ML/AI Tools to Advance Genomic Translational Research (MAGen) - Development Sites (UG3/UH3 Clinical Trials Not Allowed) | RFA HG 24 004

Opportunity Information: Apply for RFA HG 24 004

The National Human Genome Research Institute (NHGRI), part of the National Institutes of Health (NIH), is soliciting applications for the "ML/AI Tools to Advance Genomic Translational Research (MAGen) - Development Sites" program (Funding Opportunity Number RFA-HG-24-004). This opportunity supports the creation of a research consortium focused on testing whether machine learning and artificial intelligence methods can be developed and used in practical translational genomics settings to better predict if, when, and how people who carry known pathogenic genetic variants will actually develop disease. The emphasis is on improving accuracy and precision in real-world prediction, not just demonstrating an algorithm in isolation. Clinical trials are not allowed under this announcement.

The scientific purpose centers on a well-known challenge in genomic medicine: two individuals can carry the same pathogenic variant yet show very different outcomes, ranging from early and severe disease to mild symptoms or no apparent disease at all. MAGen aims to address that variability by building ML/AI tools that integrate and learn from existing multimodal datasets. These datasets can include genomic data (such as sequencing and variant annotations) alongside non-genomic factors that may shape disease expression, for example clinical history, laboratory values, imaging, demographics, environmental exposures, lifestyle factors, and social or behavioral determinants of health, depending on what is available in existing resources. A key expected output is a clearer identification of which genomic and non-genomic modifiers are most associated with disease development among variant carriers, and how those modifiers can be used to make predictions that hold up across different settings.

A major requirement embedded in the program concept is cross-validation in genomic translational research contexts. In practice, that means models should not be built only on a single dataset and reported with internal performance metrics; they should be evaluated in ways that demonstrate generalizability and usefulness for translation, such as testing across multiple cohorts, sites, or data sources when feasible. The program is framed as a consortium effort, so awardees should be prepared to collaborate, share approaches, and align on evaluation strategies that make results comparable across participating groups. The funding mechanism is a cooperative agreement (UG3/UH3), which generally indicates substantial NIH scientific and programmatic involvement and typically supports a staged structure: an initial milestone-driven phase focused on feasibility and development (UG3) that can transition to an implementation/expansion phase (UH3) if milestones are met.

In addition to technical development and validation, MAGen explicitly integrates ethical, legal, and social implications (ELSI) work as a core part of the consortium. The opportunity calls for establishing an ELSI Framework to guide the development and potential integration of ML/AI tools into genomic medicine, along with the implementation of ELSI research projects. This focus reflects common risks and concerns when applying AI in health and genomics, including fairness and bias across ancestry and demographic groups, transparency and explainability of models used in clinical decision-making, accountability for errors, privacy and data governance for sensitive genomic and health information, and how predictions are communicated to patients and families in ways that are understandable and do not cause avoidable harm. Applicants should expect that the program will value responsible AI practices as much as predictive performance.

Eligibility is broad and includes many organization types: state, county, and local governments; special district governments; independent school districts; public and state-controlled institutions of higher education; private institutions of higher education; federally recognized Native American tribal governments; Native American tribal organizations that are not federally recognized tribal governments; public housing authorities/Indian housing authorities; nonprofit organizations with and without 501(c)(3) status (other than institutions of higher education); for-profit organizations (other than small businesses); and small businesses. The announcement also highlights additional eligible applicant categories such as Alaska Native and Native Hawaiian Serving Institutions, Asian American Native American Pacific Islander Serving Institutions (AANAPISI), Hispanic-serving Institutions, Historically Black Colleges and Universities (HBCUs), Tribally Controlled Colleges and Universities (TCCUs), faith-based or community-based organizations, eligible federal agencies, regional organizations, and U.S. territories or possessions. Foreign organizations (non-U.S. entities) are not eligible to apply, and non-domestic components of U.S. organizations are not eligible. However, foreign components, as defined in the NIH Grants Policy Statement, are allowed, meaning a U.S. applicant may include certain foreign activities or collaborations if they meet NIH policy requirements.

From an administrative standpoint, the opportunity is categorized as discretionary funding and uses the NIH cooperative agreement instrument. It falls under health-related NIH activity categories and lists CFDA numbers 93.172 and 93.310. The original application closing date is July 28, 2024, and the listed award ceiling is $1,600,000. The opportunity was created on May 10, 2024. Overall, MAGen is positioned for teams that can combine strong translational genomics expertise, access to high-quality existing multimodal datasets, robust ML/AI development and validation capability, and a serious plan for addressing ELSI issues alongside technical work, all within a collaborative consortium structure.

• The National Institutes of Health in the health sector is offering a public funding opportunity titled "ML/AI Tools to Advance Genomic Translational Research (MAGen) - Development Sites (UG3/UH3 Clinical Trials Not Allowed)" and is now available to receive applicants.
• Interested and eligible applicants and submit their applications by referencing the CFDA number(s): 93.172, 93.310.
• This funding opportunity was created on 2024-05-10.
• Applicants must submit their applications by 2024-07-28. (Agency may still review applications by suitable applicants for the remaining/unused allocated funding in 2026.)
• Each selected applicant is eligible to receive up to $1,600,000.00 in funding.
• Eligible applicants include: State governments, County governments, City or township governments, Special district governments, Independent school districts, Public and State controlled institutions of higher education, Native American tribal governments (Federally recognized), Public housing authorities/Indian housing authorities, Native American tribal organizations (other than Federally recognized tribal governments), Nonprofits having a 501 (c) (3) status with the IRS, other than institutions of higher education, Nonprofits that do not have a 501 (c) (3) status with the IRS, other than institutions of higher education, Private institutions of higher education, For-profit organizations other than small businesses, Small businesses, Others.

Apply for RFA HG 24 004

[Watch] Creating a grant proposal using the step-by-step wizard inside the applicant portal: