Machine Learning Informs RNA-Binding Chemical Space.

Small molecule targeting of RNA has emerged as a new frontier in medicinal chemistry, but compared to the protein targeting literature our understanding of chemical matter that binds to RNA is limited. In this study, we reported Repository Of BInders to Nucleic acids (ROBIN), a new library of nucleic acid binders identified by small molecule microarray (SMM) screening. The complete results of 36 individual nucleic acid SMM screens against a library of 24 572 small molecules were reported (including a total of 1 627 072 interactions assayed). A set of 2 003 RNA-binding small molecules was identified, representing the largest fully public, experimentally derived library of its kind to date. Machine learning was used to develop highly predictive and interpretable models to characterize RNA-binding molecules. This work demonstrates that machine learning algorithms applied to experimentally derived sets of RNA binders are a powerful method to inform RNA-targeted chemical space.

CATMoS: Collaborative Acute Toxicity Modeling Suite.

BACKGROUND: Humans are exposed to tens of thousands of chemical substances that need to be assessed for their potential toxicity. Acute systemic toxicity testing serves as the basis for regulatory hazard classification, labeling, and risk management. However, it is cost- and time-prohibitive to evaluate all new and existing chemicals using traditional rodent acute toxicity tests. In silico models built using existing data facilitate rapid acute toxicity predictions without using animals. OBJECTIVES: The U.S. Interagency Coordinating Committee on the Validation of Alternative Methods (ICCVAM) Acute Toxicity Workgroup organized an international collaboration to develop in silico models for predicting acute oral toxicity based on five different end points: Lethal Dose 50 (LD50 value, U.S. Environmental Protection Agency hazard (four) categories, Globally Harmonized System for Classification and Labeling hazard (five) categories, very toxic chemicals [LD50 (LD50≤50mg/kg)], and nontoxic chemicals (LD50>2,000mg/kg). METHODS: An acute oral toxicity data inventory for 11,992 chemicals was compiled, split into training and evaluation sets, and made available to 35 participating international research groups that submitted a total of 139 predictive models. Predictions that fell within the applicability domains of the submitted models were evaluated using external validation sets. These were then combined into consensus models to leverage strengths of individual approaches. RESULTS: The resulting consensus predictions, which leverage the collective strengths of each individual model, form the Collaborative Acute Toxicity Modeling Suite (CATMoS). CATMoS demonstrated high performance in terms of accuracy and robustness when compared with in vivo results. DISCUSSION: CATMoS is being evaluated by regulatory agencies for its utility and applicability as a potential replacement for in vivo rat acute oral toxicity studies. CATMoS predictions for more than 800,000 chemicals have been made available via the National Toxicology Program's Integrated Chemical Environment tools and data sets (ice.ntp.niehs.nih.gov). The models are also implemented in a free, standalone, open-source tool, OPERA, which allows predictions of new and untested chemicals to be made. https://doi.org/10.1289/EHP8495.

Machine Learning in Predictive Toxicology: Recent Applications and Future Directions for Classification Models.

In recent times, machine learning has become increasingly prominent in predictive toxicology as it has shifted from in vivo studies toward in silico studies. Currently, in vitro methods together with other computational methods such as quantitative structure-activity relationship modeling and absorption, distribution, metabolism, and excretion calculations are being used. An overview of machine learning and its applications in predictive toxicology is presented here, including support vector machines (SVMs), random forest (RF) and decision trees (DTs), neural networks, regression models, naïve Bayes, k-nearest neighbors, and ensemble learning. The recent successes of these machine learning methods in predictive toxicology are summarized, and a comparison of some models used in predictive toxicology is presented. In predictive toxicology, SVMs, RF, and DTs are the dominant machine learning methods due to the characteristics of the data available. Lastly, this review describes the current challenges facing the use of machine learning in predictive toxicology and offers insights into the possible areas of improvement in the field.

Confidence in Inactive and Active Predictions from Structural Alerts.

