Search

Regulatory agencies rely upon rodent in vivo acute oral toxicity data to determine hazard categorization, require appropriate precautionary labeling, and perform quantitative risk assessments. As the field of toxicology moves toward animal-free new approach methodologies (NAMs), there is a pressing need to develop a reliable, robust reference data set to characterize the reproducibility and inherent variability in the in vivo acute oral toxicity test method, which would serve to contextualize results and set expectations regarding NAM performance. Such a data set is also needed for training and evaluating computational models. To meet these needs, rat acute oral LD50 data from multiple databases were compiled, curated, and analyzed to characterize variability and reproducibility of results across a set of up to 2441 chemicals with multiple independent study records. Conditional probability analyses reveal that replicate studies only result in the same hazard categorization on average at 60% likelihood. Although we did not have sufficient study metadata to evaluate the impact of specific protocol components (eg, strain, age, or sex of rat, feed used, treatment vehicle, etc.), studies were assumed to follow standard test guidelines. We investigated, but could not attribute, various chemical properties as the sources of variability (ie, chemical structure, physiochemical properties, functional use). Thus, we conclude that inherent biological or protocol variability likely underlies the variance in the results. Based on the observed variability, we were able to quantify a margin of uncertainty of ±0.24 log10 (mg/kg) associated with discrete in vivo rat acute oral LD50 values.

The developmental toxicity potential (dTP) concentration from the devTOX quickPredict (devTOXVPA and a series of structural analogues were tested with the devTOXThe EAD estimates for the VPA analogues based on different PK/PBPK models were quantitatively similar to in vivo data from both rats and humans, where available, and the derived rank order of the chemicals was consistent with observed in vivo developmental toxicity. Different models were identified that provided accurate predictions for rat prenatal LELs and conservative estimates of human safe exposure. The impact of in vitro kinetics on EAD estimates is chemical-dependent. EADs from this study were within range of predicted doses from other in vitro and model organism data.This study highlights the importance of pharmacokinetic considerations when using in vitro assays and demonstrates the utility of the devTOX

The U.S. federal consortium on toxicology in the 21st century (Tox21) produces quantitative, high-throughput screening (HTS) data on thousands of chemicals across a wide range of assays covering critical biological targets and cellular pathways. Many of these assays, and those used in other in vitro screening programs, rely on luciferase and fluorescence-based readouts that can be susceptible to signal interference by certain chemical structures resulting in false positive outcomes. Included in the Tox21 portfolio are assays specifically designed to measure interference in the form of luciferase inhibition and autofluorescence via multiple wavelengths (red, blue, and green) and under various conditions (cell-free and cell-based, two cell types). Out of 8,305 chemicals tested in the Tox21 interference assays, percent actives ranged from 0.5% (red autofluorescence) to 9.9% (luciferase inhibition). Self-organizing maps and hierarchical clustering were used to relate chemical structural clusters to interference activity profiles. Multiple machine learning algorithms were applied to predict assay interference based on molecular descriptors and chemical properties. The best performing predictive models (accuracies of ~80%) have been included in a web-based tool called InterPred that will allow users to predict the likelihood of assay interference for any new chemical structure and thus increase confidence in HTS data by decreasing false positive testing results.

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.

Efforts are underway to develop and implement nonanimal approaches which can characterize acute systemic lethality. A workshop was held in October 2019 to discuss developments in the prediction of acute oral lethality for chemicals and mixtures, as well as progress and needs in the understanding and modeling of mechanisms of acute lethality. During the workshop, each speaker led the group through a series of charge questions to determine clear next steps to progress the aims of the workshop. Participants concluded that a variety of approaches will be needed and should be applied in a tiered fashion. Non-testing approaches, including waiving tests, computational models for single chemicals, and calculating the acute lethality of mixtures based on the LD50 values of mixture components, could be used for some assessments now, especially in the very toxic or non-toxic classification ranges. Agencies can develop policies indicating contexts under which mathematical approaches for mixtures assessment are acceptable; to expand applicability, poorly predicted mixtures should be examined to understand discrepancies and adapt the approach. Transparency and an understanding of the variability of in vivo approaches are crucial to facilitate regulatory application of new approaches. In a replacement strategy, mechanistically based in vitro or in silico models will be needed to support non-testing approaches especially for highly acutely toxic chemicals. The workshop discussed approaches that can be used in the immediate or near term for some applications and identified remaining actions needed to implement approaches to fully replace the use of animals for acute systemic toxicity testing.

High-throughput screening (HTS) research programs for drug development or chemical hazard assessment are designed to screen thousands of molecules across hundreds of biological targets or pathways. Most HTS platforms use fluorescence and luminescence technologies, representing more than 70% of the assays in the US Tox21 research consortium. These technologies are subject to interferent signals largely explained by chemicals interacting with light spectrum. This phenomenon results in up to 5–10% of false positive results, depending on the chemical library used. Here, we present the InterPred webserver (version 1.0), a platform to predict such interference chemicals based on the first large-scale chemical screening effort to directly characterize chemical-assay interference, using assays in the Tox21 portfolio specifically designed to measure autofluorescence and luciferase inhibition. InterPred combines 17 quantitative structure activity relationship (QSAR) models built using optimized machine learning techniques and allows users to predict the probability that a new chemical will interfere with different combinations of cellular and technology conditions. InterPred models have been applied to the entire Distributed Structure-Searchable Toxicity (DSSTox) Database (∼800,000 chemicals). The InterPred webserver is available at https://sandbox.ntp.niehs.nih.gov/interferences/.

Moving towards species-relevant chemical safety assessments and away from animal testing requires access to reliable data to develop and build confidence in new approaches. The Integrated Chemical Environment (ICE) provides tools and curated data centered around chemical safety assessment. This article describes updates to ICE, including improved accessibility and interpretability of in vitro data via mechanistic target mapping and enhanced interactive tools for in vitro to in vivo extrapolation (IVIVE). Mapping of in vitro assay targets to toxicity endpoints of regulatory importance uses literature-based mode-of-action information and controlled terminology from existing knowledge organization systems to support data interoperability with external resources. The most recent ICE update includes Tox21 high-throughput screening data curated using analytical chemistry data and assay-specific parameters to eliminate potential artifacts or unreliable activity. Also included are physicochemical/ADME parameters for over 800,000 chemicals predicted by quantitative structure-activity relationship models. These parameters are used by the new ICE IVIVE tool in combination with the U.S. Environmental Protection Agency’s httk R package to estimate in vivo exposures corresponding to in vitro bioactivity concentrations from stored or user-defined assay data. These new ICE features allow users to explore the applications of an expanded data space and facilitate building confidence in non-animal approaches.

