Base-line model for identifying the bioaccumulation potential of chemicals

Molecular designing of potential environmentally friendly PFAS based on deep learning and generative models

Perfluoroalkyl and polyfluoroalkyl substances (PFAS) are widely used across a spectrum of industrial and consumer goods. Nonetheless, their persistent nature and tendency to accumulate in biological systems pose substantial environmental and health threats. Consequently, striking a balance between maximizing product efficiency and minimizing environmental and health risks by tailoring the molecular structure of PFAS has become a pivotal challenge in the fields of environmental chemistry and sustainable development. To address this issue, a computational workflow was proposed for designing an environmentally friendly PFAS by incorporating deep learning (DL) and molecular generative models. The hybrid DL architecture MolHGT+ based on heterogeneous graph neural network with transformer-like attention was applied to predict the surface tension, bioaccumulation, and hepatotoxicity of the molecules. Through virtual screening of the PFAS master database using MolHGT+, the findings indicate that incorporating the siloxane group and betaine fragment can effectively decrease both the bioaccumulation and hepatotoxicity of PFAS while preserving low surface tension. In addition, molecular generative models were employed to create a structurally diverse pool of novel PFASs with the aforementioned hit molecules serving as the initial template structures. Overall, our study presents a promising AI-driven method for advancing the development of environmentally friendly PFAS.

Machine learning-based q-RASAR predictions of the bioconcentration factor of organic molecules estimated following the organisation for economic co-operation and development guideline 305

In this study, we utilized an innovative quantitative read-across (RA) structure-activity relationship (q-RASAR) approach to predict the bioconcentration factor (BCF) values of a diverse range of organic compounds, based on a dataset of 575 compounds tested using Organisation for Economic Co-operation and Development Test Guideline 305 for bioaccumulation in fish. Initially, we constructed the q-RASAR model using the partial least squares regression method, yielding promising statistical results for the training set (R2 =0.71, Q2LOO=0.68, mean absolute error [MAE]training=0.54). The model was further validated using the test set (Q2F1=0.77, Q2F2=0.75, MAEtest=0.51). Subsequently, we explored the q-RASAR method using other regression-based supervised machine-learning algorithms, demonstrating favourable results for the training and test sets. All models exhibited R2 and Q2F1 values exceeding 0.7, Q2LOO values greater than 0.6, and low MAE values, indicating high model quality and predictive capability for new, unidentified chemical substances. These findings represent the significance of the RASAR method in enhancing predictivity for new unknown chemicals due to the incorporation of similarity functions in the RASAR descriptors, independent of a specific algorithm.

Ecotoxicological Evaluation of Bisphenol A and Alternatives: A Comprehensive In Silico Modelling Approach

Bisphenol A (BPA), a compound widely used in industrial applications, has raised concerns due to its environmental impact. As a key component in the manufacture of polycarbonate plastics and epoxy resins used in many consumer products, concerns about potential harm to human health and the environment are unavoidable. This study seeks to address these concerns by evaluating a range of potential BPA alternatives, focusing on their ecotoxicological properties. The research examines 76 bisphenols, including BPA derivatives, using a variety of in silico ecotoxicological models, although it should be noted that these models were not developed exclusively for this particular class of compounds. Consequently, interpretations should be made with caution. The results of this study highlight specific compounds of potential environmental concern and underscore the need to develop more specific models for BPA alternatives that will allow for more accurate and reliable assessment.

How error-prone bioaccumulation experiments affect the risk assessment of hydrophobic chemicals and what could be improved

Bioaccumulation is one of the three criteria for the PBT assessment of chemicals, where P stands for persistence, B for bioaccumulation, and T for toxicity, which is a cornerstone for the "Registration, Evaluation, Authorization, and Restriction of Chemicals" (REACH) in the EU. Registrants are required by REACH to submit data on bioaccumulation if the chemical is manufactured in and/or imported to the European Economic Area at more than 100 t/year. Most of the experimental bioaccumulation studies submitted were on the bioconcentration factor (BCF) and were conducted prior to 2012, before the OECD Test Guideline 305 on Bioaccumulation in Fish was updated. An analysis of the submitted data revealed that many of the experimental data, but also the data from QSARs and other calculation methods, underestimate the actual bioaccumulation potential of hydrophobic substances considerably. One of the main reasons in the nonexperimental studies is that the BCF is related there to the total concentration of the chemical in water and not to the dissolved chemical concentration. There is therefore an urgent need to reassess the bioaccumulation potential of the hydrophobic substances registered under REACH. Based on the model calculations in the present study, between 332 and 584 substances that are registered under REACH are likely to bioaccumulate in the aquatic environment-many more than have so far been identified in the B assessment. Integr Environ Assess Manag 2023;19:792-803. © 2022 The Authors. Integrated Environmental Assessment and Management published by Wiley Periodicals LLC on behalf of Society of Environmental Toxicology & Chemistry (SETAC).

Affinity of Compounds for Phosphatydylcholine-Based Immobilized Artificial Membrane-A Measure of Their Bioconcentration in Aquatic Organisms

The BCF (bioconcentration factor) of solutes in aquatic organisms is an important parameter because many undesired chemicals enter the ecosystem and affect the wildlife. Chromatographic retention factor log k_w^IAM obtained from immobilized artificial membrane (IAM) HPLC chromatography with buffered, aqueous mobile phases and calculated molecular descriptors obtained for a group of 120 structurally unrelated compounds were used to generate useful models of log BCF. It was established that log k_w^IAM obtained in the conditions described in this study is not sufficient as a sole predictor of bioconcentration. Simple, potentially useful models based on log k_w^IAM and a selection of readily available, calculated descriptors and accounting for over 88% of total variability were generated using multiple linear regression (MLR), partial least squares (PLS) regression and artificial neural networks (ANN). The models proposed in the study were tested on an external group of 120 compounds and on a group of 40 compounds with known experimental log BCF values. It was established that a relatively simple MLR model containing four independent variables leads to satisfying BCF predictions and is more intuitive than PLS or ANN models.

Molecular docking and QSAR theoretical model for prediction of phthalazinone derivatives as new class of potent dengue virus inhibitors

Background

Dengue fever is a key public health unease in various tropical and sub-tropical regions. The improvement of existing agents that can inhibit the dengue virus is therefore of utmost importance. In this work, the QSAR study was carried out on 25 molecules of phthalazinone derivatives which have been reported to possess excellent dengue virus inhibitory activity. Density functional computational technique was used in the optimisation of the molecules with the basis set at theory level (B3LYP, 6-31G*) respectively. The multiple linear regression (MLR) model was built using genetic function approximation (GFA) in the material studio software package. Also, in this study, molecular docking simulation was carried between dengue virus serotype 2 protease (PDB CODE: 6mol) and some selected phthalazinone derivatives (compounds 1, 2, 7, 11, and 21).

Results

The model was robust as evidenced by validation and robustness statistical parameter which include predicted R2pred., adjusted R2adj., cross-validated Q2 and R2 regression coefficient, etc (R2pred. = 0.71922, R2adj. = 0.939699, Q2CV = 0.905909, R2 = 0.955567) respectively. The molecular docking studies conducted in this study have outlined the binding affinities of the selected compounds (1, 2, 7 11, and 21) which are all in good correlation with their respective pIC50 values. The free binding affinities of the selected compounds were found to be (− 8.7, − 8.8, − 8.7, − 8.3, and − 8.9 kcal/mol) respectively, compound 21 with the binding affinity of − 8.9 kcal/mol had the best binding free energy with the protease relative to other compounds under consideration.

Conclusion

The MLR-GFA model study alongside with the molecular docking analysis has essentially provided a valuable and in-depth understanding as well as knowledge for the development of novel chemical compounds with enhanced inhibitory potential against the dengue virus serotype 2 (DNV-2). Hence, the developed model can be applicable in predicting the anti-dengue activity of a new set of chemical compounds that fall within its applicability domain.

A comparison of the persistence, toxicity, and exposure to high-volume natural plant-derived and synthetic pesticides

The immobility of plants exerted evolutionary selection pressures resulting in the production of thousands of chemical substances thought to function as pesticides against predation by insects and animals. More than 10,000 plant-derived compounds have been isolated with the existence of about 100,000 such compounds postulated. In 1990, Ames et al. reported that 99.99% by weight of the pesticides ingested in a normal human diet are derived from natural plant-based sources. This surprising result raised the question as to whether these natural plant pesticides were toxic to humans. These authors examined a relatively small subset of natural pesticides and determined that their tumorigenicity in rodent cancer bioassays was similar to synthetic pesticides. In this analysis, we used standard United States Environmental Protection Agency programs to estimate the toxicity (T.E.S.T. 4.2) and persistence (EPI Suite 4.1) of a series of high-volume synthetic and natural pesticides. On average, synthetic pesticides were more persistent in the environment than were natural pesticides. This result is consistent with cost, time, and logistical constraints under which farmers apply a limited number of applications of pesticides during a crop cycle. Synthetic and natural pesticides are predicted to possess toxicities including mutagenicity and developmental toxicity. Synthetic pesticides are less often mutagenic.

QSPR models for bioconcentration factor (BCF): are they able to predict data of industrial interest?

The bioconcentration factor (BCF), a key parameter required by the REACH regulation, estimates the tendency for a xenobiotic to concentrate inside living organisms. In silico methods can be valid alternatives to costly data measurements. However, in the industrial context, these theoretical approaches may fail to predict BCF with reasonable accuracy. We analyzed whether models built on public data only have adequate performances when challenged to predict industrial compounds. A new set of 1129 compounds has been collected by merging publicly available datasets. Generative Topographic Mapping was employed to compare this chemical space with a set of new compounds issued from the industry. Some new chemotypes absent in the training set (such as siloxanes) have been detected. A new BCF model has been built using ISIDA (In SIlico design and Data Analysis) fragment descriptors, support vector regression and random forest machine-learning methods. It has been externally validated on: (i) collected data from the literature and (ii) industrial data. The latter also served as benchmark for the freely available tools VEGA, EPISuite, TEST, OPERA. New model performs (RMSE of 0.58 log BCF units) comparably to existing ones but benefits of an extended applicability, covering the industrial set chemical space (78% data coverage).

Mechanistic relationship between biodegradation and bioaccumulation. Practical outcomes

According to the REACH Regulation, for all substances manufactured or imported in amounts of 10 or more tons per year, that are not exempted from the registration requirement, a Chemical Safety Assessment (CSA) must be conducted. According to CSA criteria, for these substances persistent, bioaccumulative and toxic (PBT), and very persistent and very bioaccumulative (vPvB) assessment is requested. In order to reduce the number of applications of the expensive bioaccumulation test it seems useful to search thresholds for other related parameters above which no bioaccumulation is observed. Given the known relationship between ready biodegradability and bioaccumulation, one such parameter is biodegradation. This article addresses this relationship in searching for BOD threshold above which no vB and B chemicals could be observed. It was found that the regulatory criteria for persistency could be used for identification of not vB and B chemicals. In addition, fish liver metabolism is determined as the most significant factor in reducing of maximum bioaccumulation potential of the chemicals. It was found that parameters associated with the models simulating fish metabolism could be also used for identification of not vB and B chemicals.

Prediction of bioconcentration factors in fish and invertebrates using machine learning

The application of machine learning has recently gained interest from ecotoxicological fields for its ability to model and predict chemical and/or biological processes, such as the prediction of bioconcentration. However, comparison of different models and the prediction of bioconcentration in invertebrates has not been previously evaluated. A comparison of 24 linear and machine learning models is presented herein for the prediction of bioconcentration in fish and important factors that influenced accumulation identified. R² and root mean square error (RMSE) for the test data (n = 110 cases) ranged from 0.23-0.73 and 0.34-1.20, respectively. Model performance was critically assessed with neural networks and tree-based learners showing the best performance. An optimised 4-layer multi-layer perceptron (14 descriptors) was selected for further testing. The model was applied for cross-species prediction of bioconcentration in a freshwater invertebrate, Gammarus pulex. The model for G. pulex showed good performance with R² of 0.99 and 0.93 for the verification and test data, respectively. Important molecular descriptors determined to influence bioconcentration were molecular mass (MW), octanol-water distribution coefficient (logD), topological polar surface area (TPSA) and number of nitrogen atoms (nN) among others. Modelling of hazard criteria such as PBT, showed potential to replace the need for animal testing. However, the use of machine learning models in the regulatory context has been minimal to date and is critically discussed herein. The movement away from experimental estimations of accumulation to in silico modelling would enable rapid prioritisation of contaminants that may pose a risk to environmental health and the food chain.

Acceptable-by-design QSARs to predict the dietary biomagnification of organic chemicals in fish

This work presents the first-time QSAR approach to predict the laboratory-based fish biomagnification factor (BMF) of organic chemicals, to be used as a supporting tool for assessing bioaccumulation at the regulatory level. The developed strategy is based on 2 levels of prediction, with a varying trade-off between interpretability and performance according to the user's needs. We designed our models to be intrinsically acceptable at the regulatory level (in what we defined as "acceptable-by-design" strategy), by (i) complying with OECD principles directly in the approach development phase, (ii) choosing easy-to-apply modeling techniques, (iii) preferring simple descriptors when possible, and (iv) striving to provide data-driven mechanistic insights. Our novel tool has an error comparable to the observed experimental inter- and intraspecies variability and is stable on borderline compounds (root mean square error [RMSE] ranging from RMSE = 0.45 to RMSE = 0.45 log units on test data). Additionally, the models' molecular descriptors are carefully described and interpreted, allowing us to gather additional mechanistic insights into the structural features controlling the dietary bioaccumulation of chemicals in fish. To improve the transparency and promote the application of the model, the data set and the stand alone prediction tool are provided free of charge at https://github.com/grisoniFr/bmf_qsar Integr Environ Assess Manag 2019;15:51-63. © 2018 SETAC.

An assessment of applicability of existing approaches to predicting the bioaccumulation of conventional substances in nanomaterials

The experimental determination of bioaccumulation is challenging, and a number of approaches have been developed for its prediction. It is important to assess the applicability of these predictive approaches to nanomaterials (NMs), which have been shown to bioaccumulate. The octanol/water partition coefficient (K_OW ) may not be applicable to some NMs that are not found in either the octanol or water phases but rather are found at the interface. Thus the K_OW values obtained for certain NMs are shown not to correlate well with the experimentally determined bioaccumulation. Implementation of quantitative structure-activity relationships (QSARs) for NMs is also challenging because the bioaccumulation of NMs depends on nano-specific properties such as shape, size, and surface area. Thus there is a need to develop new QSAR models based on these new nanodescriptors; current efforts appear to focus on digital processing of NM images as well as the conversion of surface chemistry parameters into adsorption indices. Water solubility can be used as a screening tool for the exclusion of NMs with short half-lives. Adaptation of fugacity/aquivalence models, which include physicochemical properties, may give some insights into the bioaccumulation potential of NMs, especially with the addition of a biota component. The use of kinetic models, including physiologically based pharmacokinetic models, appears to be the most suitable approach for predicting bioaccumulation of NMs. Furthermore, because bioaccumulation of NMs depends on a number of biotic and abiotic factors, it is important to take these factors into account when one is modeling bioaccumulation and interpreting bioaccumulation results. Environ Toxicol Chem 2018;37:2972-2988. © 2018 SETAC.

