Predicting Organ Toxicity Using in Vitro Bioactivity Data and Chemical Structure

Systematic Approaches for the Encoding of Chemical Groups: A Case Study

Regulatory authorities aim to organize substances into groups to facilitate prioritization within hazard and risk assessment processes. Often, such chemical groupings are not explicitly defined by structural rules or physicochemical property information. This is largely due to how these groupings are developed, namely, a manual expert curation process, which in turn makes updating and refining groupings, as new substances are evaluated, a practical challenge. Herein, machine learning methods were leveraged to build models that could preliminarily assign substances to predefined groups. A set of 86 groupings containing 2,184 substances as published on the European Chemicals Agency (ECHA) website were mapped to the U.S. Environmental Protection Agency (EPA) Distributed Toxicity Structure Database (DSSTox) content to extract chemical and structural information. Substances were represented using Morgan fingerprints, and two machine learning approaches were used to classify test substances into 56 groups containing at least 10 substances with a structural representation in the data set: k-nearest neighbor (kNN) and random forest (RF), that led to mean 5-fold cross-validation test accuracies (average F1 scores) of 0.781 and 0.853, respectively. With a 9% improvement, the RF classifier was significantly more accurate than KNN (p-value = 0.001). The approach offers promise as a means of the initial profiling of new substances into predefined groups to facilitate prioritization efforts and streamline the assessment of new substances when earlier groupings are available. The algorithm to fit and use these models has been made available in the accompanying repository, thereby enabling both use of the produced models and refitting of these models, as new groupings become available by regulatory authorities or industry.

Review of machine learning and deep learning models for toxicity prediction

The ever-increasing number of chemicals has raised public concerns due to their adverse effects on human health and the environment. To protect public health and the environment, it is critical to assess the toxicity of these chemicals. Traditional in vitro and in vivo toxicity assays are complicated, costly, and time-consuming and may face ethical issues. These constraints raise the need for alternative methods for assessing the toxicity of chemicals. Recently, due to the advancement of machine learning algorithms and the increase in computational power, many toxicity prediction models have been developed using various machine learning and deep learning algorithms such as support vector machine, random forest, k-nearest neighbors, ensemble learning, and deep neural network. This review summarizes the machine learning- and deep learning-based toxicity prediction models developed in recent years. Support vector machine and random forest are the most popular machine learning algorithms, and hepatotoxicity, cardiotoxicity, and carcinogenicity are the frequently modeled toxicity endpoints in predictive toxicology. It is known that datasets impact model performance. The quality of datasets used in the development of toxicity prediction models using machine learning and deep learning is vital to the performance of the developed models. The different toxicity assignments for the same chemicals among different datasets of the same type of toxicity have been observed, indicating benchmarking datasets is needed for developing reliable toxicity prediction models using machine learning and deep learning algorithms. This review provides insights into current machine learning models in predictive toxicology, which are expected to promote the development and application of toxicity prediction models in the future.

Identification and Prioritization of Environmental Organic Pollutants: From an Analytical and Toxicological Perspective

Exposure to environmental organic pollutants has triggered significant ecological impacts and adverse health outcomes, which have been received substantial and increasing attention. The contribution of unidentified chemical components is considered as the most significant knowledge gap in understanding the combined effects of pollutant mixtures. To address this issue, remarkable analytical breakthroughs have recently been made. In this review, the basic principles on recognition of environmental organic pollutants are overviewed. Complementary analytical methodologies (i.e., quantitative structure-activity relationship prediction, mass spectrometric nontarget screening, and effect-directed analysis) and experimental platforms are briefly described. The stages of technique development and/or essential parts of the analytical workflow for each of the methodologies are then reviewed. Finally, plausible technique paths and applications of the future nontarget screening methods, interdisciplinary techniques for achieving toxicant identification, and burgeoning strategies on risk assessment of chemical cocktails are discussed.

In Silico Prediction of Human Organ Toxicity via Artificial Intelligence Methods

Unpredicted human organ level toxicity remains one of the major reasons for drug clinical failure. There is a critical need for cost-efficient strategies in the early stages of drug development for human toxicity assessment. At present, artificial intelligence methods are popularly regarded as a promising solution in chemical toxicology. Thus, we provided comprehensive in silico prediction models for eight significant human organ level toxicity end points using machine learning, deep learning, and transfer learning algorithms. In this work, our results showed that the graph-based deep learning approach was generally better than the conventional machine learning models, and good performances were observed for most of the human organ level toxicity end points in this study. In addition, we found that the transfer learning algorithm could improve model performance for skin sensitization end point using source domain of in vivo acute toxicity data and in vitro data of the Tox21 project. It can be concluded that our models can provide useful guidance for the rapid identification of the compounds with human organ level toxicity for drug discovery.

