Using Machine Learning to Classify Bioactivity for 3486 Per- and Polyfluoroalkyl Substances (PFASs) from the OECD List

Advancing toxicity studies of per- and poly-fluoroalkyl substances (pfass) through machine learning: Models, mechanisms, and future directions

Perfluorinated and perfluoroalkyl substances (PFASs), encompassing a vast array of isomeric chemicals, are recognized as typical emerging contaminants with direct or potential impacts on human health and the ecological environment. With the complex and elusive toxicological profiles of PFASs, machine learning (ML) has been increasingly employed in their toxicity studies due to its proficiency in prediction and data analytics. This integration is poised to become a predominant trend in environmental toxicology, propelled by the swift advancements in computational technology. This review diligently examines the literature to encapsulate the varied objectives of employing ML in the toxicity studies of PFASs: (1) Utilizing ML to establish Quantitative Structure-Activity Relationship (QSAR) models for PFASs with diverse toxicity endpoints, facilitating the targeted toxicity prediction of unidentified PFASs; (2) Investigating and substantiating the Adverse Outcome Pathway (AOP) through the synergy of ML and traditional toxicological methods, with this refining the toxicity assessment framework for PFASs; (3) Dissecting and elucidating the features of established ML models to advance Open Research into the toxicity of PFASs, with a primary focus on determinants and mechanisms. The discourse extends to an in-depth examination of ML studies, segregating findings based on their distinct application trajectories. Given that ML represents a nascent paradigm within PFASs research, this review delineates the collective challenges encountered in the ML-mediated study of PFAS toxicity and proffers strategic guidance for ensuing investigations.

Integrated Transfer Learning and Multitask Learning Strategies to Construct Graph Neural Network Models for Predicting Bioaccumulation Parameters of Chemicals

Accurate prediction of parameters related to the environmental exposure of chemicals is crucial for the sound management of chemicals. However, the lack of large data sets for training models may result in poor prediction accuracy and robustness. Herein, integrated transfer learning (TL) and multitask learning (MTL) was proposed for constructing a graph neural network (GNN) model (abbreviated as TL-MTL-GNN model) using n-octanol/water partition coefficients as a source domain. The TL-MTL-GNN model was trained to predict three bioaccumulation parameters based on enlarged data sets that cover 2496 compounds with at least one bioaccumulation parameter. Results show that the TL-MTL-GNN model outperformed single-task GNN models with and without the TL, as well as conventional machine learning models trained with molecular descriptors or fingerprints. Applicability domains were characterized by a state-of-the-art structure-activity landscape-based (abbreviated as ADSAL) methodology. The TL-MTL-GNN model coupled with the optimal ADSAL was employed to predict bioaccumulation parameters for around 60,000 chemicals, with more than 13,000 compounds identified as bioaccumulative chemicals. The high predictive accuracy and robustness of the TL-MTL-GNN model demonstrate the feasibility of integrating the TL and MTL strategy in modeling small-sized data sets. The strategy holds significant potential for addressing small data challenges in modeling environmental chemicals.

Toxicity and speciation of inorganic arsenics and their adverse effects on in vivo endpoints and oxidative stress in the marine medaka Oryzias melastigma

Here, we investigate the effects of acute and chronic exposure to arsenate (AsV) and arsenite (AsIII) in the marine medaka Oryzias melastigma. In vivo effects, biotransformation, and oxidative stress were studied in marine medaka exposed to the two inorganic arsenics for 4 or 28 days. An investigation of embryonic development revealed no effect on in vivo parameters, but the hatching rate increased in the group exposed to AsIII. Exposure to AsIII also caused the greatest accumulation of arsenic in medaka. For acute exposure, the ratio of AsV to AsIII was higher than that of chronic exposure, indicating that bioaccumulation of inorganic arsenic can induce oxidative stress. The largest increase in oxidative stress was observed following acute exposure to AsIII, but no significant degree of oxidative stress was induced by chronic exposure. During acute exposure to AsV, the increase in the enzymatic activity of glutathione-S-transferase (GST) was twice as high compared with exposure to AsIII, suggesting that GST plays an important role in the initial detoxification process. In addition, an RNA-seq-based ingenuity pathway analysis revealed that acute exposure to AsIII may be related to cell-cycle progression. A network analysis using differentially expressed genes also revealed a potential link between the generation of inflammatory cytokines and oxidative stress due to arsenic exposure.

Carcinogenic health outcomes associated with endocrine disrupting chemicals exposure in humans: A wide-scope analysis

Endocrine-disrupting chemicals (EDCs) are persistent and pervasive compounds that pose serious risks. Numerous studies have explored the effects of EDCs on human health, among which tumors have been the primary focus. However, because of study design flaws, lack of effective exposure levels of EDCs, and inconsistent population data and findings, it is challenging to draw clear conclusions on the effect of these compounds on tumor-related outcomes. Our study is the first to systematically integrate observational studies and randomized controlled trials from over 20 years and summarize over 300 subgroup associations. We found that most EDCs promote tumor development, and that exposure to residential environmental pollutants may be a major source of pesticide exposure. Furthermore, we found that phytoestrogens exhibit antitumor effects. The findings of this study can aid in the development of global EDCs regulatory health policies and alleviate the severe risks associated with EDCs exposure.

Machine learning-enhanced molecular network reveals global exposure to hundreds of unknown PFAS

Unknown forever chemicals like per- and polyfluoroalkyl substances (PFASs) are difficult to identify. Current platforms designed for metabolites and natural products cannot capture the diverse structural characteristics of PFAS. Here, we report an automatic PFAS identification platform (APP-ID) that screens for PFAS in environmental samples using an enhanced molecular network and identifies unknown PFAS structures using machine learning. Our networking algorithm, which enhances characteristic fragment matches, has lower false-positive rate (0.7%) than current algorithms (2.4 to 46%). Our support vector machine model identified unknown PFAS in test set with 58.3% accuracy, surpassing current software. Further, APP-ID detected 733 PFASs in real fluorochemical wastewater, 39 of which are previously unreported in environmental media. Retrospective screening of 126 PFASs against public data repository from 20 countries show PFAS substitutes are prevalent worldwide.

Building species trait-specific nano-QSARs: Model stacking, navigating model uncertainties and limitations, and the effect of dataset size

A strong need exists for broadly applicable nano-QSARs, capable of predicting toxicological outcomes towards untested species and nanomaterials, under different environmental conditions. Existing nano-QSARs are generally limited to only a few species but the inclusion of species characteristics into models can aid in making them applicable to multiple species, even when toxicity data is not available for biological species. Species traits were used to create classification- and regression machine learning models to predict acute toxicity towards aquatic species for metallic nanomaterials. Afterwards, the individual classification- and regression models were stacked into a meta-model to improve performance. Additionally, the uncertainty and limitations of the models were assessed in detail (beyond the OECD principles) and it was investigated whether models would benefit from the addition of more data. Results showed a significant improvement in model performance following model stacking. Investigation of model uncertainties and limitations highlighted the discrepancy between the applicability domain and accuracy of predictions. Data points outside of the assessed chemical space did not have higher likelihoods of generating inadequate predictions or vice versa. It is therefore concluded that the applicability domain does not give complete insight into the uncertainty of predictions and instead the generation of prediction intervals can help in this regard. Furthermore, results indicated that an increase of the dataset size did not improve model performance. This implies that larger dataset sizes may not necessarily improve model performance while in turn also meaning that large datasets are not necessarily required for prediction of acute toxicity with nano-QSARs.

Integrating high-throughput phenotypic profiling and transcriptomic analyses to predict the hepatosteatosis effects induced by per- and polyfluoroalkyl substances

Per- and polyfluoroalkyl substances (PFAS) is a large compound class (n > 12,000) that is extensively present in food, drinking water, and aquatic environments. Reduced serum triglycerides and hepatosteatosis appear to be the common phenotypes for different PFAS chemicals. However, the hepatosteatosis potential of most PFAS chemicals remains largely unknown. This study aims to investigate PFAS-induced hepatosteatosis using in vitro high-throughput phenotype profiling (HTPP) and high-throughput transcriptomic (HTTr) data. We quantified the in vitro hepatosteatosis effects and mitochondrial damage using high-content imaging, curated the transcriptomic data from the Gene Expression Omnibus (GEO) database, and then calculated the point of departure (POD) values for HTPP phenotypes or HTTr transcripts, using the Bayesian benchmark dose modeling approach. Our results indicated that PFAS compounds with fully saturated C-F bonds, sulfur- and nitrogen-containing functional groups, and a fluorinated carbon chain length greater than 8 have the potential to produce biological effects consistent with hepatosteatosis. PFAS primarily induced hepatosteatosis via disturbance in lipid transport and storage. The potency rankings of PFAS compounds are highly concordant among in vitro HTPP, HTTr, and in vivo hepatosteatosis phenotypes (ρ = 0.60-0.73). In conclusion, integrating the information from in vitro HTPP and HTTr analyses can accurately project in vivo hepatosteatosis effects induced by PFAS compounds.