Detecting the bioaccumulation patterns of chemicals through data-driven approaches

This work investigates the bioaccumulation patterns of 168 organic chemicals in fish, by comparing their bioconcentration factor (BCF), biomagnification factor (BMF) and octanol-water partitioning coefficient (K_OW). It aims to gain insights on the relationships between dietary and non-dietary bioaccumulation in aquatic environment, on the effectiveness of K_OW and BCF to detect compounds that bioaccumulate through diet, as well as to detect the presence of structure-related bioaccumulation patterns. A linear relationship between logBMF and logK_OW was observed (logBMF = 1.14·logBCF - 6.20) up to logK_OW ≈ 4, as well as between logBMF and logBCF (logBMF = 0.96·logBCF - 4.06) up to a logBCF ≈ 5. 10% of compounds do not satisfy the linear BCF-BMF relationship. The deviations from such linear relationships were further investigated with the aid of a self-organizing map and canonical correlation analysis, which allowed us to shed light on some structure-related patterns. Finally, the usage of K_OW- and BCF-based thresholds to detect compounds that accumulate through diet led to many false positives (47%-91% for K_OW), and a moderate number of false negatives (up to 5% for BCF). These results corroborate the need of using the experimental BMF for hazard assessment practices, as well as of developing computational tools for BMF prediction.

3D matters! 3D-RISM and 3D convolutional neural network for accurate bioaccumulation prediction

In this work, we present a new method for predicting complex physical-chemical properties of organic molecules. The approach utilizes 3D convolutional neural network (ActivNet4) that uses solvent spatial distributions around solutes as input. These spatial distributions are obtained by a molecular theory called three-dimensional reference interaction site model. We have shown that the method allows one to achieve a good accuracy of prediction of bioconcentration factor which is difficult to predict by direct application of methods of molecular theory or simulations. Our research demonstrates that combination of molecular theories with modern machine learning approaches can be effectively used for predicting properties that are otherwise inaccessible to purely theory-based models.

Estimating the Bioconcentration Factors of Hydrophobic Organic Compounds from Biotransformation Rates Using Rainbow Trout Hepatocytes

Determining the biotransformation potential of commercial chemicals is critical for estimating their persistence in the aquatic environment. In vitro systems are becoming increasingly important as screening methods for assessing the potential for chemical metabolism. Depletion rate constants (k_d) for several organic chemicals with high octanol-water partition coefficient (K_ow) values (9-methylanthracene, benzo(a)pyrene, chrysene, and PCB-153) in rainbow trout hepatocytes were determined to estimate biotransformation rate constants (k_MET) that were used in fish bioconcentration factor (BCF) models. Benzo[a]pyrene was rapidly biotransformed when incubated singly; however, its depletion rate constant (k_d) declined 79% in a mixture of all four chemicals. Chrysene also exhibited significant biotransformation and its depletion rate constant declined by 50% in the mixture incubation. These data indicate that biotransformation rates determined using single chemicals may overestimate metabolism in environments containing chemical mixtures. Incubations with varying cell concentrations were used to determine whether cell concentration affected k_d estimates. No statistically significant change in depletion rate constants were seen, possibly due to an increase in nonspecific binding of hydrophobic chemicals as cell density increased, decreasing overall biotransformation. A new model was used to estimate BCFs from k_MET values calculated from empirically derived k_d values. The inclusion of k_MET in models resulted in significantly lower BCF values (compared k_MET = 0). Modelled BCF values were consistent with empirically derived BCF values from the literature.

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.

Fate and exposure modeling in regulatory chemical evaluation: new directions from retrospection

The development and application of fate and exposure modeling has undergone fundamental changes over the last 20 years. This has, in part, been driven by different needs within the regulatory community to address chemicals of concern using different approaches. Here we present a retrospective look at fate and exposure model application over the last two decades keeping an international regulatory perspective and using the Government of Canada's Chemicals Management Plan to illustrate concepts. We discuss the important role fate and exposure modeling has played to help address key data gaps when evaluating the risk of chemicals for both human health and ecological reasons. Yet limitations for more widespread model application within a regulatory context remain. Consequently, we identify specific data gaps and regulatory needs with an eye towards new directions for 21^st century chemical evaluation. We suggest that one factor limiting greater model application is the need for increased awareness and agreement of what chemical exposure assessment encompasses within the risk assessment paradigm. This is of particular importance today because of the increased availability of computational and high-throughput data and methods for chemical assessment allowing evaluators to potentially examine exposure from site of release to site of toxic action, thus linking exposure with toxicology. We further suggest there is a need for discussion at a global level to promote the awareness of new tools and approaches available for fate and exposure modeling and suggest that this could be organized using the aggregate exposure pathways concept.

The use of immobilized artificial membrane chromatography to predict bioconcentration of pharmaceutical compounds

The potential of immobilized artificial membrane chromatography (IAM) to predict bioconcentration factors (BCF) of pharmaceutical compounds in aquatic organisms was studied. For this purpose, retention factors extrapolated to pure aqueous phase, logk_w(IAM), of 27 drugs were measured on an IAM stationary phase, IAM.PC.MG type. The data were combined with retention factors on two IAM columns, IAM.PC.MG and IAM.PC.DD2 types, reported previously by our research group and correlated with logBCF values predicted by Estimation Program Interface (EPI Suite) Software. Linear models were established upon exclusion of ionic or highly hydrophilic nonionic drugs, for which a constant value of logBCF equal to 0.50 was arbitrarily assigned by EPI Suite Software. As additional physicochemical parameter BioWin5 proved to be statistically significant, expressing the decrease of bioaccumulation potential as a result of biodegradation in the aquatic environment. The constructed IAM model was successfully validated by application to a set of pharmaceuticals, whose experimental BCF values are available. Better predictions compared to EPI Suite Software were achieved for the dataset under study. Since bioconcentration process involves electrostatic interactions, IAM retention may be a better measure for BCF values, especially for ionic species, compared to octanol-water partition coefficients widely implemented in environmental sciences. The developed approach can be considered as a novel tool for the prediction of bioconcentration of pharmaceutical compounds in aquatic organisms in order to minimize further experimental assays in the future.

Assessing the bioaccumulation potential of ionizable organic compounds: Current knowledge and research priorities

The objective of the present study was to review the current knowledge regarding the bioaccumulation potential of ionizable organic compounds (IOCs), with a focus on the availability of empirical data for fish. Aspects of the bioaccumulation potential of IOCs in fish that can be characterized relatively well include the pH dependence of gill uptake and elimination, uptake in the gut, and sorption to phospholipids (membrane-water partitioning). Key challenges include the lack of empirical data for biotransformation and binding in plasma. Fish possess a diverse array of proteins that may transport IOCs across cell membranes. Except in a few cases, however, the significance of this transport for uptake and accumulation of environmental contaminants is unknown. Two case studies are presented. The first describes modeled effects of pH and biotransformation on the bioconcentration of organic acids and bases, while the second employs an updated model to investigate factors responsible for accumulation of perfluorinated alkyl acids. The perfluorinated alkyl acid case study is notable insofar as it illustrates the likely importance of membrane transporters in the kidney and highlights the potential value of read-across approaches. Recognizing the current need to perform bioaccumulation hazard assessments and ecological and exposure risk assessment for IOCs, the authors provide a tiered strategy that progresses (as needed) from conservative assumptions (models and associated data) to more sophisticated models requiring chemical-specific information. Environ Toxicol Chem 2017;36:882-897. © 2016 SETAC.