Background:Endocrine disrupting chemicals (EDCs) are xenobiotics that mimic the interaction of natural hormones and alter synthesis, transport, or metabolic pathways. The prospect of EDCs causing adverse health effects in humans and wildlife has led to the development of scientific and regulatory approaches for evaluating bioactivity. This need is being addressed using high-throughput screening (HTS) in vitro approaches and computational modeling.Objectives:In support of the Endocrine Disruptor Screening Program, the U.S. Environmental Protection Agency (EPA) led two worldwide consortiums to virtually screen chemicals for their potential estrogenic and androgenic activities. Here, we describe the Collaborative Modeling Project for Androgen Receptor Activity (CoMPARA) efforts, which follows the steps of the Collaborative Estrogen Receptor Activity Prediction Project (CERAPP).Methods:The CoMPARA list of screened chemicals built on CERAPP’s list of 32,464 chemicals to include additional chemicals of interest, as well as simulated ToxCast™ metabolites, totaling 55,450 chemical structures. Computational toxicology scientists from 25 international groups contributed 91 predictive models for binding, agonist, and antagonist activity predictions. Models were underpinned by a common training set of 1,746 chemicals compiled from a combined data set of 11 ToxCast™/Tox21 HTS in vitro assays.Results:The resulting models were evaluated using curated literature data extracted from different sources. To overcome the limitations of single-model approaches, CoMPARA predictions were combined into consensus models that provided averaged predictive accuracy of approximately 80% for the evaluation set.Discussion:The strengths and limitations of the consensus predictions were discussed with example chemicals; then, the models were implemented into the free and open-source OPERA application to enable screening of new chemicals with a defined applicability domain and accuracy assessment. This implementation was used to screen the entire EPA DSSTox database of ∼875,000 chemicals, and their predicted AR activities have been made available on the EPA CompTox Chemicals dashboard and National Toxicology Program’s Integrated Chemical Environment. https://doi.org/10.1289/EHP5580

Concern that chemicals in the environment can disrupt the endocrine systems of humans and ecological species (fish, frogs, birds) has driven the development of bioassays to test for endocrine activity. In the last decade, there has been increased focus on in vitro and high-throughput screening (HTS) assays and in silico models that allow the rapid evaluation of hundreds to thousands of compounds. We review efforts at the US EPA and the National Toxicology Program to develop assays and models to predict estrogen, androgen and steroidogenesis activity. These approaches include in vitro HTS assays, computational models that link the results of multiple assays and reduce the impact of assay artifacts and noise, and structure-based QSAR models. These approaches have been shown to be robust enough to move from research to application in regulatory risk assessment.

The new paradigm of toxicity testing approaches involves rapid screening of thousands of chemicals across hundreds of biological targets through use of in vitro assays. Such assays may lead to false negatives when the complex metabolic processes that render a chemical bioactive in a living system are unable to be replicated in an in vitro environment. In the current study, a workflow is presented for complementing in vitro testing results with in silico and in vitro techniques to identify inactive parents that may produce active metabolites. A case study applying this workflow involved investigating the influence of metabolism for over 1,400 chemicals considered inactive across 18 in vitro assays related to the estrogen receptor (ER) pathway. Over 7,500 first-generation and second-generation metabolites were generated for these in vitro inactive chemicals using an in silico software program. Next, a consensus model comprised of four individual quantitative structure activity relationship (QSAR) models was used to predict ER-binding activity for each of the metabolites. Binding activity was predicted for ∼8–10% of metabolites in each generation, with these metabolites linked to 259 in vitro inactive parent chemicals. Metabolites were enriched in substructures consisting of alcohol, aromatic, and phenol bonds relative to their inactive parent chemicals, suggesting these features are potentially favorable for ER-binding. The workflow presented here can be used to identify parent chemicals that can be potentially bioactive, to aid confidence in high throughput risk screening.

High-resolution mass spectrometry (HRMS) data has revolutionized the identification of environmental contaminants through non-targeted analysis (NTA). However, chemical identification remains challenging due to the vast number of unknown molecular features typically observed in environmental samples. Advanced data processing techniques are required to improve chemical identification workflows. The ideal workflow brings together a variety of data and tools to increase the certainty of identification. One such tool is chromatographic retention time (RT) prediction, which can be used to reduce the number of possible suspect chemicals within an observed RT window. This paper compares the relative predictive ability and applicability to NTA workflows of three RT prediction models: (1) a logP (octanol-water partition coefficient)-based model using EPI Suite(TM) logP predictions; (2) a commercially available ACD/ChromGenius model; and, (3) a newly developed Quantitative Structure Retention Relationship model called OPERA-RT. Models were developed using the same training set of 78 compounds with experimental RT data and evaluated for external predictivity on an identical test set of 19 compounds. Both the ACD/ChromGenius and OPERA-RT models outperformed the EPI Suite(TM) logP-based RT model (R(2)=0.81–0.92, 0.86–0.83, 0.66–0.69 for training-test sets, respectively). Further, both OPERA-RT and ACD/ChromGenius predicted 95% of RTs within a ± 15% chromatographic time window of experimental RTs. Based on these results, we simulated an NTA workflow with a ten-fold larger list of candidate structures generated for formulae of the known test set chemicals using the U.S. EPA’s CompTox Chemistry Dashboard (https://comptox.epa.gov/dashboard), RTs for all candidates were predicted using both ACD/ChromGenius and OPERA-RT, and RT screening windows were assessed for their ability to filter out unlikely candidate chemicals and enhance potential identification. Compared to ACD/ChromGenius, OPERA-RT screened out a greater percentage of candidate structures within a 3-minute RT window (60% vs. 40%) but retained fewer of the known chemicals (42% vs. 83%). By several metrics, the OPERA-RT model, generated as a proof-of-concept using a limited set of open source data, performed as well as the commercial tool ACD/ChromGenius when constrained to the same small training and test sets. As the availability of RT data increases, we expect the OPERA-RT model’s predictive ability will increase.

