Accelerating environmental health sciences through data and metadata standards to support FAIR principles

Abstract

Environmental health sciences (EHS) generate varied and highly specialized datasets that span fields including exposure science, toxicology, epidemiology, and molecular biology. Despite their shared relevance to EHS, these data streams can be siloed due to inconsistent metadata practices, limited use of common data elements, and uneven adoption of community standards. This fragmentation limits data integration, reduces reproducibility, and hampers the development and application of emerging computational approaches, including artificial intelligence (AI) and machine learning (ML). This article provides an overview of existing data and metadata standards most relevant to EHS, including public health initiatives, controlled vocabularies and ontologies, and platforms that support findability, accessibility, interoperability, and reusability (FAIR data practices). We outline practical steps for researchers who wish to incorporate standards into their workflows, emphasizing early planning, use of established templates, and availability of institutional resources. We then provide examples of ongoing efforts across EHS domains to develop interoperable infrastructures that enable data integration and highlight how standards enable leveraging of emerging technologies such as AI/ML to enhance data use. This work makes a key contribution to the EHS field by demonstrating the importance of community engagement and active collaboration among developers and users, demystifying the processes of adoption of these tools across EHS domains. As AI/ML technologies stand poised to revolutionize cross-disciplinary science, we provide a vision for the future of an AI-ready, interoperable EHS data ecosystem to support interdisciplinary research, enhance transparency, and accelerate scientific discovery.