From Sediment to Top Predators: Broad Exposure of Polyhalogenated Carbazoles in San Francisco Bay (U.S.A.)

The present study provides the first comprehensive investigation of polyhalogenated carbazoles (PHCZs) contamination in an aquatic ecosystem. PHCZs have been found in soil and aquatic sediment from several different regions, but knowledge of their bioaccumulation and trophodynamics is extremely scarce. This work investigated a suite of 11 PHCZ congeners in San Francisco Bay (United States) sediment and organisms, including bivalves (n = 6 composites), sport fish (n = 12 composites), harbor seal blubber (n = 18), and bird eggs (n = 8 composites). The most detectable congeners included 3,6-dichlorocarbazole (36-CCZ), 3,6-dibromocarbazole (36-BCZ), 1,3,6-tribromocarbazole (136-BCZ), 1,3,6,8-tetrabromocarbazole (1368-BCZ), and 1,8-dibromo-3,6-dichlorocarbazole (18-B-36-CCZ). The median concentrations of ΣPHCZs were 9.3 ng/g dry weight in sediment and ranged from 33.7 to 164 ng/g lipid weight in various species. Biomagnification was observed from fish to harbor seal and was mainly driven by chlorinated carbazoles, particularly 36-CCZ. Congener compositions of PHCZs differed among species, suggesting that individual congeners may be subject to different bioaccumulation or metabolism in species occupying various trophic levels in the studied aquatic system. Toxic equivalent (TEQ) values of PHCZs were determined on the basis of their relative effect potencies (REP) compared to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). The median TEQ was 1.2 pg TEQ/g dry weight in sediment and 4.8-19.5 pg TEQ/g lipid weight in biological tissues. Our study demonstrated the broad exposure of PHCZs in San Francisco Bay and their characteristics of bioaccumulation and biomagnification along with dioxin-like effects. These findings raise the need for additional research to better elucidate their sources, environmental behavior, and fate in global environments.

AutoQSAR: an automated machine learning tool for best-practice quantitative structure-activity relationship modeling

AIM

We introduce AutoQSAR, an automated machine-learning application to build, validate and deploy quantitative structure-activity relationship (QSAR) models.

METHODOLOGY/RESULTS

The process of descriptor generation, feature selection and the creation of a large number of QSAR models has been automated into a single workflow within AutoQSAR. The models are built using a variety of machine-learning methods, and each model is scored using a novel approach. Effectiveness of the method is demonstrated through comparison with literature QSAR models using identical datasets for six end points: protein-ligand binding affinity, solubility, blood-brain barrier permeability, carcinogenicity, mutagenicity and bioaccumulation in fish.

CONCLUSION

AutoQSAR demonstrates similar or better predictive performance as compared with published results for four of the six endpoints while requiring minimal human time and expertise.

Investigating the mechanisms of bioconcentration through QSAR classification trees

This paper proposes a scheme to predict whether a compound (1) is mainly stored within lipid tissues, (2) has additional storage sites (e.g., proteins), or (3) is metabolized/eliminated with a reduced bioconcentration. The approach is based on two validated QSAR (Quantitative Structure-Activity Relationship) trees, whose salient features are: (a) descriptor interpretability and (b) simplicity. Trees were developed for 779 organic compounds, the TGD approach was used to quantify the lipid-driven bioconcentration, and a refined machine-learning optimization procedure was applied. We focused on molecular descriptor interpretation, which allowed us to gather new mechanistic insights into the bioconcentration mechanisms.

Evaluation of QSAR models for predicting the partition coefficient (log P) of chemicals under the REACH regulation

The partition coefficient (log P) is a physicochemical parameter widely used in environmental and health sciences and is important in REACH and CLP regulations. In this regulatory context, the number of existing experimental data on log P is negligible compared to the number of chemicals for which it is necessary. There are many models to predict log P and we have selected a number of free programs to examine how they predict the log P of chemicals registered for REACH and to evaluate wheter they can be used in place of experimental data. Some results are good, especially if the information on the applicability domain of the models is considered, with R(2) values from 0.7 to 0.8 and root mean square error (RMSE) from 0.8 to 1.5.

Integrating QSAR and read-across for environmental assessment

Read-across and QSAR have different traditions and drawbacks. We address here two main questions: (1) How do we solve the issue of the subjectivity in the evaluation of data and results, which may be particularly critical for read-across, but may have a role also for the QSAR assessment? (2) How do we take advantage of the results of both approaches to support each other? The QSAR model starts from the training set. The presence of similar chemicals with property values close to that predicted can support the result. The approach in read-across is the opposite. The assessment is focused on the few substances similar to the target. The data quality of the similar chemicals is fundamental. A risk is poor standardization in the definition of 'similarity', because different approaches may be applied. Inspired by the principles of high transparency and reproducibility, a new program for read-across, called ToxRead, has been developed and made freely available ( www.toxgate.eu ). The output of ToxRead can be compared and integrated with the output of QSAR, within a weight-of-evidence strategy. We discuss the evaluation and integration of ToxRead and QSAR with examples of the assessment of bioconcentration factors of chemicals.

Observed and modeled effects of pH on bioconcentration of diphenhydramine, a weakly basic pharmaceutical, in fathead minnows

A need exists to better understand the influence of pH on the uptake and accumulation of ionizable pharmaceuticals in fish. In the present study, fathead minnows were exposed to diphenhydramine (DPH; disassociation constant = 9.1) in water for up to 96 h at 3 nominal pH levels: 6.7, 7.7, and 8.7. In each case, an apparent steady state was reached by 24 h, allowing for direct determination of the bioconcentration factor (BCF), blood-water partitioning (PBW,TOT), and apparent volume of distribution (approximated from the whole-body-plasma concentration ratio). The BCFs and measured PBW,TOT values increased in a nonlinear manner with pH, whereas the volume of distribution remained constant, averaging 3.0 L/kg. The data were then simulated using a model that accounts for acidification of the gill surface caused by elimination of metabolically produced acid. Good agreement between model simulations and measured data was obtained for all tests by assuming that plasma binding of ionized DPH is 16% that of the neutral form. A simpler model, which ignores elimination of metabolically produced acid, performed less well. These findings suggest that pH effects on accumulation of ionizable compounds in fish are best described using a model that accounts for acidification of the gill surface. Moreover, measured plasma binding and volume of distribution data for humans, determined during drug development, may have considerable value for predicting chemical binding behavior in fish.

Evaluation and comparison of benchmark QSAR models to predict a relevant REACH endpoint: The bioconcentration factor (BCF)

The bioconcentration factor (BCF) is an important bioaccumulation hazard assessment metric in many regulatory contexts. Its assessment is required by the REACH regulation (Registration, Evaluation, Authorization and Restriction of Chemicals) and by CLP (Classification, Labeling and Packaging). We challenged nine well-known and widely used BCF QSAR models against 851 compounds stored in an ad-hoc created database. The goodness of the regression analysis was assessed by considering the determination coefficient (R(2)) and the Root Mean Square Error (RMSE); Cooper's statistics and Matthew's Correlation Coefficient (MCC) were calculated for all the thresholds relevant for regulatory purposes (i.e. 100L/kg for Chemical Safety Assessment; 500L/kg for Classification and Labeling; 2000 and 5000L/kg for Persistent, Bioaccumulative and Toxic (PBT) and very Persistent, very Bioaccumulative (vPvB) assessment) to assess the classification, with particular attention to the models' ability to control the occurrence of false negatives. As a first step, statistical analysis was performed for the predictions of the entire dataset; R(2)>0.70 was obtained using CORAL, T.E.S.T. and EPISuite Arnot-Gobas models. As classifiers, ACD and logP-based equations were the best in terms of sensitivity, ranging from 0.75 to 0.94. External compound predictions were carried out for the models that had their own training sets. CORAL model returned the best performance (R(2)ext=0.59), followed by the EPISuite Meylan model (R(2)ext=0.58). The latter gave also the highest sensitivity on external compounds with values from 0.55 to 0.85, depending on the thresholds. Statistics were also compiled for compounds falling into the models Applicability Domain (AD), giving better performances. In this respect, VEGA CAESAR was the best model in terms of regression (R(2)=0.94) and classification (average sensitivity>0.80). This model also showed the best regression (R(2)=0.85) and sensitivity (average>0.70) for new compounds in the AD but not present in the training set. However, no single optimal model exists and, thus, it would be wise a case-by-case assessment. Yet, integrating the wealth of information from multiple models remains the winner approach.