An evolving regulatory, scientific, and legislative landscape is driving a fundamental change in how chemical safety decisions are made. As we move to implement changes, regulatory agencies and industry are beginning to adopt tiered approaches, which leverage high-throughput screening technologies for prioritization and read across, followed by interrogation of "hit chemicals" with more rigorous dose-response assessment either in fit-for-purpose human cell-based assays or with traditional in vivo tests. However, to date, suitable in vitro alternatives do not exist for the vast majority of the organ toxicities that form the basis of current regulatory decisions. To successfully support safety decisions, biologically relevant, quantitative, cell-based assays that evaluate dose-response and identify regions of safety for chemical exposure are required. This review evaluates the current state of the science in the development of such assays, identifies key gaps in the current tests, and recommends areas where research efforts may be focused to help move the risk assessment community towards more wide-spread use of in vitro methods. Our analysis suggests that a key shortcoming in the current efforts is the ability to test volatile compounds and to predict pulmonary toxicity. We present a mechanistically-based path forward for the development of a fit-for-purpose lung toxicity assay.

The collection of chemical structure information and associated experimental data for quantitative structure–activity/property relationship (QSAR/QSPR) modeling is facilitated by an increasing number of public databases containing large amounts of useful data. However, the performance of QSAR models highly depends on the quality of the data and modeling methodology used. This study aims to develop robust QSAR/QSPR models for chemical properties of environmental interest that can be used for regulatory purposes. This study primarily uses data from the publicly available PHYSPROP database consisting of a set of 13 common physicochemical and environmental fate properties. These datasets have undergone extensive curation using an automated workflow to select only high-quality data, and the chemical structures were standardized prior to calculation of the molecular descriptors. The modeling procedure was developed based on the five Organization for Economic Cooperation and Development (OECD) principles for QSAR models. A weighted k-nearest neighbor approach was adopted using a minimum number of required descriptors calculated using PaDEL, an open-source software. The genetic algorithms selected only the most pertinent and mechanistically interpretable descriptors (2–15, with an average of 11 descriptors). The sizes of the modeled datasets varied from 150 chemicals for biodegradability half-life to 14,050 chemicals for logP, with an average of 3222 chemicals across all endpoints. The optimal models were built on randomly selected training sets (75%) and validated using fivefold cross-validation (CV) and test sets (25%). The CV Q2 of the models varied from 0.72 to 0.95, with an average of 0.86 and an R2 test value from 0.71 to 0.96, with an average of 0.82. Modeling and performance details are described in QSAR model reporting format and were validated by the European Commission’s Joint Research Center to be OECD compliant. All models are freely available as an open-source, command-line application called OPEn structure–activity/property Relationship App (OPERA). OPERA models were applied to more than 750,000 chemicals to produce freely available predicted data on the U.S. Environmental Protection Agency’s CompTox Chemistry Dashboard. Electronic supplementary material The online version of this article (10.1186/s13321-018-0263-1) contains supplementary material, which is available to authorized users.

Chemical ionization plays an important role in many aspects of pharmacokinetic (PK) processes such as protein binding, tissue partitioning, and apparent volume of distribution at steady state (Vdss). Here, estimates of ionization equilibrium constants (i.e., pKa) were analyzed for 8,132 pharmaceuticals and 24,281 other compounds to which humans might be exposed in the environment. Results revealed broad differences in the ionization of pharmaceutical chemicals and chemicals with either near-field (in the home) or far-field sources. The utility of these high-throughput ionization predictions was evaluated via a case-study of predicted PK Vdss for 22 compounds monitored in the blood and serum of the U.S. population by the U.S. Centers for Disease Control and Prevention National Health and Nutrition Examination Survey (NHANES). The chemical distribution ratio between water and tissue was estimated using predicted ionization states characterized by pKa. Probability distributions corresponding to ionizable atom types (IATs) were then used to analyze the sensitivity of predicted Vdss on predicted pKa using Monte Carlo methods. 8 of the 22 compounds were predicted to be ionizable. For 5 of the 8 the predictions based upon ionization are significantly different from what would be predicted for a neutral compound. For all but one (foramsulfuron), the probability distribution of predicted Vdss generated by IAT sensitivity analysis spans both the neutral prediction and the prediction using ionization. As new data sets of chemical-specific information on metabolism and excretion for hundreds of chemicals are being made available (e.g., Wetmore et al., 2015), high-throughput methods for calculating Vdss and tissue-specific PK distribution coefficients will allow the rapid construction of PK models to provide context for both biomonitoring data and high-throughput toxicity screening studies such as Tox21 and ToxCast.

New approach methodologies (NAMs) for toxicological applications such as in vitro assays and in silico models generate data that can be useful for assessing potential health impacts of chemicals. The National Toxicology Program’s (NTP’s) Integrated Chemical Environment (ICE; https://ice.ntp.niehs.nih.gov/ ) provides user-friendly access to NAM data and tools to explore and contextualize chemical bioactivity and molecular properties. ICE contains curated in vivo and in vitro toxicity testing data and experimental physicochemical property data gathered from different literature sources. ICE also contains computationally generated toxicity data and physicochemical parameter predictions. ICE provides interactive computational tools that characterize, analyze, and predict bioactivity for user-defined chemicals. ICE Search allows users to select and merge data sets for lists of chemicals and mixtures, yielding summary-level information, curated reference data, and bioactivity details mapped to mechanistic targets and modes of action. With the Curve Surfer tool, the user can explore concentration–response relationships of curated high-throughput screening assays. The Physiologically Based Pharmacokinetics (PBPK) tool predicts tissue-level concentrations resulting from in vivo doses, while the In Vitro–In Vivo Extrapolation (IVIVE) tool translates in vitro activity concentrations to equivalent in vivo dose estimates. The Chemical Characterization tool displays distributions of physicochemical properties, bioactivity- and structure-based projections, and consumer product use information. Chemical Quest, the newest ICE tool, allows users to search for structurally similar chemicals to a target chemical or substructure from within the extensive ICE database. Retrieved information on target chemicals and those with similar structures can then be used to query other ICE tools and datasets, greatly expanding data available to address the user’s question. ICE links to other NTP and U.S. Environmental Protection Agency data sources, expanding ICE’s capacity to examine chemicals based on physicochemical properties, bioactivity, and product use categories.