Navigating the bridge between wet and dry lab toxicology research to address current challenges with high-dimensional data

Toxicology research has rapidly evolved, leveraging increasingly advanced technologies in high-throughput approaches to yield important information on toxicological mechanisms and health outcomes. Data produced through toxicology studies are consequently becoming larger, often producing high-dimensional data. These types of data hold promise for imparting new knowledge, yet inherently have complexities causing them to be a rate-limiting element for researchers, particularly those that are housed in "wet lab" settings (i.e., researchers that use liquids to analyze various chemicals and biomarkers as opposed to more computationally focused, "dry lab" researchers). These types of challenges represent topics of ongoing conversation amongst our team and researchers in the field. The aim of this perspective is to i) summarize hurdles in analyzing high-dimensional data in toxicology that require improved training and translation for wet lab researchers, ii) highlight example methods that have aided in translating data analysis techniques to wet lab researchers; and iii) describe challenges that remain to be effectively addressed, to date, in toxicology research. Specific aspects include methodologies that could be introduced to wet lab researchers, including data pre-processing, machine learning, and data reduction. Current challenges discussed include model interpretability, study biases, and data analysis training. Example efforts implemented to translate these data analysis techniques are also mentioned, including online data analysis resources and hands-on workshops. Questions are also posed to continue conversation in the toxicology community. Contents of this perspective represent timely issues broadly occurring in the fields of bioinformatics and toxicology that require ongoing dialogue between wet and dry lab researchers.

Evaluating the utility of a high throughput thiol-containing fluorescent probe to screen for reactivity: A case study with the Tox21 library

High-throughput screening (HTS) assays for bioactivity in the Tox21 program aim to evaluate an array of different biological targets and pathways, but a significant barrier to interpretation of these data is the lack of high-throughput screening (HTS) assays intended to identify non-specific reactive chemicals. This is an important aspect for prioritising chemicals to test in specific assays, identifying promiscuous chemicals based on their reactivity, as well as addressing hazards such as skin sensitisation which are not necessarily initiated by a receptor-mediated effect but act through a non-specific mechanism. Herein, a fluorescence-based HTS assay that allows the identification of thiol-reactive compounds was used to screen 7,872 unique chemicals in the Tox21 10K chemical library. Active chemicals were compared with profiling outcomes using structural alerts encoding electrophilic information. Random Forest classification models based on chemical fingerprints were developed to predict assay outcomes and evaluated through 10-fold stratified cross validation (CV). The mean CV Balanced Accuracy of the validation set was 0.648. The model developed shows promise as a tool to screen untested chemicals for their potential electrophilic reactivity based solely on chemical structural features.

Accurate clinical toxicity prediction using multi-task deep neural nets and contrastive molecular explanations

Explainable machine learning for molecular toxicity prediction is a promising approach for efficient drug development and chemical safety. A predictive ML model of toxicity can reduce experimental cost and time while mitigating ethical concerns by significantly reducing animal and clinical testing. Herein, we use a deep learning framework for simultaneously modeling in vitro, in vivo, and clinical toxicity data. Two different molecular input representations are used; Morgan fingerprints and pre-trained SMILES embeddings. A multi-task deep learning model accurately predicts toxicity for all endpoints, including clinical, as indicated by the area under the Receiver Operator Characteristic curve and balanced accuracy. In particular, pre-trained molecular SMILES embeddings as input to the multi-task model improved clinical toxicity predictions compared to existing models in MoleculeNet benchmark. Additionally, our multitask approach is comprehensive in the sense that it is comparable to state-of-the-art approaches for specific endpoints in in vitro, in vivo and clinical platforms. Through both the multi-task model and transfer learning, we were able to indicate the minimal need of in vivo data for clinical toxicity predictions. To provide confidence and explain the model's predictions, we adapt a post-hoc contrastive explanation method that returns pertinent positive and negative features, which correspond well to known mutagenic and reactive toxicophores, such as unsubstituted bonded heteroatoms, aromatic amines, and Michael receptors. Furthermore, toxicophore recovery by pertinent feature analysis captures more of the in vitro (53%) and in vivo (56%), rather than of the clinical (8%), endpoints, and indeed uncovers a preference in known toxicophore data towards in vitro and in vivo experimental data. To our knowledge, this is the first contrastive explanation, using both present and absent substructures, for predictions of clinical and in vivo molecular toxicity.

Using chemical and biological data to predict drug toxicity

Various sources of information can be used to better understand and predict compound activity and safety-related endpoints, including biological data such as gene expression and cell morphology. In this review, we first introduce types of chemical, in vitro and in vivo information that can be used to describe compounds and adverse effects. We then explore how compound descriptors based on chemical structure or biological perturbation response can be used to predict safety-related endpoints, and how especially biological data can help us to better understand adverse effects mechanistically. Overall, the described applications demonstrate how large-scale biological information presents new opportunities to anticipate and understand the biological effects of compounds, and how this can support predictive toxicology and drug discovery projects.

