EUVIMEDEUVIMED entdecken
Crossref

TranscriptFormer: A generative cell atlas across 1.5 billion years of evolution

James D Pearce iD, Sara E Simmonds, Gita Mahmoudabadi iD, Lakshmi Krishnan, Giovanni Palla iD, Ana-Maria Istrate iD, Alexander Tarashansky, Benjamin Nelson iD, Omar Valenzuela iD, Donghui Li iD, Stephen R Quake iD, Theofanis Karaletsos iD

Science · 02.07.2026 · 393 (6806)

DOI: 10.1126/science.aec8514 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Science, erschienen am 02.07.2026. Verfasst von James D Pearce und 11 weiteren Autor:innen. Beteiligte Einrichtung: Biohub, Redwood City, CA, USA. Hauptthemen laut MEDLINE: Biological Evolution, Evolution, Molecular, Single-Cell Gene Expression Analysis, Transcriptome. Die Arbeit wurde bisher 30-mal zitiert (Crossref) und nennt 65 Literaturangaben. Gefördert von: Chan Zuckerberg Initiative.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Single-cell transcriptomics is revolutionizing our understanding of cellular diversity, yet comparing transcriptional programs across the tree of life remains challenging. We developed TranscriptFormer, a family of generative foundation models trained on up to 112 million cells spanning 1.53 billion years of evolution across 12 species. We demonstrate state-of-the-art performance on cell type classification, even for species separated by over 685 million years of evolution, and zero-shot disease state identification in human cells. Developmental trajectories, phylogenetic relationships, and cellular hierarchies emerge naturally in TranscriptFormer’s representations without any explicit training on these annotations. This work establishes a powerful framework for quantitative single-cell analysis and comparative cellular biology, thus demonstrating that universal principles of cellular organization can be learned and predicted across the tree of life.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Biological EvolutionEvolution, MolecularSingle-Cell Gene Expression AnalysisTranscriptomeAtlases as TopicAnimalsHumansPhylogeny

Autor:innen und Institutionen

  • Biohub, Redwood City, CA, USA.
  • Department of Bioengineering, Stanford University, Stanford, CA, USA.
  • Department of Applied Physics, Stanford University, Stanford, CA, USA.
  • Chan Zuckerberg Initiative, Redwood City, CA, USA.

ORCID-Profile: James D Pearce, Gita Mahmoudabadi, Giovanni Palla, Ana-Maria Istrate, Benjamin Nelson, Omar Valenzuela, Donghui Li, Stephen R Quake, Theofanis Karaletsos

Förderung

  • Chan Zuckerberg Initiative

Zitieren

APA

Pearce, J. D., Simmonds, S. E., Mahmoudabadi, G., Krishnan, L., Palla, G., Istrate, A. M., Tarashansky, A., Nelson, B., Valenzuela, O., Li, D., Quake, S. R., & Karaletsos, T. (2026). TranscriptFormer: A generative cell atlas across 1.5 billion years of evolution. Science, 393(6806). https://doi.org/10.1126/science.aec8514

Vancouver

Pearce JD, Simmonds SE, Mahmoudabadi G, Krishnan L, Palla G, Istrate AM, et al. TranscriptFormer: A generative cell atlas across 1.5 billion years of evolution. Science. 2026;393(6806). doi:10.1126/science.aec8514

BibTeX
@article{Pearce2026,
  author = {Pearce, James D and Simmonds, Sara E and Mahmoudabadi, Gita and Krishnan, Lakshmi and Palla, Giovanni and Istrate, Ana-Maria and Tarashansky, Alexander and Nelson, Benjamin and Valenzuela, Omar and Li, Donghui and Quake, Stephen R and Karaletsos, Theofanis},
  title = {TranscriptFormer: A generative cell atlas across 1.5 billion years of evolution},
  journal = {Science},
  year = {2026},
  volume = {393},
  number = {6806},
  doi = {10.1126/science.aec8514}
}
Kostenlos starten →