Having a measure of confidence in computational predictions of biological activity from in silico tools is vital when making predictions for new chemicals, for example, in chemical risk assessment. Where predictions of biological activity are used as an indicator of a potential hazard, false-negative predictions are the most concerning prediction; however, assigning confidence in inactive predictions is particularly challenging. How can one confidently identify the absence of activating features? In this study, we present methods for assigning confidence to both active and inactive predictions from structural alerts for protein-binding molecular initiating events (MIEs). Structural alerts were derived through an iterative statistical method. Confidence in the activity predictions is assigned by measuring the Tanimoto similarity between Morgan fingerprints of chemicals in the test set to relevant chemicals in the training set, and suitable cutoff values have been defined to give different confidence categories. To avoid a potential compound series bias in the test set and hence overestimate the performance of the method, we measured the biological activity of 27 compounds with 24 proteins, which gave us an additional 648 experimental measurements; many of the measurements are currently nonexistent in the literature and databases. This data set was complemented with newly measured biological activities published in ChEMBL25 and formed a combined independent validation data set. Applying the confidence categories to the computational predictions for the new data leads to the identification of chemicals for which one should be confident of either an inactive or active prediction, allowing model predictions to be used responsibly.

In Silico Guidance for In Vitro Androgen and Glucocorticoid Receptor ToxCast Assays.

Molecular initiating events (MIEs) are key events in adverse outcome pathways that link molecular chemistry to target biology. As they are based on chemistry, these interactions are excellent targets for computational chemistry approaches to in silico modeling. In this work, we aim to link ligand chemical structures to MIEs for androgen receptor (AR) and glucocorticoid receptor (GR) binding using ToxCast data. This has been done using an automated computational algorithm to perform maximal common substructure searches on chemical binders for each target from the ToxCast dataset. The models developed show a high level of accuracy, correctly assigning 87.20% of AR binders and 96.81% of GR binders in a 25% test set using holdout cross-validation. The 2D structural alerts developed can be used as in silico models to predict these MIEs and as guidance for in vitro ToxCast assays to confirm hits. These models can target such experimental work, reducing the number of assays to be performed to gain required toxicological insight. Development of these models has also allowed some structural alerts to be identified as predictors for agonist or antagonist behavior at the receptor target. This work represents a first step in using computational methods to guide and target experimental approaches.

The use of adverse outcome pathways in the safety evaluation of food additives.

In the last decade, adverse outcome pathways have been introduced in the fields of toxicology and risk assessment of chemicals as pragmatic tools with broad application potential. While their use in the pharmaceutical and cosmetics sectors has been well documented, their application in the food area remains largely unexplored. In this respect, an expert group of the International Life Sciences Institute Europe has recently explored the use of adverse outcome pathways in the safety evaluation of food additives. A key activity was the organization of a workshop, gathering delegates from the regulatory, industrial and academic areas, to discuss the potentials and challenges related to the application of adverse outcome pathways in the safety assessment of food additives. The present paper describes the outcome of this workshop followed by a number of critical considerations and perspectives defined by the International Life Sciences Institute Europe expert group.

Neural network activation similarity: a new measure to assist decision making in chemical toxicology.

Deep learning neural networks, constructed for the prediction of chemical binding at 79 pharmacologically important human biological targets, show extremely high performance on test data (accuracy 92.2 ± 4.2%, MCC 0.814 ± 0.093 and ROC-AUC 0.96 ± 0.04). A new molecular similarity measure, Neural Network Activation Similarity, has been developed, based on signal propagation through the network. This is complementary to standard Tanimoto similarity, and the combined use increases confidence in the computer's prediction of activity for new chemicals by providing a greater understanding of the underlying justification. The in silico prediction of these human molecular initiating events is central to the future of chemical safety risk assessment and improves the efficiency of safety decision making.

Structural Alerts and Random Forest Models in a Consensus Approach for Receptor Binding Molecular Initiating Events.

A molecular initiating event (MIE) is the gateway to an adverse outcome pathway (AOP), a sequence of events ending in an adverse effect. In silico predictions of MIEs are a vital tool in a modern, mechanism-focused approach to chemical risk assessment. For 90 biological targets representing important human MIEs, structural alert-based models have been constructed with an automated procedure that uses Bayesian statistics to iteratively select substructures. These models give impressive average performance statistics (an average of 92% correct predictions across targets), significantly improving on previous models. Random Forest models have been constructed from physicochemical features for the same targets, giving similarly impressive performance statistics (93% correct predictions). A key difference between the models is interpretation of predictions-the structural alert models are transparent and easy to interpret, while Random Forest models can only identify the most important physicochemical features for making predictions. The two complementary models have been combined in a consensus model, improving performance compared to each individual model (94% correct predictions) and increasing confidence in predictions. Variation in model performance has been explained by calculating a modelability index (MODI), using Tanimoto coefficient between Morgan fingerprints to identify nearest neighbor chemicals. This work is an important step toward building confidence in the use of in silico tools for assessment of toxicity.

Quantitative Predictions for Molecular Initiating Events Using Three-Dimensional Quantitative Structure-Activity Relationships.

