Artificial Intelligence for chemical risk assessment

Intuitive toxicology in the 21st century-Bridging the perspectives of the public and risk assessors in Europe

Three decades ago, several articles on the subjectivity in chemical risk judgments (i.e., labeled "intuitive toxicology") measured the divide between the public and toxicologists with different backgrounds regarding the validity of predicting health effects based on in vivo studies. Similar divides with impacts on societal discourse and chemical risk assessment practices might exist concerning alternative toxicity testing methods (i.e., in vitro and in silico). However, studies to date have focused either on the public's views of in vivo or stem cell testing or on experts' views of in vivo testing and potential alternatives (i.e., toxicologists and medical students), which do not allow for a direct investigation of potential divides. To fill this knowledge gap, we conducted two online surveys, involving members of the German-speaking public in Switzerland and European human health risk assessors, respectively. This article presents the results of these two surveys regarding the divide in the public's and risk assessors' perspectives on risk assessment based on in vivo, in vitro, and in silico testing. Particularly, the survey with the risk assessors highlights that, beyond scientific and regulatory barriers, alternatives to in vivo testing may encounter individual hurdles, such as higher uncertainty associated with them. Understanding and addressing these hurdles will be crucial to facilitate the integration of new approach methodologies into chemical risk assessment practices as well as a successful transition toward next-generation risk assessment, bringing us closer to a fit-for-purpose and more efficient regulatory landscape.

The role of trust in the use of artificial intelligence for chemical risk assessment

Risk assessment of chemicals is a time-consuming process and needs to be optimized to ensure all chemicals are timely evaluated and regulated. This transition could be stimulated by valuable applications of in silico Artificial Intelligence (AI)/Machine Learning (ML) models. However, implementation of AI/ML models in risk assessment is lagging behind. Most AI/ML models are considered 'black boxes' that lack mechanistical explainability, causing risk assessors to have insufficient trust in their predictions. Here, we explore 'trust' as an essential factor towards regulatory acceptance of AI/ML models. We provide an overview of the elements of trust, including technical and beyond-technical aspects, and highlight elements that are considered most important to build trust by risk assessors. The results provide recommendations for risk assessors and computational modelers for future development of AI/ML models, including: 1) Keep models simple and interpretable; 2) Offer transparency in the data and data curation; 3) Clearly define and communicate the scope/intended purpose; 4) Define adoption criteria; 5) Make models accessible and user-friendly; 6) Demonstrate the added value in practical settings; and 7) Engage in interdisciplinary settings. These recommendations should ideally be acknowledged in future developments to stimulate trust and acceptance of AI/ML models for regulatory purposes.

Glyphosate Separating and Sensing for Precision Agriculture and Environmental Protection in the Era of Smart Materials

The present article critically and comprehensively reviews the most recent reports on smart sensors for determining glyphosate (GLP), an active agent of GLP-based herbicides (GBHs) traditionally used in agriculture over the past decades. Commercialized in 1974, GBHs have now reached 350 million hectares of crops in over 140 countries with an annual turnover of 11 billion USD worldwide. However, rolling exploitation of GLP and GBHs in the last decades has led to environmental pollution, animal intoxication, bacterial resistance, and sustained occupational exposure of the herbicide of farm and companies' workers. Intoxication with these herbicides dysregulates the microbiome-gut-brain axis, cholinergic neurotransmission, and endocrine system, causing paralytic ileus, hyperkalemia, oliguria, pulmonary edema, and cardiogenic shock. Precision agriculture, i.e., an (information technology)-enhanced approach to crop management, including a site-specific determination of agrochemicals, derives from the benefits of smart materials (SMs), data science, and nanosensors. Those typically feature fluorescent molecularly imprinted polymers or immunochemical aptamer artificial receptors integrated with electrochemical transducers. Fabricated as portable or wearable lab-on-chips, smartphones, and soft robotics and connected with SM-based devices that provide machine learning algorithms and online databases, they integrate, process, analyze, and interpret massive amounts of spatiotemporal data in a user-friendly and decision-making manner. Exploited for the ultrasensitive determination of toxins, including GLP, they will become practical tools in farmlands and point-of-care testing. Expectedly, smart sensors can be used for personalized diagnostics, real-time water, food, soil, and air quality monitoring, site-specific herbicide management, and crop control.

Construction of an In Silico Structural Profiling Tool Facilitating Mechanistically Grounded Classification of Aquatic Toxicants

The performance of chemical safety assessment within the domain of environmental toxicology is often impeded by a shortfall of appropriate experimental data describing potential hazards across the many compounds in regular industrial use. In silico schemes for assigning aquatic-relevant modes or mechanisms of toxic action to substances, based solely on consideration of chemical structure, have seen widespread employment─including those of Verhaar, Russom, and later Bauer (MechoA). Recently, development of a further system was reported by Sapounidou, which, in common with MechoA, seeks to ground its classifications in understanding and appreciation of molecular initiating events. Until now, this Sapounidou scheme has not seen implementation as a tool for practical screening use. Accordingly, the primary purpose of this study was to create such a resource─in the form of a computational workflow. This exercise was facilitated through the formulation of 183 structural alerts/rules describing molecular features associated with narcosis, chemical reactivity, and specific mechanisms of action. Output was subsequently compared relative to that of the three aforementioned alternative systems to identify strengths and shortcomings as regards coverage of chemical space.

