PANGEA makes 46.5 million cells easier to compare across diseases
PANGEA, published November 12 in Nature Methods, combines 46.5 million cell profiles across 160 disease categories and 48 organs, with an open annotation toolkit for comparing cell identities.
PANGEA assembles 46.5 million single-cell profiles across 160 disease categories and 48 organs within a common annotation framework.
Published November 12 in Nature Methods, it addresses a problem in combining cell atlases: different studies can give related populations different names. [1]
The compared source populations were liver Kupffer cells, breast M2 macrophages and most myeloid cells in the heart atlas. PANGEA grouped these populations as LYVE1-positive macrophages, supported by shared LYVE1 expression. This makes the labels comparable across studies without establishing identical behavior in each tissue. [1]
The web portal and pangeapy Python package support cell annotation, gene-expression exploration and metadata prediction. Researchers can compare an existing dataset’s original labels with the new assignments and examine disagreements before combining populations in a downstream analysis. The resource is a research atlas, not a validated diagnostic test. [1][2][3]
Take it further
Explore PANGEA Public research portal; no deadline.
Sources & details 3 sources
Discuss in your voice app
Take the reporting, visual explanations and original sources into a conversation. Copy the brief, paste it into your preferred assistant, then start its voice mode.
Paste the brief into your voice app to discuss this reporting.Revised 12 November 2026. Revised visual explanations and copy; added Nvidia investigation context and clarified tool eligibility.
Reporting sources
- Nature Methods · PANGEA
November 12 original resource paper, methods, data composition and cross-organ LYVE1-positive macrophage example. Full public HTML reopened; no source figure reproduced.
- PANGEA · research portal
Public annotation and exploration entry point. Availability inspected; no private data uploaded and no model run performed.
- pangeapy · code and documentation
Public cell-annotation and metadata-prediction interfaces, tutorials and web-platform instructions. Same research team; not independent performance replication.
Images & figures
Shared LYVE1 expression supports a comparable annotation across three source atlases; the tissues and their biological differences remain distinct.
Euangelion · Source · Usage terms · Download figure