MS2Tox Machine Learning Tool for Predicting the Ecotoxicity of Unidentified Chemicals in Water by Nontarget LC-HRMS

Progress, applications, and challenges in high-throughput effect-directed analysis for toxicity driver identification - is it time for HT-EDA?

The rapid increase in the production and global use of chemicals and their mixtures has raised concerns about their potential impact on human and environmental health. With advances in analytical techniques, in particular, high-resolution mass spectrometry (HRMS), thousands of compounds and transformation products with potential adverse effects can now be detected in environmental samples. However, identifying and prioritizing the toxicity drivers among these compounds remain a significant challenge. Effect-directed analysis (EDA) emerged as an important tool to address this challenge, combining biotesting, sample fractionation, and chemical analysis to unravel toxicity drivers in complex mixtures. Traditional EDA workflows are labor-intensive and time-consuming, hindering large-scale applications. The concept of high-throughput (HT) EDA has recently gained traction as a means of accelerating these workflows. Key features of HT-EDA include the combination of microfractionation and downscaled bioassays, automation of sample preparation and biotesting, and efficient data processing workflows supported by novel computational tools. In addition to microplate-based fractionation, high-performance thin-layer chromatography (HPTLC) offers an interesting alternative to HPLC in HT-EDA. This review provides an updated perspective on the state-of-the-art in HT-EDA, and novel methods/tools that can be incorporated into HT-EDA workflows. It also discusses recent studies on HT-EDA, HT bioassays, and computational prioritization tools, along with considerations regarding HPTLC. By identifying current gaps in HT-EDA and proposing new approaches to overcome them, this review aims to bring HT-EDA a step closer to monitoring applications.

Predicting Odor Sensory Attributes of Unidentified Chemicals in Water Using Fragmentation Mass Spectra with Machine Learning Models

Knowing odor sensory attributes of odorants lies at the core of odor tracking when addressing waterborne odor issues. However, experimental determination covering tens of thousands of odorants in authentic water is not pragmatic due to the complexity of odorant identification and odor evaluation. In this study, we propose the first machine learning (ML) model to predict odor perception/threshold aiming at odorants in water, which can use either molecular structure or MS2 spectra as input features. We demonstrate that model performance using MS2 spectra is nearly as good as that using unequivocal structures, both with outstanding accuracy. We particularly show the model's robustness in predicting odor sensory attributes of unidentified chemicals by using the experimentally obtained MS2 spectra from nontarget analysis on authentic water samples. Interpreting the developed models, we identify the intricate interaction of functional groups as the predominant influence factor on odor sensory attributes. We also highlight the important roles of carbon chain length, molecular weight, etc., in the inherent olfactory mechanisms. These findings streamline the odor sensory attribute prediction and are crucial advancements toward credible tracking and efficient control of off-odors in water.

Non-target analysis of Danish wastewater treatment plant effluent: Statistical analysis of chemical fingerprinting as a step toward a future monitoring tool

In an attempt to discover and characterize the plethora of xenobiotic substances, this study investigates chemical compounds released into the environment with wastewater effluents. A novel non-targeted screening methodology based on ultra-high resolution Orbitrap mass spectrometry and nanoflow ultra-high performance liquid chromatography together with a newly optimized data-processing pipeline were applied to effluent samples from two state-of-the-art and one small wastewater treatment facility. In total, 785 molecular structures were obtained, of which 38 were identified as single compounds, while 480 structures were identified at a putative level. Most of these substances were therapeutics and drugs, present as parent compounds and metabolites. Using R packages Phyloseq and MetacodeR, originally developed for bioinformatics, significant differences in xenobiotic presence in the wastewater effluents between the three sites were demonstrated.

Grouping strategies for assessing and managing persistent and mobile substances

Background

Persistent, mobile and toxic (PMT), or very persistent and very mobile (vPvM) substances are a wide class of chemicals that are recalcitrant to degradation, easily transported, and potentially harmful to humans and the environment. Due to their persistence and mobility, these substances are often widespread in the environment once emitted, particularly in water resources, causing increased challenges during water treatment processes. Some PMT/vPvM substances such as GenX and perfluorobutane sulfonic acid have been identified as substances of very high concern (SVHCs) under the European Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) regulation. With hundreds to thousands of potential PMT/vPvM substances yet to be assessed and managed, effective and efficient approaches that avoid a case-by-case assessment and prevent regrettable substitution are necessary to achieve the European Union's zero-pollution goal for a non-toxic environment by 2050.

