Chemical Domain of QSAR Models from Atom-Centered Fragments

Unveiling G-protein coupled receptor kinase-5 inhibitors for chronic degenerative diseases: Multilayered prioritization employing explainable machine learning-driven multi-class QSAR, ligand-based pharmacophore and free energy-inspired molecular simulation

GRK5 holds a pivotal role in cellular signaling pathways, with its overexpression in cardiomyocytes, neuronal cells, and tumor cells strongly associated with various chronic degenerative diseases, which highlights the urgent need for potential inhibitors. In this study, multiclass classification-based QSAR models were developed using diverse machine learning algorithms. These models were built from curated compounds with experimentally derived GRK5 inhibitory activity. Additionally, a pharmacophore model was constructed using active compounds from the dataset. Among the models, the SVM-based approach proved most effective and was initially used to screen DrugBank compounds within the applicability domain. Compounds showing significant GRK5 inhibitory potential underwent evaluation for key pharmacophoric features. Prospective compounds were subjected to molecular docking to assess binding affinity towards GRK5's key active site amino acid residues. Stability at the binding site was analyzed through 200 ns molecular dynamics simulations. MM-GBSA analysis quantified individual free energy components contributing to the total binding energy with respect to binding site residues. Metadynamics analysis, including PCA, FEL, and PDF, provided crucial insights into conformational changes of both apo and holo forms of GRK5 at defined energy states. The study identifies DB02844 (S-Adenosyl-1,8-Diamino-3-Thiooctane) and DB13155 (Esculin) as promising GRK5 inhibitors, warranting further in vitro and in vivo validation studies.

In silico bioavailability triggers applied to direct and indirect thyroid hormone disruptors

Among endocrine disruption, interference with the thyroid hormone (TH) regulation is of increasing concern. Respective compounds encode through their structural features both the potential for TH disruption, and the bioavailability mitigating the toxicological effect. The aim of this study is to provide a substructure-based screening-level QSAR (quantitative structure-activity relationship) that discriminates bioavailable TH disruptors from not bioavailable counterparts, covering both direct and indirect (retinoid- and AhR-mediated) modes of action. The QSAR has been derived from literature data for 1642 compounds, and takes into account Lipinski's rule-of-five and the brain/blood partition coefficient Kbrain/blood. For its validation, an external test set of 145 substances has been used. For 1787 compounds meeting the model application domain, the model yields only one false negative. The discussion addresses the mechanistic meaning of the bioavailability triggers molecular weight, H-bond donor and acceptor, hydrophobicity (log Kow), and the physicochemical properties underlying log Kbrain/blood. The model may serve as bioavailability-screening step within a decision support system for the predictive assessment of chemicals regarding their potential to disrupt thyroid hormone function in a direct or indirect manner.

Finding structural requirements of structurally diverse α-glucosidase and α-amylase inhibitors through validated and predictive 2D-QSAR and 3D-QSAR analyses

Diabetes mellitus (DM) is a chronic metabolic disorder characterized by hyperglycemic state. The α-glucosidase and α-amylase are considered two major targets for the management of Type 2 DM due to their ability of metabolizing carbohydrates into simpler sugars. In the current study, cheminformatics analyses were performed to develop validated and predictive models with a dataset of 187 α-glucosidase and α-amylase dual inhibitors. Separate linear, interpretable and statistically robust 2D-QSAR models were constructed with datasets containing the activities of α-glucosidase and α-amylase inhibitors with an aim to explain the crucial structural and physicochemical attributes responsible for higher activity towards these targets. Consequently, some descriptors of the models pointed out the importance of specific structural moieties responsible for the higher activities for these targets and on the other hand, properties such as ionization potential and mass of the compounds as well as number of hydrogen bond donors in molecules were found to be crucial in determining the binding potentials of the dataset compounds. Statistically significant 3D-QSAR models were developed with both α-glucosidase and α-amylase inhibition datapoints to estimate the importance of 3D electrostatic and steric fields for improved potentials towards these two targets. Molecular docking performed with selected compounds with homology model of α-glucosidase and X-ray crystal structure of α-amylase largely supported the interpretations obtained from the cheminformatic analyses. The current investigation should serve as important guidelines for the design of future α-glucosidase and α-amylase inhibitors. Besides, the current investigation is entirely performed by using non-commercial open-access tools to ensure easy accessibility and reproducibility of the investigation which may help researchers throughout the world to work more on drug design and discovery.

Applicability Domains Based on Molecular Graph Contrastive Learning Enable Graph Attention Network Models to Accurately Predict 15 Environmental End Points

In silico models for predicting physicochemical properties and environmental fate parameters are necessary for the sound management of chemicals. This study employed graph attention network (GAT) algorithms to construct such models on 15 end points. The results showed that the GAT models outperformed the previous state-of-the-art models, and their performance was not influenced by the presence or absence of compounds with certain structures. Molecular similarity density (ρs) was found to be a key metrics characterizing data set modelability, in addition to the proportion of compounds at activity cliffs. By introducing molecular graph (MG) contrastive learning, MG-based ρs and molecular inconsistency in activities (IA) were calculated and employed for characterizing the structure-activity landscape (SAL)-based applicability domain ADSAL{ρs, IA}. The GAT models coupled with ADSAL{ρs, IA} significantly improved the prediction coefficient of determination (R2) on all the end points by an average of 14.4% and enabled all the end points to have R2 > 0.9, which could hardly be achieved previously. The models were employed to screen persistent, mobile, and/or bioaccumulative chemicals from inventories consisting of about 106 chemicals. Given the current state-of-the-art model performance and coverage of the various environmental end points, the constructed models with ADSAL{ρs, IA} may serve as benchmarks for future efforts to improve modeling efficacy.