Classifying bio-concentration factor with random forest algorithm, influence of the bio-accumulative vs. non-bio-accumulative compound ratio to modelling result, and applicability domain for random forest model

In environmental risk assessment, the bio-concentration factor (BCF) is a widely used parameter in the estimation of the bio-accumulation potential of chemicals. BCF data often have an uneven distribution of classes (bio-accumulative vs. non-bio-accumulative), which could severely bias the classification results towards the prevailing class. The present study focuses on the influence of uneven distribution of the classes in training phase of Random Forest (RF) classification models. Three different training set designs were used and descriptors selected to the models based on the occurrence frequency in RF trees and considering the mechanistic aspects they reflect. Models were compared and their classification performance was analysed, indicating good predictive characteristics (sensitivity = 0.90 and specificity = 0.83) for the balanced set; also imbalanced sets have their strengths in certain application scenarios. The confidence of classifications was assessed with a new schema for the applicability domain that makes use of the RF proximity matrix by analysing the similarity between the predicted compound and the training set of the model. All developed models were made available in the transparent, accessible and reproducible way in QsarDB repository (http://dx.doi.org/10.15152/QDB.116).

Development of a list of reference chemicals for evaluating alternative methods to in vivo fish bioaccumulation tests

The aim to reduce the number of animals in experiments has highlighted the need to develop and validate nonanimal methods as alternatives to bioaccumulation studies using fish. The present study details a novel 3-tier approach to develop a list of reference compounds to aid this process. The approach was based on 1) the inclusion of relevant chemical classes supported by high-quality in vivo data for the bioconcentration factor (BCF), whole-body biotransformation rates (K(met)), and metabolism characterization for rainbow trout (Oncorhynchus mykiss) and common carp (Cyprinus carpio) (tiers I and II); and 2) the refinement to ensure a broad coverage of hydrophobicity, bioconcentration potential, molecular weight, maximum molecular diameter, whole-body biotransformation half-lives, and metabolic pathways (tier III). In silico techniques were employed to predict maximal log BCF and molecular and metabolic properties. Of the 157 compounds considered as reference compounds, 144 were supported by high-quality BCF data, 8 were supported by K(met) data, and 5 were supported by in vivo metabolism data. Additional criteria for refinement of the list of reference compounds were suggested to aid practical implementation in experimental efforts. The present list of reference compounds is anticipated to facilitate the development of alternative approaches, enhance understanding of in vivo and in vitro bioaccumulation relationships, and refine in silico BCF and metabolism predictions.

Integrated testing strategy (ITS) for bioaccumulation assessment under REACH

REACH (registration, evaluation, authorisation and restriction of chemicals) regulation requires that all the chemicals produced or imported in Europe above 1 tonne/year are registered. To register a chemical, physicochemical, toxicological and ecotoxicological information needs to be reported in a dossier. REACH promotes the use of alternative methods to replace, refine and reduce the use of animal (eco)toxicity testing. Within the EU OSIRIS project, integrated testing strategies (ITSs) have been developed for the rational use of non-animal testing approaches in chemical hazard assessment. Here we present an ITS for evaluating the bioaccumulation potential of organic chemicals. The scheme includes the use of all available data (also the non-optimal ones), waiving schemes, analysis of physicochemical properties related to the end point and alternative methods (both in silico and in vitro). In vivo methods are used only as last resort. Using the ITS, in vivo testing could be waived for about 67% of the examined compounds, but bioaccumulation potential could be estimated on the basis of non-animal methods. The presented ITS is freely available through a web tool.

Qualitative environmental risk assessment of photolytic transformation products of iodinated X-ray contrast agent diatrizoic acid

Recent studies have confirmed that the aquatic ecosystem is being polluted with an unknown cocktail of pharmaceuticals, their metabolites and/or their transformation products (TPs). Although individual chemicals are typically present at low concentrations, they can interact with each other resulting in additive or potentially even synergistic mixture effects. Therefore it is necessary to assess the environmental risk caused by these chemicals. Data on exposure is required for quantitative risk assessment of TPs and/or metabolites. Such data are mostly missing because of the non-availability of TPs and very often metabolites for experimental testing. This study demonstrates the application of different in silico tools for qualitative risk assessment using the example of photodegradation TPs (photo-TPs) of diatrizoic acid (DIAT), which itself is not readily biodegradable. Its photolytic transformation was studied and the photodegradation pathway was established. The aerobic biodegradability of photo-TPs under the conditions of an aquatic environment was assessed using standardized OECD tests. The qualitative risk assessment of DIAT and selected photo-TPs was performed by the PBT approach (i.e. Persistence, Bioaccumulation and Toxicity), using experimental biodegradation test assays, applying different QSAR models with several different toxicological endpoints and in silico read-across approaches. The qualitative risk assessment pointed out that the photo-TPs were less persistent compared to DIAT and none of them possessed any bioaccumulation threat. However, a few photo-TPs were predicted to be active for mutagenicity and genotoxicity, which indicate the need for further testing to confirm these predictions. The present study demonstrates that in silico qualitative risk assessment analysis can increase the knowledge space about the environmental fate of TPs.

Estimation of the bioaccumulation potential of a nonchlorinated bisphenol and an ionogenic xanthene dye to Eisenia andrei in field-collected soils, in conjunction with predictive in silico profiling

In silico-based model predictions, originating from structural and mechanistic (e.g., transport, bioavailability, reactivity, and binding potential) profiling, were compared against laboratory-derived data to estimate the bioaccumulation potential in earthworms of 2 organic substances (1 neutral, 1 ionogenic) known to primarily partition to soil. Two compounds representative of specific classes of chemicals were evaluated: a nonchlorinated bisphenol containing an -OH group (4,4′-methylenebis[2,6-di-tert-butylphenol] [Binox]), and an ionogenic xanthene dye (2′,4′,5′,7′-tetrabromo-4,5,6,7-tetrachloro-3′,6′-dihydroxy-, disodium salt [Phloxine B]). Soil bioaccumulation studies were conducted using Eisenia andrei and 2 field-collected soils (a clay loam and a sandy soil). In general, the in silico structural and mechanistic profiling was consistent with the observed soil bioaccumulation tests. Binox did not bioaccumulate to a significant extent in E. andrei in either soil type; however, Phloxine B not only accumulated within tissue, but was not depurated from the earthworms during the course of the elimination phase. Structural and mechanistic profiling demonstrated the binding and reactivity potential of Phloxine B; this would not be accounted for using traditional bioaccumulation metrics, which are founded on passive-based diffusion mechanisms. This illustrates the importance of profiling for reactive ionogenic substances; even limited bioavailability combined with reactivity can result in exposures to a hazardous substance not predictable by traditional in silico modeling methods.