Main

Substance grouping has helped global regulation of some highly hazardous chemicals, e.g., through the Montreal Protocol and the Stockholm Convention. This article explores the potential of grouping strategies for identifying, assessing and managing PMT/vPvM substances. The aim is to facilitate early identification of lesser-known or new substances that potentially meet PMT/vPvM criteria, prompt additional testing, avoid regrettable use or substitution, and integrate into existing risk management strategies. Thus, this article provides an overview of PMT/vPvM substances and reviews the definition of PMT/vPvM criteria and various lists of PMT/vPvM substances available. It covers the current definition of groups, compares the use of substance grouping for hazard assessment and regulation, and discusses the advantages and disadvantages of grouping substances for regulation. The article then explores strategies for grouping PMT/vPvM substances, including read-across, structural similarity and commonly retained moieties, as well as the potential application of these strategies using cheminformatics to predict P, M and T properties for selected examples.

Conclusions

Effective substance grouping can accelerate the assessment and management of PMT/vPvM substances, especially for substances that lack information. Advances to read-across methods and cheminformatics tools are needed to support efficient and effective chemical management, preventing broad entry of hazardous chemicals into the global market and favouring safer and more sustainable alternatives.

Deep Learning Bridged Bioactivity, Structure, and GC-HRMS-Readable Evidence to Decipher Nontarget Toxicants in Sediments

Identifying causative toxicants in mixtures is critical, but this task is challenging when mixtures contain multiple chemical classes. Effect-based methods are used to complement chemical analyses to identify toxicants, yet conventional bioassays typically rely on an apical and/or single endpoint, providing limited diagnostic potential to guide chemical prioritization. We proposed an event-driven taxonomy framework for mixture risk assessment that relied on high-throughput screening bioassays and toxicant identification integrated by deep learning. In this work, the framework was evaluated using chemical mixtures in sediments eliciting aryl-hydrocarbon receptor activation and oxidative stress response. Mixture prediction using target analysis explained <10% of observed sediment bioactivity. To identify additional contaminants, two deep learning models were developed to predict fingerprints of a pool of bioactive substances (event driver fingerprint, EDFP) and convert these candidates to MS-readable information (event driver ion, EDION) for nontarget analysis. Two libraries with 121 and 118 fingerprints were established, and 247 bioactive compounds were identified at confidence level 2 or 3 in sediment extract using GC-qToF-MS. Among them, 12 toxicants were analytically confirmed using reference standards. Collectively, we present a "bioactivity-signature-toxicant" strategy to deconvolute mixtures and to connect patchy data sets and guide nontarget analysis for diverse chemicals that elicit the same bioactivity.

Predicting the Activity of Unidentified Chemicals in Complementary Bioassays from the HRMS Data to Pinpoint Potential Endocrine Disruptors

The majority of chemicals detected via nontarget liquid chromatography high-resolution mass spectrometry (HRMS) in environmental samples remain unidentified, challenging the capability of existing machine learning models to pinpoint potential endocrine disruptors (EDs). Here, we predict the activity of unidentified chemicals across 12 bioassays related to EDs within the Tox21 10K dataset. Single- and multi-output models, utilizing various machine learning algorithms and molecular fingerprint features as an input, were trained for this purpose. To evaluate the models under near real-world conditions, Monte Carlo sampling was implemented for the first time. This technique enables the use of probabilistic fingerprint features derived from the experimental HRMS data with SIRIUS+CSI:FingerID as an input for models trained on true binary fingerprint features. Depending on the bioassay, the lowest false-positive rate at 90% recall ranged from 0.251 (sr.mmp, mitochondrial membrane potential) to 0.824 (nr.ar, androgen receptor), which is consistent with the trends observed in the models' performances submitted for the Tox21 Data Challenge. These findings underscore the informativeness of fingerprint features that can be compiled from HRMS in predicting the endocrine-disrupting activity. Moreover, an in-depth SHapley Additive exPlanations analysis unveiled the models' ability to pinpoint structural patterns linked to the modes of action of active chemicals. Despite the superior performance of the single-output models compared to that of the multi-output models, the latter's potential cannot be disregarded for similar tasks in the field of in silico toxicology. This study presents a significant advancement in identifying potentially toxic chemicals within complex mixtures without unambiguous identification and effectively reducing the workload for postprocessing by up to 75% in nontarget HRMS.