Octanol/Air Partition Coefficient─A General-Purpose Fragment Model to Predict Log Koa from Molecular Structure

The octanol/air partition coefficient K_oa is important for assessing the bioconcentration of airborne xenobiotics in foliage and in air-breathing organisms. Moreover, K_oa informs about compound partitioning to aerosols and indoor dust, and complements the octanol/water partition coefficient K_ow and the air/water partition coefficient K_aw for multimedia fate modeling. Experimental log K_oa at 25 °C has been collected from literature for 2161 compounds with molecular weights from 16 to 959 Da. The curated data set covers 18.2 log units (from -1.0 to 17.2). A newly developed fragment model for predicting log K_oa from molecular structure outperforms COSMOtherm, EPI-Suite KOAWIN, OPERA, and linear solvation energy relationships (LSERs) regarding the root-mean-squared error (rms) and the maximum negative and positive errors (mne and mpe) (rms: 0.57 vs 0.86 vs 1.09 vs 1.19 vs 1.05-1.53, mne: -2.55 vs -3.95 vs -7.51 vs -7.54 vs (-5.63) - (-7.34), mpe: 2.91 vs 5.97 vs 7.54 vs 4.24 vs 6.89-10.2 log units). The prediction capability, statistical robustness, and sound mechanistic basis are demonstrated through initial separation into a training and prediction set (80:20%), mutual leave-50%-out validation, and target value scrambling in terms of temporarily wrong compound-K_oa allocations. The new general-purpose model is implemented in a fully automatized form in the ChemProp software available to the public. Regarding K_oa indirectly determined through K_ow and K_aw, a new approach is developed to convert from wet to dry octanol, enabling higher consistency in experimental (and thus also predicted) K_oa.

Henry's Law Constant─A General-Purpose Fragment Model to Predict Log Kaw from Molecular Structure

Henry's law constant is important for assessing the environmental fate of organic compounds, including polar accumulation, indoor contamination, and the impact of airborne predominance on persistence. Moreover, it can be used in the context of alternative 3R bioassays to inform about the compound loss through volatilization as a confounding factor. For 2636 compounds, curated experimental log K_aw (air/water partition coefficient) data at 25° covering 23.6 orders of magnitude (from -18.6 to 5.0) have been collected from the literature. Subsequently, a new fragment model for predicting log K_aw from molecular structures has been developed. According to the root-mean-squared error (rms) and the maximum negative and positive errors (mne and mpe), this general-purpose model outperforms COSMOtherm, EPISuite HENRYWIN, OPERA, and LSER with calculated input parameters significantly (rms 0.50 vs 0.92 vs 1.25 vs 1.28 vs 1.38, mne -2.74 vs -6.78 vs -9.11 vs -6.24 vs -6.27, mpe 2.25 vs 6.22 vs 8.27 vs 11.5 vs 7.69 log units). Initial separation into a training and prediction set (80%:20%), mutual leave-50%-out validation, and target value scrambling (temporarily wrong compound-K_aw allocations) demonstrate the prediction capability, statistical robustness, and mechanistically sound basis of the fragment scheme. The new model is available to the public in fully computerized form through the ChemProp software, and can be combined with a separate existing model to extend the log K_aw prediction to temperatures different from 25 °C.

Computational prediction of potential inhibitors for SARS-COV-2 main protease based on machine learning, docking, MM-PBSA calculations, and metadynamics

The development of new drugs is a very complex and time-consuming process, and for this reason, researchers have been resorting heavily to drug repurposing techniques as an alternative for the treatment of various diseases. This approach is especially interesting when it comes to emerging diseases with high rates of infection, because the lack of a quickly cure brings many human losses until the mitigation of the epidemic, as is the case of COVID-19. In this work, we combine an in-house developed machine learning strategy with docking, MM-PBSA calculations, and metadynamics to detect potential inhibitors for SARS-COV-2 main protease among FDA approved compounds. To assess the ability of our machine learning strategy to retrieve potential compounds we calculated the Enrichment Factor of compound datasets for three well known protein targets: HIV-1 reverse transcriptase (PDB 4B3P), 5-HT2A serotonin receptor (PDB 6A94), and H1 histamine receptor (PDB 3RZE). The Enrichment Factor for each target was, respectively, 102.5, 12.4, 10.6, which are considered significant values. Regarding the identification of molecules that can potentially inhibit the main protease of SARS-COV-2, compounds output by the machine learning step went through a docking experiment against SARS-COV-2 Mpro. The best scored poses were the input for MM-PBSA calculations and metadynamics using CHARMM and AMBER force fields to predict the binding energy for each complex. Our work points out six molecules, highlighting the strong interaction obtained for Mpro-mirabegron complex. Among these six, to the best of our knowledge, ambenonium has not yet been described in the literature as a candidate inhibitor for the SARS-COV-2 main protease in its active pocket.