Read-across is an important data gap filling technique used within category and analog approaches for regulatory hazard identification and risk assessment. Although much technical guidance is available that describes how to develop category/analog approaches, practical principles to evaluate and substantiate analog validity (suitability) are still lacking. This case study uses hindered phenols as an example chemical class to determine: (1) the capability of three structure fingerprint/descriptor methods (PubChem, ToxPrints and MoSS MCSS) to identify analogs for read-across to predict Estrogen Receptor (ER) binding activity and, (2) the utility of data confidence measures, physicochemical properties, and chemical R-group properties as filters to improve ER binding predictions. The training dataset comprised 462 hindered phenols and 257 non- hindered phenols. For each chemical of interest (target), source analogs were identified from two datasets (hindered and non-hindered phenols) that had been characterized by a fingerprint/descriptor method and by two cut-offs: (1) minimum similarity distance (range: 0.1 – 0.9) and, (2) N closest analogs (range: 1 – 10). Analogs were then filtered using: (1) physicochemical properties of the phenol (termed global filtering) and, (2) physicochemical properties of the R-groups neighboring the active hydroxyl group (termed local filtering). A read-across prediction was made for each target chemical on the basis of a majority vote of the N closest analogs. The results demonstrate that: (1) concordance in ER activity increases with structural similarity, regardless of the structure fingerprint/descriptor method, (2) increased data confidence significantly improves read-across predictions, and (3) filtering analogs using global and local properties can help identify more suitable analogs. This case study illustrates that the quality of the underlying experimental data and use of endpoint relevant chemical descriptors to evaluate source analogs are critical to achieving robust read-across predictions.

In an effort to address a major challenge in chemical safety assessment, alternative approaches for characterizing systemic effect levels, a predictive model was developed. Systemic effect levels were curated from ToxRefDB, HESS-DB and COSMOS-DB from numerous study types totaling 4379 in vivo studies for 1247 chemicals. Observed systemic effects in mammalian models are a complex function of chemical dynamics, kinetics, and inter- and intra-individual variability. To address this complex problem, systemic effect levels were modeled at the study-level by leveraging study covariates (e.g., study type, strain, administration route) in addition to multiple descriptor sets, including chemical (ToxPrint, PaDEL, and Physchem), biological (ToxCast), and kinetic descriptors. Using random forest modeling with cross-validation and external validation procedures, study-level covariates alone accounted for approximately 15% of the variance reducing the root mean squared error (RMSE) from 0.96 log10 to 0.85 log10 mg/kg/day, providing a baseline performance metric (lower expectation of model performance). A consensus model developed using a combination of study-level covariates, chemical, biological, and kinetic descriptors explained a total of 43% of the variance with an RMSE of 0.69 log10 mg/kg/day. A benchmark model (upper expectation of model performance) was also developed with an RMSE of 0.5 log10 mg/kg/day by incorporating study-level covariates and the mean effect level per chemical. To achieve a representative chemical-level prediction, the minimum study-level predicted and observed effect level per chemical were compared reducing the RMSE from 1.0 to 0.73 log10 mg/kg/day, equivalent to 87% of predictions falling within an order-of-magnitude of the observed value. Although biological descriptors did not improve model performance, the final model was enriched for biological descriptors that indicated xenobiotic metabolism gene expression, oxidative stress, and cytotoxicity, demonstrating the importance of accounting for kinetics and non-specific bioactivity in predicting systemic effect levels. Herein, we generated an externally predictive model of systemic effect levels for use as a safety assessment tool and have generated forward predictions for over 30,000 chemicals. Electronic supplementary material The online version of this article (doi:10.1007/s00204-017-2067-x) contains supplementary material, which is available to authorized users.

There are little available toxicity data on the vast majority of chemicals in commerce. High-throughput screening (HTS) studies, such as being carried out by the U.S. Environmental Protection Agency (EPA) ToxCast program in partnership with the federal Tox21 research program, can generate biological data to inform models for predicting potential toxicity. However, physicochemical properties are also needed to model environmental fate and transport, as well as exposure potential. The purpose of the present study was to generate an open-source Quantitative Structure-Property Relationship (QSPR) workflow to predict a variety of physicochemical properties that would have cross-platform compatibility to integrate into existing cheminformatics workflows. In this effort, decades-old experimental property data sets available within EPA EPI Suite™ were reanalyzed using modern cheminformatics workflows to build updated QSPR models capable of supplying computationally efficient, open, and transparent HTS property predictions in support of environmental modeling efforts. Models were built using updated EPI Suite data sets for the prediction of six physicochemical properties: octanol-water partition coefficient (log P), water solubility (log S), boiling point (BP), melting point (MP), vapor pressure (log VP) and bioconcentration factor (log BCF). The coefficient of determination (R2) between the estimated values and experimental data for the six predicted properties ranged from 0.826 (MP) to 0.965 (BP), with model performance for five of the six properties exceeding those from the original EPI Suite™ models. The newly derived models can be employed for rapid estimation of physicochemical properties within an open-source HTS workflow to inform fate and toxicity prediction models of environmental chemicals.

The median lethal dose for rodent oral acute toxicity (LD50) is a standard piece of information required to categorize chemicals in terms of the potential hazard posed to human health after acute exposure. The exclusive use of in vivo testing is limited by the time and costs required for performing experiments and by the need to sacrifice a number of animals. (Quantitative) structure–activity relationships [(Q)SAR] proved a valid alternative to reduce and assist in vivo assays for assessing acute toxicological hazard. In the framework of a new international collaborative project, the NTP Interagency Center for the Evaluation of Alternative Toxicological Methods and the U.S. Environmental Protection Agency’s National Center for Computational Toxicology compiled a large database of rat acute oral LD50 data, with the aim of supporting the development of new computational models for predicting five regulatory relevant acute toxicity endpoints. In this article, a series of regression and classification computational models were developed by employing different statistical and knowledge-based methodologies. External validation was performed to demonstrate the real-life predictability of models. Integrated modeling was then applied to improve performance of single models. Statistical results confirmed the relevance of developed models in regulatory frameworks, and confirmed the effectiveness of integrated modeling. The best integrated strategies reached RMSEs lower than 0.50 and the best classification models reached balanced accuracies over 0.70 for multi-class and over 0.80 for binary endpoints. Computed predictions will be hosted on the EPA’s Chemistry Dashboard and made freely available to the scientific community.