Advancing exposure data analytics and repositories as part of the European Exposure Science Strategy 2020-2030

High-quality and comprehensive exposure-related data are critical for different decision contexts, including environmental and human health monitoring, and chemicals risk assessment and management. However, exposure-related data are currently scattered, frequently of unclear quality and structure, not readily accessible, and stored in various-partly overlapping-data repositories, leading to inefficient and ineffective data usage in Europe and globally. We propose strategic guidance for an integrated European exposure data production and management framework for use in science and policy, building on current and future data analysis and digitalization trends. We map the existing exposure data landscape to requirements for data analytics and repositories across European policies and regulations. We further identify needs and ways forward for improving data generation, sharing, and usage, and translate identified needs into an operational action plan for European and global advancement of exposure data for policies and regulations. Identified key areas of action are to develop consistent exposure data standards and terminology for data production and reporting, increase data transparency and availability, enhance data storage and related infrastructure, boost automation in data management, increase data integration, and advance tools for innovative data analysis. Improving and streamlining exposure data generation and uptake into science and policy is crucial for the European Chemicals Strategy for Sustainability and European Digital Strategy, in line with EU Data policies on data management and interoperability.

Developing human biomonitoring as a 21st century toolbox within the European exposure science strategy 2020-2030

Human biomonitoring (HBM) is a crucial approach for exposure assessment, as emphasised in the European Commission's Chemicals Strategy for Sustainability (CSS). HBM can help to improve chemical policies in five major key areas: (1) assessing internal and aggregate exposure in different target populations; 2) assessing exposure to chemicals across life stages; (3) assessing combined exposure to multiple chemicals (mixtures); (4) bridging regulatory silos on aggregate exposure; and (5) enhancing the effectiveness of risk management measures. In this strategy paper we propose a vision and a strategy for the use of HBM in chemical regulations and public health policy in Europe and beyond. We outline six strategic objectives and a roadmap to further strengthen HBM approaches and increase their implementation in the regulatory risk assessment of chemicals to enhance our understanding of exposure and health impacts, enabling timely and targeted policy interventions and risk management. These strategic objectives are: 1) further development of sampling strategies and sample preparation; 2) further development of chemical-analytical HBM methods; 3) improving harmonisation throughout the HBM research life cycle; 4) further development of quality control / quality assurance throughout the HBM research life cycle; 5) obtain sustained funding and reinforcement by legislation; and 6) extend target-specific communication with scientists, policymakers, citizens and other stakeholders. HBM approaches are essential in risk assessment to address scientific, regulatory and societal challenges. HBM requires full and strong support from the scientific and regulatory domain to reach its full potential in public and occupational health assessment and in regulatory decision-making.

Artificial Intelligence-Based Toxicity Prediction of Environmental Chemicals: Future Directions for Chemical Management Applications

Recently, research on the development of artificial intelligence (AI)-based computational toxicology models that predict toxicity without the use of animal testing has emerged because of the rapid development of computer technology. Various computational toxicology techniques that predict toxicity based on the structure of chemical substances are gaining attention, including the quantitative structure-activity relationship. To understand the recent development of these models, we analyzed the databases, molecular descriptors, fingerprints, and algorithms considered in recent studies. Based on a selection of 96 papers published since 2014, we found that AI models have been developed to predict approximately 30 different toxicity end points using more than 20 toxicity databases. For model development, molecular access system and extended-connectivity fingerprints are the most commonly used molecular descriptors. The most used algorithm among the machine learning techniques is the random forest, while the most used algorithm among the deep learning techniques is a deep neural network. The use of AI technology in the development of toxicity prediction models is a new concept that will aid in achieving a scientific accord and meet regulatory applications. The comprehensive overview provided in this study will provide a useful guide for the further development and application of toxicity prediction models.

A matter of trust: Learning lessons about causality will make qAOPs credible

•

Quantitative AOPs are toxicodynamic models based on Adverse Outcome Pathways.

•

Mathematical models, e.g. Bayesian networks, can inform the quantification of AOPs.

•

Model credibility is enhanced by applying the principles of causality theory.

•

A causal diagram is illustrated for an AOP for Parkinsonian motor deficits.

•

Computational resources to help assess causality are listed.