Prioritization of Mycotoxins Based on Their Genotoxic Potential with an In Silico-In Vitro Strategy

Humans are widely exposed to a great variety of mycotoxins and their mixtures. Therefore, it is important to design strategies that allow prioritizing mycotoxins based on their toxic potential in a time and cost-effective manner. A strategy combining in silico tools (Phase 1), including an expert knowledge-based (DEREK Nexus®, Lhasa Limited, Leeds, UK) and a statistical-based platform (VEGA QSAR©, Mario Negri Institute, Milan, Italy), followed by the in vitro SOS/umu test (Phase 2), was applied to a set of 12 mycotoxins clustered according to their structure into three groups. Phase 1 allowed us to clearly classify group 1 (aflatoxin and sterigmatocystin) as mutagenic and group 3 (ochratoxin A, zearalenone and fumonisin B1) as non-mutagenic. For group 2 (trichothecenes), contradictory conclusions were obtained between the two in silico tools, being out of the applicability domain of many models. Phase 2 confirmed the results obtained in the previous phase for groups 1 and 3. It also provided extra information regarding the role of metabolic activation in aflatoxin B1 and sterigmatocystin mutagenicity. Regarding group 2, equivocal results were obtained in few experiments; however, the group was finally classified as non-mutagenic. The strategy used correlated with the published Ames tests, which detect point mutations. Few alerts for chromosome aberrations could be detected. The SOS/umu test appeared as a good screening test for mutagenicity that can be used in the absence and presence of metabolic activation and independently of Phase 1, although the in silico-in vitro combination gave more information for decision making.

ERGO: Breaking Down the Wall between Human Health and Environmental Testing of Endocrine Disrupters

ERGO (EndocRine Guideline Optimization) is the acronym of a European Union-funded research and innovation action, that aims to break down the wall between mammalian and non-mammalian vertebrate regulatory testing of endocrine disruptors (EDs), by identifying, developing and aligning thyroid-related biomarkers and endpoints (B/E) for the linkage of effects between vertebrate classes. To achieve this, an adverse outcome pathway (AOP) network covering various modes of thyroid hormone disruption (THD) in multiple vertebrate classes will be developed. The AOP development will be based on existing and new data from in vitro and in vivo experiments with fish, amphibians and mammals, using a battery of different THDs. This will provide the scientifically plausible and evidence-based foundation for the selection of B/E and assays in lower vertebrates, predictive of human health outcomes. These assays will be prioritized for validation at OECD (Organization for Economic Cooperation and Development) level. ERGO will re-think ED testing strategies from in silico methods to in vivo testing and develop, optimize and validate existing in vivo and early life-stage OECD guidelines, as well as new in vitro protocols for THD. This strategy will reduce requirements for animal testing by preventing duplication of testing in mammals and non-mammalian vertebrates and increase the screening capacity to enable more chemicals to be tested for ED properties.

Compartment-Specific Screening Tools for Persistence: Potential Role and Application in the Regulatory Context

The persistence assessment under the European Union regulation Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) relies on compartment-specific degradation half-lives derived from laboratory simulation studies with surface water, aquatic sediment, or soil. Although these data are given priority, they are not available for most of the compounds. Therefore, according to the Integrated Assessment and Testing Strategy (ITS) for persistence assessment, results from ready biodegradability tests (RBTs) are used within a persistence screening to decide whether a substance is considered as "not persistent" or "potentially persistent." However, ready biodegradability is currently tested only in water. Consequently, there is a lack of approaches that include the soil and sediment compartments for persistence assessment at the screening level. In previous studies, compartment-specific screening tools for water-sediment (Water-Sediment Screening Tool [WSST]) and soil (Soil Screening Tool [SST]) were developed based on the existing test guideline Organisation for Economic Development and Co-operation (OECD TG 301C [MITI (Ministry of International Trade and Industry, Japan) test]). The test systems MITI, WSST, and SST were successfully applied to determine sound and reliable biodegradation data for 15 test compounds. In the present study, these results are used within the scope of a new alternative persistence screening approach, the Compartment-Specific Persistence Screening (CSPS). Compared to the persistence screening under REACH, the CSPS is a more conservative approach that provides additional reasonable results, particularly for compounds that sorb to sediment and soil, and for which the current standard persistence screening might be insufficient. Thus, the CSPS can be used to identify potentially persistent and nonpersistent compounds in the regulatory context by a comprehensive assessment that includes water, soil, and sediment. Moreover, experimentally determined half-lives from the compartment-specific screening tools can be used as input for multimedia models that estimate, for example, overall persistence (P_ov ). The application of fixed half-life factors to extrapolate from water to soil and sediment, which is here demonstrated to be inappropriate, can thereby be avoided. Integr Environ Assess Manag 2019;00:000-000. © 2019 SETAC.

