What is MAGen?
MAGen (ML/AI Tools to Advance Genomic Translational Research) is an NIH-supported research consortium that develops and validates machine learning (ML) and super intelligence (AI) tools to improve prediction of disease manifestation in people who carry pathogenic genetic variants. MAGen combines genomic, clinical, environmental, and lifestyle data to advance translational genomic research and support more personalized approaches to disease prediction.
What problem is MAGen trying to solve?
Genetic testing can identify disease-associated variants, but it often cannot accurately predict whether a person will develop disease, when symptoms may occur, or how severe the condition may be. MAGen is developing ML/AI tools that integrate multiple sources of data to better answer these important questions and improve risk prediction.
What types of data does MAGen use?
MAGen uses multimodal data, which may include genomic and other omics data, electronic health record (EHR) data, demographic and ancestry-related information, environmental exposures, social determinants of health (SDOH), biomarkers, and wearable device data. Available data sources vary by participating site and project.
How is MAGen different from other AI research initiatives?
A defining feature of MAGen is its commitment to independent cross-site validation. ML/AI tools developed at one consortium site are tested by other sites using different datasets and populations. This helps ensure that the tools are robust, reproducible, and generalizable across diverse health systems and patient populations.
Who participates in the MAGen consortium?
MAGen consists of three Development Sites and a Coordinating Center. Participating institutions include Vanderbilt Health, Broad Institute, Cedars-Sinai, Harvard University, Brigham and Women’s Hospital, Boston Children’s Hospital, Mount Sinai, and Columbia University. The consortium is also guided by NIH program leadership and an independent External Scientific Panel.
What is meant by “translational genomics”?
Translational genomics focuses on turning discoveries from genomic research into practical tools, knowledge, and approaches that can improve healthcare. MAGen seeks to bridge the gap between genetic discoveries and real-world applications by developing predictive tools that can ultimately support clinical decision-making and public health efforts.
How does MAGen address ethics and responsible AI?
MAGen integrates Ethical, Legal, and Social Implications (ELSI) research into all stages of tool development. This includes evaluating and mitigating bias, protecting privacy, supporting responsible data governance, improving transparency, and assessing how AI tools perform across different populations to help reduce the risk of inequitable outcomes.
Does MAGen share patient data across institutions?
Each participating institution maintains stewardship and governance of its own data in accordance with institutional policies, regulatory requirements, and ethical standards. MAGen promotes collaboration through shared definitions, standardized models, common protocols, and cross-site validation approaches while respecting local data governance requirements.
What resources and tools will MAGen produce?
MAGen aims to produce validated ML/AI tool packages, model documentation, phenotype definitions, data dictionaries, evaluation frameworks, cross-site validation protocols, ELSI resources, publications, and dissemination materials. Availability of specific resources may vary by project and stage of development.
How can researchers or organizations collaborate with MAGen?
Researchers, institutions, and other stakeholders interested in collaboration can contact the consortium through the MAGen contact form. Prospective collaborators are encouraged to provide a brief description of their interests, expertise, or intended use of MAGen resources to help identify potential opportunities for engagement.