Screening certain environmental chemicals for their ability to interact with endocrine targets, including the androgen receptor (AR), is an important global concern. We previously developed a model using a battery of eleven in vitro AR assays to predict in vivo AR activity. Here we describe a revised mathematical modeling approach that also incorporates data from newly available assays and demonstrate that subsets of assays can provide close to the same level of predictivity. These subset models are evaluated against the full model using 1820 chemicals, as well as in vitro and in vivo reference chemicals from the literature. Agonist batteries of as few as six assays and antagonist batteries of as few as five assays can yield balanced accuracies of 95% or better relative to the full model. Balanced accuracy for predicting reference chemicals is 100%. An approach is outlined for researchers to develop their own subset batteries to accurately detect AR activity using assays that map to the pathway of key molecular and cellular events involved in chemical-mediated AR activation and transcriptional activity. This work indicates in vitro bioactivity and in silico predictions that map to the AR pathway could be used in an integrated approach to testing and assessment for identifying chemicals that interact directly with the mammalian AR.

Chemical risk assessment relies on toxicity tests that require significant numbers of animals, time and costs. For the >30,000 chemicals in commerce, the current scale of animal testing is insufficient to address chemical safety concerns as regulatory and product stewardship considerations evolve to require more comprehensive understanding of potential biological effects, conditions of use, and associated exposures. We demonstrate the use of a multi-level new approach methodology (NAMs) strategy for hazard- and risk-based prioritization to reduce animal testing. A Level 1/2 chemical prioritization based on estrogen receptor (ER) activity and metabolic activation using ToxCast data was used to select 112 chemicals for testing in a Level 3 human uterine cell estrogen response assay (IKA assay). The Level 3 data were coupled with quantitative in vitro to in vivo extrapolation (Q-IVIVE) to support bioactivity determination (as a surrogate for hazard) in a tissue-specific context. Assay AC50s and Q-IVIVE were used to estimate human equivalent doses (HEDs), and HEDs were compared to rodent uterotrophic assay in vivo-derived points of departure (PODs). For substances active both in vitro and in vivo, IKA assay-derived HEDs were lower or equivalent to in vivo PODs for 19/23 compounds (83%). Activity exposure relationships were calculated, and the IKA assay was as or more protective of human health than the rodent uterotrophic assay for all IKA-positive compounds. This study demonstrates the utility of biologically relevant fit-for-purpose assays and supports the use of a multi-level strategy for chemical risk assessment.

Humans are exposed to large numbers of environmental chemicals, some of which potentially interfere with the endocrine system. The identification of potential endocrine disrupting chemicals (EDCs) has gained increasing priority in the assessment of environmental hazards. The U.S. Environmental Protection Agency (U.S. EPA) has developed the Endocrine Disruptor Screening Program (EDSP) which aims to prioritize and screen potential EDCs. The Toxicity Forecaster (ToxCast) program has generated data using in vitro high-throughput screening (HTS) assays measuring activity of chemicals at multiple points along the androgen receptor (AR) activity pathway. In the present study, using a large and diverse data set of 1667 chemicals provided by the U.S. EPA from the combined ToxCast AR assays in the framework of the Collaborative Modeling Project for Androgen Receptor Activity (CoMPARA). Two models were built using ADMET Predictor™; one is based on Artificial Neural Networks (ANNs) technology and the other uses a Support Vector Machine (SVM) algorithm; one model is a Decision Tree (DT) developed in R; and two models make use of differently combined sets of structural alerts (SAs) automatically extracted by SARpy. We used two strategies to integrate predictions from single models; one is based on a majority vote approach and the other on prediction convergence. These strategies led to enhanced statistical performance in most cases. Moreover, the majority vote approach improved prediction coverage when one or more single models were not able to provide any estimations. This study integrates multiple in silico approaches as a virtual screening tool for use in risk assessment of endocrine disrupting chemicals.

In early 2018, the Interagency Coordinating Committee for the Validation of Alternative Methods (ICCVAM) published the "Strategic Roadmap for Establishing New Approaches to Evaluate the Safety of Chemicals and Medical Products in the United States" (ICCVAM 2018). Cross-agency federal workgroups have been established to implement this roadmap for various toxicological testing endpoints, with an initial focus on acute toxicity testing. The ICCVAM acute toxicity workgroup (ATWG) helped organize a global collaboration to build predictive in silico models for acute oral systemic toxicity, based on a large dataset of rodent studies and targeted towards regulatory needs identified across federal agencies. Thirty-two international groups across government, industry, and academia participated in the project, culminating in a workshop in April 2018 held at the National Institutes of Health (NIH). At the workshop, computational modelers and regulatory decision makers met to discuss the feasibility of using predictive model outputs for regulatory use in lieu of acute oral systemic toxicity testing. The models were combined to yield consensus predictions which demonstrated excellent performance when compared to the animal data, and workshop outcomes and follow-up activities to make these tools available and put them into practice are discussed here.