Species sensitivity distributions for use in environmental protection, assessment, and management of aquatic ecosystems for 12 386 chemicals

The present study considers the collection and use of ecotoxicity data for risk assessment with species sensitivity distributions (SSDs) of chemical pollution in surface water, which are used to quantify the likelihood that critical effect levels are exceeded. This fits the European Water Framework Directive, which suggests using models to assess the likelihood that chemicals affect water quality for management prioritization. We derived SSDs based on chronic and acute ecotoxicity test data for 12 386 compounds. The log-normal SSDs are characterized by the median and the standard deviation of log-transformed ecotoxicity data and by a quality score. A case study illustrates the utility of SSDs for water quality assessment and management prioritization. We quantified the chronic and acute mixture toxic pressure of mixture exposures for >22 000 water bodies in Europe for 1760 chemicals for which we had both exposure and hazard data. The results show the likelihood of mixture exposures exceeding a negligible effect level and increasing species loss. The SSDs in the present study represent a versatile and comprehensive approach to prevent, assess, and manage chemical pollution problems. Environ Toxicol Chem 2019;38:905-917. © 2019 SETAC.

Integrating bioassays, chemical analysis and in silico techniques to identify genotoxicants in surface water

Identification of hazardous compounds, as the first step of water protection and regulation, is still challenged by the difficulty to establish a linkage between toxic effects and suspected contaminants. Genotoxic compounds are one type of highly relevant toxicants in surface water, which may attack the DNA and lead to cancer in individual organism, or even damaged germ cells to be passed on to future generations. Thus, the establishment of a linkage between genotoxic effects and genotoxicant is important for environmental toxicologists and chemists. For this purpose, in the present study in silico methods were integrated with bioassays, chemical analysis and literature information to identify genotoxicants in surface water. Large volume water samples from 22 sampling sites of the Danube were collected and subjected to biological and chemical analysis. Samples from the most toxic sites (JDS32, JDS44 and JDS63) induced significant genotoxic effects in the micronucleus assay, and two of them caused mutagenicity in the Ames fluctuation assay. Chemical analysis showed that 68 chemicals were detected in these most toxic samples. Literature findings and in silico techniques using the OECD QSAR Toolbox and the ChemProp software package revealed genotoxic potentials for 29 compounds out of 68 targeted chemicals. To confirm the integrative technical data, the micronucleus assay and the Ames fluctuation assay were applied with artificial mixtures of those compounds and the raw water sample extracts. The results showed that 18 chemicals explained 48.5% of the genotoxicity in the micronucleus assay. This study highlights the capability of in silico techniques in linking adverse biological effect to suspicious hazardous compounds for the identification of toxicity drivers, and demonstrates the genotoxic potential of pollutants in the Danube.

QSAR Development for Plasma Protein Binding: Influence of the Ionization State

Purpose

This study explored several strategies to improve the performance of literature QSAR models for plasma protein binding (PPB), such as a suitable endpoint transformation, a correct representation of chemicals, more consistency in the dataset, and a reliable definition of the applicability domain.

Methods

We retrieved human fraction unbound (Fu) data for 670 compounds from the literature and carefully checked them for consistency. Descriptors were calculated taking account of the ionization state of molecules at physiological pH (7.4), in order to better estimate the affinity of molecules to blood proteins. We used different algorithms and chemical descriptors to explore the most suitable strategy for modeling the endpoint. SMILES (simplified molecular input line entry system)-based string descriptors were also tested with the CORAL software (CORelation And Logic). We did an outlier analysis to establish the models to use (or not to use) in case of well recognized families.

Results

Internal validation of the selected models returned Q² values close to 0.60. External validation also gave r² values always greater than 0.60. The CORAL descriptor based model for √fu was the best, with r² 0.74 in external validation.

Conclusions

Performance in prediction confirmed the robustness of all the derived models and their suitability for real-life purposes, i.e. screening chemicals for their ADMET profiling. Optimization of descriptors can be useful in order to obtain the correct results with a ionized molecule.

Electronic supplementary material

The online version of this article (10.1007/s11095-018-2561-8) contains supplementary material, which is available to authorized users.

PBT assessment under REACH: Screening for low aquatic bioaccumulation with QSAR classifications based on physicochemical properties to replace BCF in vivo testing on fish

