, the chalcones represent attractive candidates for hit-to-lead optimization in the search of new anti-ZIKV therapeutics.

Machine Learning Models for Mycobacterium tuberculosisIn Vitro Activity: Prediction and Target Visualization

Tuberculosis (TB) is a major global health challenge, with approximately 1.4 million deaths per year. There is still a need to develop novel treatments for patients infected with Mycobacterium tuberculosis (Mtb). There have been many large-scale phenotypic screens that have led to the identification of thousands of new compounds. Yet, there is very limited investment in TB drug discovery which points to the need for new methods to increase the efficiency of drug discovery against Mtb. We have used machine learning approaches to learn from the public Mtb data, resulting in many data sets and models with robust enrichment and hit rates leading to the discovery of new active compounds. Recently, we have curated predominantly small-molecule Mtb data and developed new machine learning classification models with 18 886 molecules at different activity cutoffs. We now describe the further validation of these Bayesian models using a library of over 1000 molecules synthesized as part of EU-funded New Medicines for TB and More Medicines for TB programs. We highlight molecular features which are enriched in these active compounds. In addition, we provide new regression and classification models that can be used for scoring compound libraries or used to design new molecules. We have also visualized these molecules in the context of known molecular targets and identified clusters in chemical property space, which may aid in future target identification efforts. Finally, we are also making these data sets publicly available, representing a significant increase to the available Mtb inhibition data in the public domain.

Tuberculosis drug resistance profiling based on machine learning: A literature review

Tuberculosis (TB), caused by Mycobacterium tuberculosis (MTB), is one of the top 10 causes of death worldwide. Drug-resistant tuberculosis (DR-TB) poses a major threat to the World Health Organization's "End TB" strategy which has defined its target as the year 2035. In 2019, there were close to 0.5 million cases of DRTB, of which 78% were resistant to multiple TB drugs. The traditional culture-based drug susceptibility test (DST - the current gold standard) often takes multiple weeks and the necessary laboratory facilities are not readily available in low-income countries. Whole genome sequencing (WGS) technology is rapidly becoming an important tool in clinical and research applications including transmission detection or prediction of DR-TB. For the latter, many tools have recently been developed using curated database(s) of known resistance conferring mutations. However, documenting all the mutations and their effect is a time-taking and a continuous process and therefore Machine Learning (ML) techniques can be useful for predicting the presence of DR-TB based on WGS data. This can pave the way to an earlier detection of drug resistance and consequently more efficient treatment when compared to the traditional DST.

Mycobacterium abscessus drug discovery using machine learning

The prevalence of infections by nontuberculous mycobacteria is increasing, having surpassed tuberculosis in the United States and much of the developed world. Nontuberculous mycobacteria occur naturally in the environment and are a significant problem for patients with underlying lung diseases such as bronchiectasis, chronic obstructive pulmonary disease, and cystic fibrosis. Current treatment regimens are lengthy, complicated, toxic and they are often unsuccessful as seen by disease recurrence. Mycobacterium abscessus is one of the most commonly encountered organisms in nontuberculous mycobacteria disease and it is the most difficult to eradicate. There is currently no systematically proven regimen that is effective for treating M. abscessus infections. Our approach to drug discovery integrates machine learning, medicinal chemistry and in vitro testing and has been previously applied to Mycobacterium tuberculosis. We have now identified several novel 1-(phenylsulfonyl)-1H-benzimidazol-2-amines that have weak activity on M. abscessus in vitro but may represent a starting point for future further medicinal chemistry optimization. We also address limitations still to be overcome with the machine learning approach for M. abscessus.

UV-adVISor: Attention-Based Recurrent Neural Networks to Predict UV-Vis Spectra

Ultraviolet-visible (UV-Vis) absorption spectra are routinely collected as part of high-performance liquid chromatography (HPLC) analysis systems and can be used to identify chemical reaction products by comparison to the reference spectra. Here, we present UV-adVISor as a new computational tool for predicting the UV-Vis spectra from a molecule's structure alone. UV-Vis prediction was approached as a sequence-to-sequence problem. We utilized Long-Short Term Memory and attention-based neural networks with Extended Connectivity Fingerprint Diameter 6 or molecule SMILES to generate predictive models for the UV spectra. We have produced two spectrum datasets (dataset I, N = 949, and dataset II, N = 2222) using different compound collections and spectrum acquisition methods to train, validate, and test our models. We evaluated the prediction accuracy of the complete spectra by the correspondence of wavelengths of absorbance maxima and with a series of statistical measures (the best test set median model parameters are in parentheses for model II), including RMSE (0.064), R² (0.71), and dynamic time warping (DTW, 0.194) of the entire spectrum curve. Scrambling molecule structures with the experimental spectra during training resulted in a degraded R², confirming the utility of the approaches for prediction. UV-adVISor is able to provide fast and accurate predictions for libraries of compounds.