Advances in computational methods along the exposure to toxicological response paradigm

Human health risk assessment is a function of chemical toxicity, bioavailability to reach target biological tissues, and potential environmental exposure. These factors are complicated by many physiological, biochemical, physical and lifestyle factors. Furthermore, chemical health risk assessment is challenging in view of the large, and continually increasing, number of chemicals found in the environment. These challenges highlight the need to prioritize resources for the efficient and timely assessment of those environmental chemicals that pose greatest health risks. Computational methods, either predictive or investigative, are designed to assist in this prioritization in view of the lack of cost prohibitive in vivo experimental data. Computational methods provide specific and focused toxicity information using in vitro high throughput screening (HTS) assays. Information from the HTS assays can be converted to in vivo estimates of chemical levels in blood or target tissue, which in turn are converted to in vivo dose estimates that can be compared to exposure levels of the screened chemicals. This manuscript provides a review for the landscape of computational methods developed and used at the U.S. Environmental Protection Agency (EPA) highlighting their potentials and challenges.

Artificial Intelligence-Based Toxicity Prediction of Environmental Chemicals: Future Directions for Chemical Management Applications

Recently, research on the development of artificial intelligence (AI)-based computational toxicology models that predict toxicity without the use of animal testing has emerged because of the rapid development of computer technology. Various computational toxicology techniques that predict toxicity based on the structure of chemical substances are gaining attention, including the quantitative structure-activity relationship. To understand the recent development of these models, we analyzed the databases, molecular descriptors, fingerprints, and algorithms considered in recent studies. Based on a selection of 96 papers published since 2014, we found that AI models have been developed to predict approximately 30 different toxicity end points using more than 20 toxicity databases. For model development, molecular access system and extended-connectivity fingerprints are the most commonly used molecular descriptors. The most used algorithm among the machine learning techniques is the random forest, while the most used algorithm among the deep learning techniques is a deep neural network. The use of AI technology in the development of toxicity prediction models is a new concept that will aid in achieving a scientific accord and meet regulatory applications. The comprehensive overview provided in this study will provide a useful guide for the further development and application of toxicity prediction models.

Machine Learning Models for Predicting Liver Toxicity

Liver toxicity is a major adverse drug reaction that accounts for drug failure in clinical trials and withdrawal from the market. Therefore, predicting potential liver toxicity at an early stage in drug discovery is crucial to reduce costs and the potential for drug failure. However, current in vivo animal toxicity testing is very expensive and time consuming. As an alternative approach, various machine learning models have been developed to predict potential liver toxicity in humans. This chapter reviews current advances in the development and application of machine learning models for prediction of potential liver toxicity in humans and discusses possible improvements to liver toxicity prediction.

In silico approaches in organ toxicity hazard assessment: current status and future needs in predicting liver toxicity

Hepatotoxicity is one of the most frequently observed adverse effects resulting from exposure to a xenobiotic. For example, in pharmaceutical research and development it is one of the major reasons for drug withdrawals, clinical failures, and discontinuation of drug candidates. The development of faster and cheaper methods to assess hepatotoxicity that are both more sustainable and more informative is critically needed. The biological mechanisms and processes underpinning hepatotoxicity are summarized and experimental approaches to support the prediction of hepatotoxicity are described, including toxicokinetic considerations. The paper describes the increasingly important role of in silico approaches and highlights challenges to the adoption of these methods including the lack of a commonly agreed upon protocol for performing such an assessment and the need for in silico solutions that take dose into consideration. A proposed framework for the integration of in silico and experimental information is provided along with a case study describing how computational methods have been used to successfully respond to a regulatory question concerning non-genotoxic impurities in chemically synthesized pharmaceuticals.

Multitask CapsNet: An Imbalanced Data Deep Learning Method for Predicting Toxicants

Drug development has a high failure rate, with safety properties constituting a considerable challenge. To reduce risk, in silico tools, including various machine learning methods, have been applied for toxicity prediction. However, these approaches often confront a serious problem: the training data sets are usually biased (imbalanced positive and negative samples), which would result in model training difficulty and unsatisfactory prediction accuracy. Multitask networks obtained significantly better predictive accuracies than single-task methods, and capsule neural networks showed excellent performance in sparse data sets in previous studies. In this study, we developed a new multitask framework based on a capsule neural network (multitask CapsNet) to measure 12 different toxic effects simultaneously. We found that multitask CapsNet excelled in toxicity prediction and outperformed many other computational approaches using the multitask strategy. Only after training on biased data sets did multitask CapsNet achieve significantly improved prediction accuracy on the Tox21 Data Challenge, which gave the largest ratio of highest accuracy (8/12) among compared models. Our model gave a prediction accuracy of 96.6% for the target NR.PPAR.gamma, whose ratio of negative to positive samples was up to 36:1. These results suggested that multitask CapsNet could overcome the bias problems and would provide a novel, accurate, and efficient approach for predicting the toxicities of compounds.