Aquatic bioconcentration factors (BCFs) are critical in PBT (persistent, bioaccumulative, toxic) and risk assessment of chemicals. High costs and use of more than 100 fish per standard BCF study (OECD 305) call for alternative methods to replace as much in vivo testing as possible. The BCF waiving scheme is a screening tool combining QSAR classifications based on physicochemical properties related to the distribution (hydrophobicity, ionisation), persistence (biodegradability, hydrolysis), solubility and volatility (Henry's law constant) of substances in water bodies and aquatic biota to predict substances with low aquatic bioaccumulation (nonB, BCF<2000). The BCF waiving scheme was developed with a dataset of reliable BCFs for 998 compounds and externally validated with another 181 substances. It performs with 100% sensitivity (no false negatives), >50% efficacy (waiving potential), and complies with the OECD principles for valid QSARs. The chemical applicability domain of the BCF waiving scheme is given by the structures of the training set, with some compound classes explicitly excluded like organometallics, poly- and perfluorinated compounds, aromatic triphenylphosphates, surfactants. The prediction confidence of the BCF waiving scheme is based on applicability domain compliance, consensus modelling, and the structural similarity with known nonB and B/vB substances. Compounds classified as nonB by the BCF waiving scheme are candidates for waiving of BCF in vivo testing on fish due to low concern with regard to the B criterion. The BCF waiving scheme supports the 3Rs with a possible reduction of >50% of BCF in vivo testing on fish. If the target chemical is outside the applicability domain of the BCF waiving scheme or not classified as nonB, further assessments with in silico, in vitro or in vivo methods are necessary to either confirm or reject bioaccumulative behaviour.

Applying Mondrian Cross-Conformal Prediction To Estimate Prediction Confidence on Large Imbalanced Bioactivity Data Sets

Conformal prediction has been proposed as a more rigorous way to define prediction confidence compared to other application domain concepts that have earlier been used for QSAR modeling. One main advantage of such a method is that it provides a prediction region potentially with multiple predicted labels, which contrasts to the single valued (regression) or single label (classification) output predictions by standard QSAR modeling algorithms. Standard conformal prediction might not be suitable for imbalanced data sets. Therefore, Mondrian cross-conformal prediction (MCCP) which combines the Mondrian inductive conformal prediction with cross-fold calibration sets has been introduced. In this study, the MCCP method was applied to 18 publicly available data sets that have various imbalance levels varying from 1:10 to 1:1000 (ratio of active/inactive compounds). Our results show that MCCP in general performed well on bioactivity data sets with various imbalance levels. More importantly, the method not only provides confidence of prediction and prediction regions compared to standard machine learning methods but also produces valid predictions for the minority class. In addition, a compound similarity based nonconformity measure was investigated. Our results demonstrate that although it gives valid predictions, its efficiency is much worse than that of model dependent metrics.

Classification of baseline toxicants for QSAR predictions to replace fish acute toxicity studies

Fish acute toxicity studies are required for environmental hazard and risk assessment of chemicals by national and international legislations such as REACH, the regulations of plant protection products and biocidal products, or the GHS (globally harmonised system) for classification and labelling of chemicals. Alternative methods like QSARs (quantitative structure-activity relationships) can replace many ecotoxicity tests. However, complete substitution of in vivo animal tests by in silico methods may not be realistic. For the so-called baseline toxicants, it is possible to predict the fish acute toxicity with sufficient accuracy from log K_ow and, hence, valid QSARs can replace in vivo testing. In contrast, excess toxicants and chemicals not reliably classified as baseline toxicants require further in silico, in vitro or in vivo assessments. Thus, the critical task is to discriminate between baseline and excess toxicants. For fish acute toxicity, we derived a scheme based on structural alerts and physicochemical property thresholds to classify chemicals as either baseline toxicants (=predictable by QSARs) or as potential excess toxicants (=not predictable by baseline QSARs). The step-wise approach identifies baseline toxicants (true negatives) in a precautionary way to avoid false negative predictions. Therefore, a certain fraction of false positives can be tolerated, i.e. baseline toxicants without specific effects that may be tested instead of predicted. Application of the classification scheme to a new heterogeneous dataset for diverse fish species results in 40% baseline toxicants, 24% excess toxicants and 36% compounds not classified. Thus, we can conclude that replacing about half of the fish acute toxicity tests by QSAR predictions is realistic to be achieved in the short-term. The long-term goals are classification criteria also for further groups of toxicants and to replace as many in vivo fish acute toxicity tests as possible with valid QSAR predictions.

Screening for potential endocrine disruptors in fish: evidence from structural alerts and in vitro and in vivo toxicological assays

BACKGROUND

The European chemicals' legislation REACH aims to protect man and the environment from substances of very high concern (SVHC). Chemicals like endocrine disruptors (EDs) may be subject to authorization. Identification of (potential) EDs with regard to the environment is limited because specific experimental assessments are not standard requirements under REACH. Evidence is based on a combination of in vitro and in vivo experiments (if available), expert judgement, and structural analogy with known EDs.

OBJECTIVES

The objectives of this study are to review and refine structural alerts for the indication of potential estrogenic and androgenic endocrine activities based on in vitro studies; to analyze in vivo mammalian long-term reproduction studies with regard to estrogen- and androgen-sensitive endpoints in order to identify potential indicators for endocrine activity with regard to the environment; to assess the consistency of potential estrogenic and androgenic endocrine activities based on in vitro assays and in vivo mammalian long-term reproduction studies and fish life-cycle tests; and to evaluate structural alerts, in vitro assays, and in vivo mammalian long-term reproduction studies for the indication of potential estrogenic and androgenic endocrine disruptors in fish.

