Nature Biotechnology paper challenges gene perturbation model benchmarks

A newly peer-reviewed Nature Biotechnology paper argues that benchmarks for gene expression foundation models may be uninformative, suggesting prior findings that deep learning perturbation models fail to beat uninformative baselines stem from flawed metrics, and proposes positive controls to reassess model performance.

2026-10-02 ~ 2026-10-02 · 2 related posts

• Peer-reviewed paper questions whether gene expression model benchmarks are informative — simocristea · 2026-10-02
• Miscalibrated metrics skewed benchmarks: deep learning gene perturbation models do beat baselines — simocristea · 2026-10-02