ChemBioSim: Enhancing Conformal Prediction of In Vivo Toxicity by Use of Predicted Bioactivities

Computational methods such as machine learning approaches have a strong track record of success in predicting the outcomes of in vitro assays. In contrast, their ability to predict in vivo endpoints is more limited due to the high number of parameters and processes that may influence the outcome. Recent studies have shown that the combination of chemical and biological data can yield better models for in vivo endpoints. The ChemBioSim approach presented in this work aims to enhance the performance of conformal prediction models for in vivo endpoints by combining chemical information with (predicted) bioactivity assay outcomes. Three in vivo toxicological endpoints, capturing genotoxic (MNT), hepatic (DILI), and cardiological (DICC) issues, were selected for this study due to their high relevance for the registration and authorization of new compounds. Since the sparsity of available biological assay data is challenging for predictive modeling, predicted bioactivity descriptors were introduced instead. Thus, a machine learning model for each of the 373 collected biological assays was trained and applied on the compounds of the in vivo toxicity data sets. Besides the chemical descriptors (molecular fingerprints and physicochemical properties), these predicted bioactivities served as descriptors for the models of the three in vivo endpoints. For this study, a workflow based on a conformal prediction framework (a method for confidence estimation) built on random forest models was developed. Furthermore, the most relevant chemical and bioactivity descriptors for each in vivo endpoint were preselected with lasso models. The incorporation of bioactivity descriptors increased the mean F1 scores of the MNT model from 0.61 to 0.70 and for the DICC model from 0.72 to 0.82 while the mean efficiencies increased by roughly 0.10 for both endpoints. In contrast, for the DILI endpoint, no significant improvement in model performance was observed. Besides pure performance improvements, an analysis of the most important bioactivity features allowed detection of novel and less intuitive relationships between the predicted biological assay outcomes used as descriptors and the in vivo endpoints. This study presents how the prediction of in vivo toxicity endpoints can be improved by the incorporation of biological information-which is not necessarily captured by chemical descriptors-in an automated workflow without the need for adding experimental workload for the generation of bioactivity descriptors as predicted outcomes of bioactivity assays were utilized. All bioactivity CP models for deriving the predicted bioactivities, as well as the in vivo toxicity CP models, can be freely downloaded from https://doi.org/10.5281/zenodo.4761225.

Artificial Intelligence in Drug Discovery: A Comprehensive Review of Data-driven and Machine Learning Approaches

As expenditure on drug development increases exponentially, the overall drug discovery process requires a sustainable revolution. Since artificial intelligence (AI) is leading the fourth industrial revolution, AI can be considered as a viable solution for unstable drug research and development. Generally, AI is applied to fields with sufficient data such as computer vision and natural language processing, but there are many efforts to revolutionize the existing drug discovery process by applying AI. This review provides a comprehensive, organized summary of the recent research trends in AI-guided drug discovery process including target identification, hit identification, ADMET prediction, lead optimization, and drug repositioning. The main data sources in each field are also summarized in this review. In addition, an in-depth analysis of the remaining challenges and limitations will be provided, and proposals for promising future directions in each of the aforementioned areas.

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.

A comprehensive view on mechanistic approaches for cancer risk assessment of non-genotoxic agrochemicals

Currently the only methods for non-genotoxic carcinogenic hazard assessment accepted by most regulatory authorities are lifetime carcinogenicity studies. However, these involve the use of large numbers of animals and the relevance of their predictive power and results has been scientifically challenged. With increased availability of innovative test methods and enhanced understanding of carcinogenic processes, it is believed that tumour formation can now be better predicted using mechanistic information. A workshop organised by the European Partnership on Alternative Approaches to Animal Testing brought together experts to discuss an alternative, mechanism-based approach for cancer risk assessment of agrochemicals. Data from a toolbox of test methods for detecting modes of action (MOAs) underlying non-genotoxic carcinogenicity are combined with information from subchronic toxicity studies in a weight-of-evidence approach to identify carcinogenic potential of a test substance. The workshop included interactive sessions to discuss the approach using case studies. These showed that fine-tuning is needed, to build confidence in the proposed approach, to ensure scientific correctness, and to address different regulatory needs. This novel approach was considered realistic, and its regulatory acceptance and implementation can be facilitated in the coming years through continued dialogue between all stakeholders and building confidence in alternative approaches.