The aim of human toxicity risk assessment is to determine a safe dose or exposure to a chemical for humans. This requires an understanding of the exposure of a person to a chemical and how much of the chemical is required to cause an adverse effect. To do this computationally, we need to understand how much of a chemical is required to perturb normal biological function in an adverse outcome pathway (AOP). The molecular initiating event (MIE) is the first step in an adverse outcome pathway and can be considered as a chemical interaction between a chemical toxicant and a biological molecule. Key chemical characteristics can be identified and used to model the chemistry of these MIEs. In this study, we do just this by using chemical substructures to categorize chemicals and 3D quantitative structure-activity relationships (QSARs) based on comparative molecular field analysis (CoMFA) to calculate molecular activity. Models have been constructed across a variety of human biological targets, the glucocorticoid receptor, mu opioid receptor, cyclooxygenase-2 enzyme, human ether-à-go-go related gene channel, and dopamine transporter. These models tend to provide molecular activity estimation well within one log unit and electronic and steric fields that can be visualized to better understand the MIE and biological target of interest. The outputs of these fields can be used to identify key aspects of a chemical's chemistry which can be changed to reduce its ability to activate a given MIE. With this methodology, the quantitative chemical activity can be predicted for a wide variety of MIEs, which can feed into AOP-based chemical risk assessments, and understanding of the chemistry behind the MIE can be gained.

Vision of a near future: Bridging the human health-environment divide. Toward an integrated strategy to understand mechanisms across species for chemical safety assessment.

There is a growing recognition that application of mechanistic approaches to understand cross-species shared molecular targets and pathway conservation in the context of hazard characterization, provide significant opportunities in risk assessment (RA) for both human health and environmental safety. Specifically, it has been recognized that a more comprehensive and reliable understanding of similarities and differences in biological pathways across a variety of species will better enable cross-species extrapolation of potential adverse toxicological effects. Ultimately, this would also advance the generation and use of mechanistic data for both human health and environmental RA. A workshop brought together representatives from industry, academia and government to discuss how to improve the use of existing data, and to generate new NAMs data to derive better mechanistic understanding between humans and environmentally-relevant species, ultimately resulting in holistic chemical safety decisions. Thanks to a thorough dialogue among all participants, key challenges, current gaps and research needs were identified, and potential solutions proposed. This discussion highlighted the common objective to progress toward more predictive, mechanistically based, data-driven and animal-free chemical safety assessments. Overall, the participants recognized that there is no single approach which would provide all the answers for bridging the gap between mechanism-based human health and environmental RA, but acknowledged we now have the incentive, tools and data availability to address this concept, maximizing the potential for improvements in both human health and environmental RA.

Characterizing the coverage of critical effects relevant in the safety evaluation of food additives by AOPs.

There is considerable interest in adverse outcome pathways (AOPs) as a means of organizing biological and toxicological information to assist in data interpretation and method development. While several chemical sectors have shown considerable progress in applying this approach, this has not been the case in the food sector. In the present study, safety evaluation reports of food additives listed in Annex II of Regulation (EC) No 1333/2008 of the European Union were screened to qualitatively and quantitatively characterize toxicity induced in laboratory animals. The resulting database was used to identify the critical adverse effects used for risk assessment and to investigate whether food additives share common AOPs. Analysis of the database revealed that often such scrutiny of AOPs was not possible or necessary. For 69% of the food additives, the report did not document any adverse effects in studies based on which the safety evaluation was performed. For the remaining 31% of the 326 investigated food additives, critical adverse effects and related points of departure for establishing health-based guidance values could be identified. These mainly involved effects on the liver, kidney, cardiovascular system, lymphatic system, central nervous system and reproductive system. AOPs are available for many of these apical endpoints, albeit to different degrees of maturity. For other adverse outcomes pertinent to food additives, including gastrointestinal irritation and corrosion, AOPs are lacking. Efforts should focus on developing AOPs for these particular endpoints.

Using 2D Structural Alerts to Define Chemical Categories for Molecular Initiating Events.

