Linking exposome and health using computational approaches [HBNI Th293]

Abstract

Human health is shaped by both genetics and lifetime environmental exposures. The exposome framework systematically captures these environmental exposures and links them to health outcomes. The chemical exposome is of particular concern, as thousands of new chemicals are introduced into the global market annually, with many inevitably finding their way into the environment, yet their effects on human health remain poorly understood. Traditional toxicity testing relied heavily on slow and expensive animal studies, resulting in only a fraction of these chemicals being evaluated. In recent years, computational approaches have emerged as vital alternatives. They help organize toxicity data to prioritize chemicals for testing and provide mechanistic insights into their adverse effects. Therefore, this thesis aims to link the chemical exposome to human health by systematically characterizing the chemical exposome and gaining mechanistic insights into chemical-induced adverse effects through diverse computational approaches.

As the first objective, this thesis focuses on characterizing the chemical exposome associated with specific human health outcomes. The chemical exposome of vitiligo is investigated by curating a dedicated knowledgebase of its environmental chemical triggers. The chemical space is then explored to assess regulatory coverage, and skin-specific transcriptomic signatures are utilized to gain mechanistic insights into the otherwise less understood etiology of this disease. Building on this, the focus is extended to endocrine-disrupting chemicals (EDCs), a broader class of environmental chemicals associated with a wide range of adverse health effects. The Database of Endocrine Disrupting Chemicals and their Toxicity profiles (DEDuCT) is updated by integrating data from diverse toxicological resources, providing a single-window access to all EDC-relevant information. These chemicals are further mapped to Adverse Outcome Pathways (AOPs), and all compiled data are organized into a large-scale toxicology knowledge graph, termed DEDuCT-KG, which is utilized to gain mechanistic insights into EDC-associated obesity and neurodegenerative disorders.

As the second objective, the scope of the computational investigation is expanded beyond human health, employing diverse network toxicology frameworks to assess the impacts of inorganic arsenic and cadmium on human and ecosystem health from a One Health perspective. Aggregate Exposure Pathway (AEP) networks are constructed for the first time in the Indian context and integrated with stressor-AOP networks to trace complete source-to-outcome pathways. Species Sensitivity Distributions (SSDs) are computed from aquatic toxicity data to estimate hazard concentrations, and these are integrated into stressor-species networks to identify particularly vulnerable species and support ecological risk assessment. Finally, Risk Quotients (RQs) are calculated to assess the ecological risks associated with arsenic and cadmium in Indian rivers.

Overall, the resources developed and frameworks utilized in this thesis provide findings and insights that are readily accessible to the scientific community, laying the groundwork for future efforts towards a more comprehensive characterization of the chemical exposome. The structured network toxicology investigation provides a basis for exploring the links between environmental chemical stressors and human and ecosystem health from a One Health perspective.