Novel approach for identifying VOC emission characteristics based on mobile monitoring platform data and deep learning: Application of source apportionment in a chemical industrial park

Refined volatile organic compound (VOC) emission characteristics are crucial for accurate source apportionment in chemical industrial parks. The data from mobile monitoring platforms in chemical industrial parks contain pollution information that is not intuitively displayed, requiring further excavation. A novel approach was proposed to identify VOC emission characteristics using the class activation map (CAM) technology of convolutional neural network (CNN), which was applied on the mobile monitoring platform data (MD) derived from a typical fine chemical industrial park. It converts a large amount of monitoring data with high spatiotemporal complexity into simple and interpretable characteristic maps, effectively improving the identification effect of VOC emission characteristics, supporting more accurate source apportionment of VOC pollution around the park. Using this method, the VOC emission characteristics of eight key factories were identified. VOC source apportionment in the park was conducted for one day using a positive matrix factorization (PMF) model and seven combined factor profiles (CFPs) were calculated. Based on the identified VOC emission characteristics, the main pollution sources and their contributions to surrounding schools and residential areas were determined, revealing that one pesticide factory (named LKA) had the highest contribution ratio. The source apportionment results indicated that the impact of the chemical industrial park on the surrounding areas varied from morning to afternoon, which to some extent reflected the intermittent production methods employed for fine chemicals.

Quantitative cross-species comparison of serum albumin binding of per- and polyfluoroalkyl substances from five structural classes

Per- and polyfluoroalkyl substances (PFAS) are a class of over 8000 chemicals, many of which are persistent, bioaccumulative, and toxic to humans, livestock, and wildlife. Serum protein binding affinity is instrumental in understanding PFAS toxicity, yet experimental binding data is limited to only a few PFAS congeners. Previously, we demonstrated the usefulness of a high-throughput, in vitro differential scanning fluorimetry assay for determination of relative binding affinities of human serum albumin for 24 PFAS congeners from 6 chemical classes. In the current study, we used this assay to comparatively examine differences in human, bovine, porcine, and rat serum albumin binding of 8 structurally informative PFAS congeners from 5 chemical classes. With the exception of the fluorotelomer alcohol 1H, 1H, 2H, 2H-perfluorooctanol (6:2 FTOH), each PFAS congener bound by human serum albumin was also bound by bovine, porcine, and rat serum albumin. The critical role of the charged functional headgroup in albumin binding was supported by the inability of albumin of each species tested to bind 6:2 FTOH. Significant interspecies differences in serum albumin binding affinities were identified for each of the bound PFAS congeners. Relative to human albumin, perfluoroalkyl carboxylic and sulfonic acids were bound with greater affinity by porcine and rat serum albumin, and the perfluoroalkyl ether acid congener bound with lower affinity to porcine and bovine serum albumin. These comparative affinity data for PFAS binding by serum albumin from human, experimental model, and livestock species reduce critical interspecies uncertainty and improve accuracy of predictive bioaccumulation and toxicity assessments for PFAS.

An artificial intelligence platform for automated PFAS subgroup classification: A discovery tool for PFAS screening

Since structural analyses and toxicity assessments have not been able to keep up with the discovery of unknown per- and polyfluoroalkyl substances (PFAS), there is an urgent need for effective categorization and grouping of PFAS. In this study, we presented PFAS-Atlas, an artificial intelligence-based platform containing a rule-based automatic classification system and a machine learning-based grouping model. Compared with previously developed classification software, the platform's classification system follows the latest Organization for Economic Co-operation and Development (OECD) definition of PFAS and reduces the number of uncategorized PFAS. In addition, the platform incorporates deep unsupervised learning models to visualize the chemical space of PFAS by clustering similar structures and linking related classes. Through real-world use cases, we demonstrate that PFAS-Atlas can rapidly screen for relationships between chemical structure and persistence, bioaccumulation, or toxicity data for PFAS. The platform can also guide the planning of the PFAS testing strategy by showing which PFAS classes urgently require further attention. Ultimately, the release of PFAS-Atlas will benefit both the PFAS research and regulation communities.

POPs identification using simple low-code machine learning

Effectively identifying persistent organic pollutants (POPs) with extensive organic chemical datasets poses a formidable challenge but is of utmost importance. Leveraging machine learning techniques can enhance this process, but previous models often demanded advanced programming skills and high-end computing resources. In this study, we harnessed the simplicity of PyCaret, a Python-based package, to construct machine-learning models for POP screening based on 2D molecular descriptors. We compared the performance of these models against a deep convolutional neural network (DCNN) model. Utilising minimal Python code, we generated several models that exhibited superior or comparable performance to the DCNN. The most outstanding performer, the Light Gradient Boosting Machine (LGBM), achieved an accuracy of 96.20 %, an AUC of 97.70 %, and an F1 score of 82.58 %. This model outshone the DCNN model. Furthermore, it excelled in identifying POPs within the REACH PBT and compiled industrial chemical lists. Our findings highlight the accessibility and simplicity of PyCaret, requiring only a few lines of code, rendering it suitable for non-computing professionals in environmental sciences. The ability of low code machine learning tools (e.g. PyCaret) to facilitate model comparison and interpretation holds promise, encouraging prompt assessment and management of chemical substances.

Graph Convolutional Network-Enhanced Model for Screening Persistent, Mobile, and Toxic and Very Persistent and Very Mobile Substances

The global management for persistent, mobile, and toxic (PMT) and very persistent and very mobile (vPvM) substances has been further strengthened with the rapid increase of emerging contaminants. The development of a ready-to-use and publicly available tool for the high-throughput screening of PMT/vPvM substances is thus urgently needed. However, the current model building with the coupling of conventional algorithms, small-scale data set, and simplistic features hinders the development of a robust model for screening PMT/vPvM with wide application domains. Here, we construct a graph convolutional network (GCN)-enhanced model with feature fusion of a molecular graph and molecular descriptors to effectively utilize the significant correlation between critical descriptors and PMT/vPvM substances. The model is built with 213,084 substances following the latest PMT classification criteria. The application domains of the GCN-enhanced model assessed by kernel density estimation demonstrate the high suitability for high-throughput screening PMT/vPvM substances with both a high accuracy rate (86.6%) and a low false-negative rate (6.8%). An online server named PMT/vPvM profiler is further developed with a user-friendly web interface (http://www.pmt.zj.cn/). Our study facilitates a more efficient evaluation of PMT/vPvM substances with a globally accessible screening platform.

Suspect, Nontarget Screening, and Toxicity Prediction of Per- and Polyfluoroalkyl Substances in the Landfill Leachate

Landfills are the final stage of urban wastes containing perfluoroalkyl and polyfluoroalkyl substances (PFASs). PFASs in the landfill leachate may contaminate the surrounding groundwater. As major environmental pollutants, emerging PFASs have raised global concern. Besides the widely reported legacy PFASs, the distribution and potential toxic effects of numerous emerging PFASs remain unclear, and unknown PFASs still need discovery and characterization. This study proposed a comprehensive method for PFAS screening in leachate samples using suspect and nontarget analysis. A total of 48 PFASs from 10 classes were identified; nine novel PFASs including eight chloroperfluoropolyether carboxylates (Cl-PFPECAs) and bistriflimide (HNTf2) were reported for the first time in the leachate, where Cl-PFPECA-3,1 and Cl-PFPECA-2,2 were first reported in environmental media. Optimized molecular docking models were established for prioritizing the PFASs with potential activity against peroxisome proliferator-activated receptor α and estrogen receptor α. Our results indicated that several emerging PFASs of N-methyl perfluoroalkyl sulfonamido acetic acids (N-MeFASAAs), n:3 fluorotelomer carboxylic acid (n:3 FTCA), and n:2 fluorotelomer sulfonate (n:2 FTSA) have potential health risks that cannot be ignored.