Automated machine learning for endemic active tuberculosis prediction from multiplex serological data

Serological diagnosis of active tuberculosis (TB) is enhanced by detection of multiple antibodies due to variable immune responses among patients. Clinical interpretation of these complex datasets requires development of suitable algorithms, a time consuming and tedious undertaking addressed by the automated machine learning platform MILO (Machine Intelligence Learning Optimizer). MILO seamlessly integrates data processing, feature selection, model training, and model validation to simultaneously generate and evaluate thousands of models. These models were then further tested for generalizability on out-of-sample secondary and tertiary datasets. Out of 31 antigens evaluated, a 23-antigen model was the most robust on both the secondary dataset (TB vs healthy) and the tertiary dataset (TB vs COPD) with sensitivity of 90.5% and respective specificities of 100.0% and 74.6%. MILO represents a user-friendly, end-to-end solution for automated generation and deployment of optimized models, ideal for applications where rapid clinical implementation is critical such as emerging infectious diseases.

Machine Learning Models Identify Inhibitors of SARS-CoV-2

With the rapidly evolving SARS-CoV-2 variants of concern, there is an urgent need for the discovery of further treatments for the coronavirus disease (COVID-19). Drug repurposing is one of the most rapid strategies for addressing this need, and numerous compounds have already been selected for in vitro testing by several groups. These have led to a growing database of molecules with in vitro activity against the virus. Machine learning models can assist drug discovery through prediction of the best compounds based on previously published data. Herein, we have implemented several machine learning methods to develop predictive models from recent SARS-CoV-2 in vitro inhibition data and used them to prioritize additional FDA-approved compounds for in vitro testing selected from our in-house compound library. From the compounds predicted with a Bayesian machine learning model, lumefantrine, an antimalarial was selected for testing and showed limited antiviral activity in cell-based assays while demonstrating binding (K_d 259 nM) to the spike protein using microscale thermophoresis. Several other compounds which we prioritized have since been tested by others and were also found to be active in vitro. This combined machine learning and in vitro testing approach can be expanded to virtually screen available molecules with predicted activity against SARS-CoV-2 reference WIV04 strain and circulating variants of concern. In the process of this work, we have created multiple iterations of machine learning models that can be used as a prioritization tool for SARS-CoV-2 antiviral drug discovery programs. The very latest model for SARS-CoV-2 with over 500 compounds is now freely available at www.assaycentral.org.

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.

Recent advances in drug repurposing using machine learning

Drug repurposing aims to find new uses for already existing and approved drugs. We now provide a brief overview of recent developments in drug repurposing using machine learning alongside other computational approaches for comparison. We also highlight several applications for cancer using kinase inhibitors, Alzheimer's disease as well as COVID-19.

Quantum Machine Learning Algorithms for Drug Discovery Applications

The growing quantity of public and private data sets focused on small molecules screened against biological targets or whole organisms provides a wealth of drug discovery relevant data. This is matched by the availability of machine learning algorithms such as Support Vector Machines (SVM) and Deep Neural Networks (DNN) that are computationally expensive to perform on very large data sets with thousands of molecular descriptors. Quantum computer (QC) algorithms have been proposed to offer an approach to accelerate quantum machine learning over classical computer (CC) algorithms, however with significant limitations. In the case of cheminformatics, which is widely used in drug discovery, one of the challenges to overcome is the need for compression of large numbers of molecular descriptors for use on a QC. Here, we show how to achieve compression with data sets using hundreds of molecules (SARS-CoV-2) to hundreds of thousands of molecules (whole cell screening data sets for plague and M. tuberculosis) with SVM and the data reuploading classifier (a DNN equivalent algorithm) on a QC benchmarked against CC and hybrid approaches. This study illustrates the steps needed in order to be "quantum computer ready" in order to apply quantum computing to drug discovery and to provide the foundation on which to build this field.

Structure-based in silico approaches for drug discovery against Mycobacterium tuberculosis

Mycobacterium tuberculosis is the causative agent of TB and was estimated to cause 1.4 million death in 2019, alongside 10 million new infections. Drug resistance is a growing issue, with multi-drug resistant infections representing 3.3% of all new infections, hence novel antimycobacterial drugs are urgently required to combat this growing health emergency. Alongside this, increased knowledge of gene essentiality in the pathogenic organism and larger compound databases can aid in the discovery of new drug compounds. The number of protein structures, X-ray based and modelled, is increasing and now accounts for greater than > 80% of all predicted M. tuberculosis proteins; allowing novel targets to be investigated. This review will focus on structure-based in silico approaches for drug discovery, covering a range of complexities and computational demands, with associated antimycobacterial examples. This includes molecular docking, molecular dynamic simulations, ensemble docking and free energy calculations. Applications of machine learning onto each of these approaches will be discussed. The need for experimental validation of computational hits is an essential component, which is unfortunately missing from many current studies. The future outlooks of these approaches will also be discussed.

Comparing the Pfizer Central Nervous System Multiparameter Optimization Calculator and a BBB Machine Learning Model

The ability to calculate whether small molecules will cross the blood-brain barrier (BBB) is an important task for companies working in neuroscience drug discovery. For a decade, scientists have relied on relatively simplistic rules such as Pfizer's central nervous system multiparameter optimization models (CNS-MPO) for guidance during the drug selection process. In parallel, there has been a continued development of more sophisticated machine learning models that utilize different molecular descriptors and algorithms; however, these models represent a "black box" and are generally less interpretable. In both cases, these methods predict the ability of small molecules to cross the BBB using the molecular structure information on its own without in vitro or in vivo data. We describe here the implementation of two versions of Pfizer's algorithm (Pf-MPO.v1 and Pf-MPO.v2) and compare it with a Bayesian machine learning model of BBB penetration trained on a data set of 2296 active and inactive compounds using extended connectivity fingerprint descriptors. The predictive ability of these approaches was compared with 40 known CNS active drugs initially used by Pfizer as their positive set for validation of the Pf-MPO.v1 score. 37/40 (92.5%) compounds were predicted as active by the Bayesian model, while only 30/40 (75%) received a desirable Pf-MPO.v1 score ≥4 and 33/40 (82.5%) received a desirable Pf-MPO.v2 score ≥4, suggesting the Bayesian model is more accurate than MPO algorithms. This also indicates machine learning models are more flexible and have better predictive power for BBB penetration than simple rule sets that require multiple, accurate descriptor calculations. Our machine learning model statistics are comparable to recent published studies. We describe the implications of these findings and how machine learning may have a role alongside more interpretable methods.

