Skip to content
PaperFren

Gene expression

Simple baselines beat cell foundation models on Perturb-seq

Csendes G, Sanz G, Szalay KZ, et al. · BMC genomics · 2025

Open access · cc by · source: Europe PMC

Benchmarking scGPT and scFoundation against baselines finds that training-set mean and basic ML with biological features outperform the foundation models on post-perturbation expression prediction.

Key findings

Even taking the mean of training examples beat scGPT and scFoundation. Biologically featured ML models outperformed scGPT by a large margin. Results expose weaknesses in current post-perturbation benchmarks/metrics and foundation-model claims.

Methodology

Csendes et al. fine-tuned/evaluated scGPT and scFoundation against train-mean and classical regressors (e.g., elastic net, kNN, random forest) using GO/feature embeddings on Adamson, Norman, and Replogle Perturb-seq datasets, scoring Pearson correlation in raw and differential expression spaces.

Limitations

Does not prove foundation models useless for all tasks—only that they underperform on these prediction setups. Future architectures/metrics might reverse the ranking.

How this study connects

Role on claims

Each row is a claim on a concept or method page where this paper supports, challenges, or qualifies the statement. Roles are hand-checked — not a model guess.

  • Foundation cell models for post-perturbation prediction need careful benchmarks.

    Benchmarking foundation cell models for post-perturbation RNA-seq prediction tests how well learned representations generalize after QC and perturbation—downstream of filtering choices.

    Evidence for the claim as stated.

  • A mito-filter methods result, a foundation-model benchmark, and a disease atlas answer different layers of the single-cell stack.

    Evidence for the claim as stated.

Open questions

Tensions this paper is part of

From concept pages' “where studies disagree.” Disagreement means the same question; scope means different assays, populations, or outcomes.

Related papers in this topic

Same topic cluster — not a recommendation engine.