Prediction of the Interactions of a Large Number of Per- and Poly-Fluoroalkyl Substances with Ten Nuclear Receptors

Per- and poly-fluoroalkyl substances (PFASs) are persistent, toxic chemicals that pose significant hazards to human health and the environment. Screening large numbers of chemicals for their ability to act as endocrine disruptors by modulating the activity of nuclear receptors (NRs) is challenging because of the time and cost of in vitro and in vivo experiments. For this reason, we need computational approaches to screen these chemicals and quickly prioritize them for further testing. Here, we utilized molecular modeling and machine-learning predictions to identify potential interactions between 4545 PFASs with ten different NRs. The results show that some PFASs can bind strongly to several receptors. Further, PFASs that bind to different receptors can have very different structures spread throughout the chemical space. Biological validation of these in silico findings should be a high priority.

Deciphering exogenous chemical carcinogenicity through interpretable deep learning: A novel approach for evaluating atmospheric pollutant hazards

Cancer remains a significant global health concern, with millions of deaths attributed to it annually. Environmental pollutants play a pivotal role in cancer etiology and contribute to the growing prevalence of this disease. The carcinogenic assessment of these pollutants is crucial for chemical health evaluation and environmental risk assessments. Traditional experimental methods are expensive and time-consuming, prompting the development of alternative approaches such as in silico methods. In this regard, deep learning (DL) has shown potential but lacks optimal performance and interpretability. This study introduces an interpretable DL model called CarcGC for chemical carcinogenicity prediction, utilizing a graph convolutional neural network (GCN) that employs molecular structural graphs as inputs. Compared to existing models, CarcGC demonstrated enhanced performance, with the area under the receiver operating characteristic curve (AUCROC) reaching 0.808 on the test set. Due to air pollution is closely related to the incidence of lung cancers, we applied the CarcGC to predict the potential carcinogenicity of chemicals listed in the United States Environmental Protection Agency's Hazardous Air Pollutants (HAPs) inventory, offering a foundation for environmental carcinogenicity screening. This study highlights the potential of artificially intelligent methods in carcinogenicity prediction and underscores the value of CarcGC interpretability in revealing the structural basis and molecular mechanisms underlying chemical carcinogenicity.

Environmental impact of PFAS: Filling data gaps using theoretical quantum chemistry and QSPR modeling

Per- and polyfluorinated alkyl substances (PFAS), known for their widespread environmental presence and slow degradation, pose significant concerns. Of the approximately 10,000 known PFAS, only a few have undergone comprehensive testing, resulting in limited experimental data. In this study, we employed a combination of physics-based methods and data-driven models to address gaps in PFAS bioaccumulation potential. Using the COnductor-like Screening MOdel for Realistic Solvents (COSMO-RS) method, we predicted n-octanol/water partition coefficients (logKOW), crucial for PFAS bioaccumulation. Our developed Quantitative Structure-Property Relationship (QSPR) model exhibited high accuracy (R2 = 0.95, RMSEC = 0.75) and strong predictive ability (Q2LOO = 0.93, RMSECV = 0.83). Leveraging the extensive NORMAN, we predicted logKOW for over 4,000 compounds, identifying 244 outliers out of 4519. Further categorizing the database into eight Organisation for Economic Co-operation and Development (OECD) categories, we confirmed fluorine atoms role in enhanced bioaccumulation. Utilizing predicted logKOW, water solubility logSW, and vapor pressure logVP values, we calculated additional physicochemical properties that are responsible for the transport and dispersion of PFAS in the environment. Parameters such as Henry's Law (kH), air-water partition coefficient (KAW), octanol-air coefficient (KOA), and soil adsorption coefficient (KOC) exhibited favorable correlations with literature data (R2 > 0.66). Our study successfully filled data gaps, contributing to the understanding of ubiquitous PFAS in the environment and estimating missing physicochemical data for these compounds.

Toward a comprehensive understanding of alicyclic compounds: Bio-effects perspective and deep learning approach

The escalating use of alicyclic compounds in modern industrial production has led to a rapid increase of these substances in the environment, posing significant health hazards. Addressing this challenge necessitates a comprehensive understanding of these compounds, which can be achieved through the deep learning approach. Graph neural networks (GNN) known for its' extraordinary ability to process graph data with rich relationships, have been employed in various molecular prediction tasks. In this study, alicyclic molecules screened from PCBA, Toxcast and Tox21 are made as general bioactivity and biological targets' activity prediction datasets. GNN-based models are trained on the two datasets, while the Attentive FP and PAGTN achieve best performance individually. In addition, alicyclic carbon atoms make the greatest contribution to biological activity, which indicate that the alicycle structures have significant impact on the carbon atoms' contribution. Moreover, there are terrific number of active molecules in other public datasets, indicates that alicyclic compounds deserve more attention in POPs control. This study uncovered deeper structural-activity relationships within these compounds, offering new perspectives and methodologies for academic research in the field.

Elucidating Key Characteristics of PFAS Binding to Human Peroxisome Proliferator-Activated Receptor Alpha: An Explainable Machine Learning Approach

Per- and polyfluoroalkyl substances (PFAS) are widely employed anthropogenic fluorinated chemicals known to disrupt hepatic lipid metabolism by binding to human peroxisome proliferator-activated receptor alpha (PPARα). Therefore, screening for PFAS that bind to PPARα is of critical importance. Machine learning approaches are promising techniques for rapid screening of PFAS. However, traditional machine learning approaches lack interpretability, posing challenges in investigating the relationship between molecular descriptors and PPARα binding. In this study, we aimed to develop a novel, explainable machine learning approach to rapidly screen for PFAS that bind to PPARα. We calculated the PPARα-PFAS binding score and 206 molecular descriptors for PFAS. Through systematic and objective selection of important molecular descriptors, we developed a machine learning model with good predictive performance using only three descriptors. The molecular size (b_single) and electrostatic properties (BCUT_PEOE_3 and PEOE_VSA_PPOS) are important for PPARα-PFAS binding. Alternative PFAS are considered safer than their legacy predecessors. However, we found that alternative PFAS with many carbon atoms and ether groups exhibited a higher affinity for PPARα. Therefore, confirming the toxicity of these alternative PFAS compounds with such characteristics through biological experiments is important.

Global classification models for predicting acute toxicity of chemicals towards Daphnia magna

Molecular descriptors reflecting structural information on hydrophobicity, reactivity, polarizability, hydrogen bond and charged groups, were used to predict the toxicity (pLC50) of chemicals towards Daphnia magna with global quantitative structure-activity/toxicity relationship (QSAR/QSTR) models. A sufficiently large dataset including 1517 chemical toxicity to Daphnia magna was divided into a training set (758 pLC50) and a test set (759 pLC50). By applying random forest algorithm, two classification models, Class Model A and Class Model B were developed, having prediction accuracy, sensitivity and specificity above 85% for Class 1 (with pLC50 ≤ 4.48) and Class 2 (with pLC50 > 4.48). The Class Model A was based on nine molecular descriptors and RF parameters of nodesize = 1, ntree = 80 and mtry = 2, and yielded accuracy of 92.3% (training set), 85.6% (test set) and 88.9% (total data set). Class Model B was based on ten descriptors and parameters, nodesize = 1, ntree = 90 and mtry = 2, produced accuracy of 88.3% (training set), 86.8% (test set) and 87.5% (total data set). The two classification models were satisfactory compared with other classification model reported in the literature, although classification models in this work dealt with more samples. Thus, the two classification models with a larger applicability domain provided efficient tools for assessing chemical aquatic toxicity towards Daphnia magna.