The EU-ToxRisk method documentation, data processing and chemical testing pipeline for the regulatory use of new approach methods

Hazard assessment, based on new approach methods (NAM), requires the use of batteries of assays, where individual tests may be contributed by different laboratories. A unified strategy for such collaborative testing is presented. It details all procedures required to allow test information to be usable for integrated hazard assessment, strategic project decisions and/or for regulatory purposes. The EU-ToxRisk project developed a strategy to provide regulatorily valid data, and exemplified this using a panel of > 20 assays (with > 50 individual endpoints), each exposed to 19 well-known test compounds (e.g. rotenone, colchicine, mercury, paracetamol, rifampicine, paraquat, taxol). Examples of strategy implementation are provided for all aspects required to ensure data validity: (i) documentation of test methods in a publicly accessible database; (ii) deposition of standard operating procedures (SOP) at the European Union DB-ALM repository; (iii) test readiness scoring accoding to defined criteria; (iv) disclosure of the pipeline for data processing; (v) link of uncertainty measures and metadata to the data; (vi) definition of test chemicals, their handling and their behavior in test media; (vii) specification of the test purpose and overall evaluation plans. Moreover, data generation was exemplified by providing results from 25 reporter assays. A complete evaluation of the entire test battery will be described elsewhere. A major learning from the retrospective analysis of this large testing project was the need for thorough definitions of the above strategy aspects, ideally in form of a study pre-registration, to allow adequate interpretation of the data and to ensure overall scientific/toxicological validity.

KnowTox: pipeline and case study for confident prediction of potential toxic effects of compounds in early phases of development

Risk assessment of newly synthesised chemicals is a prerequisite for regulatory approval. In this context, in silico methods have great potential to reduce time, cost, and ultimately animal testing as they make use of the ever-growing amount of available toxicity data. Here, KnowTox is presented, a novel pipeline that combines three different in silico toxicology approaches to allow for confident prediction of potentially toxic effects of query compounds, i.e. machine learning models for 88 endpoints, alerts for 919 toxic substructures, and computational support for read-across. It is mainly based on the ToxCast dataset, containing after preprocessing a sparse matrix of 7912 compounds tested against 985 endpoints. When applying machine learning models, applicability and reliability of predictions for new chemicals are of utmost importance. Therefore, first, the conformal prediction technique was deployed, comprising an additional calibration step and per definition creating internally valid predictors at a given significance level. Second, to further improve validity and information efficiency, two adaptations are suggested, exemplified at the androgen receptor antagonism endpoint. An absolute increase in validity of 23% on the in-house dataset of 534 compounds could be achieved by introducing KNNRegressor normalisation. This increase in validity comes at the cost of efficiency, which could again be improved by 20% for the initial ToxCast model by balancing the dataset during model training. Finally, the value of the developed pipeline for risk assessment is discussed using two in-house triazole molecules. Compared to a single toxicity prediction method, complementing the outputs of different approaches can have a higher impact on guiding toxicity testing and de-selecting most likely harmful development-candidate compounds early in the development process.

Predictive Models for Human Organ Toxicity Based on In Vitro Bioactivity Data and Chemical Structure

Traditional toxicity testing reliant on animal models is costly and low throughput, posing a significant challenge with the increasing numbers of chemicals that humans are exposed to in the environment. The purpose of this investigation was to build optimal prediction models for various human in vivo/organ-level toxicity end points (extracted from ChemIDPlus) using chemical structure and Tox21 in vitro quantitative high-throughput screening (qHTS) bioactivity assay data. Several supervised machine learning algorithms were applied to model 14 human toxicity end points pertaining to vascular, kidney, ureter and bladder, and liver organ systems. Three metrics were used to evaluate model performance: area under the receiver operating characteristic curve (AUC-ROC), balanced accuracy (BA), and Matthews correlation coefficient (MCC). The top four models, with AUC-ROC values >0.8, were derived for endocrine (0.90 ± 0.00), musculoskeletal (0.88 ± 0.02), peripheral nerve and sensation (0.85 ± 0.01), and brain and coverings (0.83 ± 0.02) toxicities, whereas the best model AUC-ROC values were >0.7 for the remaining 10 toxicities. Model performance was found to be dependent on the specific data set, model type, and feature selection method used. In addition, chemical structure and assay data showed different levels of contribution to the prediction of different toxicity end points. Although in vitro assay data, when combined with chemical structure, slightly improved the predictive accuracy for most end points (11 out of 14), a noteworthy finding was the near equal success of the structure-only models, which do not require Tox21 qHTS screening data, and the relatively poor performance of assay-only models. Thus, the top-performing structure-only models from this study could be applied for hazard screening of large sets of chemicals for potential human toxicity, whereas the largest assay contributions to models (i.e., cellular targets) could be used, along with the top-contributing structural features, to provide insight into toxicity mechanisms.

