Universal cell embedding provides a foundation model for cell biology
From the article
Nature, Published online: 08 July 2026; doi:10.1038/s41586-026-10689-z The universal cell embedding foundation model learns to capture the organization and variation of cells by training on 36 million cells from hundreds of experiments, dozens of tissues and eight species.
Continue reading on Nature CommunicationsYou might also wanna read
One billion cells
cells.andersmurphy.com·1y ago
LOGOS: A Unified Generative Foundation Model for the Natural Sciences Using Shared Scientific Grammar
This report introduces LOGOS (Language Of Generative Objects in Science), a scientific generative language model that unifies diverse tasks

University of Michigan researchers develop technique to map cell-to-cell communication using spatial transcriptomics
A new University of Michigan study published in Nature Genetics introduces a computational technique that uses spatial transcriptomics to in

TranscriptFormer: Generative AI model maps cellular evolution across 1.5 billion years
Researchers developed TranscriptFormer, a generative AI model that creates a cell atlas spanning 1.5 billion years of evolution by analyzing
University of Minnesota scientists create synthetic 'SpudCells' with no evolutionary ancestors
Scientists at the University of Minnesota have created synthetic cells called "SpudCells" — lab-built cells with no evolutionary ancestors.
Single-cell map of the healthy human immune system across the lifespan reveals unique infant immune signatures
nature.com·14d ago

Comments
Sign in to join the conversation.
No comments yet. Be the first.