Investigation of the Binding Fraction of PFAS in Human Plasma and Underlying Mechanisms Based on Machine Learning and Molecular Dynamics Simulation

More than 7000 per- and polyfluorinated alkyl substances (PFAS) have been documented in the U.S. Environmental Protection Agency's CompTox Chemicals database. These PFAS can be used in a broad range of industrial and consumer applications but may pose potential environmental issues and health risks. However, little is known about emerging PFAS bioaccumulation to assess their chemical safety. This study focuses specifically on the large and high-quality data set of fluorochemicals from the related environmental and pharmaceutical chemicals databases, and machine learning (ML) models were developed for the classification prediction of the unbound fraction of compounds in plasma. A comprehensive evaluation of the ML models shows that the best blending model yields an accuracy of 0.901 for the test set. The predictions suggest that most PFAS (∼92%) have a high binding fraction in plasma. Introduction of alkaline amino groups is likely to reduce the binding affinities of PFAS with plasma proteins. Molecular dynamics simulations indicate a clear distinction between the high and low binding fractions of PFAS. These computational workflows can be used to predict the bioaccumulation of emerging PFAS and are also helpful for the molecular design of PFAS to prevent the release of high-bioaccumulation compounds into the environment.

Harnessing Semi-Supervised Machine Learning to Automatically Predict Bioactivities of Per- and Polyfluoroalkyl Substances (PFASs)

Many per- and polyfluoroalkyl substances (PFASs) pose significant health hazards due to their bioactive and persistent bioaccumulative properties. However, assessing the bioactivities of PFASs is both time-consuming and costly due to the sheer number and expense of in vivo and in vitro biological experiments. To this end, we harnessed new unsupervised/semi-supervised machine learning models to automatically predict bioactivities of PFASs in various human biological targets, including enzymes, genes, proteins, and cell lines. Our semi-supervised metric learning models were used to predict the bioactivity of PFASs found in the recent Organisation of Economic Co-operation and Development (OECD) report list, which contains 4730 PFASs used in a broad range of industries and consumers. Our work provides the first semi-supervised machine learning study of structure-activity relationships for predicting possible bioactivities in a variety of PFAS species.

Quantitative Cross-Species Comparison of Serum Albumin Binding of Per- and Polyfluoroalkyl Substances from Five Structural Classes

Per- and polyfluoroalkyl substances (PFAS) are a class of over 8,000 chemicals that are persistent, bioaccumulative, and toxic to humans, livestock, and wildlife. Serum protein binding affinity is instrumental in understanding PFAS toxicity, yet experimental binding data is limited to only a few PFAS congeners. Previously, we demonstrated the usefulness of a high-throughput, in vitro differential scanning fluorimetry assay for determination of relative binding affinities of human serum albumin for 24 PFAS congeners from 6 chemical classes. In the current study, we used this differential scanning fluorimetry assay to comparatively examine differences in human, bovine, porcine, and rat serum albumin binding of 8 structurally informative PFAS congeners from 5 chemical classes. With the exception of the fluorotelomer alcohol 1H,1H,2H,2H-perfluorooctanol (6:2 FTOH), each PFAS congener bound by human serum albumin was also bound by bovine, porcine, and rat serum albumin. The critical role of the charged functional headgroup in albumin binding was supported by the inability of serum albumin of each species tested to bind 6:2 FTOH. Significant interspecies differences in serum albumin binding affinities were identified for each of the bound PFAS congeners. Relative to human albumin, perfluoroalkyl carboxylic and sulfonic acids were bound with greater affinity by porcine and rat serum albumin, and perfluoroalkyl ether congeners bound with lower affinity to porcine and bovine serum albumin. These comparative affinity data for PFAS binding by serum albumin from human, experimental model and livestock species reduce critical interspecies uncertainty and improve accuracy of predictive toxicity assessments for PFAS.

Research Priorities for the Environmental Risk Assessment of Per- and Polyfluorinated Substances

Per- and polyfluorinated substances (PFAS) are a group of thousands of ubiquitously applied persistent industrial chemicals. The field of PFAS environmental research is developing rapidly, but suffers from substantial biases toward specific compounds, environmental compartments, and organisms. The aim of our study was therefore to highlight current developments and to identify knowledge gaps and subsequent research needs that would contribute to a comprehensive environmental risk assessment for PFAS. To this end, we consulted the open literature and databases and found that knowledge of the environmental fate of PFAS is based on the analysis of <1% of the compounds categorized as PFAS. Moreover, soils and suspended particulate matter remain largely understudied. The bioavailability, bioaccumulation, and food web transfer studies of PFAS also focus on a very limited number of compounds and are biased toward aquatic biota, predominantly fish, and less frequently aquatic invertebrates and macrophytes. The available ecotoxicity data revealed that only a few PFAS have been well studied for their environmental hazards, and that PFAS ecotoxicity data are also strongly biased toward aquatic organisms. Ecotoxicity studies in the terrestrial environment are needed, as well as chronic, multigenerational, and community ecotoxicity research, in light of the persistency and bioaccumulation of PFAS. Finally, we identified an urgent need to unravel the relationships among sorption, bioaccumulation, and ecotoxicity on the one hand and molecular descriptors of PFAS chemical structures and physicochemical properties on the other, to allow predictions of exposure, bioaccumulation, and toxicity. Environ Toxicol Chem 2023;42:2302-2316. © 2023 The Authors. Environmental Toxicology and Chemistry published by Wiley Periodicals LLC on behalf of SETAC.

A Review on Per- and Polyfluoroalkyl Substances in Pregnant Women: Maternal Exposure, Placental Transfer, and Relevant Model Simulation

Per- and polyfluoroalkyl substances (PFASs) are important and ubiquitous environmental contaminants worldwide. These novel contaminants can enter human bodies via various pathways, subsequently posing risks to the ecosystem and human health. The exposure of pregnant women to PFASs might pose risks to the health of mothers and the growth and development of fetuses. However, little information is available about the placental transfer of PFASs from mothers to fetuses and the related mechanisms through model simulation. In the present study, based upon a review of previously published literature, we initially summarized the exposure pathways of PFASs in pregnant women, factors affecting the efficiency of placental transfer, and mechanisms associated with placental transfer; outlined simulation analysis approaches using molecular docking and machine learning to reveal the mechanisms of placental transfer; and finally highlighted future research emphases that need to be focused on. Consequently, it was notable that the binding of PFASs to proteins during placental transfer could be simulated by molecular docking and that the placental transfer efficiency of PFASs could also be predicted by machine learning. Therefore, future research on the maternal-fetal transfer mechanisms of PFASs with the benefit of simulation analysis approaches is warranted to provide a scientific basis for the health effects of PFASs on newborns.

Insight into the binding model of per- and polyfluoroalkyl substances to proteins and membranes

Legacy per- and polyfluoroalkyl substances (PFASs) have elicited much concern because of their ubiquitous distribution in the environment and the potential hazards they pose to wildlife and human health. Although an increasing number of effective PFAS alternatives are available in the market, these alternatives bring new challenges. This paper comprehensively reviews how PFASs bind to transport proteins (e.g., serum albumin, liver fatty acid transport proteins and organic acid transporters), nuclear receptors (e.g., peroxisome proliferator activated receptors, thyroid hormone receptors and reproductive hormone receptors) and membranes (e.g., cell membrane and mitochondrial membrane). Briefly, the hydrophobic fluorinated carbon chains of PFASs occupy the binding cavities of the target proteins, and the acid groups of PFASs form hydrogen bonds with amino acid residues. Various structural features of PFAS alternatives such as chlorine atom substitution, oxygen atom insertion and a branched structure, introduce variations in their chain length and hydrophobicity, which potentially change the affinity of PFAS alternatives for endogenous proteins. The toxic effects and mechanisms of action of legacy PFASs can be demonstrated and compared with their alternatives using binding models. In future studies, in vitro experiments and in silico quantitative structure-activity relationship modeling should be better integrated to allow more reliable toxicity predictions for both legacy and alternative PFASs.

Warming Affects Bioconcentration and Bioaccumulation of Per- and Polyfluoroalkyl Substances by Pelagic and Benthic Organisms in a Water-Sediment System

Warming and exposure to emerging global pollutants, such as per- and polyfluoroalkyl substances (PFAS), are significant stressors in the aquatic ecosystem. However, little is known about the warming effect on the bioaccumulation of PFAS in aquatic organisms. In this study, the pelagic organisms Daphnia magna and zebrafish, and the benthic organism Chironomus plumosus were exposed to 13 PFAS in a sediment-water system with a known amount of each PFAS at different temperatures (16, 20, and 24 °C). The results showed that the steady-state body burden (C_b-ss) of PFAS in pelagic organisms increased with increasing temperatures, mainly attributed to increased water concentrations. The uptake rate constant (k_u) and elimination rate constant (k_e) in pelagic organisms increased with increasing temperature. In contrast, warming did not significantly change or even mitigate C_b-ss of PFAS in the benthic organism Chironomus plumosus, except for PFPeA and PFHpA, which was consistent with declined sediment concentrations. The mitigation could be explained by the decreased bioaccumulation factor due to a more significant percent increase in k_e than k_u, especially for long-chain PFAS. This study suggests that the warming effect on the PFAS concentration varies among different media, which should be considered for their ecological risk assessment under climate change.