Innovative analytical methodologies for characterizing chemical exposure with a view to next-generation risk assessment

The chemical burden on the environment and human population is increasing. Consequently, regulatory risk assessment must keep pace to manage, reduce, and prevent adverse impacts on human and environmental health associated with hazardous chemicals. Surveillance of chemicals of known, emerging, or potential future concern, entering the environment-food-human continuum is needed to document the reality of risks posed by chemicals on ecosystem and human health from a one health perspective, feed into early warning systems and support public policies for exposure mitigation provisions and safe and sustainable by design strategies. The use of less-conventional sampling strategies and integration of full-scan, high-resolution mass spectrometry and effect-directed analysis in environmental and human monitoring programmes have the potential to enhance the screening and identification of a wider range of chemicals of known, emerging or potential future concern. Here, we outline the key needs and recommendations identified within the European Partnership for Assessment of Risks from Chemicals (PARC) project for leveraging these innovative methodologies to support the development of next-generation chemical risk assessment.

Online and Offline Prioritization of Chemicals of Interest in Suspect Screening and Non-targeted Screening with High-Resolution Mass Spectrometry

Recent advances in high-resolution mass spectrometry (HRMS) have enabled the detection of thousands of chemicals from a single sample, while computational methods have improved the identification and quantification of these chemicals in the absence of reference standards typically required in targeted analysis. However, to determine the presence of chemicals of interest that may pose an overall impact on ecological and human health, prioritization strategies must be used to effectively and efficiently highlight chemicals for further investigation. Prioritization can be based on a chemical's physicochemical properties, structure, exposure, and toxicity, in addition to its regulatory status. This Perspective aims to provide a framework for the strategies used for chemical prioritization that can be implemented to facilitate high-quality research and communication of results. These strategies are categorized as either "online" or "offline" prioritization techniques. Online prioritization techniques trigger the isolation and fragmentation of ions from the low-energy mass spectra in real time, with user-defined parameters. Offline prioritization techniques, in contrast, highlight chemicals of interest after the data has been acquired; detected features can be filtered and ranked based on the relative abundance or the predicted structure, toxicity, and concentration imputed from the tandem mass spectrum (MS2). Here we provide an overview of these prioritization techniques and how they have been successfully implemented and reported in the literature to find chemicals of elevated risk to human and ecological environments. A complete list of software and tools is available from https://nontargetedanalysis.org/.

Profiling emerging micropollutants in urban stormwater runoff using suspect and non-target screening via high-resolution mass spectrometry

Urban surface runoff contains chemicals that can negatively affect water quality. Urban runoff studies have determined the transport dynamics of many legacy pollutants. However, less attention has been paid to determining the first-flush effects (FFE) of emerging micropollutants using suspect and non-target screening (SNTS). Therefore, this study employed suspect and non-target analyses using liquid chromatography-high resolution mass spectrometry to detect emerging pollutants in urban receiving waters during stormwater events. Time-interval sampling was used to determine occurrence trends during stormwater events. Suspect screening tentatively identified 65 substances, then, their occurrence trend was grouped using correlation analysis. Non-target peaks were prioritized through hierarchical cluster analysis, focusing on the first flush-concentrated peaks. This approach revealed 38 substances using in silico identification. Simultaneously, substances identified through homologous series observation were evaluated for their observed trends in individual events using network analysis. The results of SNTS were normalized through internal standards to assess the FFE, and the most of tentatively identified substances showed observed FFE. Our findings suggested that diverse pollutants that could not be covered by target screening alone entered urban water through stormwater runoff during the first flush. This study showcases the applicability of the SNTS in evaluating the FFE of urban pollutants, offering insights for first-flush stormwater monitoring and management.

Machine Learning-Based Hazard-Driven Prioritization of Features in Nontarget Screening of Environmental High-Resolution Mass Spectrometry Data

Nontarget high-resolution mass spectrometry screening (NTS HRMS/MS) can detect thousands of organic substances in environmental samples. However, new strategies are needed to focus time-intensive identification efforts on features with the highest potential to cause adverse effects instead of the most abundant ones. To address this challenge, we developed MLinvitroTox, a machine learning framework that uses molecular fingerprints derived from fragmentation spectra (MS2) for a rapid classification of thousands of unidentified HRMS/MS features as toxic/nontoxic based on nearly 400 target-specific and over 100 cytotoxic endpoints from ToxCast/Tox21. Model development results demonstrated that using customized molecular fingerprints and models, over a quarter of toxic endpoints and the majority of the associated mechanistic targets could be accurately predicted with sensitivities exceeding 0.95. Notably, SIRIUS molecular fingerprints and xboost (Extreme Gradient Boosting) models with SMOTE (Synthetic Minority Oversampling Technique) for handling data imbalance were a universally successful and robust modeling configuration. Validation of MLinvitroTox on MassBank spectra showed that toxicity could be predicted from molecular fingerprints derived from MS2 with an average balanced accuracy of 0.75. By applying MLinvitroTox to environmental HRMS/MS data, we confirmed the experimental results obtained with target analysis and narrowed the analytical focus from tens of thousands of detected signals to 783 features linked to potential toxicity, including 109 spectral matches and 30 compounds with confirmed toxic activity.