The Role of Machine Learning in Centralized Authorization Process of Nanomedicines in European Union

BACKGROUND

Machine Learning (ML) has experienced an increasing use, given the possibilities to expand the scientific knowledge of different disciplines, such as nanotechnology. This has allowed the creation of Cheminformatic models capable of predicting biological activity and physicochemical characteristics of new components with high success rates in training and test partitions. Given the current gaps of scientific knowledge and the need for efficient application of medicines products law, this paper analyzes the position of regulators for marketing medicinal nanoproducts in the European Union and the role of ML in the authorization process.

METHODS

In terms of methodology, a dogmatic study of the European regulation and the guidance of the European Medicine Agency on the use of predictive models for nanomaterials was carried out. The study has, as the framework of reference, the European Regulation 726/2004 and has focused on the analysis of how ML processes are contemplated in the regulations.

RESULTS

As a result, we present a discussion of the information that must be provided for every case for simulation methods. The results show a favorable and flexible position for the development of the use of predictive models to complement the applicant's information.

CONCLUSION

It is concluded that Machine Learning has the capacity to help improve the application of nanotechnology medicine products regulation. Future regulations should promote this kind of information given the advanced state of the art in terms of algorithms that are able to build accurate predictive models. This especially applies to methods, such as Perturbation Theory Machine Learning (PTML), given that it is aligned with principles promoted by the standards of Organization for Economic Co-operation and Development (OECD), European Union regulations, and European Authority Medicine. To our best knowledge, this is the first study focused on nanotechnology medicine products and machine learning used to support technical European public assessment reports (EPAR) for complementary information.

Identification of active molecules against Mycobacterium tuberculosis through machine learning

Tuberculosis (TB) is an infectious disease caused by Mycobacterium tuberculosis (Mtb) and it has been one of the top 10 causes of death globally. Drug-resistant tuberculosis (XDR-TB), extensively resistant to the commonly used first-line drugs, has emerged as a major challenge to TB treatment. Hence, it is quite necessary to discover novel drug candidates for TB treatment. In this study, based on different types of molecular representations, four machine learning (ML) algorithms, including support vector machine, random forest (RF), extreme gradient boosting (XGBoost) and deep neural networks (DNN), were used to develop classification models to distinguish Mtb inhibitors from noninhibitors. The results demonstrate that the XGBoost model exhibits the best prediction performance. Then, two consensus strategies were employed to integrate the predictions from multiple models. The evaluation results illustrate that the consensus model by stacking the RF, XGBoost and DNN predictions offers the best predictions with area under the receiver operating characteristic curve of 0.842 and 0.942 for the 10-fold cross-validated training set and external test set, respectively. Besides, the association between the important descriptors and the bioactivities of molecules was interpreted by using the Shapley additive explanations method. Finally, an online webserver called ChemTB (http://cadd.zju.edu.cn/chemtb/) was developed, and it offers a freely available computational tool to detect potential Mtb inhibitors.

Use of Artificial Intelligence and Machine Learning for Discovery of Drugs for Neglected Tropical Diseases

Neglected tropical diseases continue to create high levels of morbidity and mortality in a sizeable fraction of the world’s population, despite ongoing research into new treatments. Some of the most important technological developments that have accelerated drug discovery for diseases of affluent countries have not flowed down to neglected tropical disease drug discovery. Pharmaceutical development business models, cost of developing new drug treatments and subsequent costs to patients, and accessibility of technologies to scientists in most of the affected countries are some of the reasons for this low uptake and slow development relative to that for common diseases in developed countries. Computational methods are starting to make significant inroads into discovery of drugs for neglected tropical diseases due to the increasing availability of large databases that can be used to train ML models, increasing accuracy of these methods, lower entry barrier for researchers, and widespread availability of public domain machine learning codes. Here, the application of artificial intelligence, largely the subset called machine learning, to modelling and prediction of biological activities and discovery of new drugs for neglected tropical diseases is summarized. The pathways for the development of machine learning methods in the short to medium term and the use of other artificial intelligence methods for drug discovery is discussed. The current roadblocks to, and likely impacts of, synergistic new technological developments on the use of ML methods for neglected tropical disease drug discovery in the future are also discussed.

