NIH partners to create SI-ready datasets for models that predict human biology

October 7, 2026

National Institutes Of Health

Overview

NIH announced a collaboration with the U.S. Department of Energy, Biohub, and other partners to prepare biomedical datasets for Super Intelligence models that predict cellular and biological-system responses to disease and interventions. The effort will draw on existing NIH repositories, infrastructure, and research programs, while supporting dataset standardization and expanded measurement of cellular responses.

Background

Predictive models of biology require high-quality data describing how cells respond to interventions across different cell types and conditions. NIH identifies national biomedical repositories and Common Fund programs developing biological atlases, shared data standards, and SI-ready datasets as resources for this work. The announcement places the collaboration within the Bio Genesis Mission and the Predicting Living Systems National Science and Technology Challenge. Related NIH initiatives described in the supplied context address human-based research infrastructure and technologies.

Data Highlights

The announcement describes a data and infrastructure initiative; it reports no numerical study results or clinical trial outcomes.

Area	Details reported
Partners	NIH, the U.S. Department of Energy, Biohub, and additional partners
Data sources	NIH-catalogued national biomedical repositories, including resources associated with NLM and NCBI, and NIH Common Fund programs
Planned work	Standardize suitable datasets and support measurements of cellular responses across more cell types and conditions
Intended use	Train predictive models of cellular and biological-system responses to disease and possible interventions

Key Findings

NIH is working with DOE, Biohub, and additional partners to develop SI-ready resources for predictive models of human biology.
The Bio Genesis Mission will integrate existing biomedical datasets, national data infrastructure, and research programs for use by the wider scientific community.
NIH identifies NLM- and NCBI-associated repositories and Common Fund programs developing biological atlases, common data standards, and SI-ready datasets as relevant resources.
NIH and Biohub plan to standardize suitable datasets for model training; the announcement also identifies a need for cellular-response measurements across substantially more cell types and conditions than studied to date.
NIH Deputy Director Nicole Kleinstreuer said combined resources and expertise might accelerate development of universal cell models capable of predicting responses to interventions.
The related NIH announcement on human-based research infrastructure describes five major initiatives intended to advance biomedical research beyond animal models; the supplied context does not provide further details about those initiatives.

Clinical Implications

The announcement describes a research-resource effort, not a clinical evaluation or a recommendation for patient care. NIH states that predictive models could help researchers investigate biological questions computationally, identify promising drug targets and interventions, and prioritize concepts for laboratory and clinical evaluation.

Conclusion

The NIH-led collaboration aims to make existing and future biomedical data more suitable for training predictive models of living systems. Its stated goals include enabling computational investigation of biology and supporting subsequent laboratory and clinical evaluation of promising concepts.