Efficient high-throughput transcriptomics (HTT) tools promise inexpensive, rapid assessment of possible biological consequences of human and environmental exposures to tens of thousands of chemicals in commerce. HTT systems have used relatively small sets of gene expression measurements coupled with mathematical prediction methods to estimate genome-wide gene expression and are often trained and validated using pharmaceutical compounds. It is unclear whether these training sets are suitable for general toxicity testing applications and the more diverse chemical space represented by commercial chemicals and environmental contaminants. In this work, we built predictive computational models that inferred whole genome transcriptional profiles from a smaller sample of surrogate genes. The model was trained and validated using a large scale toxicogenomics database with gene expression data from exposure to heterogeneous chemicals from a wide range of classes (the Open TG-GATEs data base). The method of predictor selection was designed to allow high fidelity gene prediction from any pre-existing gene expression data set, regardless of animal species or data measurement platform. Predictive qualitative models were developed with this TG-GATES data that contained gene expression data of human primary hepatocytes with over 941 samples covering 158 compounds. A sequential forward search-based greedy algorithm, combining different fitting approaches and machine learning techniques, was used to find an optimal set of surrogate genes that predicted differential expression changes of the remaining genome. We then used pathway enrichment of up-regulated and down-regulated genes to assess the ability of a limited gene set to determine relevant patterns of tissue response. In addition, we compared prediction performance using the surrogate genes found from our greedy algorithm (referred to as the SV2000) with the landmark genes provided by existing technologies such as L1000 (Genometry) and S1500 (Tox21), finding better predictive performance for the SV2000. The ability of these predictive algorithms to predict pathway level responses is a positive step toward incorporating mode of action (MOA) analysis into the high throughput prioritization and testing of the large number of chemicals in need of safety evaluation.

Chemical database searching has become a fixture in many non-targeted identification workflows based on high-resolution mass spectrometry (HRMS). However, the form of a chemical structure observed in HRMS does not always match the form stored in a database (e.g., the neutral form versus a salt; one component of a mixture rather than the mixture form used in a consumer product). Linking the form of a structure observed via HRMS to its related form(s) within a database will enable the return of all relevant variants of a structure, as well as the related metadata, in a single query. A Konstanz Information Miner (KNIME) workflow has been developed to produce structural representations observed using HRMS (“MS-Ready structures”) and links them to those stored in a database. These MS-Ready structures, and associated mappings to the full chemical representations, are surfaced via the US EPA’s Chemistry Dashboard (https://comptox.epa.gov/dashboard/). This article describes the workflow for the generation and linking of ~ 700,000 MS-Ready structures (derived from ~ 760,000 original structures) as well as download, search and export capabilities to serve structure identification using HRMS. The importance of this form of structural representation for HRMS is demonstrated with several examples, including integration with the in silico fragmentation software application MetFrag. The structures, search, download and export functionality are all available through the CompTox Chemistry Dashboard, while the MetFrag implementation can be viewed at https://msbi.ipb-halle.de/MetFragBeta/. Electronic supplementary material The online version of this article (10.1186/s13321-018-0299-2) contains supplementary material, which is available to authorized users.

In August 2015, the US Environmental Protection Agency (EPA) convened a workshop entitled "Advancing non-targeted analyses of xenobiotic chemicals in environmental and biological media." The purpose of the workshop was to bring together the foremost experts in non-targeted analysis (NTA) to discuss the state-of-the-science for generating, interpreting, and exchanging NTA measurement data. During the workshop, participants discussed potential designs for a collaborative project that would use EPA resources, including the ToxCast library of chemical substances, the DSSTox database, and the CompTox Chemicals Dashboard, to evaluate cutting-edge NTA methods. That discussion was the genesis of EPA's Non-Targeted Analysis Collaborative Trial (ENTACT). Nearly 30 laboratories have enrolled in ENTACT and used a variety of chromatography, mass spectrometry, and data processing approaches to characterize ten synthetic chemical mixtures, three standardized media (human serum, house dust, and silicone band) extracts, and thousands of individual substances. Initial results show that nearly all participants have detected and reported more compounds in the mixtures than were intentionally added, with large inter-lab variability in the number of reported compounds. A comparison of gas and liquid chromatography results shows that the majority (45.3%) of correctly identified compounds were detected by only one method and 15.4% of compounds were not identified. Finally, a limited set of true positive identifications indicates substantial differences in observable chemical space when employing disparate separation and ionization techniques as part of NTA workflows. This article describes the genesis of ENTACT, all study methods and materials, and an analysis of results submitted to date. Graphical abstract ᅟ.

Because of their broad biological coverage and increasing affordability transcriptomic technologies have increased our ability to evaluate cellular response to chemical stressors, providing a potential means of evaluating chemical response while decreasing dependence on apical endpoints derived from traditional long-term animal studies. It has recently been suggested that dose-response modeling of transcriptomic data may be incorporated into risk assessment frameworks as a means of approximating chemical hazard. However, identification of mode of action from transcriptomics lacks a similar systematic framework. To this end, we developed a web-based interactive browser—MoAviz—that allows visualization of perturbed pathways. We populated this browser with expression data from a large public toxicogenomic database (TG-GATEs). We evaluated the extent to which gene expression changes from in-life exposures could be associated with mode of action by developing a novel similarity index—the Modified Jaccard Index (MJI)—that provides a quantitative description of genomic pathway similarity (rather than gene level comparison). While typical compound-compound similarity is low (median MJI = 0.026), clustering of the TG-GATES compounds identifies groups of similar chemistries. Some clusters aggregated compounds with known similar modes of action, including PPARa agonists (median MJI = 0.315) and NSAIDs (median MJI = 0.322). Analysis of paired in vitro (hepatocyte)-in vivo (liver) experiments revealed systematic patterns in the responses of model systems to chemical stress. Accounting for these model-specific, but chemical-independent, differences improved pathway concordance by 36% between in vivo and in vitro models.

The ToxCast program has generated in vitro screening data on over a thousand chemicals to assess potential disruption of important biological processes and assist in hazard identification and chemical testing prioritization. Few results have been reported for complex mixtures. To extend these ToxCast efforts to mixtures, we tested extracts from 30 organically grown fruits and vegetables in concentration-response in the BioMAP® assays. BioMAP systems use human primary cells primed with endogenous pathway activators to identify phenotypic perturbations related to proliferation, inflammation, immunomodulation, and tissue remodeling. Clustering of bioactivity profiles revealed separation of these produce extracts and ToxCast chemicals. Produce extracts elicited 87 assay endpoint responses per item compared to 20 per item for ToxCast chemicals. On a molar basis, the produce extracts were 10 to 50-fold less potent and when constrained to the maximum testing concentration of the ToxCast chemicals, the produce extracts did not show activity in as many assay endpoints. Using intake adjusted measures of dose, the bioactivity potential was higher for produce extracts than for agrichemicals, as expected based on the comparatively small amounts of agrichemical residues present on conventionally grown produce. The evaluation of BioMAP readouts and the dose responses for produce extracts showed qualitative and quantitative differences from results with single chemicals, highlighting challenges in the interpretation of bioactivity data and dose-response from complex mixtures.