Flavonoids from Pterogyne nitens as Zika virus NS2B-NS3 protease inhibitors

Although the widespread epidemic of Zika virus (ZIKV) and its neurological complications are well-known there are still no approved drugs available to treat this arboviral disease or vaccine to prevent the infection. Flavonoids from Pterogyne nitens have already demonstrated anti-flavivirus activity, although their target is unknown. In this study, we virtually screened an in-house database of 150 natural and semi-synthetic compounds against ZIKV NS2B-NS3 protease (NS2B-NS3p) using docking-based virtual screening, as part of the OpenZika project. As a result, we prioritized three flavonoids from P. nitens, quercetin, rutin and pedalitin, for experimental evaluation. We also used machine learning models, built with Assay Central® software, for predicting the activity and toxicity of these flavonoids. Biophysical and enzymatic assays generally agreed with the in silico predictions, confirming that the flavonoids inhibited ZIKV protease. The most promising hit, pedalitin, inhibited ZIKV NS2B-NS3p with an IC50 of 5 μM. In cell-based assays, pedalitin displayed significant activity at 250 and 500 µM, with slight toxicity in Vero cells. The results presented here demonstrate the potential of pedalitin as a candidate for hit-to-lead (H2L) optimization studies towards the discovery of antiviral drug candidates to treat ZIKV infections.

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.

Using Bibliometric Analysis and Machine Learning to Identify Compounds Binding to Sialidase-1

Rare diseases impact hundreds of millions of individuals worldwide. However, few therapies exist to treat the rare disease population because financial resources are limited, the number of patients affected is low, bioactivity data is often nonexistent, and very few animal models exist to support preclinical development efforts. Sialidosis is an ultrarare lysosomal storage disorder in which mutations in the NEU1 gene result in the deficiency of the lysosomal enzyme sialidase-1. This enzyme catalyzes the removal of sialic acid moieties from glycoproteins and glycolipids. Therefore, the defective or deficient protein leads to the buildup of sialylated glycoproteins as well as several characteristic symptoms of sialidosis including visual impairment, ataxia, hepatomegaly, dysostosis multiplex, and developmental delay. In this study, we used a bibliometric tool to generate links between lysosomal storage disease (LSD) targets and existing bioactivity data that could be curated in order to build machine learning models and screen compounds in silico. We focused on sialidase as an example, and we used the data curated from the literature to build a Bayesian model which was then used to score compound libraries and rank these molecules for in vitro testing. Two compounds were identified from in vitro testing using microscale thermophoresis, namely sulfameter (K d 2.15 ± 1.02 μM) and mexenone (K d 8.88 ± 4.02 μM), which validated our approach to identifying new molecules binding to this protein, which could represent possible drug candidates that can be evaluated further as potential chaperones for this ultrarare lysosomal disease for which there is currently no treatment. Combining bibliometric and machine learning approaches has the ability to assist in curating small molecule data and model building, respectively, for rare disease drug discovery. This approach also has the capability to identify new compounds that are potential drug candidates.

The Antiviral Drug Tilorone Is a Potent and Selective Inhibitor of Acetylcholinesterase

Acetylcholinesterase (AChE) is an important drug target in neurological disorders like Alzheimer's disease, Lewy body dementia, and Parkinson's disease dementia as well as for other conditions like myasthenia gravis and anticholinergic poisoning. In this study, we have used a combination of high-throughput screening, machine learning, and docking to identify new inhibitors of this enzyme. Bayesian machine learning models were generated with literature data from ChEMBL for eel and human AChE inhibitors as well as butyrylcholinesterase inhibitors (BuChE) and compared with other machine learning methods. High-throughput screens for the eel AChE inhibitor model identified several molecules including tilorone, an antiviral drug that is well-established outside of the United States, as a newly identified nanomolar AChE inhibitor. We have described how tilorone inhibits both eel and human AChE with IC₅₀'s of 14.4 nM and 64.4 nM, respectively, but does not inhibit the closely related BuChE IC₅₀ > 50 μM. We have docked tilorone into the human AChE crystal structure and shown that this selectivity is likely due to the reliance on a specific interaction with a hydrophobic residue in the peripheral anionic site of AChE that is absent in BuChE. We also conducted a pharmacological safety profile (SafetyScreen44) and kinase selectivity screen (SelectScreen) that showed tilorone (1 μM) only inhibited AChE out of 44 toxicology target proteins evaluated and did not appreciably inhibit any of the 485 kinases tested. This study suggests there may be a potential role for repurposing tilorone or its derivatives in conditions that benefit from AChE inhibition.

Bioactivity Comparison across Multiple Machine Learning Algorithms Using over 5000 Datasets for Drug Discovery

Machine learning methods are attracting considerable attention from the pharmaceutical industry for use in drug discovery and applications beyond. In recent studies, we and others have applied multiple machine learning algorithms and modeling metrics and, in some cases, compared molecular descriptors to build models for individual targets or properties on a relatively small scale. Several research groups have used large numbers of datasets from public databases such as ChEMBL in order to evaluate machine learning methods of interest to them. The largest of these types of studies used on the order of 1400 datasets. We have now extracted well over 5000 datasets from CHEMBL for use with the ECFP6 fingerprint and in comparison of our proprietary software Assay Central with random forest, k-nearest neighbors, support vector classification, naïve Bayesian, AdaBoosted decision trees, and deep neural networks (three layers). Model performance was assessed using an array of fivefold cross-validation metrics including area-under-the-curve, F1 score, Cohen's kappa, and Matthews correlation coefficient. Based on ranked normalized scores for the metrics or datasets, all methods appeared comparable, while the distance from the top indicated that Assay Central and support vector classification were comparable. Unlike prior studies which have placed considerable emphasis on deep neural networks (deep learning), no advantage was seen in this case. If anything, Assay Central may have been at a slight advantage as the activity cutoff for each of the over 5000 datasets representing over 570,000 unique compounds was based on Assay Central performance, although support vector classification seems to be a strong competitor. We also applied Assay Central to perform prospective predictions for the toxicity targets PXR and hERG to further validate these models. This work appears to be the largest scale comparison of these machine learning algorithms to date. Future studies will likely evaluate additional databases, descriptors, and machine learning algorithms and further refine the methods for evaluating and comparing such models.