RESULTS

Screening for potential endocrine activities in fish via estrogenic and androgenic modes of action based on structural alerts provides similar information as in vitro receptor-mediated assays. Additional evidence can be obtained from in vivo mammalian long-term reproduction studies. Conclusive confirmation is possible with fish life-cycle tests. Application of structural alerts to the more than 33,000 discrete organic compounds of the EINECS inventory indicated 3585 chemicals (approx. 11%) as potential candidates for estrogenic and androgenic effects that should be further investigated. Endocrine activities of the remaining substances cannot be excluded; however, because the structural alerts perform much better for substances with (very) high estrogenic and androgenic activities, there is reasonable probability that the most hazardous candidates have been identified.

CONCLUSIONS

The combination of structural alerts, in vitro receptor-based assays, and in vivo mammalian studies may support the priority setting for further assessments of chemicals with potential environmental hazards due to estrogenic and androgenic activities.

Mind the Gap! A Journey towards Computational Toxicology

Computational methods have advanced toxicology towards the development of target-specific models based on a clear cause-effect rationale. However, the predictive potential of these models presents strengths and weaknesses. On the good side, in silico models are valuable cheap alternatives to in vitro and in vivo experiments. On the other, the unconscious use of in silico methods can mislead end-users with elusive results. The focus of this review is on the basic scientific and regulatory recommendations in the derivation and application of computational models. Attention is paid to examine the interplay between computational toxicology and drug discovery and development. Avoiding the easy temptation of an overoptimistic future, we report our view on what can, or cannot, realistically be done. Indeed, studies of safety/toxicity represent a key element of chemical prioritization programs carried out by chemical industries, and primarily by pharmaceutical companies.

Comparison of prediction methods for octanol-air partition coefficients of diverse organic compounds

The octanol-air partition coefficient (KOA) is needed for assessing multimedia transport and bioaccumulability of organic chemicals in the environment. As experimental determination of KOA for various chemicals is costly and laborious, development of KOA estimation methods is necessary. We investigated three methods for KOA prediction, conventional quantitative structure-activity relationship (QSAR) models based on molecular structural descriptors, group contribution models based on atom-centered fragments, and a novel model that predicts KOA via solvation free energy from air to octanol phase (ΔGO(0)), with a collection of 939 experimental KOA values for 379 compounds at different temperatures (263.15-323.15 K) as validation or training sets. The developed models were evaluated with the OECD guidelines on QSAR models validation and applicability domain (AD) description. Results showed that although the ΔGO(0) model is theoretically sound and has a broad AD, the prediction accuracy of the model is the poorest. The QSAR models perform better than the group contribution models, and have similar predictability and accuracy with the conventional method that estimates KOA from the octanol-water partition coefficient and Henry's law constant. One QSAR model, which can predict KOA at different temperatures, was recommended for application as to assess the long-range transport potential of chemicals.

Mechanism-specific and whole-organism ecotoxicity of mono-rhamnolipids

Biosurfactants like rhamnolipids are promising alternatives to chemical surfactants in a range of applications. A wider use requires an analysis of their environmental fate and their ecotoxicological potential. In the present study mono-rhamnolipids produced by a recombinant Pseudomonas putida strain were analyzed using the Green Toxicology concept for acute and mechanism-specific toxicity in an ecotoxicological test battery. Acute toxicity tests with the invertebrate Daphnia magna and with zebrafish embryos (Danio rerio) were performed. In addition, microbial and fungicidal effectiveness was investigated. Mutagenicity of the sample was tested by means of the Ames fluctuation assay. A selected mono-rhamnolipid was used for model simulations regarding mutagenicity and estrogenic activity. Our results indicate that mono-rhamnolipids cause acute toxicity to daphnids and zebrafish embryos comparable to or even lower than chemical surfactants. Rhamnolipids showed very low toxicity to the germination of Aspergillus niger spores and the growth of Candida albicans. No frameshift mutation or base substitutions were observed using the Ames fluctuation assay with the two tester strains TA98 and TA100. This result was confirmed by model simulations. Likewise it was computed that rhamnolipids have no estrogenic potential. In conclusion, mono-rhamnolipids are an environmental friendly alternative to chemical surfactants as the ecotoxicological potential is low.

Inhalation threshold of toxicological concern (TTC) - Structural alerts discriminate high from low repeated-dose inhalation toxicity

The threshold of toxicological concern (TTC) of a compound represents an exposure value below which the associated human health risk is considered negligible. As such, this approach offers assessing the risk of potential toxicants when little or no toxicological information is available. For the inhalation repeated-dose TTC, the goal was to derive structural alerts that discriminate between high- and low-toxic compounds. A further aim was to identify physicochemical parameters related to the inhalation-specific bioavailability of the compounds, and to explore their use as predictors of high vs low toxicity. 296 compounds with subacute, subchronic and chronic inhalation toxicity NOEC (no-observed effect concentration) values were subdivided into three almost equal-sized high-, medium- and low-toxic (HTox, MTox, LTox) potency classes. Whereas the derived 14 HTox and 7 LTox structural alerts yield an only moderate discrimination between these three groups, the high-toxic vs low-toxic mis-classification is very low: LTox-predicted compounds are not HTox to 97.5%, and HTox-predicted compounds not LTox to 88.6%. The probability of a compound being HTox vs LTox is triggered further by physicochemical properties encoding the tendency to evaporate from blood. The new structural alerts may aid in the predictive inhalation toxicity assessment of compounds as well as in designing low-toxicity chemicals, and provide a rationale for the chemistry underlying the toxicological outcome that can also be used for scoping targeted experimental studies.