The structures and physicochemical properties of chemicals are important for determining their potential toxicological effects, toxicokinetics, and route(s) of exposure. These data are needed to prioritize the risk for thousands of environmental chemicals, but experimental values are often lacking. In an attempt to efficiently fill data gaps in physicochemical property information, we generated new data for 200 structurally diverse compounds, which were rigorously selected from the USEPA ToxCast chemical library, and whose structures are available within the Distributed Structure-Searchable Toxicity Database (DSSTox). This pilot study evaluated rapid experimental methods to determine five physicochemical properties, including the log of the octanol:water partition coefficient (known as log(Kow) or logP), vapor pressure, water solubility, Henry's law constant, and the acid dissociation constant (pKa). For most compounds, experiments were successful for at least one property; log(Kow) yielded the largest return (176 values). It was determined that 77 ToxPrint structural features were enriched in chemicals with at least one measurement failure, indicating which features may have played a role in rapid method failures. To gauge consistency with traditional measurement methods, the new measurements were compared with previous measurements (where available). Since quantitative structure-activity/property relationship (QSAR/QSPR) models are used to fill gaps in physicochemical property information, 5 suites of QSPRs were evaluated for their predictive ability and chemical coverage or applicability domain of new experimental measurements. The ability to have accurate measurements of these properties will facilitate better exposure predictions in two ways: 1) direct input of these experimental measurements into exposure models; and 2) construction of QSPRs with a wider applicability domain, as their predicted physicochemical values can be used to parameterize exposure models in the absence of experimental data.

International audience; Nowadays, the massive use of new heterogeneous mobile devices and technologies for discovering, selecting and composing cloud smart services has led a trade-off between costs and improved quality of services (e.g., fast response time, low cost, improved security, the reduction of energy consumption, and considerable emissions of carbon). This trade-off has led most cloud service providers to call for new intelligent, faster, and energy-saving services selection and composition solutions. In order to make the cloud computing more attractive by the smart application, it is compulsory to provide best services that users can be satisfied once using them. This paper aims to propose a new generic green cloud service context-aware ontology to manage a large number of heterogeneous cloud services that are grouped semantically according to their service category, functional, and QoS descriptions. We propose also dynamic trust semantic-based bio-inspired selection algorithm that fits user's functional needs and QoS preferences. It focuses on composition process adaptation to context changes (evolution of user's needs and their preferences, energy saving and its execution environment). Also, our approach targets to determine optimal composite service from several relevant cloud smart services results from the selection phase in order to respect the users' global user's needs, energy saving, and quality services' experiences.

We present a practical and easy-to-run in silico workflow exploiting a structure-based strategy making use of docking simulations to derive highly predictive classification models of the androgenic potential of chemicals. Models were trained on a high-quality chemical collection comprising 1689 curated compounds made available within the CoMPARA consortium from the US Environmental Protection Agency and were integrated with a two-step applicability domain whose implementation had the effect of improving both the confidence in prediction and statistics by reducing the number of false negatives. Among the nine androgen receptor X-ray solved structures, the crystal 2PNU (entry code from the Protein Data Bank) was associated with the best performing structure-based classification model. Three validation sets comprising each 2590 compounds extracted by the DUD-E collection were used to challenge model performance and the effectiveness of Applicability Domain implementation. Next, the 2PNU model was applied to screen and prioritize two collections of chemicals. The first is a small pool of 12 representative androgenic compounds that were accurately classified based on outstanding rationale at the molecular level. The second is a large external blind set of 55450 chemicals with potential for human exposure. We show how the use of molecular docking provides highly interpretable models and can represent a real-life option as an alternative nontesting method for predictive toxicology.

Monitored contaminants in drinking water represent a small portion of the total compounds present, many of which may be relevant to human health. To understand the totality of human exposure to compounds in drinking water, broader monitoring methods are imperative. In an effort to more fully characterize the drinking water exposome, point-of-use water filtration devices (Brita(®) filters) were employed to collect time-integrated drinking water samples in a pilot study of nine North Carolina homes. A suspect screening analysis was performed by matching high resolution mass spectra of unknown features to molecular formulas from EPA’s DSSTox database. Candidate compounds with those formulas were retrieved from the EPA’s CompTox Chemistry Dashboard, a recently developed data hub for approximately 720,000 compounds. To prioritize compounds into those most relevant for human health, toxicity data from the US federal collaborative Tox21 program and the EPA ToxCast program, as well as exposure estimates from EPA’s ExpoCast program, were used in conjunction with sample detection frequency and abundance to calculate a “ToxPi” score for each candidate compound. From ∼15,000 molecular features in the raw data, 91 candidate compounds were ultimately grouped into the highest priority class for follow up study. Fifteen of these compounds were confirmed using analytical standards including the highest priority compound, 1,2-Benzisothiazolin-3-one, which appeared in 7 out of 9 samples. The majority of the other high priority compounds are not targets of routine monitoring, highlighting major gaps in our understanding of drinking water exposures. General product-use categories from EPA’s CPCat database revealed that several of the high priority chemicals are used in industrial processes, indicating the drinking water in central North Carolina may be impacted by local industries.

The European REACH regulation requires information on ready biodegradation, which is a screening test to assess the biodegradability of chemicals. At the same time REACH encourages the use of alternatives to animal testing which includes predictions from quantitative structure-activity relationship (QSAR) models. The aim of this study was to build QSAR models to predict ready biodegradation of chemicals by using different modeling methods and types of molecular descriptors. Particular attention was given to data screening and validation procedures in order to build predictive models. Experimental values of 1055 chemicals were collected from the webpage of the National Institute of Technology and Evaluation of Japan (NITE): 837 and 218 molecules were used for calibration and testing purposes, respectively. In addition, models were further evaluated using an external validation set consisting of 670 molecules. Classification models were produced in order to discriminate biodegradable and nonbiodegradable chemicals by means of different mathematical methods: k nearest neighbors, partial least squares discriminant analysis, and support vector machines, as well as their consensus models. The proposed models and the derived consensus analysis demonstrated good classification performances with respect to already published QSAR models on biodegradation. Relationships between the molecular descriptors selected in each QSAR model and biodegradability were evaluated. © 2013 American Chemical Society.