Modeling bioconcentration factor (BCF) using mechanistically interpretable descriptors computed from open source tool "PaDEL-Descriptor"

Predictive regression-based models for bioconcentration factor (BCF) have been developed using mechanistically interpretable descriptors computed from open source tool PaDEL-Descriptor ( http://padel.nus.edu.sg/software/padeldescriptor/ ). A data set of 522 diverse chemicals has been used for this modeling study, and extended topochemical atom (ETA) indices developed by the present authors' group were chosen as the descriptors. Due to the importance of lipohilicity in modeling BCF, XLogP (computed partition coefficient) was also tried as an additional descriptor. Genetic function approximation followed by multiple linear regression algorithm was applied to select descriptors, and subsequent partial least squares analyses were performed to establish mathematical equations for BCF prediction. The model generated from only ETA indices shows importance of seven descriptors in model development, while the model generated from ETA descriptors along with XlogP shows importance of four descriptors in model development. In general, BCF depends on lipophilicity, presence of heteroatoms, presence of halogens, fused ring system, hydrogen bonding groups, etc. The developed models show excellent statistical qualities and predictive ability. The developed models were used also for prediction of an external data set available from the literature, and good quality of predictions (R (2) pred = 0.812 and 0.826) was demonstrated. Thus, BCF can be predicted using ETA and XlogP descriptors calculated from open source PaDEL-Descriptor software in the context of aquatic chemical toxicity management.

The distribution, fate, and effects of propylene glycol substances in the environment

The propylene glycol substances comprise a homologous family of synthetic organic molecules that have widespread use and very high production volumes across the globe. The information presented and summarized here is intended to provide an overview of the most current and reliable information available for assessing the potential environmental exposures and impacts of these substances across the manufacture, use, and disposal phases of their product life cycles.The PG substances are characterized as being miscible in water, having very low octanol-water partition coefficients (log Pow) and exhibiting low potential to volatilize from water or soil in both pure and dissolved forms. The combination of these properties dictates that, almost regardless of the mode of their initial emission, they will ultimately associate with surface water, soil, and the related groundwater compartments in the environment. These substances have low affinity for soil and sediment particles, and thus will remain mobile and bio-available within these media.In the atmosphere, the PG substances are demonstrated to have short lifetimes(I. 7-11 h), due to rapid reaction with photochemically-generated hydroxyl radicals.This reactivity, combined with efficient wet deposition of their vapor and aerosol forms, lends to their very low potential for long-range transport via the atmosphere.In the aquatic and terrestrial compartments of the environment, the PG substances are rapidly and ultimately biodegraded under both aerobic and anaerobic conditions by a wide variety of microorganisms, regardless of prior adaptation to the substances.Except for the TePG substance, the propylene glycol substances meet the OECD definition of "readily biodegradable", and according to this definition are not expected to persist in either aquatic or terrestrial environments. The TePG exhibits inherent biodegradability, is not regarded to be persistent, and is expected to ultimately biodegrade in the environment, albeit at a somewhat slower rate. The apparent ease with which microorganisms and higher organisms can metabolize the PG substances, along with their low log Pow and very high water solubility values, portends them to have very low potential for bioaccumulation and/or biomagnification in aquatic and terrestrial organisms. These same properties, along with their neutral structures and lack of biological reactivity, are the reasons for which the PG substances exhibit a base-line, non-polar narcosis mode of toxicity.The PG substances have been shown here to be practically non-toxic to essentially every aquatic and terrestrial animal and plant species tested. Collectively, the available wealth of information relating to persistence, bioaccumulation, and eco-toxicity of these substances allows a definitive conclusion of their categorization as not being PBT (i.e., persistently bioaccumulative/toxic). The PBT screening and categorization of substances on the Canadian Domestic Substances List (DSL) by Environment Canada has formally concluded that each member of this substance family is "not P", "not B", and "not T' according to their associated PBT criteria.Similarly, the preceding evaluations of these high production volume substances within the OECD SIDS program concluded that MPG, DPG, and TPG are low priorities for further examination of potential impacts to humans and the environment.More extensive evaluations of potential risks to human health and the environment were recently completed by industry, as required for their registration under the European Union REACh legislation; each evaluation demonstrated that current uses, associated exposures, and controls thereof, will not result in exposures that exceed predicted no effect concentrations in the environment.

Biotransformation model of neutral and weakly polar organic compounds in fish incorporating internal partitioning

A model for whole-body in vivo biotransformation of neutral and weakly polar organic chemicals in fish is presented. It considers internal chemical partitioning and uses Abraham solvation parameters as reactivity descriptors. It assumes that only chemicals freely dissolved in the body fluid may bind with enzymes and subsequently undergo biotransformation reactions. Consequently, the whole-body biotransformation rate of a chemical is retarded by the extent of its distribution in different biological compartments. Using a randomly generated training set (n = 64), the biotransformation model is found to be: log (HLφfish ) = 2.2 (±0.3)B - 2.1 (±0.2)V - 0.6 (±0.3) (root mean square error of prediction [RMSE] = 0.71), where HL is the whole-body biotransformation half-life in days, φfish is the freely dissolved fraction in body fluid, and B and V are the chemical's H-bond acceptance capacity and molecular volume. Abraham-type linear free energy equations were also developed for lipid-water (Klipidw ) and protein-water (Kprotw ) partition coefficients needed for the computation of φfish from independent determinations. These were found to be 1) log Klipidw  = 0.77E - 1.10S - 0.47A - 3.52B + 3.37V + 0.84 (in Lwat /kglipid ; n = 248, RMSE = 0.57) and 2) log Kprotw  = 0.74E - 0.37S - 0.13A - 1.37B + 1.06V - 0.88 (in Lwat /kgprot ; n = 69, RMSE = 0.38), where E, S, and A quantify dispersive/polarization, dipolar, and H-bond-donating interactions, respectively. The biotransformation model performs well in the validation of HL (n = 424, RMSE = 0.71). The predicted rate constants do not exceed the transport limit due to circulatory flow. Furthermore, the model adequately captures variation in biotransformation rate between chemicals with varying log octanol-water partitioning coefficient, B, and V and exhibits high degree of independence from the choice of training chemicals. The present study suggests a new framework for modeling chemical reactivity in biological systems.

Integration of QSAR models for bioconcentration suitable for REACH

QSAR (Quantitative Structure Activity Relationship) models can be a valuable alternative method to replace or reduce animal test required by REACH. In particular, some endpoints such as bioconcentration factor (BCF) are easier to predict and many useful models have been already developed. In this paper we describe how to integrate two popular BCF models to obtain more reliable predictions. In particular, the herein presented integrated model relies on the predictions of two among the most used BCF models (CAESAR and Meylan), together with the Applicability Domain Index (ADI) provided by the software VEGA. Using a set of simple rules, the integrated model selects the most reliable and conservative predictions and discards possible outliers. In this way, for the prediction of the 851 compounds included in the ANTARES BCF dataset, the integrated model discloses a R(2) (coefficient of determination) of 0.80, a RMSE (Root Mean Square Error) of 0.61 log units and a sensitivity of 76%, with a considerable improvement in respect to the CAESAR (R(2)=0.63; RMSE=0.84 log units; sensitivity 55%) and Meylan (R(2)=0.66; RMSE=0.77 log units; sensitivity 65%) without discarding too many predictions (118 out of 851). Importantly, considering solely the compounds within the new integrated ADI, the R(2) increased to 0.92, and the sensitivity to 85%, with a RMSE of 0.44 log units. Finally, the use of properly set safety thresholds applied for monitoring the so called "suspicious" compounds, which are those chemicals predicted in proximity of the border normally accepted to discern non-bioaccumulative from bioaccumulative substances, permitted to obtain an integrated model with sensitivity equal to 100%.

