Skip to content

Single-Cell Foundation Model

Definition
AI-generated

A single-cell foundation model is a large machine-learning model pretrained on massive single-cell transcriptomic (or multi-omic) corpora—often with self-supervised objectives—so that embeddings or heads transfer to tasks such as cell-type annotation, batch correction, gene regulatory prediction, or perturbation response forecasting.

Why it matters in GWAS

These models increasingly mediate how atlases are queried and how in silico interventions are proposed for GWAS genes; claims require scrutiny against simple baselines and independent benchmarks, as in broader foundation model debates.

Example usage

"The discussion compared a single-cell foundation model’s zero-shot cell annotation to marker-based labels in the eQTL panel."

References

  • Cui H, et al. (2024). scGPT: toward building a foundation model for single-cell multi-omics using generative AI. Nat Methods.
  • Hao M, et al. (2024). Large-scale foundation model on single-cell transcriptomics. Nat Methods.
  • Dimitrov D, Schrod S, Rohbeck M, et al. (2026). Interpretation, extrapolation and perturbation of single cells. Nat Rev Genet. https://doi.org/10.1038/s41576-025-00920-4

Last updated (UTC · Git history)