Federated data access and analysis for autoimmune disease research

National Library Of Medicine

Investigators

Abstract

The National Library of Medicine (NLM), in collaboration with the Office of Autoimmune Disease Research (OADR), is spearheading the development of a federated research environment that will significantly advance the integration and analysis of data in autoimmune disease research. This project addresses the inherent challenges in studying autoimmune diseases, which impact a broad spectrum of medical specialties and affect various organs and systems. The complexity of these diseases necessitates a data-driven approach to uncover underlying mechanisms and improve therapeutic strategies. Autoimmune diseases affect millions globally, and recent studies have shown that autoimmune antibodies can be detected years before clinical symptoms appear, highlighting the importance of early detection. However, current research is often hindered by fragmented data sources, limiting the ability to perform comprehensive analyses. This fragmentation impedes the identification of disease subtypes (endotypes) and the development of personalized treatment strategies. A significant barrier to progress in autoimmune disease research is the lack of integration across existing data systems, which often operate in isolation. The proposed federated data integration platform will enable the harmonization of diverse datasets without requiring data to leave its original environment. Importantly, no data will be transferred to Lifebit; all data remains under the control of its respective steward. This federated approach ensures that technology vendors have no control or ownership over the data or the results generated from its analysis. All outcomes and standardized data—such as those mapped to the OMOP Common Data Model—will be stored within the environment of the data steward, ensuring full compliance with data governance policies. The OMOP mappings, a key deliverable of this project, will be made available to the scientific community as open-source resources, providing transparency and supporting broader research efforts. Additionally, any analysis pipelines developed during this project will be based on protocols provided by collaborators and released under the MIT open-source license. These pipelines will be deposited in public repositories, such as GitHub, ensuring that the tools developed benefit the wider research community. The ownership of these pipelines will be vested with the collaborators, who will have discretion over their release. This pilot project is designed to yield significant outcomes, including the development of robust governance frameworks and security protocols for managing federated data architectures for the study of autoimmune diseases to identify and characterize disease endotypes and their diagnostic biomarkers. Technical documentation, policy guidelines, and governance frameworks will be among the key deliverables, ensuring that the federated environment is managed securely and effectively. The insights gained from this pilot will not only benefit the immediate project but will also inform broader efforts to enhance data integration and analysis in autoimmune disease research. In alignment with the goals of this initiative, all models, software, and methods developed will be made available to the autoimmune research community through public repositories. This commitment ensures that the outcomes of this pilot will benefit the entire community, promoting transparency, collaboration, and innovation. Specific Aim 1. Demonstrate Federated Data Query to Identify Cohorts of Autoimmune Disease Patients with Common Characteristics Across Data Resources Specific Aim 2. Explore the Use of Topic Modeling and Swarm Learning Methods to Identify Autoimmune Disease Endotypes Using Clinical and Laboratory Data Across Data Resources Specific Aim 3. Identify Molecular Biomarkers of Autoimmune Disease Endotypes Using Omics Data Across Data Resources