How the Structure of Per- and Polyfluoroalkyl Substances (PFAS) Influences Their Binding Potency to the Peroxisome Proliferator-Activated and Thyroid Hormone Receptors-An In Silico Screening Study

In this study, we investigated PFAS (per- and polyfluoroalkyl substances) binding potencies to nuclear hormone receptors (NHRs): peroxisome proliferator-activated receptors (PPARs) α, β, and γ and thyroid hormone receptors (TRs) α and β. We have simulated the docking scores of 43 perfluoroalkyl compounds and based on these data developed QSAR (Quantitative Structure-Activity Relationship) models for predicting the binding probability to five receptors. In the next step, we implemented the developed QSAR models for the screening approach of a large group of compounds (4464) from the NORMAN Database. The in silico analyses indicated that the probability of PFAS binding to the receptors depends on the chain length, the number of fluorine atoms, and the number of branches in the molecule. According to the findings, the considered PFAS group bind to the PPARα, β, and γ only with low or moderate probability, while in the case of TR α and β it is similar except that those chemicals with longer chains show a moderately high probability of binding.

Machine Learning-Based Models with High Accuracy and Broad Applicability Domains for Screening PMT/vPvM Substances

Persistent, mobile, and toxic (PMT) substances and very persistent and very mobile (vPvM) substances can transport over long distances from various sources, increasing the public health risk. A rapid and high-throughput screening of PMT/vPvM substances is thus warranted to the risk prevention and mitigation measures. Herein, we construct a machine learning-based screening system integrated with five models for high-throughput classification of PMT/vPvM substances. The models are constructed with 44 971 substances by conventional learning, deep learning, and ensemble learning algorithms, among which, LightGBM and XGBoost outperform other algorithms with metrics exceeding 0.900. Good model interpretability is achieved through the number of free halogen atoms (fr_halogen) and the logarithm of partition coefficient (MolLogP) as the two most critical molecular descriptors representing the persistence and mobility of substances, respectively. Our screening system exhibits a great generalization capability with area under the receiver operating characteristic curve (AUROC) above 0.951 and is successfully applied to the persistent organic pollutants (POPs), prioritized PMT/vPvM substances, and pesticides. The screening system constructed in this study can serve as an efficient and reliable tool for high-throughput risk assessment and the prioritization of managing emerging contaminants.

XGBoost model as an efficient machine learning approach for PFAS removal: Effects of material characteristics and operation conditions

Due to the implications of poly- and perfluoroalkyl substances (PFAS) on the environment and public health, great attention has been recently made to finding innovative materials and methods for PFAS removal. In this work, PFAS is considered universal contamination which can be found in many wastewater streams. Conventional materials and processes used to remove and degrade PFAS do not have enough competence to address the issue particularly when it comes to eliminating short-chain PFAS. This is mainly due to the large number of complex parameters that are involved in both material and process designs. Here, we took the advantage of artificial intelligence to introduce a model (XGBoost) in which material and process factors are considered simultaneously. This research applies a machine learning approach using data collected from reported articles to predict the PFAS removal factors. The XGBoost modeling provided accurate adsorption capacity, equilibrium, and removal estimates with the ability to predict the adsorption mechanisms. The performance comparison of adsorbents and the role of AI in one dominant are studied and reviewed for the first time, even though many studies have been carried out to develop PFAS removal through various adsorption methods such as ion exchange, nanofiltration, and activated carbon (AC). The model showed that pH is the most effective parameter to predict PFAS removal. The proposed model in this work can be extended for other micropollutants and can be used as a basic framework for future adsorbent design and process optimization.

Comprehensive Exposure Studies of Per- and Polyfluoroalkyl Substances in the General Population: Target, Nontarget Screening, and Toxicity Prediction

Novel per- and polyfluoroalkyl substances (PFASs) in the environment and populations have received extensive attention; however, their distribution and potential toxic effects in the general population remain unclear. Here, a comprehensive study on PFAS screening was carried out in serum samples of 202 individuals from the general population in four cities in China. A total of 165 suspected PFASs were identified using target and nontarget analysis, including seven identified PFAS homolog series, of which 16 PFASs were validated against standards, and seven PFASs [4:2 chlorinated polyfluorinated ether sulfonate (4:2 Cl-PFESA), 7:2 chlorinated polyfluorinated ether sulfonate (7:2 Cl-PFESA), hydrosubstituted perfluoroheptanoate (H-PFHpA), chlorine-substituted perfluorooctanoate (Cl-PFOA), chlorine-substituted perfluorononanate (Cl-PFNA), chlorine-substituted perfluorodecanoate (Cl-PFDA), and perfluorodecanedioic acid (PFLDCA n = 8)] were reported for the first time in human serum. The Tox21-GCN model (a graph convolutional neural network model based on the Tox21 database) was established to predict the toxicity of the discovered PFASs, revealing that PFASs containing sulfonic acid groups exhibited multiple potential toxic effects, such as estrogenic effects and stress responses. Our study indicated that the general population was exposed to various PFASs, and the toxicity prediction results of individual PFASs suggested potential health risks that could not be ignored.

Molecular Screening and Toxicity Estimation of 260,000 Perfluoroalkyl and Polyfluoroalkyl Substances (PFASs) through Machine Learning

Perfluoroalkyl and polyfluoroalkyl substances (PFASs) are a class of chemicals widely used in industrial applications due to their exceptional properties and stability. However, they do not readily degrade in the environment and are linked to contamination and adverse health effects in humans and wildlife. To find alternatives for the most commonly used PFAS molecules that maintain their desirable chemical properties but are not adverse to biological lifeforms, a novel approach based upon machine learning is utilized. The machine learning model is trained on an existing set of PFAS molecules to generate over 260,000 novel PFAS molecules, which we dub PFAS-AI-Gen. Using molecular descriptors with known relationships to toxicity and industrial suitability followed by molecular docking and molecular dynamics simulations, this set of molecules is screened. In this manner, increasingly complex calculations are performed only for candidate molecules that are most likely to yield the desired properties of low binding affinity toward two selected protein receptors, the human pregnane x receptor (hPXR) and peroxisome proliferator-activated receptor γ (PPAR-γ), and high industrial suitability, defined by critical micelle concentration (CMC). The selection criteria of low binding affinity and high industrial suitability are relative to the popular PFAS alternative GenX. hPXR and PPAR-γ are selected as they are PFAS targets and facilitate a variety of functions, such as drug metabolism and glucose regulation, respectively. Through this approach, 22 promising new PFAS substitutes that may warrant experimental investigation are identified. This integrated approach of molecular screening and toxicity estimation may be applicable to other chemical classes.

A new prediction method of industrial atmospheric pollutant emission intensity based on pollutant emission standard quantification

Industrial emissions are the main source of atmospheric pollutants in China. Accurate and reasonable prediction of the emission of atmospheric pollutants from single enterprise can determine the exact source of atmospheric pollutants and control atmospheric pollution precisely. Based on China's coking enterprises in 2020, we proposed a quantitative method for pollutant emission standards and introduced the quantification results of pollutant emission standards (QRPES) into the construction of support vector regression (SVR) and random forest regression (RFR) prediction methods for SO₂ emission of coking enterprises in China. The results show that, affected by the types of coke ovens and regions, China's current coking enterprises have implemented a total of 21 emission standards, with marked differences. After adding QRPES, it was found that the root mean squared error (RMSE) of SVR and RFR decreased from 0.055 kt/a and 0.059 kt/a to 0.045 kt/a and 0.039 kt/a, and the R ² increased from 0.890 and 0.881 to 0.926 and 0.945, respectively. This shows that the QRPES can greatly improve the prediction accuracy, and the SO₂ emissions of each enterprise are highly correlated with the strictness of standards. The predicted result shows that 45% of SO₂ emissions from Chinese coking enterprises are concentrated in Shanxi, Shaanxi and Hebei provinces in central China. The method created in this paper fills in the blank of forecasting method of air pollutant emission intensity of single enterprise and is of great help to the accurate control of air pollutants.

