AblateCell: A Reproduce-then-Ablate Agent for Virtual Cell Repositories

TL;DR

AblateCell is an agent that reproduces and systematically ablates virtual cell repositories, improving verification and understanding of biological AI models through automated reproduction and targeted component testing.

Contribution

It introduces a reproduce-then-ablate framework that automates reproduction and component ablation in biological AI repositories, addressing standardization and verification challenges.

Findings

Achieved 88.9% success rate in reproducing baseline workflows.

Recovered ground-truth critical components with 93.3% accuracy.

Outperformed heuristic methods by 53.3% in component identification.

Abstract

Systematic ablations are essential to attribute performance gains in AI Virtual Cells, yet they are rarely performed because biological repositories are under-standardized and tightly coupled to domain-specific data and formats. While recent coding agents can translate ideas into implementations, they typically stop at producing code and lack a verifier that can reproduce strong baselines and rigorously test which components truly matter. We introduce AblateCell, a reproduce-then-ablate agent for virtual cell repositories that closes this verification gap. AblateCell first reproduces reported baselines end-to-end by auto-configuring environments, resolving dependency and data issues, and rerunning official evaluations while emitting verifiable artifacts. It then conducts closed-loop ablation by generating a graph of isolated repository mutations and adaptively selecting experiments under a reward that trades off performance impact and execution cost. Evaluated on three single-cell perturbation prediction repositories (CPA, GEARS, BioLORD), AblateCell achieves 88.9% (+29.9% to human expert) end-to-end workflow success and 93.3% (+53.3% to heuristic) accuracy in recovering ground-truth critical components. These results enable scalable, repository-grounded verification and attribution directly on biological codebases.