Prioritizing molecular formulae identified by non-target analysis through high-throughput modelling: application to identify compounds with high human accumulation potential from house dust

Because it is typically not possible to pursue compound identification efforts for all chemical features detected during non-target analysis (NTA), the need for prioritization arises. Here we propose a strategy that ranks chemical features detected in environmental samples based on a model-derived metric that quantifies a feature's attribute that makes it desirable to elucidate its structure, e.g., a high potential for bioaccumulation in humans or wildlife. The procedure involves the identification of isomers that could plausibly represent the molecular formulae assigned to NTA-detected chemical features. For each isomer, the prioritization metric is calculated using properties predicted with high-throughput methods. After the molecular formulae are ranked based on the average values of the prioritization metric calculated for all isomers assigned to a formula, the highest ranked molecular formulae are prioritized for structure elucidation. We applied this workflow to features identified in house dust, using the ratio of chemical intake through dust ingestion to chemical concentration in blood (dose-to-concentration ratio, DCR) as the prioritization metric. Collections of isomers for the molecular formulae were assembled from the PubChem database and DCR was estimated using partitioning and biotransformation properties predicted for each isomer using quantitative structure property relationships. The ten top-ranked molecular formulae with notably lower average DCR-values represented mostly compounds already known to be indoor pollutants of concern, such as two polybrominated diphenyl ethers, bis(2-ethylhexyl) tetrabromophthalate, tetrabromobisphenol A, tris(1,3-dichloroisopropyl)phosphate and the azo dye disperse blue 373.

Predicting the Diagnostic Information of Tandem Mass Spectra of Environmentally Relevant Compounds Using Machine Learning

Acquisition and processing of informative tandem mass spectra (MS2) is crucial for numerous applications, including library-based (tentative) identification, feature prioritization, and prediction of chemical and toxicological characteristics. However, for environmentally relevant compounds, approaches to automatically assess the quality of the MS2 spectra are missing. This work focused on developing a machine learning-based approach to automatically evaluate the diagnostic information of MS2 spectra (e.g., number, distribution, and intensity of diagnostic fragments) of environmentally relevant compounds analyzed with electrospray ionization. For this, approximately 1400 MS2 spectra of 204 environmental contaminants, acquired with different collision energies using liquid chromatography coupled to high-resolution mass spectrometry, were used to train a random forest classifier to distinguish between spectra providing good or poor diagnostic information. Prior to training, validation, and testing, spectra were manually labeled based on criteria such as number, intensity, range of fragments present, molecular ion intensity, and noise levels. Subsequently, feature engineering and selection were applied to retrieve relevant variables from raw MS2 spectra as inputs for the classifier. The optimal set of features based on model performances was selected and used to train a final model, which showed an accuracy of 84%, a precision of 88%, and a recall of 75%. Results show that the combination of selected features and the machine learning model used here can effectively distinguish between MS2 spectra providing good or poor diagnostic information according to the defined criteria. The developed model has the potential to improve a broad range of applications that rely on MS2 data.

Bypassing the Identification: MS2Quant for Concentration Estimations of Chemicals Detected with Nontarget LC-HRMS from MS2 Data

Nontarget analysis by liquid chromatography-high-resolution mass spectrometry (LC-HRMS) is now widely used to detect pollutants in the environment. Shifting away from targeted methods has led to detection of previously unseen chemicals, and assessing the risk posed by these newly detected chemicals is an important challenge. Assessing exposure and toxicity of chemicals detected with nontarget HRMS is highly dependent on the knowledge of the structure of the chemical. However, the majority of features detected in nontarget screening remain unidentified and therefore the risk assessment with conventional tools is hampered. Here, we developed MS2Quant, a machine learning model that enables prediction of concentration from fragmentation (MS2) spectra of detected, but unidentified chemicals. MS2Quant is an xgbTree algorithm-based regression model developed using ionization efficiency data for 1191 unique chemicals that spans 8 orders of magnitude. The ionization efficiency values are predicted from structural fingerprints that can be computed from the SMILES notation of the identified chemicals or from MS2 spectra of unidentified chemicals using SIRIUS+CSI:FingerID software. The root mean square errors of the training and test sets were 0.55 (3.5×) and 0.80 (6.3×) log-units, respectively. In comparison, ionization efficiency prediction approaches that depend on assigning an unequivocal structure typically yield errors from 2× to 6×. The MS2Quant quantification model was validated on a set of 39 environmental pollutants and resulted in a mean prediction error of 7.4×, a geometric mean of 4.5×, and a median of 4.0×. For comparison, a model based on PaDEL descriptors that depends on unequivocal structural assignment was developed using the same dataset. The latter approach yielded a comparable mean prediction error of 9.5×, a geometric mean of 5.6×, and a median of 5.2× on the validation set chemicals when the top structural assignment was used as input. This confirms that MS2Quant enables to extract exposure information for unidentified chemicals which, although detected, have thus far been disregarded due to lack of accurate tools for quantification. The MS2Quant model is available as an R-package in GitHub for improving discovery and monitoring of potentially hazardous environmental pollutants with nontarget screening.

