Geneformer
Theodoris Lab (Harvard / MIT)
Transformer-based foundation model pre-trained on ~30M single-cell transcriptomes. Learns context-dependent gene network dynamics and transfers to diverse downstream tasks including disease modeling and therapeutic target prioritization.
Best For
Transfer learning for gene network analysis; in silico perturbation prediction; disease gene prioritization
License
Open Source (Apache 2.0)
Strengths
- +Pre-trained on ~30M cells
- +In silico perturbation
- +Network-aware gene ranking
- +Apache 2.0 license
Limitations
- −Large compute for fine-tuning
- −Transcriptome-only (no multi-omics)
- −Gene tokenization limits to measured genes
R&D Pipeline Coverage
Related Tools
scGPT
Bo Wang Lab (University of Toronto)
Foundation model for single-cell multi-omics built on generative pre-training of ~33M cells. Fine-tunes to SOTA on cell type annotation, multi-batch integration, perturbation prediction, and gene network inference.
CellTypist
Teichmann Lab (Wellcome Sanger Institute)
Automated cell type annotation tool for scRNA-seq data using logistic regression models trained on curated cross-tissue immune cell atlases. Provides a growing encyclopedia of pre-trained cell type models.
More in Single-Cell
scGPT
Bo Wang Lab (University of Toronto)
Foundation model for single-cell multi-omics built on generative pre-training of ~33M cells. Fine-tunes to SOTA on cell type annotation, multi-batch integration, perturbation prediction, and gene network inference.
CellTypist
Teichmann Lab (Wellcome Sanger Institute)
Automated cell type annotation tool for scRNA-seq data using logistic regression models trained on curated cross-tissue immune cell atlases. Provides a growing encyclopedia of pre-trained cell type models.
Stay updated on Geneformer
Weekly newsletter covering AI tool releases, benchmarks, and what practitioners actually use.