The NIH funds 18 Common Fund programs — each generating valuable datasets, each with its own data formats, metadata standards, and portals. Individually, they’re useful. Together, they could be transformative. The Common Fund Data Ecosystem (CFDE) was built to make that integration real.
This preprint describes the evolution, architecture, and practical outcomes of CFDE: a collaborative infrastructure that links Common Fund programs and makes their data findable, accessible, and reusable across program boundaries.
The HuBMAP Data Portal is the public face of the Human BioMolecular Atlas Program — the place where the data actually lands and where the broader research community can access it. This preprint describes the portal’s architecture, capabilities, and current scale.
As of October 2025, the portal holds 5,032 datasets spanning 22 data types across 27 organ classes from 310 donors. That’s not a static archive: it’s a queryable, visualizable, analysis-ready resource.
A reference atlas is only as useful as the coordinate system it’s built on. For the human body — with its 37 trillion cells, dozens of organs, and enormous variation across individuals — building that coordinate system from scratch is one of the most ambitious undertakings in modern biology.
This paper describes HuBMAP’s 3D Human Reference Atlas (HRA) v2.0: what it contains, how it was built, and how researchers can use it to map their own data.
Whole-brain microscopy datasets are enormous — often terabytes per image — and the field has been generating thousands of them. The bottleneck is no longer acquisition; it’s storage, sharing, and making all of that data usable by researchers who weren’t part of the original experiment.
The Brain Image Library (BIL) was built to solve that problem.
BIL is a public, persistent repository for brain microscopy data, hosted at the Pittsburgh Supercomputing Center and serving the broader neuroscience community. Rather than requiring researchers to download multi-terabyte datasets before they can work with them, BIL provides integrated analysis and visualization tools that let users explore data in place — directly through the repository.
The Human BioMolecular Atlas Program set out to do something audacious: map the healthy human body at single-cell resolution, across all major organs, from diverse populations, and make everything freely available to the research community.
The first phase was about building the foundation — ontologies, standardized protocols, analytical pipelines, and the infrastructure needed to support a project at this scale. That work is done. The program has now entered its production phase.
Senescent cells — cells that have permanently stopped dividing in response to stress — are a fundamental feature of aging, yet remarkably little is known about where they are, how many exist, or how they change across a human lifespan. The NIH SenNet Consortium was established to answer those questions at scale.
This Perspective lays out the goals, approach, and infrastructure of SenNet: a Common Fund initiative to comprehensively map senescent cells across 18 human tissues and build a publicly available atlas.
Spatial simulations of cell processes require realistic cell geometries — accurate representations of where organelles are, how they’re shaped, and how they vary from cell to cell. Building those geometries by hand doesn’t scale. CellOrganizer was designed to learn them directly from microscope images.
This chapter describes the full workflow: from image preparation through model training, quality assessment, geometry sampling, and integration with biochemical simulation frameworks.
CellOrganizer learns generative statistical models of cell spatial organization from fluorescence microscopy images. A trained model captures not just a single representative cell shape, but the full distribution of variation — in overall cell architecture, organelle count, organelle size and shape, and spatial positioning — across a population of imaged cells.
Searching a database of fluorescence microscopy images by typing keywords works well enough when annotations are complete and consistent. In practice, they rarely are. Images get uploaded with minimal metadata, terminology varies across labs, and no text description fully captures what a subcellular pattern actually looks like.
The real question is: given this image, which other images in the database look like it?
That’s the problem we set out to solve.