Comparing Machine Learning Models for Aromatase (P450 19A1)

Aromatase, or cytochrome P450 19A1, catalyzes the aromatization of androgens to estrogens within the body. Changes in the activity of this enzyme can produce hormonal imbalances that can be detrimental to sexual and skeletal development. Inhibition of this enzyme can occur with drugs and natural products as well as environmental chemicals. Therefore, predicting potential endocrine disruption via exogenous chemicals requires that aromatase inhibition be considered in addition to androgen and estrogen pathway interference. Bayesian machine learning methods can be used for prospective prediction from the molecular structure without the need for experimental data. Herein, the generation and evaluation of multiple machine learning models utilizing different sources of aromatase inhibition data are described. These models are applied to two test sets for external validation with molecules relevant to drug discovery from the public domain. In addition, the performance of multiple machine learning algorithms was evaluated by comparing internal five-fold cross-validation statistics of the training data. These methods to predict aromatase inhibition from molecular structure, when used in concert with estrogen and androgen machine learning models, allow for a more holistic assessment of endocrine-disrupting potential of chemicals with limited empirical data and enable the reduction of the use of hazardous substances.

Predicting Drug Interactions with Human Equilibrative Nucleoside Transporters 1 and 2 Using Functional Knockout Cell Lines and Bayesian Modeling

Equilibrative nucleoside transporters (ENTs) 1 and 2 facilitate nucleoside transport across the blood-testis barrier (BTB). Improving drug entry into the testes with drugs that use endogenous transport pathways may lead to more effective treatments for diseases within the reproductive tract. In this study, CRISPR/CRISPR-associated protein 9 was used to generate HeLa cell lines in which ENT expression was limited to ENT1 or ENT2. We characterized uridine transport in these cell lines and generated Bayesian models to predict interactions with the ENTs. Quantification of [³H]uridine uptake in the presence of the ENT-specific inhibitor S-(4-nitrobenzyl)-6-thioinosine (NBMPR) demonstrated functional loss of each transporter. Nine nucleoside reverse-transcriptase inhibitors and 37 nucleoside/heterocycle analogs were evaluated to identify ENT interactions. Twenty-one compounds inhibited uridine uptake and abacavir, nevirapine, ticagrelor, and uridine triacetate had different IC₅₀ values for ENT1 and ENT2. Total accumulation of four identified inhibitors was measured with and without NBMPR to determine whether there was ENT-mediated transport. Clofarabine and cladribine were ENT1 and ENT2 substrates, whereas nevirapine and lexibulin were ENT1 and ENT2 nontransported inhibitors. Bayesian models generated using Assay Central machine learning software yielded reasonably high internal validation performance (receiver operator characteristic > 0.7). ENT1 IC₅₀-based models were generated from ChEMBL; subvalidations using this training data set correctly predicted 58% of inhibitors when analyzing activity by percent uptake and 63% when using estimated-IC₅₀ values. Determining drug interactions with these transporters can be useful in identifying and predicting compounds that are ENT1 and ENT2 substrates and can thereby circumvent the BTB through this transepithelial transport pathway in Sertoli cells. SIGNIFICANCE STATEMENT: This study is the first to predict drug interactions with equilibrative nucleoside transporter (ENT) 1 and ENT2 using Bayesian modeling. Novel CRISPR/CRISPR-associated protein 9 functional knockouts of ENT1 and ENT2 in HeLa S3 cells were generated and characterized. Determining drug interactions with these transporters can be useful in identifying and predicting compounds that are ENT1 and ENT2 substrates and can circumvent the blood-testis barrier through this transepithelial transport pathway in Sertoli cells.

Computational Approaches to Identify Molecules Binding to Mycobacterium tuberculosis KasA

Tuberculosis is caused by Mycobacterium tuberculosis (Mtb) and is a deadly disease resulting in the deaths of approximately 1.5 million people with 10 million infections reported in 2018. Recently, a key condensation step in the synthesis of mycolic acids was shown to require β-ketoacyl-ACP synthase (KasA). A crystal structure of KasA with the small molecule DG167 was recently described, which provided a starting point for using computational structure-based approaches to identify additional molecules binding to this protein. We now describe structure-based pharmacophores, docking and machine learning studies with Assay Central as a computational tool for the identification of small molecules targeting KasA. We then tested these compounds using nanoscale differential scanning fluorimetry and microscale thermophoresis. Of note, we identified several molecules including the Food and Drug Administration (FDA)-approved drugs sildenafil and flubendazole with K _d values between 30-40 μM. This may provide additional starting points for further optimization.

