Professor Yale School of Medicine New Haven, Connecticut, United States
Disclosure(s):
Steven Kleinstein, PhD: No relevant disclosure to display
Introduction/Rationale: Immunology research generates diverse molecular and clinical datasets spanning autoimmunity, infectious diseases, allergy, transplantation, vaccine response and beyond. Decades of accumulated data in generalist and domain-specific repositories present an unprecedented opportunity for machine learning and artificial intelligence (AI) - powered analyses of large-scale, multi-modal datasets, accelerating scientific discovery.
Methods: ImmPort, the NIH-NIAID sponsored open-access immunology data portal, houses over 1,300 studies, 5,000 experiments, and over 7.5 million data points, supporting conventional and emerging technologies for capturing molecular and clinical datasets at subject level. To fully unlock this potential, ImmPort is deploying AI to semi-automate repository operations, reformatting data for AI-readiness, and establishing integrated AI ecosystems that maximize the value of existing immunology data for reproducible research and discoveries.
Results: AI-driven automation streamlines data curation and enhances metadata quality through intelligent extraction and validation. Standardized provenance frameworks, ontologies, machine-readable metadata, and Application Programming Interface (APIs) transform data into computationally accessible resources. ImmPort has developed methods for integrating Common Data Elements (CDEs) and piloted automated pipelines to facilitate systematic mapping for promoting Findability, Accessibility, Interoperability, and Reuse (FAIR) principles. ImmPort disseminates Fast Healthcare Interoperability Resources (FHIR) formatted data and is piloting Data Mesh architecture using ImmPort as a domain-oriented, interoperable data node to enable AI-assisted biomedical discovery.
Conclusion: The ImmPort AI ecosystem enables immunology-specific foundation models trained on multimodal data to capture immune system dynamics, transforming repositories into AI-enabled platforms that accelerate biomarker discovery and advance personalized, translational research.
