In silico screening of estrogen-like chemicals based on different nonlinear classification models

Identification of estrogen receptor agonists among hydroxylated polychlorinated biphenyls using classification-based quantitative structure-activity relationship models

Identification of estrogen receptor (ER) agonists among environmental toxicants is essential for assessing the potential impact of toxicants on human health. Using 2D autocorrelation descriptors as predictor variables, two binary logistic regression models were developed to identify active ER agonists among hydroxylated polychlorinated biphenyls (OH-PCBs). The classifications made by the two models on the training set compounds resulted in accuracy, sensitivity and specificity of 95.9 %, 93.9 % and 97.6 % for ERα dataset and 91.9 %, 90.9 % and 92.7 % for ERβ dataset. The areas under the ROC curves, constructed with the training set data, were found to be 0.985 and 0.987 for the two models. Predictions made by models I and II correctly classified 84.0 % and 88.0 % of the test set compounds and 89.8 % and 85.8% of the cross-validation set compounds respectively. The two classification-based QSAR models proposed in this paper are considered robust and reliable for rapid identification of ERα and ERβ agonists among OH-PCB congeners.

Principles of QSAR Modeling

At the end of her academic career, the author summarizes the main aspects of QSAR modeling, giving comments and suggestions according to her 23 years' experience in QSAR research on environmental topics. The focus is mainly on Multiple Linear Regression, particularly Ordinary Least Squares, using a Genetic Algorithm for variable selection from various theoretical molecular descriptors, but the comments can be useful also for other QSAR methods. The need for rigorous validation, also external, and for applicability domain check to guarantee predictivity and reliability of QSAR models is particularly highlighted. The commented approach is the “predictive” one, based on chemometrics, and is usefully applied to the prioritization of environmental pollutants. All the discussed points and the author's ideas are implemented in the software QSARINS, as a legacy to the QSAR community.

Bio-Based Aromatic Epoxy Monomers for Thermoset Materials

The synthesis of polymers from renewable resources is a burning issue that is actively investigated. Polyepoxide networks constitute a major class of thermosetting polymers and are extensively used as coatings, electronic materials, adhesives. Owing to their outstanding mechanical and electrical properties, chemical resistance, adhesion, and minimal shrinkage after curing, they are used in structural applications as well. Most of these thermosets are industrially manufactured from bisphenol A (BPA), a substance that was initially synthesized as a chemical estrogen. The awareness on BPA toxicity combined with the limited availability and volatile cost of fossil resources and the non-recyclability of thermosets implies necessary changes in the field of epoxy networks. Thus, substitution of BPA has witnessed an increasing number of studies both from the academic and industrial sides. This review proposes to give an overview of the reported aromatic multifunctional epoxide building blocks synthesized from biomass or from molecules that could be obtained from transformed biomass. After a reminder of the main glycidylation routes and mechanisms and the recent knowledge on BPA toxicity and legal issues, this review will provide a brief description of the main natural sources of aromatic molecules. The different epoxy prepolymers will then be organized from simple, mono-aromatic di-epoxy, to mono-aromatic poly-epoxy, to di-aromatic di-epoxy compounds, and finally to derivatives possessing numerous aromatic rings and epoxy groups.

Prediction of in vitro and in vivo oestrogen receptor activity using hierarchical clustering

In this study, hierarchical clustering classification models were developed to predict in vitro and in vivo oestrogen receptor (ER) activity. Classification models were developed for binding, agonist, and antagonist in vitro ER activity and for mouse in vivo uterotrophic ER binding. In vitro classification models yielded balanced accuracies ranging from 0.65 to 0.85 for the external prediction set. In vivo ER classification models yielded balanced accuracies ranging from 0.72 to 0.83. If used as additional biological descriptors for in vivo models, in vitro scores were found to increase the prediction accuracy of in vivo ER models. If in vitro activity was used directly as a surrogate for in vivo activity, the results were poor (balanced accuracy ranged from 0.49 to 0.72). Under-sampling negative compounds in the training set was found to increase the coverage (fraction of chemicals which can be predicted) and increase prediction sensitivity.

Use of artificial neural networks in the QSAR prediction of physicochemical properties and toxicities for REACH legislation

With the introduction of the REACH legislation in the European Union, there is a requirement for property and toxicity data on chemicals produced in or imported into the EU at levels of 1 tonne/year or more. This has meant an increase in the in silico prediction of such data. One of the chief predictive approaches is QSAR (quantitative structure-activity relationships), which is widely used in many fields. A QSAR approach that is increasingly being used is that of artificial neural networks (ANNs), and this chapter discusses its application to the range of physicochemical properties and toxicities required by REACH. ANNs generally outperform the main QSAR approach of multiple linear regression (MLR), although other approaches such as support vector machines sometimes outperform ANNs. Most ANN QSARs reported to date comply with only two of the five OECD Guidelines for the Validation of (Q)SARs.