Dispirotripiperazine-core compounds, their biological activity with a focus on broad antiviral property, and perspectives in drug design (mini-review)

Viruses are obligate intracellular parasites and have evolved to enter the host cell. To gain access they come into contact with the host cell through an initial adhesion, and some viruses from different genus may use heparan sulfate proteoglycans for it. The successful inhibition of this early event of the infection by synthetic molecules has always been an attractive target for medicinal chemists. Numerous reports have yielded insights into the function of compounds based on the dispirotripiperazine scaffold. Analysis suggests that this is a structural requirement for inhibiting the interactions between viruses and cell-surface heparan sulfate proteoglycans, thus preventing virus entry and replication. This review summarizes our current knowledge about the early history of development, synthesis, structure-activity relationships and antiviral evaluation of dispirotripiperazine-based compounds and where they are going in the future.

Mycobacterial drug discovery

Mycobacterium tuberculosis is the causative pathogen of the pulmonary disease tuberculosis. Despite the availability of effective treatment programs, there is a global pursuit of new anti-tubercular agents to respond to the developing threat of drug resistance, in addition to reducing the extensive duration of chemotherapy and any associated toxicity. The route to mycobacterial drug discovery can be considered from two directions: target-to-drug and drug-to-target. The former approach uses conventional methods including biochemical assays along with innovative computational screens, but is yet to yield any drug candidates to the clinic, with a high attrition rate owing to lack of whole cell activity. In the latter approach, compound libraries are screened for efficacy against the bacilli or model organisms, ensuring whole cell activity, but here subsequent target identification is the rate-limiting step. Advances in a variety of scientific fields have enabled the amalgamation of aspects of both approaches in the development of novel drug discovery tools, which are now primed to accelerate the discovery of novel hits and leads with known targets and whole cell activity. This review discusses these traditional and innovative techniques, which are widely used in the quest for new anti-tubercular compounds.

Innovations in mycobacterial drug discovery to accelerate the identification of new drug candidates with confirmed targets and whole cell activity.

Identification of early liver toxicity gene biomarkers using comparative supervised machine learning

Screening agrochemicals and pharmaceuticals for potential liver toxicity is required for regulatory approval and is an expensive and time-consuming process. The identification and utilization of early exposure gene signatures and robust predictive models in regulatory toxicity testing has the potential to reduce time and costs substantially. In this study, comparative supervised machine learning approaches were applied to the rat liver TG-GATEs dataset to develop feature selection and predictive testing. We identified ten gene biomarkers using three different feature selection methods that predicted liver necrosis with high specificity and selectivity in an independent validation dataset from the Microarray Quality Control (MAQC)-II study. Nine of the ten genes that were selected with the supervised methods are involved in metabolism and detoxification (Car3, Crat, Cyp39a1, Dcd, Lbp, Scly, Slc23a1, and Tkfc) and transcriptional regulation (Ablim3). Several of these genes are also implicated in liver carcinogenesis, including Crat, Car3 and Slc23a1. Our biomarker gene signature provides high statistical accuracy and a manageable number of genes to study as indicators to potentially accelerate toxicity testing based on their ability to induce liver necrosis and, eventually, liver cancer.

Comparison of Machine Learning Models for the Androgen Receptor

The androgen receptor (AR) is a target of interest for endocrine disruption research, as altered signaling can affect normal reproductive and neurological development for generations. In an effort to prioritize compounds with alternative methodologies, the U.S. Environmental Protection Agency (EPA) used in vitro data from 11 assays to construct models of AR agonist and antagonist signaling pathways. While these EPA ToxCast AR models require in vitro data to assign a bioactivity score, Bayesian machine learning methods can be used for prospective prediction from molecule structure alone. This approach was applied to multiple types of data corresponding to the EPA's AR signaling pathway with proprietary software, Assay Central. The training performance of all machine learning models, including six other algorithms, was evaluated by internal 5-fold cross-validation statistics. Bayesian machine learning models were also evaluated with external predictions of reference chemicals to compare prediction accuracies to published results from the EPA. The machine learning model group selected for further studies of endocrine disruption consisted of continuous AC50 data from the February 2019 release of ToxCast/Tox21. These efforts demonstrate how machine learning can be used to predict AR-mediated bioactivity and can also be applied to other targets of endocrine disruption.

Machine Learning for Discovery of GSK3β Inhibitors

Alzheimer's disease (AD) is the most common cause of dementia, affecting approximately 35 million people worldwide. The current treatment options for people with AD consist of drugs designed to slow the rate of decline in memory and cognition, but these treatments are not curative, and patients eventually suffer complete cognitive injury. With the substantial amounts of published data on targets for this disease, we proposed that machine learning software could be used to find novel small-molecule treatments that can supplement the AD drugs currently on the market. In order to do this, we used publicly available data in ChEMBL to build and validate Bayesian machine learning models for AD target proteins. The first AD target that we have addressed with this method is the serine-threonine kinase glycogen synthase kinase 3 beta (GSK3β), which is a proline-directed serine-threonine kinase that phosphorylates the microtubule-stabilizing protein tau. This phosphorylation prompts tau to dissociate from the microtubule and form insoluble oligomers called paired helical filaments, which are one of the components of the neurofibrillary tangles found in AD brains. Using our Bayesian machine learning model for GSK3β consisting of 2368 molecules, this model produced a five-fold cross validation ROC of 0.905. This model was also used for virtual screening of large libraries of FDA-approved drugs and clinical candidates. Subsequent testing of selected compounds revealed a selective small-molecule inhibitor, ruboxistaurin, with activity against GSK3β (avg IC50 = 97.3 nM) and GSK3α (IC50 = 695.9 nM). Several other structurally diverse inhibitors were also identified. We are now applying this machine learning approach to additional AD targets to identify approved drugs or clinical trial candidates that can be repurposed as AD therapeutics. This represents a viable approach to accelerate drug discovery and do so at a fraction of the cost of traditional high throughput screening.

