Transformative Artificial Intelligence and Machine Learning Based Strategies to Identify Determinants of Exceptional Health and Life Span (R21/R33 Clinical Trial Not Allowed) | RFA AG 22 022

FAQs: NIH RFA-AG-22-022 (R21/R33) Transformative AI/ML Strategies for Exceptional Health and Life Span

What is the focus of this NIH grant opportunity?

This opportunity supports the creation of new, high-impact artificial intelligence (AI) and machine learning (ML) methods to make sense of complex biological data related to exceptional longevity and healthy aging. The emphasis is on computational innovation that can integrate and interpret massive, heterogeneous multi-omic datasets and produce biologically meaningful, interpretable patterns tied to exceptional healthspan and lifespan.

What is the main scientific goal of the FOA?

The central goal is to accelerate the development of transformative computational approaches that connect multiple layers of biology (for example, genome, epigenome, transcriptome, proteome, metabolome, microbiome, and phenome) and extract signals that help explain why some humans experience exceptional longevity with relatively good health and why some animal species live much longer than others.

Is this grant meant to fund wet-lab or clinical intervention research?

No. The FOA is not intended to fund traditional wet-lab discovery or clinical intervention studies. It is specifically aimed at computational method development, AI/ML strategy creation, and automation pipelines for data integration and interpretation.

Are clinical trials allowed under this funding opportunity?

No. This FOA is designated "Clinical Trial Not Allowed," meaning proposed work must not include clinical trial activities as defined by NIH. The focus should remain on computational approaches and data-driven inference rather than testing interventions in humans.

What types of data are expected to be addressed by proposed methods?

Projects are expected to handle complex, heterogeneous biological data, especially multi-omic data. The FOA highlights integrating layers such as genome, epigenome, transcriptome, proteome, metabolome, microbiome, and phenome, and relating molecular and cellular variables to aging-relevant outcomes.

What does "integrative, cross-domain analysis" mean in this context?

It refers to approaches that link multiple biological measurement domains (DNA, RNA, proteins, metabolites, and other cellular measurements) to outcomes relevant to aging, including disease risk and resilience. A key requirement is connecting patterns across these layers rather than analyzing each omic layer in isolation.

Does the FOA emphasize exceptional healthy aging specifically?

Yes. The FOA particularly focuses on exceptionally healthy aging rather than aging characterized by heavy disease burden. Proposed methods should help identify determinants of disease resistance and resilience associated with exceptional healthspan and lifespan.

Does the opportunity encourage using both human and non-human data?

Yes. The FOA explicitly encourages work leveraging human exceptional longevity cohorts and comparative biology across multiple non-human species with wide variation in lifespan. Cross-species signals may help distinguish conserved mechanisms of lifespan and healthspan from signals that are specific to a population or environment.

Why is cross-species analysis considered valuable for this program?

Because cross-species comparisons can help identify deep, conserved biological mechanisms related to lifespan and healthspan. By contrasting species with dramatically different lifespans, applicants may be able to separate fundamental aging biology from cohort-specific or environment-specific effects.

What kinds of computational deliverables are expected?

The FOA describes deliverables such as transformative AI/ML strategies, automated workflows or pipelines, usable software, models, or analytic frameworks that can scale to large datasets, generalize across cohorts and species, and produce interpretable outputs tied to aging biology.

What funding mechanism is used for this opportunity?

The mechanism is the NIH Phased Innovation Award (R21/R33). It is a two-stage structure designed to support early development and proof-of-concept work first, followed by expansion and more robust development once milestones are met.

What is the purpose of the R21 phase in the R21/R33 mechanism?

The R21 phase is generally meant for early, exploratory development and proof-of-concept activities. Examples described include designing new algorithms, building automated workflows, demonstrating feasibility on representative datasets, and defining performance metrics.

What is the purpose of the R33 phase in the R21/R33 mechanism?

The R33 phase supports expansion and more robust development after predefined milestones are met. The FOA describes activities such as scaling methods to larger datasets, improving generalizability across cohorts and species, strengthening validation and benchmarking, and delivering usable tools or frameworks.

How does a project transition from the R21 phase to the R33 phase?

Transition is tied to achieving predefined milestones. The excerpt notes that the listing does not include the detailed milestone expectations, so applicants would typically consult the full FOA for the specific milestones and transition requirements.

Who is eligible to apply as the lead applicant organization?

Eligibility is broad and includes many types of U.S.-based organizations and governmental entities. Examples listed include state, county, and local governments; special district governments; independent school districts; public and state-controlled institutions of higher education; private institutions of higher education; tribal governments and tribal organizations; public housing authorities; nonprofits with or without 501(c)(3) status (other than institutions of higher education); for-profit organizations (including other than small businesses); small businesses; and other organizations that meet NIH eligibility rules.

Are minority-serving institutions and community-based organizations included in eligible applicant types?

Yes. The announcement calls out additional eligible categories including HBCUs, Hispanic-serving Institutions, TCCUs, Alaska Native and Native Hawaiian Serving Institutions, AANAPISIs, as well as faith-based or community-based organizations, eligible federal agencies, regional organizations, and U.S. territories or possessions.

Can a non-U.S. organization apply as the primary applicant?

No. Non-domestic (non-U.S.) entities and non-U.S. foreign institutions are not eligible to apply as the primary applicant organization for this opportunity.

Are foreign collaborations allowed at all?

Yes, within limits. Non-domestic components of U.S. organizations may participate, and foreign components (as NIH defines them) are allowed. The key point is that the lead applicant must be an eligible domestic entity, while foreign components can be included when justified and compliant with NIH policy.

What is the CFDA number and general category of this opportunity?

The opportunity is listed as a discretionary NIH grant in the health category with CFDA 93.866.

What are the key administrative dates mentioned in the listing?

The excerpt states a creation date of July 6, 2021, and an original closing date of October 28, 2021.

Does the excerpt specify the award ceiling or expected number of awards?

No. The excerpt provided does not include an award ceiling or expected number of awards. It notes that applicants would typically consult the full FOA text for budget limits, project period guidance, review criteria, and milestone expectations.

What kinds of technical challenges is the FOA expecting teams to address?

The FOA highlights challenges such as integrating massive and heterogeneous multi-omic datasets; harmonizing or modeling differences in data structure, measurement platforms, and biological context across cohorts and species; and still producing interpretable, biologically grounded insights tied to exceptional health and lifespan.

What would a strong-fit proposal generally look like based on the excerpt?

Based on the description, strong-fit proposals would likely push beyond standard bioinformatics by delivering genuinely new AI/ML approaches and automated systems that can handle multi-omic scale and complexity, reconcile differences across cohorts and species, include rigorous benchmarking and validation, and generate interpretable outputs useful to aging researchers.

What is the official title and identifier of the funding announcement?

The title is "Transformative Artificial Intelligence and Machine Learning Based Strategies to Identify Determinants of Exceptional Health and Life Span (R21/R33 Clinical Trial Not Allowed)" and the FOA identifier is RFA-AG-22-022.