Computational toxicology, friend or foe?

Computational approaches offer the attraction of being both fast and cheap to run being able to process thousands of chemical structures in a few minutes. As with all new technology, there is a tendency for these approaches to be hyped up and claims of reliability and performance may be exaggerated. So just how good are these computational methods?

ChemStable: a web server for rule-embedded naïve Bayesian learning approach to predict compound stability

Predicting compound chemical stability is important because unstable compounds can lead to either false positive or to false negative conclusions in bioassays. Experimental data (COMDECOM) measured from DMSO/H2O solutions stored at 50 °C for 105 days were used to predicted stability by applying rule-embedded naïve Bayesian learning, based upon atom center fragment (ACF) features. To build the naïve Bayesian classifier, we derived ACF features from 9,746 compounds in the COMDECOM dataset. By recursively applying naïve Bayesian learning from the data set, each ACF is assigned with an expected stable probability (p(s)) and an unstable probability (p(uns)). 13,340 ACFs, together with their p(s) and p(uns) data, were stored in a knowledge base for use by the Bayesian classifier. For a given compound, its ACFs were derived from its structure connection table with the same protocol used to drive ACFs from the training data. Then, the Bayesian classifier assigned p(s) and p(uns) values to the compound ACFs by a structural pattern recognition algorithm, which was implemented in-house. Compound instability is calculated, with Bayes' theorem, based upon the p(s) and p(uns) values of the compound ACFs. We were able to achieve performance with an AUC value of 84% and a tenfold cross validation accuracy of 76.5%. To reduce false negatives, a rule-based approach has been embedded in the classifier. The rule-based module allows the program to improve its predictivity by expanding its compound instability knowledge base, thus further reducing the possibility of false negatives. To our knowledge, this is the first in silico prediction service for the prediction of the stabilities of organic compounds.

Three dimensional quantitative structure-toxicity relationship modeling and prediction of acute toxicity for organic contaminants to algae

Although numerous chemicals have been identified to have significant toxicological effect on aquatic organisms, there is still lack of a reliable, high-throughput approach to evaluate, screen and monitor the presence of organic contaminants in aquatic system. In the current study, we proposed a synthetic pipeline to automatically model and predict the acute toxicity of chemicals to algae. In the procedure, a new alignment-free three dimensional (3D) structure characterization method was described and, with this method, several 3D-quantitative structure-toxicity relationship (3D-QSTR) models were developed, from which two were found to exhibit strong internal fitting ability and high external predictive power. The best model was established by Gaussian process (GP), which was further employed to perform extrapolation on a random compound library consisting of 1014 virtually generated substituted benzenes. It was found that (i) substitution number can only exert slight influence on chemical׳s toxicity, but low-substituted benzenes seem to have higher toxicity than those of high-substituted entities, and (ii) benzenes substituted by nitro group and halogens exhibit high acute toxicity as compared to other substituents such as methyl and carboxyl groups. Subsequently, several promising candidates suggested by computational prediction were assayed by using a standard algal growth inhibition test. Consequently, four substituted benzenes, namely 2,3-dinitrophenol, 2-chloro-4-nitroaniline, 1,2,3-trinitrobenzene and 3-bromophenol, were determined to have high acute toxicity to Scenedesmus obliquus, with their EC50 values of 2.5±0.8, 10.5±2.1, 1.4±0.2 and 42.7±5.4μmol/L, respectively.

A Combinational Strategy of Model Disturbance and Outlier Comparison to Define Applicability Domain in Quantitative Structural Activity Relationship

In order to define an applicability domain for quantitative structure-activity relationship modeling, a combinational strategy of model disturbance and outlier comparison is developed. An indicator named model disturbance index was defined to estimate the prediction error. Moreover, the information of the outliers in the training set was used to filter the unreliable samples in the test set based on "structural similarity". Chromatography retention indices data were used to investigate this approach. The relationship between model disturbance index and prediction error can be found. Also, the comparison between the outlier set and the test set could provide additional information about which unknown samples should be paid more attentions. A novel technique based on model population analysis was used to evaluate the validity of applicability domain. Finally, three commonly used methods, i.e. Leverage, descriptor range-based and model perturbation method, were compared with the proposed approach.