Virtual Molecular Projections and Convolutional Neural Networks for the End-to-End Modeling of Nanoparticle Activities and Properties

Digitalizing complex nanostructures into data structures suitable for machine learning modeling without losing nanostructure information has been a major challenge. Deep learning frameworks, particularly convolutional neural networks (CNNs), are especially adept at handling multidimensional and complex inputs. In this study, CNNs were applied for the modeling of nanoparticle activities exclusively from nanostructures. The nanostructures were represented by virtual molecular projections, a multidimensional digitalization of nanostructures, and used as input data to train CNNs. To this end, 77 nanoparticles with various activities and/or physicochemical property results were used for modeling. The resulting CNN model predictions show high correlations with the experimental results. An analysis of a trained CNN quantitatively showed that neurons were able to recognize distinct nanostructure features critical to activities and physicochemical properties. This "end-to-end" deep learning approach is well suited to digitalize complex nanostructures for data-driven machine learning modeling and can be broadly applied to rationally design nanoparticles with desired activities.

Machine Learning Models for Estrogen Receptor Bioactivity and Endocrine Disruption Prediction

The U.S. Environmental Protection Agency (EPA) periodically releases in vitro data across a variety of targets, including the estrogen receptor (ER). In 2015, the EPA used these data to construct mathematical models of ER agonist and antagonist pathways to prioritize chemicals for endocrine disruption testing. However, mathematical models require in vitro data prior to predicting estrogenic activity, but machine learning methods are capable of prospective prediction from the molecular structure alone. The current study describes the generation and evaluation of Bayesian machine learning models grouped by the EPA's ER agonist pathway model using multiple data types with proprietary software, Assay Central. External predictions with three test sets of in vitro and in vivo reference chemicals with agonist activity classifications were compared to previous mathematical model publications. Training data sets were subjected to additional machine learning algorithms and compared with rank normalized scores of internal five-fold cross-validation statistics. External predictions were found to be comparable or superior to previous studies published by the EPA. When assessing six additional algorithms for the training data sets, Assay Central performed similarly at a reduced computational cost. This study demonstrates that machine learning can prioritize chemicals for future in vitro and in vivo testing of ER agonism.

Repurposing Pyramax®, quinacrine and tilorone as treatments for Ebola virus disease

We have recently identified three molecules (tilorone, quinacrine and pyronaridine tetraphosphate) which all demonstrated efficacy in the mouse model of infection with mouse-adapted Ebola virus (EBOV) model of disease and had similar in vitro inhibition of an Ebola pseudovirus (VSV-EBOV-GP), suggesting they interfere with viral entry. Using a machine learning model to predict lysosomotropism these compounds were evaluated for their ability to possess a lysosomotropic mechanism in vitro. We now demonstrate in vitro that pyronaridine tetraphosphate is an inhibitor of Lysotracker accumulation in lysosomes (IC50 = 0.56 μM). Further, we evaluated antiviral synergy between pyronaridine and artesunate (Pyramax®), which are used in combination to treat malaria. Artesunate was not found to have lysosomotropic activity in vitro and the combination effect on EBOV inhibition was shown to be additive. Pyramax® may represent a unique example of the repurposing of a combination product for another disease.

Advancing computer-aided drug discovery (CADD) by big data and data-driven machine learning modeling

Advancing a new drug to market requires substantial investments in time as well as financial resources. Crucial bioactivities for drug candidates, including their efficacy, pharmacokinetics (PK), and adverse effects, need to be investigated during drug development. With advancements in chemical synthesis and biological screening technologies over the past decade, a large amount of biological data points for millions of small molecules have been generated and are stored in various databases. These accumulated data, combined with new machine learning (ML) approaches, such as deep learning, have shown great potential to provide insights into relevant chemical structures to predict in vitro, in vivo, and clinical outcomes, thereby advancing drug discovery and development in the big data era.

Synergistic drug combinations and machine learning for drug repurposing in chordoma

Chordoma is a devastating rare cancer that affects one in a million people. With a mean-survival of just 6 years and no approved medicines, the primary treatments are surgery and radiation. In order to speed new medicines to chordoma patients, a drug repurposing strategy represents an attractive approach. Drugs that have already advanced through human clinical safety trials have the potential to be approved more quickly than de novo discovered medicines on new targets. We have taken two strategies to enable this: (1) generated and validated machine learning models of chordoma inhibition and screened compounds of interest in vitro. (2) Tested combinations of approved kinase inhibitors already being individually evaluated for chordoma. Several published studies of compounds screened against chordoma cell lines were used to generate Bayesian Machine learning models which were then used to score compounds selected from the NIH NCATS industry-provided assets. Out of these compounds, the mTOR inhibitor AZD2014, was the most potent against chordoma cell lines (IC50 0.35 µM U-CH1 and 0.61 µM U-CH2). Several studies have shown the importance of the mTOR signaling pathway in chordoma and suggest it as a promising avenue for targeted therapy. Additionally, two currently FDA approved drugs, afatinib and palbociclib (EGFR and CDK4/6 inhibitors, respectively) demonstrated synergy in vitro (CI50 = 0.43) while AZD2014 and afatanib also showed synergy (CI50 = 0.41) against a chordoma cell in vitro. These findings may be of interest clinically, and this in vitro- and in silico approach could also be applied to other rare cancers.