Molecular initiating events (MIEs) are important concepts for in silico predictions. They can be used to link chemical characteristics to biological activity through an adverse outcome pathway (AOP). In this work, we capture chemical characteristics in 2D structural alerts, which are then used as models to predict MIEs. An automated procedure has been used to identify these alerts, and the chemical categories they define have been used to provide quantitative predictions for the activity of molecules that contain them. This has been done across a diverse group of 39 important pharmacological human targets using open source data. The alerts for each target combine into a model for that target, and these models are joined into a tool for MIE prediction with high average model performance (sensitivity = 82%, specificity = 93%, overall quality = 93%, Matthews correlation coefficient = 0.57). The result is substantially improved from our previous study (Allen, T. E. H., Goodman, J. M., Gutsell, S., and Russell, P. J. 2016. A history of the molecular initiating event. Chem. Res. Toxicol. 29, 2060-2070) for which the mean sensitivity for each target was only 58%. This tool provides the first step in an AOP-based risk assessment, linking chemical structure to toxicity endpoint.

Using Transition State Modeling To Predict Mutagenicity for Michael Acceptors.

The Ames mutagenicity assay is a long established in vitro test to measure the mutagenicity potential of a new chemical used in regulatory testing globally. One of the key computational approaches to modeling of the Ames assay relies on the formation of chemical categories based on the different electrophilic compounds that are able to react directly with DNA and form a covalent bond. Such approaches sometimes predict false positives, as not all Michael acceptors are found to be Ames-positive. The formation of such covalent bonds can be explored computationally using density functional theory transition state modeling. We have applied this approach to mutagenicity, allowing us to calculate the activation energy required for α,ß-unsaturated carbonyls to react with a model system for the guanine nucleobase of DNA. These calculations have allowed us to identify that chemical compounds with activation energies greater than or equal to 25.7 kcal/mol are not able to bind directly to DNA. This allows us to reduce the false positive rate for computationally predicted mutagenicity assays. This methodology can be used to investigate other covalent-bond-forming reactions that can lead to toxicological outcomes and learn more about experimental results.

A History of the Molecular Initiating Event.

The adverse outcome pathway (AOP) framework provides an alternative to traditional in vivo experiments for the risk assessment of chemicals. AOPs consist of a number of key events (KEs) linked by key event relationships across a range of biological organization backed by scientific evidence. The first KE in the pathway is the molecular initiating event (MIE)-the initial chemical trigger that starts an AOP. Over the past 3 years the AOP conceptual framework has gained a large amount of momentum in toxicology as an alternative to animal methods, and so the MIE has come into the spotlight. What is an MIE? How can MIEs be measured or predicted? What research is currently contributing to our understanding of MIEs? In this Perspective we outline answers to these key questions.

Using Molecular Initiating Events To Generate 2D Structure-Activity Relationships for Toxicity Screening.

Molecular initiating events (MIEs) can be boiled down to chemical interactions. Chemicals that interact must have intrinsic properties that allow them to exhibit this behavior, be these properties stereochemical, electronic, or otherwise. In an attempt to discover some of these chemical characteristics, we have constructed structural alert-style structure-activity relationships (SARs) to computationally predict MIEs. This work utilizes chemical informatics approaches, searching the ChEMBL database for molecules that bind to a number of pharmacologically important human toxicology targets, including G-protein coupled receptors, enzymes, ion channels, nuclear receptors, and transporters. By screening these compounds to find common 2D fragments and combining this approach with a good understanding of the literature, bespoke 2D structural alerts have been written. These SARs form the beginning of a tool for screening novel chemicals to establish the kind of interactions that they may be able to make in humans. These SARs have been run through an internal validation to test their quality, and the results of this are also discussed. MIEs have proven to be difficult to find and characterize, but we believe we have taken a key first step with this work.

Defining molecular initiating events in the adverse outcome pathway framework for risk assessment.

Consumer and environmental safety decisions are based on exposure and hazard data, interpreted using risk assessment approaches. The adverse outcome pathway (AOP) conceptual framework has been presented as a logical sequence of events or processes within biological systems which can be used to understand adverse effects and refine current risk assessment practices in ecotoxicology. This framework can also be applied to human toxicology and is explored on the basis of investigating the molecular initiating events (MIEs) of compounds. The precise definition of the MIE has yet to reach general acceptance. In this work we present a unified MIE definition: an MIE is the initial interaction between a molecule and a biomolecule or biosystem that can be causally linked to an outcome via a pathway. Case studies are presented, and issues with current definitions are addressed. With the development of a unified MIE definition, the field can look toward defining, classifying, and characterizing more MIEs and using knowledge of the chemistry of these processes to aid AOP research and toxicity risk assessment. We also present the role of MIE research in the development of in vitro and in silico toxicology and suggest how, by using a combination of biological and chemical approaches, MIEs can be identified and characterized despite a lack of detailed reports, even for some of the most studied molecules in toxicology.