Predicting drug-induced hepatotoxicity based on biological feature maps and diverse classification strategies

Identifying hepatotoxicity as early as possible is significant in drug development. In this study, we developed a drug-induced hepatotoxicity prediction model taking account of both the biological context and the computational efficacy based on toxicogenomics data. Specifically, we proposed a novel gene selection algorithm considering gene's participation, named BioCB, to choose the discriminative genes and make more efficient prediction. Then instead of using the raw gene expression levels to characterize each drug, we developed a two-dimensional biological process feature pattern map to represent each drug. Then we employed two strategies to handle the maps and identify the hepatotoxicity, the direct use of maps, named Two-dim branch, and vectorization of maps, named One-dim branch. The two strategies subsequently used the deep convolutional neural networks and LightGBM as predictors, respectively. Additionally, we here for the first time proposed a stacked vectorized gene matrix, which was more predictive than the raw gene matrix. Results validated on both in vivo and in vitro data from two public data sets, the TG-GATES and DrugMatrix, show that the proposed One-dim branch outperforms the deep framework, the Two-dim branch, and has achieved high accuracy and efficiency. The implementation of the proposed method is available at https://github.com/RanSuLab/Hepatotoxicity.

ToxRefDB version 2.0: Improved utility for predictive and retrospective toxicology analyses

The Toxicity Reference Database (ToxRefDB) structures information from over 5000 in vivo toxicity studies, conducted largely to guidelines or specifications from the US Environmental Protection Agency and the National Toxicology Program, into a public resource for training and validation of predictive models. Herein, ToxRefDB version 2.0 (ToxRefDBv2) development is described. Endpoints were annotated (e.g. required, not required) according to guidelines for subacute, subchronic, chronic, developmental, and multigenerational reproductive designs, distinguishing negative responses from untested. Quantitative data were extracted, and dose-response modeling for nearly 28,000 datasets from nearly 400 endpoints using Benchmark Dose (BMD) Modeling Software were generated and stored. Implementation of controlled vocabulary improved data quality; standardization to guideline requirements and cross-referencing with United Medical Language System (UMLS) connects ToxRefDBv2 observations to vocabularies linked to UMLS, including PubMed medical subject headings. ToxRefDBv2 allows for increased connections to other resources and has greatly enhanced quantitative and qualitative utility for predictive toxicology.

Incorporating new approach methodologies in toxicity testing and exposure assessment for tiered risk assessment using the RISK21 approach: Case studies on food contact chemicals

Programs including the ToxCast project have generated large amounts of in vitro high‒throughput screening (HTS) data, and best approaches for the interpretation and use of HTS data, including for chemical safety assessment, remain to be evaluated. To fill this gap, we conducted case studies of two indirect food additive chemicals where ToxCast data were compared with in vivo toxicity data using the RISK21 approach. Two food contact substances, sodium (2-pyridylthio)-N-oxide and dibutyltin dichloride, were selected, and available exposure data, toxicity data, and model predictions were compiled and assessed. Oral equivalent doses for the ToxCast bioactivity data were determined by in-vitro in-vivo extrapolation (IVIVE). For sodium (2-pyridylthio)-N-oxide, bioactive concentrations in ToxCast assays corresponded to low-and no-observed adverse effect levels in animal studies. For dibutyltin dichloride, the ToxCast bioactive concentrations were below the dose range that demonstrated toxicity in animals; however, this was confounded by the lack of toxicokinetic data, necessitating the use of conservative toxicokinetic parameter estimates for IVIVE calculations. This study highlights the potential utility of the RISK21 approach for interpretation of the ToxCast HTS data, as well as the challenges involved in integrating in vitro HTS data into safety assessments.

Open-source QSAR models for pKa prediction using multiple machine learning approaches

BACKGROUND

The logarithmic acid dissociation constant pKa reflects the ionization of a chemical, which affects lipophilicity, solubility, protein binding, and ability to pass through the plasma membrane. Thus, pKa affects chemical absorption, distribution, metabolism, excretion, and toxicity properties. Multiple proprietary software packages exist for the prediction of pKa, but to the best of our knowledge no free and open-source programs exist for this purpose. Using a freely available data set and three machine learning approaches, we developed open-source models for pKa prediction.

METHODS

The experimental strongest acidic and strongest basic pKa values in water for 7912 chemicals were obtained from DataWarrior, a freely available software package. Chemical structures were curated and standardized for quantitative structure-activity relationship (QSAR) modeling using KNIME, and a subset comprising 79% of the initial set was used for modeling. To evaluate different approaches to modeling, several datasets were constructed based on different processing of chemical structures with acidic and/or basic pKas. Continuous molecular descriptors, binary fingerprints, and fragment counts were generated using PaDEL, and pKa prediction models were created using three machine learning methods, (1) support vector machines (SVM) combined with k-nearest neighbors (kNN), (2) extreme gradient boosting (XGB) and (3) deep neural networks (DNN).

RESULTS

The three methods delivered comparable performances on the training and test sets with a root-mean-squared error (RMSE) around 1.5 and a coefficient of determination (R2) around 0.80. Two commercial pKa predictors from ACD/Labs and ChemAxon were used to benchmark the three best models developed in this work, and performance of our models compared favorably to the commercial products.