Computational models to predict endocrine-disrupting chemical binding with androgen or oestrogen receptors

Rapidly and correctly identifying endocrine-disrupting chemicals (EDCs) is an important issue in environmental risk assessment. Major EDCs are associated with the androgen receptor (AR) and oestrogen receptors (ERs). Because of the high cost and time-consuming nature of experimental tests, in silico methods are valuable alternative tools for the identification of EDCs. In this study, a large dataset related to EDCs was constructed. Each molecule was represented with seven fingerprints, and computational models were subsequently developed to predict AR and ER binders via machine learning methods including k-nearest neighbour (kNN), C4.5 decision tree (C4.5 DT), naïve Bayes (NB), and support vector machine (SVM) algorithms. The best model for predicting AR binders was PubChem Fingerprint-SVM, which exhibited an accuracy of 0.84. For ER binders, the best method was Extended Fingerprint-SVM with an accuracy of 0.79. Moreover, several representative substructure alerts for characterizing EDCs, such as phenol, trifluoromethyl, and annelated rings, were identified using the combination of information gain and substructure frequency analysis. Our study involved a systematic computational assessment of EDCs related to AR and ERs, and provides significant information on the structural characteristics of these chemicals, which are a great help in identifying EDCs.

QSAR prediction of the competitive interaction of emerging halogenated pollutants with human transthyretin

The determination of the potential endocrine disruption (ED) activity of chemicals such as poly/perfluorinated compounds (PFCs) and brominated flame retardants (BFRs) is still hindered by a limited availability of experimental data. Quantitative structure-activity relationship (QSAR) strategies can be applied to fill this data gap, help in the characterization of the ED potential, and screen PFCs and BFRs with a hazardous toxicological profile. This paper proposes the modelling of T4-TTR (thyroxin-transthyretin) competing potency and relative binding potency toward T4 (logT4-REP) of PFCs and BFRs by regression and classification QSAR models. This study is a follow up of a former work, which analysed separately the interaction of BFRs and PFCs with the carrier TTR. The new results demonstrate the possibility of developing robust and predictive QSARs, which include both BFRs and PFCs in the training set, obtaining larger applicability domains than the existing models developed separately for BFRs and PFCs. The selection of modelling molecular descriptors confirms the importance of structural features, such as the aromatic OH or the molecular length, to increase the binding of the studied chemicals to TTR. Additionally, the need of experimental tests for some chemicals, and in particular for some of the BFRs, is highlighted.

QSAR classification models for the prediction of endocrine disrupting activity of brominated flame retardants

The identification of potential endocrine disrupting (ED) chemicals is an important task for the scientific community due to their diffusion in the environment; the production and use of such compounds will be strictly regulated through the authorization process of the REACH regulation. To overcome the problem of insufficient experimental data, the quantitative structure-activity relationship (QSAR) approach is applied to predict the ED activity of new chemicals. In the present study QSAR classification models are developed, according to the OECD principles, to predict the ED potency for a class of emerging ubiquitary pollutants, viz. brominated flame retardants (BFRs). Different endpoints related to ED activity (i.e. aryl hydrocarbon receptor agonism and antagonism, estrogen receptor agonism and antagonism, androgen and progesterone receptor antagonism, T4-TTR competition, E2SULT inhibition) are modeled using the k-NN classification method. The best models are selected by maximizing the sensitivity and external predictive ability. We propose simple QSARs (based on few descriptors) characterized by internal stability, good predictive power and with a verified applicability domain. These models are simple tools that are applicable to screen BFRs in relation to their ED activity, and also to design safer alternatives, in agreement with the requirements of REACH regulation at the authorization step.

QSAR classification of estrogen receptor binders and pre-screening of potential pleiotropic EDCs

Endocrine disrupting chemicals (EDCs) are suspected of posing serious threats to human and wildlife health through a variety of mechanisms, these being mainly receptor-mediated modes of action. It is reported that some EDCs exhibit dual activities as estrogen receptor (ER) and androgen receptor (AR) binders. Indeed, such compounds can affect the normal endocrine system through a dual complex mechanism, so steps should be taken not only to identify them a priori from their chemical structure, but also to prioritize them for experimental tests in order to reduce and even forbid their usage. To date, very few EDCs with dual activities have been identified. The present research uses QSARs, to investigate what, so far, is the largest and most heterogeneous ER binder data set (combined METI and EDKB databases). New predictive classification models were derived using different modelling methods and a consensus approach, and these were used to virtually screen a large AR binder data set after strict validation. As a result, 46 AR antagonists were predicted from their chemical structure to also have potential ER binding activities, i.e. pleiotropic EDCs. In addition, 48 not yet recognized ER binders were in silico identified, which increases the number of potential EDCs that are substances of very high concern (SVHC) in REACH. Thus, the proposed screening models, based only on structure information, have the main aim to prioritize experimental tests for the highlighted compounds with potential estrogenic activities and also to design safer alternatives.