The increasing availability of large collections of chemical structures and associated experimental data provides an opportunity to build robust QSAR models for applications in different fields. One common concern is the quality of both the chemical structure information and associated experimental data. Here we describe the development of an automated KNIME workflow to curate and correct errors in the structure and identity of chemicals using the publicly available PHYSPROP physicochemical properties and environmental fate datasets. The workflow first assembles structure-identity pairs using up to four provided chemical identifiers, including chemical name, CASRNs, SMILES, and MolBlock. Problems detected included errors and mismatches in chemical structure formats, identifiers and various structure validation issues, including hypervalency and stereochemistry descriptions. Subsequently, a machine learning procedure was applied to evaluate the impact of this curation process. The performance of QSAR models built on only the highest-quality subset of the original dataset was compared with the larger curated and corrected dataset. The latter showed statistically improved predictive performance. The final workflow was used to curate the full list of PHYSPROP datasets, and is being made publicly available for further usage and integration by the scientific community.

The US Environmental Protection Agency's (EPA) Endocrine Disruptor Screening Program (EDSP) is using in vitro data generated from ToxCast/Tox21 high-throughput screening assays to assess the endocrine activity of environmental chemicals. Considering that in vitro assays may have limited metabolic capacity, inactive chemicals that are biotransformed into metabolites with endocrine bioactivity may be missed for further screening and testing. Therefore, there is a value in developing novel approaches to account for metabolism and endocrine activity of both parent chemicals and their associated metabolites. We used commercially available software to predict metabolites of 50 parent compounds, out of which 38 chemicals are known to have estrogenic metabolites, and 12 compounds and their metabolites are negative for estrogenic activity. Three ER QSAR models were used to determine potential estrogen bioactivity of the parent compounds and predicted metabolites, the outputs of the models were averaged, and the chemicals were then ranked based on the total estrogenicity of the parent chemical and metabolites. The metabolite prediction software correctly identified known estrogenic metabolites for 26 out of 27 parent chemicals with associated metabolite data, and 39 out of 46 estrogenic metabolites were predicted as potential biotransformation products derived from the parent chemical. The QSAR models estimated stronger estrogenic activity for the majority of the known estrogenic metabolites compared to their parent chemicals. Finally, the three models identified a similar set of parent compounds as top ranked chemicals based on the estrogenicity of putative metabolites. This proposed in silico approach is an inexpensive and rapid strategy for the detection of chemicals with estrogenic metabolites and may reduce potential false negative results from in vitro assays.

The U.S. Environmental Protection Agency's (EPA) ToxCast program is testing a large library of Agency-relevant chemicals using in vitro high-throughput screening (HTS) approaches to support the development of improved toxicity prediction models. Launched in 2007, Phase I of the program screened 310 chemicals, mostly pesticides, across hundreds of ToxCast assay end points. In Phase II, the ToxCast library was expanded to 1878 chemicals, culminating in the public release of screening data at the end of 2013. Subsequent expansion in Phase III has resulted in more than 3800 chemicals actively undergoing ToxCast screening, 96% of which are also being screened in the multi-Agency Tox21 project. The chemical library unpinning these efforts plays a central role in defining the scope and potential application of ToxCast HTS results. The history of the phased construction of EPA's ToxCast library is reviewed, followed by a survey of the library contents from several different vantage points. CAS Registry Numbers are used to assess ToxCast library coverage of important toxicity, regulatory, and exposure inventories. Structure-based representations of ToxCast chemicals are then used to compute physicochemical properties, substructural features, and structural alerts for toxicity and biotransformation. Cheminformatics approaches using these varied representations are applied to defining the boundaries of HTS testability, evaluating chemical diversity, and comparing the ToxCast library to potential target application inventories, such as used in EPA's Endocrine Disruption Screening Program (EDSP). Through several examples, the ToxCast chemical library is demonstrated to provide comprehensive coverage of the knowledge domains and target inventories of potential interest to EPA. Furthermore, the varied representations and approaches presented here define local chemistry domains potentially worthy of further investigation (e.g., not currently covered in the testing library or defined by toxicity "alerts") to strategically support data mining and predictive toxicology modeling moving forward.

The US EPA's ToxCast program is screening thousands of chemicals of environmental interest in hundreds of in vitro high-throughput screening (HTS) assays. One goal is to prioritize chemicals for more detailed analyses based on activity in assays that target molecular initiating events (MIEs) of adverse outcome pathways (AOPs). However, the chemical space of interest for environmental exposure is much wider than ToxCast's chemical library. In silico methods such as quantitative structure-activity relationships (QSARs) are proven and cost-effective approaches to predict biological activity for untested chemicals. However, empirical data is needed to build and validate QSARs. ToxCast has developed datasets for about 2000 chemicals ideal for training and testing QSAR models. The overall goal of the present work was to develop QSAR models to fill the data gaps in larger environmental chemical lists. The specific aim of the current work was to build QSAR models for 18 G-protein-coupled receptor (GPCR) assays, part of the aminergic family. Two QSAR modeling strategies were adopted: classification models were developed to separate chemicals into active/non-active classes, and then regression models were built to predict the potency values of the bioassays for the active chemicals. Multiple software programs were used to calculate constitutional, topological, and substructural molecular descriptors from two-dimensional (2D) chemical structures. Model-fitting methods included PLSDA (partial least square discriminant analysis), SVMs (support vector machines), kNNs (k-nearest neighbors), and PLSs (partial least squares). Genetic algorithms (GAs) were applied as a variable selection technique to select the most predictive molecular descriptors for each assay. N-fold cross-validation (CV) coupled with multi-criteria decision-making fitting criteria was used to evaluate the models. Finally, the models were applied to make predictions within the established chemical space limits. The most accurate model was for the bovine nonselective dopamine receptor (bDR_NS) GPCR assay, for which the classification balanced accuracy reached 0.96 in fitting and 0.95 in fivefold CV, with only two latent variables. These results demonstrate the accuracy of QSAR models to predict the biological activity of chemicals specifically for each one of the studied assays.