CONCLUSIONS

This work provides multiple QSAR models to predict the strongest acidic and strongest basic pKas of chemicals, built using publicly available data, and provided as free and open-source software on GitHub.

Drug safety Africa: An overview of safety pharmacology & toxicology in South Africa

This meeting report is based on presentations given at the first Drug Safety Africa Meeting in Potchefstroom, South Africa from November 20-22, 2018 at the North-West University campus. There were 134 attendees (including 26 speakers and 34 students) from the pharmaceutical industry, academia, regulatory agencies as well as 6 exhibitors. These meeting proceedings are designed to inform the content that was presented in terms of Safety Pharmacology (SP) and Toxicology methods and models that are used by the pharmaceutical industry to characterize the safety profile of novel small chemical or biological molecules. The first part of this report includes an overview of the core battery studies defined by cardiovascular, central nervous system (CNS) and respiratory studies. Approaches to evaluating drug effects on the renal and gastrointestinal systems and murine phenotyping were also discussed. Subsequently, toxicological approaches were presented including standard strategies and options for early identification and characterization of risks associated with a novel therapeutic, the types of toxicology studies conducted and relevance to risk assessment supporting first-in-human (FIH) clinical trials and target organ toxicity. Biopharmaceutical development and principles of immunotoxicology were discussed as well as emerging technologies. An additional poster session was held that included 18 posters on advanced studies and topics by South African researchers, postgraduate students and postdoctoral fellows.

Development of quantitative structure-activity relationship models to predict potential nephrotoxic ingredients in traditional Chinese medicines

The broad use of traditional Chinese medicines (TCMs) and the accompanied incidences of kidney injury have attracted considerable interest in investigating the responsible toxic ingredients. It is challenging to evaluate toxicity of TCMs since they contain complex mixtures of phytochemicals. Quantitative structure-activity relationship (QSAR) is an efficient tool to predict toxicity and QSAR study on TCMs-induced nephrotoxicity remains lacked. We developed QSAR models using three datasets of 609 compounds: natural products, drugs, and mixed (contained both kinds of data) datasets. Each dataset was used for modelling by utilizing artificial neural networks (ANN) and support vector machines (SVM) algorithms separately. Both internal and external validations were performed on each model. Six QSAR models were developed and yielded reliable performance in the internal validation. For external validation, 30 ingredients in the TCMs were predicted well by the natural product models (accuracy: ANN 96.7%, SVM 93.3%). The mixed models (accuracy: ANN 76.7%, SVM 66.7%) showed a better performance than the drug models (accuracy: ANN 50%, SVM 53.3%). Particularly, natural product models produced the most reliable results. It has the application not only on screening the nephrotoxic ingredients in TCMs, but it is also helpful at prioritizing the subsequent toxicity testing of natural products.

Stacked Generalization with Applicability Domain Outperforms Simple QSAR on in Vitro Toxicological Data

The development of in silico tools able to predict bioactivity and toxicity of chemical substances is a powerful solution envisioned to assess toxicity as early as possible. To enable the development of such tools, the ToxCast program has generated and made publicly available in vitro bioactivity data for thousands of compounds. The goal of the present study is to characterize and explore the data from ToxCast in terms of Machine Learning capability. For this, a large scale analysis on the entire database has been performed to build models to predict bioactivities measured in in vitro assays. Simple classical QSAR algorithms (ANN, SVM, LDA, random forest, and Bayesian) were first applied on the data, and the results of these algorithms suggested that they do not seem to be well-suited for data sets with a high proportion of inactive compounds. The study then showed for the first time that the use of an ensemble method named "Stacked generalization" could improve the model performance on this type of data. Indeed, for 61% of 483 models, the Stacked method led to models with higher performance. Moreover, the combination of this ensemble method with an applicability domain filter allows one to assess the reliability of the predictions for further compound prioritization. In particular we showed that for 50% of the models, the ROC score is better if we do not consider the compounds that are not within the applicability domain.

Computational in Vitro Toxicology Uncovers Chemical Structures Impairing Mitochondrial Membrane Potential

Technological advances in molecular biology have enabled high-throughput screening (HTS) of large chemical libraries. These approaches have provided valuable toxicity data for many physiological responses, including mitochondrial dysfunction. While several quantitative structure-activity relationship (QSAR) models have been developed for mitochondrial dysfunction, there remains a need to identify specific chemical features associated with this response. Thus, the objective of this study was to identify chemical structures associated with altered mitochondrial membrane potential (MMP). To achieve this, we developed computational models to examine the relationship between specific chemotypes (e.g., ToxPrints) and bioactivity in ToxCast/Tox21 HTS assays for altered MMP. The analysis revealed that the "bond:COH_alcohol_aromatic", "bond:COH_alcohol_aromatic_phenol", and "ring:aromatic_benzene" ToxPrints had the highest average correlations (phi coefficient) with ToxCast/Tox21 assay component endpoints for decreased MMP. These structures also constituted a "core" group of ToxPrints for decreased MMP in a force-directed network model and were the most important chemotypes in a random forest (RF) classification model for the "TOX21_MMP_ratio_down" assay component endpoint. Based on multiple lines of evidence, these structures, which are present in numerous chemicals (e.g., aromatic hydrocarbons, pesticides, and industrial chemicals) are likely involved in mitochondrial dysfunction. Because of the hierarchical structure of ToxPrints, these chemotypes were highly convergent and, when excluded from training data, had limited effects on the classification performance as related structures compensated for predictor loss. These results highlight the flexibility of the RF algorithm and ToxPrints for QSAR modeling, which is useful to identify chemicals affecting mitochondrial function.

