A field guide to the cell-scale AI model zoo

What it does

This is a curated awesome-list that tracks the rapidly expanding literature on foundation models for single-cell genomics. It sorts papers into unimodal, multimodal, and LLM-based models, plus surrounding topics like virtual cells, perturbation benchmarks, and AI agents. The taxonomy follows the authors’ own review, so the repo acts as a living companion bibliography rather than a codebase.

The interesting bit

The maintainers include an evolutionary tree diagram that attempts to chart how dozens of models relate to one another—an unusually visual way to impose order on a field currently drowning in preprints.

Key highlights

• Organizes models by modality: unimodal (scRNA-seq, scATAC-seq), multimodal, and LLM-based.
• Surveys adjacent infrastructure: pretraining datasets, benchmarks, AI agents, and lab-in-the-loop discovery platforms.
• Explicitly invites pull requests for missing papers and broken links, admitting the list decays without community upkeep.
• Tied directly to the review “The landscape of single-cell foundation models: design principles, applications, and open challenges”.

Caveats

• This is strictly a reading list; there are no installable tools or reproducible pipelines here.
• The README itself warns that broken links are an expected hazard in such a fast-moving bibliography.
• Several entries are dated 2026, indicating aggressive tracking of preprints—verify publication status before treating them as settled science.

Verdict

Computational biologists and ML researchers who need to keep pace with single-cell foundation models should bookmark this. If you are hunting for a drop-in Python package, look elsewhere; this is purely a literature map.