Using random forest to model the domain applicability of another random forest model

In QSAR, a statistical model is generated from a training set of molecules (represented by chemical descriptors) and their biological activities. We will call this traditional type of QSAR model an "activity model". The activity model can be used to predict the activities of molecules not in the training set. A relatively new subfield for QSAR is domain applicability. The aim is to estimate the reliability of prediction of a specific molecule on a specific activity model. A number of different metrics have been proposed in the literature for this purpose. It is desirable to build a quantitative model of reliability against one or more of these metrics. We can call this an "error model". A previous publication from our laboratory (Sheridan J. Chem. Inf. Model., 2012, 52, 814-823.) suggested the simultaneous use of three metrics would be more discriminating than any one metric. An error model could be built in the form of a three-dimensional set of bins. When the number of metrics exceeds three, however, the bin paradigm is not practical. An obvious solution for constructing an error model using multiple metrics is to use a QSAR method, in our case random forest. In this paper we demonstrate the usefulness of this paradigm, specifically for determining whether a useful error model can be built and which metrics are most useful for a given problem. For the ten data sets and for the seven metrics we examine here, it appears that it is possible to construct a useful error model using only two metrics (TREE_SD and PREDICTED). These do not require calculating similarities/distances between the molecules being predicted and the molecules used to build the activity model, which can be rate-limiting.

Perspectives for integrating human and environmental risk assessment and synergies with socio-economic analysis

For more than a decade, the integration of human and environmental risk assessment (RA) has become an attractive vision. At the same time, existing European regulations of chemical substances such as REACH (EC Regulation No. 1907/2006), the Plant Protection Products Regulation (EC regulation 1107/2009) and Biocide Regulation (EC Regulation 528/2012) continue to ask for sector-specific RAs, each of which have their individual information requirements regarding exposure and hazard data, and also use different methodologies for the ultimate risk quantification. In response to this difference between the vision for integration and the current scientific and regulatory practice, the present paper outlines five medium-term opportunities for integrating human and environmental RA, followed by detailed discussions of the associated major components and their state of the art. Current hazard assessment approaches are analyzed in terms of data availability and quality, and covering non-test tools, the integrated testing strategy (ITS) approach, the adverse outcome pathway (AOP) concept, methods for assessing uncertainty, and the issue of explicitly treating mixture toxicity. With respect to exposure, opportunities for integrating exposure assessment are discussed, taking into account the uncertainty, standardization and validation of exposure modeling as well as the availability of exposure data. A further focus is on ways to complement RA by a socio-economic assessment (SEA) in order to better inform about risk management options. In this way, the present analysis, developed as part of the EU FP7 project HEROIC, may contribute to paving the way for integrating, where useful and possible, human and environmental RA in a manner suitable for its coupling with SEA.

Time-split cross-validation as a method for estimating the goodness of prospective prediction

Cross-validation is a common method to validate a QSAR model. In cross-validation, some compounds are held out as a test set, while the remaining compounds form a training set. A model is built from the training set, and the test set compounds are predicted on that model. The agreement of the predicted and observed activity values of the test set (measured by, say, R(2)) is an estimate of the self-consistency of the model and is sometimes taken as an indication of the predictivity of the model. This estimate of predictivity can be optimistic or pessimistic compared to true prospective prediction, depending how compounds in the test set are selected. Here, we show that time-split selection gives an R(2) that is more like that of true prospective prediction than the R(2) from random selection (too optimistic) or from our analog of leave-class-out selection (too pessimistic). Time-split selection should be used in addition to random selection as a standard for cross-validation in QSAR model building.

QSAR with experimental and predictive distributions: an information theoretic approach for assessing model quality

We propose that quantitative structure–activity relationship (QSAR) predictions should be explicitly represented as predictive (probability) distributions. If both predictions and experimental measurements are treated as probability distributions, the quality of a set of predictive distributions output by a model can be assessed with Kullback–Leibler (KL) divergence: a widely used information theoretic measure of the distance between two probability distributions. We have assessed a range of different machine learning algorithms and error estimation methods for producing predictive distributions with an analysis against three of AstraZeneca’s global DMPK datasets. Using the KL-divergence framework, we have identified a few combinations of algorithms that produce accurate and valid compound-specific predictive distributions. These methods use reliability indices to assign predictive distributions to the predictions output by QSAR models so that reliable predictions have tight distributions and vice versa. Finally we show how valid predictive distributions can be used to estimate the probability that a test compound has properties that hit single- or multi- objective target profiles.

Quantitative read-across for predicting the acute fish toxicity of organic compounds

Read-across enables the interpolation of a property for a target chemical from respective experimental data of sufficiently similar compounds. Employing a set of 692 organic compounds with experimental values for the 96 h fish toxicity toward the fathead minnow in terms of LC(50) (lethal concentration 50%) values, a read-across method has been developed that is based on atom-centered fragments (ACFs) for evaluating chemical similarity. Prediction of log LC(50) proceeds through reading across the toxicity enhancement over predicted narcosis-level toxicity in terms of the respective logarithmic ratio, log T(e), and adding the respective baseline narcosis LC(50) estimated from log K(ow) (octanol/water partition coefficient). Depending on the minimum similarity imposed on a compound to serve as read-across basis for the target chemical, three different standard settings have been introduced, allowing one to perform screening-level estimations as well as predictions with intermediate and good confidence. The respective squared correlation coefficients (r(2)) are 0.73, 0.78, and 0.87, with root-mean square errors (rms) of 0.73, 0.60, and 0.39 log units, respectively. As a general trend, increasing the ACF minimum similarity increases the prediction quality at the cost of decreasing the application range. The method has the potential to assist in the predictive evaluation of fish toxicity for regulatory purposes such as under the REACH legislation.