ELECTRONIC SUPPLEMENTARY MATERIAL

Supplementary material is available in the online version of this article at 10.1007/s11783-023-1608-1 and is accessible for authorized users.

PFAS risk propagation terminology in spatial and temporal scales: Risk intensification, risk attenuation, and risk amplification

Poly- and per fluorinated alkyl substances (PFAS) are man-made chemicals that are used in many industrial applications to improve performance and durability of products. The CF bond is one of the strongest bonds in organic chemistry which makes PFAS highly persistent in the environment. Therefore, PFAS levels have increased in different environmental compartments (air, water, soil) at global scale over time since the early 1950s. Terminology used for describing potential risks and those used in risk communication can be confusing as different disciplines use risk concepts and vocabulary that are different from those used for exposure and/or health risk assessment for hazardous materials. For example, terms such as emergent risk, emerging risk, risk intensification, risk awareness, risk perception, risk attenuation, risk amplification, and risk absorption are often misused or misinterpreted, especially in describing and communicating risks associated with PFAS in the environment or PFAS exposure. In addition, appropriate risk terms associated with psychological, social, institutional, and cultural elements are often misused. Here, appropriate risk terminology for describing and quantifying risks and health risks in reference to PFAS exposure and PFAS related risk propagation in spatial and temporal scales are explained with examples.

Machine learning and artificial intelligence in toxicological sciences

Machine learning and artificial intelligence approaches have revolutionized multiple disciplines, including toxicology. This review summarizes representative recent applications of machine learning and artificial intelligence approaches in different areas of toxicology, including physiologically based pharmacokinetic (PBPK) modeling, quantitative structure-activity relationship modeling for toxicity prediction, adverse outcome pathway analysis, high-throughput screening, toxicogenomics, big data and toxicological databases. By leveraging machine learning and artificial intelligence approaches, now it is possible to develop PBPK models for hundreds of chemicals efficiently, to create in silico models to predict toxicity for a large number of chemicals with similar accuracies compared to in vivo animal experiments, and to analyze a large amount of different types of data (toxicogenomics, high-content image data, etc.) to generate new insights into toxicity mechanisms rapidly, which was impossible by manual approaches in the past. To continue advancing the field of toxicological sciences, several challenges should be considered: (1) not all machine learning models are equally useful for a particular type of toxicology data, and thus it is important to test different methods to determine the optimal approach; (2) current toxicity prediction is mainly on bioactivity classification (yes/no), so additional studies are needed to predict the intensity of effect or dose-response relationship; (3) as more data become available, it is crucial to perform rigorous data quality check and develop infrastructure to store, share, analyze, evaluate, and manage big data; and (4) it is important to convert machine learning models to user-friendly interfaces to facilitate their applications by both computational and bench scientists.

A high-throughput assay for screening the abilities of per- and polyfluoroalkyl substances in inducing plasma kallikrein-like activity

The plasma consists of multiple functional serine zymogens, such as plasma kallikrein-kinin system (KKS), which are vulnerable to exogenous chemical exposure, and may closely relate to the deleterious effects. Testing whether the anthropogenic chemicals could increase the kallikrein-like activity in plasma or not would be of great help to understand their potentials in triggering the cascade activation of the plasma zymogens and explain the corresponding hematotoxicity. In this study, a novel high-throughput ex vivo assay was established to screen the abilities of emerging chemicals like per- and polyfluoroalkyl substances (PFASs) in inducing kallikrein-like activities on basis of using rat plasma as the protease zymogen source. Upon the optimization of the conditions in the test system, the assay gave sensitive fluorescent response to the stimulation of the positive control, dextran sulfate, and the dose-response showed a typical S-shaped curve with EC50 of 0.24 mg/L. The intra-plate and inter-plate relative standard deviations (RSDs) were less than 10% in the quantitative range of dextran sulfate, indicating a good reliability and repeatability of this newly-established assay. Using this method, several alternatives or congeners of perfluorooctanesulfonic acid (PFOS) and perfluorooctanoic acid (PFOA), including 6:2 chlorinated polyfluoroalkyl ether sulfonate (6:2 Cl-PFESA), Ag-PFOA, K-PFOA, Na-PFOA and ammonium pentadecafluorooctanoate (APFO), were further screened, and their capabilities in inducing kallikrein-like activities were identified. The ex vivo assay newly-developed in the present study would be promising in high-throughput screening of the hematological effects of emerging chemicals of concern.

Legacy and Emerging Per- and Polyfluoroalkyl Substances Behave Distinctly in Spatial Distribution and Multimedia Partitioning: A Case Study in the Pearl River, China

Per- and polyfluoroalkyl substances (PFASs) have attracted worldwide attention due to their ubiquitous occurrence, bioaccumulation, and toxicological effects, yet the fate of PFASs in a lotic ecosystem is largely unknown. To elucidate spatial distribution and multimedia partitioning of legacy and emerging PFASs in a lotic river flowing into an estuary, PFASs were synchronously analyzed in water, suspended particulate matter (SPM), sediment, and biota samples collected along Guangzhou reach of the Pearl River, South China. Geographically, the concentrations of PFASs in the water phase showed a decreasing trend from the upper and middle sections (urban area) to the down section (suburban area close to estuary) of the river. While perfluorooctanoic acid predominated in water and SPM, more diverse compositions were observed in sediment and biota with the increase in contributions of long-chain PFASs. Field-derived sediment-water partitioning coefficients (K_d) and bioaccumulation factors (BAFs) of PFASs increased with the increase in perfluorinated carbons. Besides hydrophobicity, water pH and salinity significantly affected the multimedia partitioning of PFASs in a lotic ecosystem. In addition, 87 homologues (63 classes) were identified as emerging PFASs in four media using suspect analysis. Interestingly, K_d and BAF of the emerging PFASs were often higher than legacy PFASs containing the same perfluorinated carbons, raising a special concern on the environmental risk of emerging PFASs.

Effect of Enterohepatic Circulation on the Accumulation of Per- and Polyfluoroalkyl Substances: Evidence from Experimental and Computational Studies

The pharmacokinetic characteristics of per- and polyfluoroalkyl substances (PFAS) affect their distribution and bioaccumulation in biological systems. The enterohepatic circulation leads to reabsorption of certain chemicals from bile back into blood and the liver and thus influences their elimination, yet its influence on PFAS bioaccumulation remains unclear. We explored the role of enterohepatic circulation in PFAS bioaccumulation by examining tissue distribution of various PFAS in wild fish and a rat model. Computational models were used to determine the reabsorbed fractions of PFAS by calculating binding affinities of PFAS for key transporter proteins of enterohepatic circulation. The results indicated that higher concentrations were observed in blood, the liver, and bile compared to other tissues for some PFAS in fish. Furthermore, exposure to a PFAS mixture on the rat model showed that the reabsorption phenomenon appeared during 8-12 h for most long-chain PFAS. Molecular docking calculations suggest that PFAS can bind to key transporter proteins via electrostatic and hydrophobic interactions. Further regression analysis adds support to the hypothesis that binding affinity of the apical sodium-dependent bile acid transporter is the most important variable to predict the human half-lives of PFAS. This study demonstrated the critical role of enterohepatic circulation in reabsorption, distribution, and accumulation of PFAS.