Toxicology in the 21st Century has seen a shift from chemical risk assessment based on traditional animal tests, identifying apical endpoints and doses that are “safe”, to the prospect of Next Generation Risk Assessment based on non-animal methods. Increasingly, large and high throughput in vitro datasets are being generated and exploited to develop computational models. This is accompanied by an increased use of machine learning approaches in the model building process. A potential problem, however, is that such models, while robust and predictive, may still lack credibility from the perspective of the end-user. In this commentary, we argue that the science of causal inference and reasoning, as proposed by Judea Pearl, will facilitate the development, use and acceptance of quantitative AOP models. Our hope is that by importing established concepts of causality from outside the field of toxicology, we can be “constructively disruptive” to the current toxicological paradigm, using the “Causal Revolution” to bring about a “Toxicological Revolution” more rapidly.

Praegnatio Perturbatio-Impact of Endocrine-Disrupting Chemicals

The burden of adverse pregnancy outcomes such as preterm birth and low birth weight is considerable across the world. Several risk factors for adverse pregnancy outcomes have been identified. One risk factor for adverse pregnancy outcomes receiving considerable attention in recent years is gestational exposure to endocrine-disrupting chemicals (EDCs). Humans are exposed to a multitude of environmental chemicals with known endocrine-disrupting properties, and evidence suggests exposure to these EDCs have the potential to disrupt the maternal-fetal environment culminating in adverse pregnancy and birth outcomes. This review addresses the impact of maternal and fetal exposure to environmental EDCs of natural and man-made chemicals in disrupting the maternal-fetal milieu in human leading to adverse pregnancy and birth outcomes-a risk factor for adult-onset noncommunicable diseases, the role lifestyle and environmental factors play in mitigating or amplifying the effects of EDCs, the underlying mechanisms and mediators involved, and the research directions on which to focus future investigations to help alleviate the adverse effects of EDC exposure.

Determination of "fitness-for-purpose" of quantitative structure-activity relationship (QSAR) models to predict (eco-)toxicological endpoints for regulatory use

In silico models are used to predict toxicity and molecular properties in chemical safety assessment, gaining widespread regulatory use under a number of legislations globally. This study has rationalised previously published criteria to evaluate quantitative structure-activity relationships (QSARs) in terms of their uncertainty, variability and potential areas of bias, into ten assessment components, or higher level groupings. The components have been mapped onto specific regulatory uses (i.e. data gap filling for risk assessment, classification and labelling, and screening and prioritisation) identifying different levels of uncertainty that may be acceptable for each. Twelve published QSARs were evaluated using the components, such that their potential use could be identified. High uncertainty was commonly observed with the presentation of data, mechanistic interpretability, incorporation of toxicokinetics and the relevance of the data for regulatory purposes. The assessment components help to guide strategies that can be implemented to improve acceptability of QSARs through the reduction of uncertainties. It is anticipated that model developers could apply the assessment components from the model design phase (e.g. through problem formulation) through to their documentation and use. The application of the components provides the possibility to assess QSARs in a meaningful manner and demonstrate their fitness-for-purpose against pre-defined criteria.

Knowledge Organization Systems for Systematic Chemical Assessments

BACKGROUND

Although the implementation of systematic review and evidence mapping methods stands to improve the transparency and accuracy of chemical assessments, they also accentuate the challenges that assessors face in ensuring they have located and included all the evidence that is relevant to evaluating the potential health effects an exposure might be causing. This challenge of information retrieval can be characterized in terms of "semantic" and "conceptual" factors that render chemical assessments vulnerable to the streetlight effect.

OBJECTIVES

This commentary presents how controlled vocabularies, thesauruses, and ontologies contribute to overcoming the streetlight effect in information retrieval, making up the key components of Knowledge Organization Systems (KOSs) that enable more systematic access to assessment-relevant information than is currently achievable. The concept of Adverse Outcome Pathways is used to illustrate what a general KOS for use in chemical assessment could look like.

DISCUSSION

Ontologies are an underexploited element of effective knowledge organization in the environmental health sciences. Agreeing on and implementing ontologies in chemical assessment is a complex but tractable process with four fundamental steps. Successful implementation of ontologies would not only make currently fragmented information about health risks from chemical exposures vastly more accessible, it could ultimately enable computational methods for chemical assessment that can take advantage of the full richness of data described in natural language in primary studies. https://doi.org/10.1289/EHP6994.

Computational modelling for the sustainable management of chemicals

This commentary explores the contribution of computational toxicology to chemical safety assessment in the context of the broad policy challenges faced by the European Union. The state of the European Environment is considered from the perspective of chemical contributions to the burden of disease and ecosystem damage. This sets the scene for highlighting research and innovation opportunities to further develop computational approaches for assessing the human health and environmental effects of chemicals. Emphasis is placed on focus topics that are particularly relevant to the political priorities of the new European Commission. In particular, two of the six priorities are discussed - “The European Green Deal” and “A Europe fit for a Digital Age”. The former includes the zero pollution ambition for a toxic-free environment, including the need to develop safe and sustainable chemicals, while the latter includes the challenges and opportunities posed by Artificial Intelligence. This commentary is based on a presentation given at the 19th meeting of The Italian Society of Toxicology (SITOX), held in Bologna, Italy, in February 2020.