Complementary methods for structural assignment of isomeric candidate structures in non-target liquid chromatography ion mobility high-resolution mass spectrometric analysis

Non-target screening with LC/IMS/HRMS is increasingly employed for detecting and identifying the structure of potentially hazardous chemicals in the environment and food. Structural assignment relies on a combination of multidimensional instrumental methods and computational methods. The candidate structures are often isomeric, and unfortunately, assigning the correct structure among a number of isomeric candidate structures still is a key challenge both instrumentally and computationally. While practicing non-target screening, it is usually impossible to evaluate separately the limitations arising from (1) the inability of LC/IMS/HRMS to resolve the isomeric candidate structures and (2) the uncertainty of in silico methods in predicting the analytical information of isomeric candidate structures due to the lack of analytical standards for all candidate structures. Here we evaluate the feasibility of structural assignment of isomeric candidate structures based on in silico-predicted retention time and database collision cross-section (CCS) values as well as based on matching the empirical analytical properties of the detected feature with those of the analytical standards. For this, we investigated 14 candidate structures corresponding to five features detected with LC/HRMS in a spiked surface water sample. Considering the predicted retention times and database CCS values with the accompanying uncertainty, only one of the isomeric candidate structures could be deemed as unlikely; therefore, the annotation of the LC/IMS/HRMS features remained ambiguous. To further investigate if unequivocal annotation is possible via analytical standards, the reversed-phase LC retention times and low- and high-resolution ion mobility spectrometry separation, as well as high-resolution MS² spectra of analytical standards were studied. Reversed-phase LC separated the highest number of candidate structures while low-resolution ion mobility and high-resolution MS² spectra provided little means for pinpointing the correct structure among the isomeric candidate structures even if analytical standards were available for comparison. Furthermore, the question arises which prediction accuracy is required from the in silico methods to par the analytical separation. Based on the experimental data of the isomeric candidate structures studied here and previously published in the literature (516 retention time and 569 CCS values), we estimate that to reduce the candidate list by 95% of the structures, the confidence interval of the predicted retention times would need to decrease to below 0.05 min for a 15-min gradient while that of CCS values would need to decrease to 0.15%. Hereby, we set a clear goal to the in silico methods for retention time and CCS prediction.

Identifying potential toxic organic substances in leachates from tire wear particles and their mechanisms of toxicity to Scenedesmus obliquus

Tire wear particles (TWPs) are increasingly being found in the aquatic environment. However, there is limited information available on the environmental consequences of TWP constituents that may be release into water. In this study, TWP leachate samples were obtained by immersing TWPs in ultrapure water. Using high-resolution mass spectrometry and toxicity identification, we identified potentially toxic organic substances in the TWP leachates. Additionally, we investigated their toxicity and underlying mechanisms. Through our established workflow, we structurally identified 13 substances using reference standards. The median effective concentration (EC₅₀) of TWP leachates on Scenedesmus obliquus growth was comparable to that of simulated TWP leachates prepared with consistent concentrations of the 13 identified substances, indicating their dominance in the toxicity of TWP leachates. Among these substances, cyclic amines (EC₅₀: 1.04-3.65 mg/L) were found to be toxic to S. obliquus. We observed significant differential metabolites in TWP leachate-exposed S. obliquus, primarily associated with linoleic acid metabolism and purine metabolism. Oxidative stress was identified as a crucial factor in algal growth inhibition. Our findings shed light on the risk posed by TWP leachable substances to aquatic organisms.

Machine Learning and Nontargeted Liquid Chromatography–Mass Spectrometry to Assess Ecotoxicity

Anneli Kruve and Pilleriin Peets from Stockholm University in Sweden, discuss their latest research in machine learning and nontargeted liquid chromatography–mass spectrometry (LC–MS) to assess ecotoxicity.