Classification and virtual screening of androgen receptor antagonists

Computational tools, such as quantitative structure-activity relationship (QSAR), are highly useful as screening support for prioritization of substances of very high concern (SVHC). From the practical point of view, QSAR models should be effective to pick out more active rather than inactive compounds, expressed as sensitivity in classification works. This research investigates the classification of a big data set of endocrine-disrupting chemicals (EDCs)-androgen receptor (AR) antagonists, mainly aiming to improve the external sensitivity and to screen for potential AR binders. The kNN, lazy IB1, and ADTree methods and the consensus approach were used to build different models, which improve the sensitivity on external chemicals from 57.1% (literature) to 76.4%. Additionally, the models' predictive abilities were further validated on a blind collected data set (sensitivity: 85.7%). Then the proposed classifiers were used: (i) to distinguish a set of AR binders into antagonists and agonists; (ii) to screen a combined estrogen receptor binder database to find out possible chemicals that can bind to both AR and ER; and (iii) to virtually screen our in-house environmental chemical database. The in silico screening results suggest: (i) that some compounds can affect the normal endocrine system through a complex mechanism binding both to ER and AR; (ii) new EDCs, which are nonER binders, but can in silico bind to AR, are recognized; and (iii) about 20% of compounds in a big data set of environmental chemicals are predicted as new AR antagonists. The priority should be given to them to experimentally test the binding activities with AR.

QSAR models for reproductive toxicity and endocrine disruption activity

Reproductive toxicity is an important regulatory endpoint, which is required in registration procedures of chemicals used for different purposes (for example pesticides). The in vivo tests are expensive, time consuming and require large numbers of animals, which must be sacrificed. Therefore an effort is ongoing to develop alternative In vitro and in silico methods to evaluate reproductive toxicity. In this review we describe some modeling approaches. In the first example we describe the CAESAR model for prediction of reproductive toxicity; the second example shows a classification model for endocrine disruption potential based on counter propagation artificial neural networks; the third example shows a modeling of relative binding affinity to rat estrogen receptor, and the fourth one shows a receptor dependent modeling experiment.

Current mathematical methods used in QSAR/QSPR studies

This paper gives an overview of the mathematical methods currently used in quantitative structure-activity/property relationship (QASR/QSPR) studies. Recently, the mathematical methods applied to the regression of QASR/QSPR models are developing very fast, and new methods, such as Gene Expression Programming (GEP), Project Pursuit Regression (PPR) and Local Lazy Regression (LLR) have appeared on the QASR/QSPR stage. At the same time, the earlier methods, including Multiple Linear Regression (MLR), Partial Least Squares (PLS), Neural Networks (NN), Support Vector Machine (SVM) and so on, are being upgraded to improve their performance in QASR/QSPR studies. These new and upgraded methods and algorithms are described in detail, and their advantages and disadvantages are evaluated and discussed, to show their application potential in QASR/QSPR studies in the future.

Evaluation and QSAR modeling on multiple endpoints of estrogen activity based on different bioassays

There is a great need for an effective means of rapidly assessing endocrine-disrupting activity, especially estrogen-simulating activity, due to the large number of chemicals that have serious adverse effects on the environment. Many approaches using a variety of biological screening assays are used to identify endocrine disrupting chemicals. The present investigation analyzes the consistency and peculiarity of information from different experimental assays collected from a literature survey, by studying the correlation of the different endpoints. In addition, the activity values of more widely used selected bioassays have been combined by principle components analysis (PCA) to build one cumulative endpoint, the estrogen activity index (EAI), for priority setting to identify chemicals most likely possessing estrogen activity for early entry into screening. This index was then modeled using only a few theoretical molecular descriptors. The constructed MLR-QSAR model has been statistically validated for its predictive power, and can be proposed as a preliminary evaluative method to screen/prioritize estrogens according to their integrated estrogen activity, just starting from molecular structure.

Binary classification models for endocrine disrupter effects mediated through the estrogen receptor

Endocrine disrupters (EDs) form an interesting field of application attracting great attention in the recent years. They represent a number of exogenous substances interfering with the function of the endocrine system, including the interfering with developmental processes. In particular EDs are mentioned as substances requiring a more detailed control and specific authorization within REACH, the new European legislation on chemicals, together with other groups of chemicals of particular concern. QSAR represents a challenging method to approach data gap which is foreseen by REACH. The aim of this study was to provide an insight into the use of QSAR models to address ED effects mediated through the estrogen receptor (ER). New predictive models were derived to assess estrogenicity for a very large and heterogeneous dataset of chemical compounds. QSAR binary classifiers were developed based on different data mining techniques such as classification trees, decision forest, fuzzy logic, neural networks and support vector machines. The focus was given to multiple endpoints to better characterize the effects of EDs evaluating both binding (RBA) and transcriptional activity (RA). A possible combination of the models was also explored. A very good accuracy was reached for both RA and RBA models (higher than 80%).