Using conditional inference trees and random forests to predict the bioaccumulation potential of organic chemicals

The present study presents a data-oriented, tiered approach to assessing the bioaccumulation potential of chemicals according to the European chemicals regulation on Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH). The authors compiled data for eight physicochemical descriptors (partition coefficients, degradation half-lives, polarity, and so forth) for a set of 713 organic chemicals for which experimental values of the bioconcentration factor (BCF) are available. The authors employed supervised machine learning methods (conditional inference trees and random forests) to derive relationships between the physicochemical descriptors and the BCF values. In a first tier, the authors established rules for classifying a chemical as bioaccumulative (B) or nonbioaccumulative (non-B). In a second tier, the authors developed a new tool for estimating numerical BCF values. For both cases the optimal set of relevant descriptors was determined; these are biotransformation half-life and octanol-water distribution coefficient (log D) for the classification rules and log D, biotransformation half-life, and topological polar surface area for the BCF estimation tool. The uncertainty of the BCF estimates obtained with the new estimation tool was quantified by comparing the estimated and experimental BCF values of the 713 chemicals. Comparison with existing BCF estimation methods indicates that the performance of this new BCF estimation tool is at least as high as that of existing methods. The authors recommend the present study's classification rules and BCF estimation tool for a consensus application in combination with existing BCF estimation methods.

Comparative analysis of local and consensus quantitative structure-activity relationship approaches for the prediction of bioconcentration factor

Quantitative structure-activity relationships (QSARs) are broadly classified as global or local, depending on their molecular constitution. Global models use large and diverse training sets covering a wide range of chemical space. Local models focus on smaller structurally or chemically similar subsets that are conventionally selected by human experts or alternatively using clustering analysis. The current study focuses on the comparative analysis of different clustering algorithms (expectation-maximization, K-means and hierarchical) for seven different descriptor sets as structural characteristics and two rule-based approaches to select subsets for designing local QSAR models. A total of 111 local QSAR models are developed for predicting bioconcentration factor. Predictions from local models were compared with corresponding predictions from the global model. The comparison of coefficients of determination (r(2)) and standard deviations for local models with similar subsets from the global model show improved prediction quality in 97% of cases. The descriptor content of derived QSARs is discussed and analyzed. Local QSAR models were further consolidated within the framework of consensus approach. All different consensus approaches increased performance over the global and local models. The consensus approach reduced the number of strongly deviating predictions by evening out prediction errors, which were produced by some local QSARs.

Resampling the bioconcentration factors data from Japan's chemical substances control law database to simulate and evaluate the bioconcentration factors derived from minimized aqueous exposure tests

Existing standard bioconcentration tests (e.g., the Organization for Economic Cooperation and Development [OECD] test guideline 305) require large numbers of test animals and resources. The minimized aqueous exposure test is a new approach based on the standard bioconcentration test but allows estimation of bioconcentration factor (BCF) by minimized sampling of the test fish. The authors collected BCF data (298 curves from 155 chemicals, using common carp as test species) from Japan's Chemical Substances Control Law database and resampled the data to simulate the calculation of BCF that would be obtained if studies had been designed to obtain kinetic BCF derived from minimized aqueous exposure tests (BCF(km)). The correlation was high (r(2)  = 0.967) between BCF derived from standard bioconcentration tests (BCF(full)) and BCF(km). The average value of the BCF(full) to BCF(km) ratio (BCF(full):BCF(km)) was 1.04 and ranged from 0.54 to 1.93, the 5th and 95th percentiles being 0.74 and 1.45, respectively. The results based on the 5th and 95th percentiles of the BCF(full):BCF(km) ratio suggest that BCF(full) 2,000 corresponds to BCF(km) 1,400 to 2,700, whereas BCF(full) 5,000 corresponds to BCF(km) 3,400 to 6,800. The authors also emphasize that the standard bioconcentration test should be performed when the resulting BCF(km) is in the region of regulatory concern.

Development and evaluation of a mechanistic bioconcentration model for ionogenic organic chemicals in fish

A mechanistic mass balance bioconcentration model is developed and parameterized for ionogenic organic chemicals (IOCs) in fish and evaluated against a compilation of empirical bioconcentration factors (BCFs). The model is subsequently applied to a set of perfluoroalkyl acids. Key aspects of model development include revised methods to estimate the chemical absorption efficiency of IOCs at the respiratory surface (E(W) ) and the use of distribution ratios to characterize the overall sorption capacity of the organism. Membrane-water distribution ratios (D(MW) ) are used to characterize sorption to phospholipids instead of only considering the octanol-water distribution ratio (D(OW) ). Modeled BCFs are well correlated with the observations (e.g., r(2)  = 0.68 and 0.75 for organic acids and bases, respectively) and accurate to within a factor of three on average. Model prediction errors appear to be largely the result of uncertainties in the biotransformation rate constant (k(M) ) estimates and the generic approaches for estimating sorption capacity (e.g., D(MW) ). Model performance for the set of perfluoroalkyl acids considered is highly dependent on the input parameters describing hydrophobicity (i.e., log K(OW) of the neutral form). The model applications broadly support the hypothesis that phospholipids contribute substantially to the sorption capacity of fish, particularly for compounds that exhibit a high degree of ionization at biologically relevant pH. Additional empirical data on biotransformation and sorption to phospholipids and subsequent incorporation into property estimation approaches (e.g., k(M) , D(MW) ) are priorities with respect to improving model performance.

Tools and perspectives for assessing chemical mixtures and multiple stressors

The present paper summarizes the most important insights and findings of the EU NoMiracle project with a focus on (1) risk assessment of chemical mixtures, (2) combinations of chemical and natural stressors, and (3) the receptor-oriented approach in cumulative risk assessment. The project aimed at integration of methods for human and ecological risk assessment. A mechanistically based model, considering uptake and toxicity as a processes in time, has demonstrated considerable potential for predicting mixture effects in ecotoxicology, but requires the measurement of toxicity endpoints at different moments in time. Within a novel framework for risk assessment of chemical mixtures, the importance of environmental factors on toxicokinetic processes is highlighted. A new paradigm for applying personal characteristics that determine individual exposure and sensitivity in human risk assessment is suggested. The results are discussed in the light of recent developments in risk assessment of mixtures and multiple stressors.

Does rational selection of training and test sets improve the outcome of QSAR modeling?

Prior to using a quantitative structure activity relationship (QSAR) model for external predictions, its predictive power should be established and validated. In the absence of a true external data set, the best way to validate the predictive ability of a model is to perform its statistical external validation. In statistical external validation, the overall data set is divided into training and test sets. Commonly, this splitting is performed using random division. Rational splitting methods can divide data sets into training and test sets in an intelligent fashion. The purpose of this study was to determine whether rational division methods lead to more predictive models compared to random division. A special data splitting procedure was used to facilitate the comparison between random and rational division methods. For each toxicity end point, the overall data set was divided into a modeling set (80% of the overall set) and an external evaluation set (20% of the overall set) using random division. The modeling set was then subdivided into a training set (80% of the modeling set) and a test set (20% of the modeling set) using rational division methods and by using random division. The Kennard-Stone, minimal test set dissimilarity, and sphere exclusion algorithms were used as the rational division methods. The hierarchical clustering, random forest, and k-nearest neighbor (kNN) methods were used to develop QSAR models based on the training sets. For kNN QSAR, multiple training and test sets were generated, and multiple QSAR models were built. The results of this study indicate that models based on rational division methods generate better statistical results for the test sets than models based on random division, but the predictive power of both types of models are comparable.

