Linking chemical data from the Comparative Toxicogenomics Database with adverse outcome pathways from the AOP-Wiki: a mechanistic data-oriented approach to help inform environmental health

Background: Chemicals can perturb gene functions to affect chronic human diseases, and a significant amount of biological knowledge involved in environmental health is available in public databases. Combining information across resources can assist in the discovery of novel testable hypotheses related to how chemical exposures influence human diseases, such as autism. Methods: The Comparative Toxicogenomics Database (CTD) is a public resource that provides curated content for chemicals, genes, phenotypes, diseases, and exposures. The AOP-Wiki is a repository of adverse outcome pathways (AOPs) that provide defined biological frameworks describing disease processes. Here, we intersect CTD toxicogenomic content with the AOP-Wiki to identify environmental chemicals that could potentially modulate key steps in autism. Results: We identify numerous chemical stressors that intersect with the individual events of the autism AOP, including bisphenol compounds, per/polyfluoroalkyl substances, pesticides, metals, and air pollutants, suggesting a wide range of environmental factors that could synergize to potentially affect autism. By integrating additional CTD curated content for three autism-associated chemicals (bisphenol A, particulate matter, and valproic acid), we discover other mechanisms, including specific genes (e.g., SLC1A1, GSTP1, CNTNAP2) and phenotypes (e.g., lipid metabolism, inflammatory response, social behavior) that can be used to help refine or expand this AOP or create an entirely new pathway for autism. Furthermore, related diseases are identified to build interconnected networks, mechanistically linking autism to fatty liver disease, intellectual disability, and cancer. Conclusions: We demonstrate the value of integrating content from different resources to address environmental health questions related to autism etiology and co-morbidities. Importantly, our methodology is easily adapted for any AOP in the AOP-Wiki to identify potential environmental influences on the disease process and help support or refine AOPs. This analysis underscores the importance of standardizing public databases to make them efficiently interoperable for enhanced shared utility across the numerous bioknowledge digital landscapes.