Distribution, source identification and health risk assessment of PFASs in groundwater from Jiangxi Province, China

There is an urgent need to investigate on the distribution and fate of short-chain analogues and emerging per- and polyfluoroalkyl substances (PFASs) in groundwater, and little research on their source apportionment and health risks through the drinking water exposure pathway has been carried out. In present study, the concentration and source of 22 PFASs, including five alternatives: 6:2 fluorotelomer sulfonate (6:2 FTS), potassium 9-chlorohexadecafluoro-3-oxanonane-1-sulfonate (F-53B), hexafluoropropylene oxide trimer acid (HFPO-TA), hexafluoropropylene oxide dimer acid (HFPO-DA) and ammonium 4, 8-dioxa-3H-perfluorononanoate (ADONA), were analyzed in 88 groundwater samples from wells in Jiangxi Province, southeastern China. The total PFASs concentration (Σ₁₈PFASs) in groundwater varied from 1.27 to 381.00 ng/L (mean 47.60 ng/L). Short-chain perfluorobutanoic acid (PFBA) and perfluoropentanoic acid (PFPeA) were the most abundant perfluorinated carboxylic acids (PFCAs), and short-chain perfluorobutanesulfonate (PFBS) was the most abundant perfluorinated sulfonic acids (PFSAs) in groundwater samples. The quantitative source apportionment by nonnegative matrix/tensor factorization coupled with k-means clustering (NMFk) model suggested that short-chain homologues and emerging alternatives have been used as substitutes for legacy PFOS and PFOA. Furthermore, the human risk assessment results showed that the estimated daily intakes (EDIs) for short-chain PFCAs were higher than that of PFOA, whereas the EDIs of PFBS, 6:2 FTS and F-53B were comparable to that of PFOS.

Evaluating the capability of municipal solid waste separation in China based on AHP-EWM and BP neural network

With the increase in municipal solid waste (MSW), most cities face solid waste management issues. In this study, the analytic hierarchy process (AHP) and artificial neural network (ANN) models were improved to assess the MSW separation capability based on 18 selected indicators of solid waste separation in 15 cities in China. The entropy weight method (EWM) was used in AHP to optimize and determine the indicators and then evaluate their weights, which showed that the general public budget expenditure had the highest weight (0.5239). This implied that the MSW separation capability could be mainly influenced by government financial support. ANN based on scan optimization and machine learning methods were established (R² = 0.9992) to predict the missing indicators. The mapping relationship between MSW separation indicators and capabilities was also significantly improved from R² = 0.5317 to R² = 0.9993, thereby increasing the prediction accuracy of MSW separation capabilities to 95.15%. Thus, this research provides a new avenue for MSW separation and establishes a combined model to predict the separation capability in practical applications.

Uncertainty-Informed Deep Transfer Learning of Perfluoroalkyl and Polyfluoroalkyl Substance Toxicity

Perfluoroalkyl and polyfluoroalkyl substances (PFAS) pose a significant hazard because of their widespread industrial uses, environmental persistence, and bioaccumulation. A growing, increasingly diverse inventory of PFAS, including 8163 chemicals, has recently been updated by the U.S. Environmental Protection Agency. However, with the exception of a handful of well-studied examples, little is known about their human toxicity potential because of the substantial resources required for in vivo toxicity experiments. We tackle the problem of expensive in vivo experiments by evaluating multiple machine learning (ML) methods, including random forests, deep neural networks (DNN), graph convolutional networks, and Gaussian processes, for predicting acute toxicity (e.g., median lethal dose, or LD₅₀) of PFAS compounds. To address the scarcity of toxicity information for PFAS, publicly available datasets of oral rat LD₅₀ for all organic compounds are aggregated and used to develop state-of-the-art ML source models for transfer learning. A total of 519 fluorinated compounds containing two or more C-F bonds with known toxicity are used for knowledge transfer to ensembles of the best-performing source model, DNN, to generate the target models for the PFAS domain with access to uncertainty. This study predicts toxicity for PFAS with a defined chemical structure. To further inform prediction confidence, the transfer-learned model is embedded within a SelectiveNet architecture, where the model is allowed to identify regions of prediction with greater confidence and abstain from those with high uncertainty using a calibrated cutoff rate.

Environmental Sources, Chemistry, Fate, and Transport of Per- and Polyfluoroalkyl Substances: State of the Science, Key Knowledge Gaps, and Recommendations Presented at the August 2019 SETAC Focus Topic Meeting

A Society of Environmental Toxicology and Chemistry (SETAC) Focused Topic Meeting (FTM) on the environmental management of per- and polyfluoroalkyl substances (PFAS) convened during August 2019 in Durham, North Carolina (USA). Experts from around the globe were brought together to critically evaluate new and emerging information on PFAS including chemistry, fate, transport, exposure, and toxicity. After plenary presentations, breakout groups were established and tasked to identify and adjudicate via panel discussions overarching conclusions and relevant data gaps. The present review is one in a series and summarizes outcomes of presentations and breakout discussions related to (1) primary sources and pathways in the environment, (2) sorption and transport in porous media, (3) precursor transformation, (4) practical approaches to the assessment of source zones, (5) standard and novel analytical methods with implications for environmental forensics and site management, and (6) classification and grouping from multiple perspectives. Outcomes illustrate that PFAS classification will continue to be a challenge, and additional pressing needs include increased availability of analytical standards and methods for assessment of PFAS and fate and transport, including precursor transformation. Although the state of the science is sufficient to support a degree of site-specific and flexible risk management, effective source prioritization tools, predictive fate and transport models, and improved and standardized analytical methods are needed to guide broader policies and best management practices. Environ Toxicol Chem 2021;40:3234-3260. © 2021 The Authors. Environmental Toxicology and Chemistry published by Wiley Periodicals LLC on behalf of SETAC.

Thermodynamics-Based Model Construction for the Accurate Prediction of Molecular Properties From Partition Coefficients

Developing models for predicting molecular properties of organic compounds is imperative for drug development and environmental safety; however, development of such models that have high predictive power and are independent of the compounds used is challenging. To overcome the challenges, we used a thermodynamics-based theoretical derivation to construct models for accurately predicting molecular properties. The free energy change that determines a property equals the sum of the free energy changes (ΔG_Fs) caused by the factors affecting the property. By developing or selecting molecular descriptors that are directly proportional to ΔG_Fs, we built a general linear free energy relationship (LFER) for predicting the property with the molecular descriptors as predictive variables. The LFER can be used to construct models for predicting various specific properties from partition coefficients. Validations show that the models constructed according to the LFER have high predictive power and their performance is independent of the compounds used, including the models for the properties having little correlation with partition coefficients. The findings in this study are highly useful for applications in drug development and environmental safety.

Predicting Micropollutant Removal by Reverse Osmosis and Nanofiltration Membranes: Is Machine Learning Viable?

Predictive models for micropollutant removal by membrane separation are highly desirable for the design and selection of appropriate membranes. While machine learning (ML) models have been applied for such purposes, their reliability might be compromised by data leakage due to inappropriate data splitting. More importantly, whether ML models can truly understand the mechanisms of membrane separation has not been revealed. In this study, we evaluate the capability of the XGBoost model to predict micropollutant removal efficiencies of reverse osmosis and nanofiltration membranes. Our results demonstrate that data leakage leads to falsely high prediction accuracy. By utilizing a model interpretation method based on the cooperative game theory, we test the knowledge of XGBoost on the mechanisms of membrane separation via quantifying the contributions of input variables to the model predictions. We reveal that XGBoost possesses an adequate understanding of size exclusion, but its knowledge of electrostatic interactions and adsorption is limited. Our findings suggest that future work should focus more on avoiding data leakage and evaluating the mechanistic knowledge of ML models. In addition, high-quality data from more diverse experimental conditions, as well as more informative variables, are needed to improve the accuracy of ML models for predicting membrane performance.

Predicting the risk of GenX contamination in private well water using a machine-learned Bayesian network model

Per- and polyfluoroalkyl substances (PFAS) are emerging contaminants that pose significant challenges in mechanistic fate and transport modeling due to their diverse and complex chemical characteristics. Machine learning provides a novel approach for predicting the spatial distribution of PFAS in the environment. We used spatial location information to link PFAS measurements from 1207 private drinking water wells around a fluorochemical manufacturing facility to a mechanistic model of PFAS air deposition and to publicly available data on soil, land use, topography, weather, and proximity to multiple PFAS sources. We used the resulting linked data set to train a Bayesian network model to predict the risk that GenX, a member of the PFAS class, would exceed a state provisional health goal (140 ng/L) in private well water. The model had high accuracy (ROC curve index for five-fold cross-validation of 0.85, 90% CI 0.84-0.87). Among factors significantly associated with GenX risk in private wells, the most important was the historic rate of atmospheric deposition of GenX from the fluorochemical manufacturing facility. The model output was used to generate spatial risk predictions for the study area to aid in risk assessment, environmental investigations, and targeted public health interventions.