Identifying Attributes That Influence In Vitro-to-In Vivo Concordance by Comparing In Vitro Tox21 Bioactivity Versus In Vivo DrugMatrix Transcriptomic Responses Across 130 Chemicals

Recent efforts aimed at integrating in vitro high-throughput screening (HTS) data into chemical toxicity assessments are necessitating increased understanding of concordance between chemical-induced responses observed in vitro versus in vivo. This investigation set out to (1) measure concordance between in vitro HTS data and transcriptomic responses observed in vivo, focusing on the liver, and (2) identify attributes that can influence concordance. Signal response profiles from 130 substances were compared between in vitro data produced through Tox21 and liver transcriptomic data through DrugMatrix, collected from rats exposed to a chemical for ≤5 days. A global in vitro-to-in vivo comparative analysis based on pathway-level responses resulted in an overall average percent agreement of 79%, ranging on a per-chemical basis between 41% and 100%. Whereas concordance amongst inactive chemicals was high (89%), concordance amongst chemicals showing in vitro activity was only 13%, suggesting that follow-up in vivo and/or orthogonal in vitro assays would improve interpretations of in vitro activity. Attributes identified to influence concordance included experimental design attributes (eg, cell type), target pathways, and physicochemical properties (eg, logP). The attribute that most consistently increased concordance was dose applicability, evaluated by filtering for experimental doses administered to rats that were within 10-fold of those related to likely bioactivity, derived using Tox21 data and high-throughput toxicokinetic modeling. Together, findings suggest that in vitro screening approaches to predict in vivo toxicity are viable particularly when certain attributes are considered, including whether activity versus inactivity is observed, experimental design, chemical properties, and dose applicability.

Development of an Infrastructure for the Prediction of Biological Endpoints in Industrial Environments. Lessons Learned at the eTOX Project

In silico methods are increasingly being used for assessing the chemical safety of substances, as a part of integrated approaches involving in vitro and in vivo experiments. A paradigmatic example of these strategies is the eTOX project http://www.etoxproject.eu, funded by the European Innovative Medicines Initiative (IMI), which aimed at producing high quality predictions of in vivo toxicity of drug candidates and resulted in generating about 200 models for diverse endpoints of toxicological interest. In an industry-oriented project like eTOX, apart from the predictive quality, the models need to meet other quality parameters related to the procedures for their generation and their intended use. For example, when the models are used for predicting the properties of drug candidates, the prediction system must guarantee the complete confidentiality of the compound structures. The interface of the system must be designed to provide non-expert users all the information required to choose the models and appropriately interpret the results. Moreover, procedures like installation, maintenance, documentation, validation and versioning, which are common in software development, must be also implemented for the models and for the prediction platform in which they are implemented. In this article we describe our experience in the eTOX project and the lessons learned after 7 years of close collaboration between industrial and academic partners. We believe that some of the solutions found and the tools developed could be useful for supporting similar initiatives in the future.

Relevance and Application of Read-Across - Mini Review of European Consensus Platform for Alternatives and Scandinavian Society for Cell Toxicology 2017 Workshop Session

Structure activity relationship (SAR)-based read-across is an effective approach for addressing data gaps in human health risk assessment for 'data-poor' chemicals. In read-across, available data on chemical structural analogues are used to predict the toxicity potential of the data-poor chemical. This approach has long been recognized by regulatory agencies and used by industry to evaluate the hazards of chemicals for which there are limited direct data. Construction of a scientifically robust SAR-based read-across hazard assessment is a complex and iterative process involving multiple considerations in each step. Traditional in vivo data generated using regulatory guideline compliant study designs typically forms the basis for read-across assessments. Recently, however, new data streams have been explored and incorporated to enhance read-across predictivity. These in vitro and omics data streams may be used in different ways involving identification of hazards or to provide insight into modes of action for observed toxicological responses. These 'new approach methods' can also enable comparison of biological responses across analogues or a category of structurally related chemicals in order to establish a pattern of biological similarity in addition chemical similarity and/or to help address potency differences across a category. The purpose of this workshop session was to inform on practical considerations in conducting SAR-based read-across assessments and to review recent activities related to application of new approach methods to the practice of read-across.