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.

PTML Model for Selection of Nanoparticles, Anticancer Drugs, and Vitamins in the Design of Drug-Vitamin Nanoparticle Release Systems for Cancer Cotherapy

Nanosystems are gaining momentum in pharmaceutical sciences because of the wide variety of possibilities for designing these systems to have specific functions. Specifically, studies of new cancer cotherapy drug-vitamin release nanosystems (DVRNs) including anticancer compounds and vitamins or vitamin derivatives have revealed encouraging results. However, the number of possible combinations of design and synthesis conditions is remarkably high. In addition, a large number of anticancer and vitamin derivatives have been already assayed, but a notably less number of cases of DVRNs were assayed as a whole (with the anticancer compound and the vitamin linked to them). Our approach combines with the perturbation theory and machine learning (PTML) model to predict the probability of obtaining an interesting DVRN by changing the anticancer compound and/or the vitamin present in a DVRN that is already tested for other anticancer compounds or vitamins that have not been tested yet as part of a DVRN. In a previous work, we built a linear PTML model useful for the design of these nanosystems. In doing so, we used information fusion (IF) techniques to carry out data enrichment of DVRN data compiled from the literature with the data for preclinical assays of vitamins from the ChEMBL database. The design features of DVRNs and the assay conditions of nanoparticles (NPs) and vitamins were included as multiplicative PT operators (PTOs) to the system, which indicates the importance of these variables. However, the previous work omitted experiments with nonlinear ML techniques and different types of PTOs such as metric-based PTOs. More importantly, the previous work does not consider the structure of the anticancer drug to be included in the new DVRNs. In this work, we are going to accomplish three main objectives (tasks). In the first task, we found a new model, alternative to the one published before, for the rational design of DVRNs using metric-based PTOs. The most accurate PTML model was the artificial neural network model, which showed values of specificity, sensitivity, and accuracy in the range of 90-95% in training and external validation series for more than 130,000 cases (DVRNs vs ChEMBL assays). Furthermore, in the second task, we used IF techniques to carry out data enrichment of our previous data set. In doing so, we constructed a new working data set of >970,000 cases with the data of preclinical assays of DVRNs, vitamins, and anticancer compounds from the ChEMBL database. All these assays have multiple continuous variables or descriptors d_k and categorical variables c_j (conditions of the assays) for drugs (d_ack, c_acj), vitamins (d_vk, c_vj), and NPs (d_nk, c_nj). These data include >20,000 potential anticancer compounds with >270 protein targets (c_ac1), >580 assay cell organisms (c_ac2), and so forth. Furthermore, we include >36,000 assay vitamin derivatives in >6200 types of cells (c_2vit), >120 assay organisms (c_3vit), >60 assay strains (c_4vit), and so forth. The enriched data set also contains >20 types of DVRNs (c_5n) with 9 NP core materials (c_4n), 8 synthesis methods (c_7n), and so forth. We expressed all this information with PTOs and developed a qualitatively new PTML model that incorporates information of the anticancer drugs. This new model presents 96-97% of accuracy for training and external validation subsets. In the last task, we carried out a comparative study of ML and/or PTML models published and described how the models we are presenting cover the gap of knowledge in terms of drug delivery. In conclusion, we present here for the first time a multipurpose PTML model that is able to select NPs, anticancer compounds, and vitamins and their conditions of assay for DVRN design.

Comparing Machine Learning Algorithms for Predicting Drug-Induced Liver Injury (DILI)

Drug-induced liver injury (DILI) is one the most unpredictable adverse reactions to xenobiotics in humans and the leading cause of postmarketing withdrawals of approved drugs. To date, these drugs have been collated by the FDA to form the DILIRank database, which classifies DILI severity and potential. These classifications have been used by various research groups in generating computational predictions for this type of liver injury. Recently, groups from Pfizer and AstraZeneca have collated DILI in vitro data and physicochemical properties for compounds that can be used along with data from the FDA to build machine learning models for DILI. In this study, we have used these data sets, as well as the Biopharmaceutics Drug Disposition Classification System data set, to generate Bayesian machine learning models with our in-house software, Assay Central. The performance of all machine learning models was assessed through both the internal 5-fold cross-validation metrics and prediction accuracy of an external test set of compounds with known hepatotoxicity. The best-performing Bayesian model was based on the DILI-concern category from the DILIRank database with an ROC of 0.814, a sensitivity of 0.741, a specificity of 0.755, and an accuracy of 0.746. A comparison of alternative machine learning algorithms, such as k-nearest neighbors, support vector classification, AdaBoosted decision trees, and deep learning methods, produced similar statistics to those generated with the Bayesian algorithm in Assay Central. This study demonstrates machine learning models grouped in a tool called MegaTox that can be used to predict early-stage clinical compounds, as well as recent FDA-approved drugs, to identify potential DILI.