Development of ecotoxicity QSAR models based on partial charge descriptors for acrylate and related compounds

Using Gasteiger's partial equalization of orbital electronegativity (PEOE) method, we constructed ecotoxicity prediction equations based on two-dimensional descriptors for α,β-unsaturated carbonyl compounds. After examining electrostatic effects on the calculated ecotoxicities of 10 α,β-unsaturated ketones and aldehydes (A-group compounds) by using the Mulliken atomic charges on the carbonyl oxygen atoms, we investigated the efficacy of the PEOE descriptors for the same 10 compounds and the correlation between the PEOE descriptors and the Mulliken charge. We then constructed QSAR models for acute fish and Daphnia toxicities by using the PEOE descriptors for acrylic acids and compounds with acrylate-like substructures (CH-group compounds). In the constructed models, the adjusted squared correlation coefficients between measured and calculated toxicities with the lowest Akaike information criterion were 0.77 and 0.79, respectively. The applicability of the constructed models was then evaluated for various methacrylates and similar compounds (CH(3)-group compounds). Both the fish and the Daphnia toxicities of some of the CH(3)-group compounds were underestimated by these models. Nevertheless, we concluded that the QSAR models based on the PEOE descriptors were practical for predicting acute toxicity, especially for α,β-unsaturated carbonyl compounds with an α-hydrogen. Combining hydrophobicity and PEOE descriptors led to accurate predictions for fish toxicity.

Quantitative consensus of bioaccumulation models for integrated testing strategies

A quantitative consensus model based on bioconcentration factor (BCF) predictions obtained from five quantitative structure-activity relationship models was developed for bioaccumulation assessment as an integrated testing approach for waiving. Three categories were considered: non-bioaccumulative, bioaccumulative and very bioaccumulative. Five in silico BCF models were selected and included into a quantitative consensus model by means of the continuous formulation of Bayes' theorem. The discrete likelihoods commonly used in the qualitative Bayesian model were substituted by probability density functions to reduce the loss of information that occurred when continuous BCF values were distributed across the three bioaccumulation categories. Results showed that the continuous Bayesian model yielded the best classification predictions compared not only to the discrete Bayesian model, but also to the individual BCF models. The proposed quantitative consensus model proved to be a suitable approach for integrated testing strategies for continuous endpoints of environmental interest.

Linear and non-linear relationships between bioconcentration and hydrophobicity: theoretical consideration

A non-linear relationship (e.g. Gaussian-type) between measured bioconcentration factor (BCF) and octanol/water partition coefficient (K(OW)) was noted many years ago. Many studies have focused on the cause of the breakdown in the log BCF/log K(OW) curve for highly hydrophobic chemicals with log K(OW)>6. However, there has been little investigation on the theoretical background of this feature for highly hydrophilic chemicals. In this paper, the cause of linear and non-linear relationships between log BCF and log K(OW) has been investigated on the basis of the partitioning-based mechanism for classified non-ionic and ionisable compounds. For highly hydrophilic compounds, lipid tissue in fish is not the major storage site of chemicals. Uptake from other tissues/organs plays a much more important role than the lipid content, leading to a variation of measured log BCF around 0.5. For hydrophobic chemicals with 0.5<log K(OW)<6, hydrophobicity is the principal driving force of bioconcentration and log BCF increases with increasing log K(OW). The log BCF/log K(OW) curve breaks down for highly hydrophobic chemicals with log K(OW)>6. The main reason for this is attributed to the reduced bioavailability of chemicals in water. A linear solvation energy relationship shows that the bioconcentration increases with increasing molecular size by increasing the dispersion interactions between the chemical and lipid content. Bioconcentration decreases with increasing the basicity of hydrophobic compounds by increasing the H-bonding of chemicals with water. Principal component analysis shows that the octanol/water system is the closest system, but not an ideal surrogate, to describe the bioconcentration for hydrophobic compounds as compared with other solvent/water partition systems.

Comparison of bioconcentration and biomagnification factors for poorly water-soluble chemicals using common carp (Cyprinus carpio L.)

Existing regulatory criteria for bioaccumulation assessment of chemicals are mainly based on a bioconcentration factors (BCF) not a biomagnification factors (BMF). We performed dietary exposure tests for nine poorly water-soluble chemicals and developed a linear regression between the 5 % lipid normalized BCF (BCF(L)) and the lipid-corrected BMF (BMF(L)). The BMF(L) of substances with BCF(L) = 5,000 was 0.31 (95 % CI 0.11-0.87), whereas the BCF(L) of substances with BMF(L) = 1 was 13,000 (95 % CI 5,600-30,000). Five substances can be considered very bioaccumulative (vB) according to the BCF end point (BCF > 5,000), but only two substances were recognized to biomagnify according to the BMF end point (BMF ≥ 1). Although our results are highly suggestive of a relationship between BCF and BMF, additional BMF and trophic magnification factor data for chemicals are required to support this relationship, and new techniques (e.g., fugacity approach) may help in resolving the apparent contradiction in hazard categorization.

Screening for PBT chemicals among the "existing" and "new" chemicals of the EU

Under the European chemicals legislation, REACH, industrial chemicals that are imported or manufactured at more than 10 t/yr need to be evaluated with respect to their persistence (P), bioaccumulation potential (B), and toxicity (T). This assessment has to be conducted for several 10,000 of chemicals but, at the same time, empirical data on degradability, bioaccumulation potential and toxicity of industrial chemicals are still scarce. Therefore, the identification of PBT chemicals among all chemicals on the market remains a challenge. We present a PBT screening of approximately 95,000 chemicals based on a comparison of estimated P, B, and T properties of each chemical with the P, B, and T thresholds defined under REACH. We also apply this screening procedure to a set of 2576 high production volume chemicals and a set of 2781 chemicals from the EU's former list of "new chemicals" (ELINCS). In the set of 95,000 chemicals, the fraction of potential PBT chemicals is around 3%, but in the ELINCS chemicals it reaches 5%. We identify the most common structural elements among the potential PBT chemicals. Analysis of the P, B, and T data for all chemicals considered here shows that the uncertainty in persistence data contributes most to the uncertainty in the number of potential PBT chemicals.

Screening of chemicals for human bioaccumulative potential with a physiologically based toxicokinetic model

Human bioaccumulative potential is an important element in the risk assessment of chemicals. Due to the high number of synthetic chemicals, there exists the need to develop prioritisation strategies. The purpose of this study was to develop a predictive tool for human bioaccumulation risk assessment that incorporates not only the chemical properties of the compounds, but also the processes that tend to decrease the concentration of the compound such as metabolisation. We used a generic physiologically based toxicokinetic model that based on in vitro human liver metabolism data, minimal renal excretion and a constant exposure was able to assess the bioaccumulative potential of a chemical. The approach has been analysed using literature data on well-known bioaccumulative compounds and liver metabolism data from the ECVAM database and a subset of the ToxCast phase I chemical library-in total 94 compounds covering pharmaceuticals, plant protection products and industrial chemicals. Our results provide further evidence that partitioning properties do not allow for a reliable screening criteria for human chemical hazard. Our model, based on a 100% intestinal absorption assumption, suggests that metabolic clearance, plasma protein-binding properties and renal excretion are the main factors in determining whether bioaccumulation will occur and its amount. It is essential that in vitro metabolic clearance tests with metabolic competent cell lines as well as plasma protein-binding assays be performed for suspected bioaccumulative compounds.