Machine Learning in Discovery of New Antivirals and Optimization of Viral Infections Therapy

Nowadays, computational approaches play an important role in the design of new drug-like compounds and optimization of pharmacotherapeutic treatment of diseases. The emerging growth of viral infections, including those caused by the Human Immunodeficiency Virus (HIV), Ebola virus, recently detected coronavirus, and some others, leads to many newly infected people with a high risk of death or severe complications. A huge amount of chemical, biological, clinical data is at the disposal of the researchers. Therefore, there are many opportunities to find the relationships between the particular features of chemical data and the antiviral activity of biologically active compounds based on machine learning approaches. Biological and clinical data can also be used for building models to predict relationships between viral genotype and drug resistance, which might help determine the clinical outcome of treatment. In the current study, we consider machine-learning approaches in the antiviral research carried out during the past decade. We overview in detail the application of machine-learning methods for the design of new potential antiviral agents and vaccines, drug resistance prediction, and analysis of virus-host interactions. Our review also covers the perspectives of using the machine-learning approaches for antiviral research, including Dengue, Ebola viruses, Influenza A, Human Immunodeficiency Virus, coronaviruses, and some others.

Assessing the Ecological Risks of Per- and Polyfluoroalkyl Substances: Current State-of-the Science and a Proposed Path Forward

Per- and poly-fluoroalkyl substances (PFAS) encompass a large, heterogenous group of chemicals of potential concern to human health and the environment. Based on information for a few relatively well-understood PFAS such as perfluorooctane sulfonate and perfluorooctanoate, there is ample basis to suspect that at least a subset can be considered persistent, bioaccumulative, and/or toxic. However, data suitable for determining risks in either prospective or retrospective assessments are lacking for the majority of PFAS. In August 2019, the Society of Environmental Toxicology and Chemistry sponsored a workshop that focused on the state-of-the-science supporting risk assessment of PFAS. The present review summarizes discussions concerning the ecotoxicology and ecological risks of PFAS. First, we summarize currently available information relevant to problem formulation/prioritization, exposure, and hazard/effects of PFAS in the context of regulatory and ecological risk assessment activities from around the world. We then describe critical gaps and uncertainties relative to ecological risk assessments for PFAS and propose approaches to address these needs. Recommendations include the development of more comprehensive monitoring programs to support exposure assessment, an emphasis on research to support the formulation of predictive models for bioaccumulation, and the development of in silico, in vitro, and in vivo methods to efficiently assess biological effects for potentially sensitive species/endpoints. Addressing needs associated with assessing the ecological risk of PFAS will require cross-disciplinary approaches that employ both conventional and new methods in an integrated, resource-effective manner. Environ Toxicol Chem 2021;40:564-605. © 2020 The Authors. Environmental Toxicology and Chemistry published by Wiley Periodicals LLC on behalf of SETAC. This article has been contributed to by US Government employees and their work is in the public domain in the USA.

PFAS Exposure Pathways for Humans and Wildlife: A Synthesis of Current Knowledge and Key Gaps in Understanding

We synthesize current understanding of the magnitudes and methods for assessing human and wildlife exposures to poly- and perfluoroalkyl substances (PFAS). Most human exposure assessments have focused on 2 to 5 legacy PFAS, and wildlife assessments are typically limited to targeted PFAS (up to ~30 substances). However, shifts in chemical production are occurring rapidly, and targeted methods for detecting PFAS have not kept pace with these changes. Total fluorine measurements complemented by suspect screening using high-resolution mass spectrometry are thus emerging as essential tools for PFAS exposure assessment. Such methods enable researchers to better understand contributions from precursor compounds that degrade into terminal perfluoroalkyl acids. Available data suggest that diet is the major human exposure pathway for some PFAS, but there is large variability across populations and PFAS compounds. Additional data on total fluorine in exposure media and the fraction of unidentified organofluorine are needed. Drinking water has been established as the major exposure source in contaminated communities. As water supplies are remediated, for the general population, exposures from dust, personal care products, indoor environments, and other sources may be more important. A major challenge for exposure assessments is the lack of statistically representative population surveys. For wildlife, bioaccumulation processes differ substantially between PFAS and neutral lipophilic organic compounds, prompting a reevaluation of traditional bioaccumulation metrics. There is evidence that both phospholipids and proteins are important for the tissue partitioning and accumulation of PFAS. New mechanistic models for PFAS bioaccumulation are being developed that will assist in wildlife risk evaluations. Environ Toxicol Chem 2021;40:631-657. © 2020 SETAC.

Per- and Polyfluoroalkyl Substance Toxicity and Human Health Review: Current State of Knowledge and Strategies for Informing Future Research

Reports of environmental and human health impacts of per- and polyfluoroalkyl substances (PFAS) have greatly increased in the peer-reviewed literature. The goals of the present review are to assess the state of the science regarding toxicological effects of PFAS and to develop strategies for advancing knowledge on the health effects of this large family of chemicals. Currently, much of the toxicity data available for PFAS are for a handful of chemicals, primarily legacy PFAS such as perfluorooctanoic acid and perfluorooctane sulfonate. Epidemiological studies have revealed associations between exposure to specific PFAS and a variety of health effects, including altered immune and thyroid function, liver disease, lipid and insulin dysregulation, kidney disease, adverse reproductive and developmental outcomes, and cancer. Concordance with experimental animal data exists for many of these effects. However, information on modes of action and adverse outcome pathways must be expanded, and profound differences in PFAS toxicokinetic properties must be considered in understanding differences in responses between the sexes and among species and life stages. With many health effects noted for a relatively few example compounds and hundreds of other PFAS in commerce lacking toxicity data, more contemporary and high-throughput approaches such as read-across, molecular dynamics, and protein modeling are proposed to accelerate the development of toxicity information on emerging and legacy PFAS, individually and as mixtures. In addition, an appropriate degree of precaution, given what is already known from the PFAS examples noted, may be needed to protect human health. Environ Toxicol Chem 2021;40:606-630. © 2020 SETAC.

Differentiating Solutes with Precise Nanofiltration for Next Generation Environmental Separations: A Review

Selective removal or enrichment of targeted solutes including micropollutants, valuable elements, and mineral scalants from complex aqueous matrices is both challenging and pivotal to the success of water purification and resource recovery from unconventional water resources. Membrane separation with precision at the subnanometer or even subangstrom scale is of paramount importance to address those challenges via enabling "fit-for-purpose" water and wastewater treatment. So far, researchers have attempted to develop novel membrane materials with precise and tailored selectivity by tuning membrane structure and chemistry. In this critical review, we first present the environmental challenges and opportunities that necessitate improved solute-solute selectivity in membrane separation. We then discuss the mechanisms and desired membrane properties required for better membrane selectivity. On the basis of the most recent progress reported in the literature, we examine the key principles of material design and fabrication, which create membranes with enhanced and more targeted selectivity. We highlight the important roles of surface engineering, nanotechnology, and molecular-level design in improving membrane selectivity. Finally, we discuss the challenges and prospects of highly selective NF membranes for practical environmental applications, identifying knowledge gaps that will guide future research to promote environmental sustainability through more precise and tunable membrane separation.

Sorption of PFOS in 114 Well-Characterized Tropical and Temperate Soils: Application of Multivariate and Artificial Neural Network Analyses

The influence of soil properties on PFOS sorption are not fully understood, particularly for variable charge soils. PFOS batch sorption isotherms were conducted for 114 temperate and tropical soils from Australia and Fiji, that were well-characterized for their soil properties, including total organic carbon (TOC), anion exchange capacity, and surface charge. In most soils, PFOS sorption isotherms were nonlinear. PFOS sorption distribution coefficients (K_d) ranged from 5 to 229 mL/g (median: 28 mL/g), with 63% of the Fijian soils and 35% of the Australian soils showing K_d values that exceeded the observed median K_d. Multiple linear regression showed that TOC, amorphous aluminum and iron oxides contents, anion exchange capacity, pH, and silt content, jointly explained about 53% of the variance in PFOS K_d in soils. Variable charge soils with net positive surface charges, and moderate to elevated TOC content, generally displayed enhanced PFOS sorption than in temperate or tropical soils with TOC as the only sorbent phase, especially at acidic pH ranges. For the first time, two artificial neural networks were developed to predict the measured PFOS K_d (R² = 0.80) in the soils. Overall, both TOC and surface charge characteristics of soils are important for describing PFOS sorption.