Correcting Cell Count Bias Improves Perturbation Effect Estimates and Virtual Cell Models

Virtual Target Identification and Double Perturbation Screens Enable Scalable Discovery of Pancreatic Cancer Vulnerabilities and KRAS Combination Opportunities

Assessing a Virtual Cell’s Utility

Masked BERT-like Pretraining of Single-cell Foundation Models Does Not Improve Virtual Cells

Computational modeling of immune cell phenotypes enhances prediction of clinical anti-cancer drug response

Benchmarking a foundational cell model for post-perturbation RNAseq prediction

The EFFECT Benchmark Suite: Measuring cancer sensitivity prediction performance – without bias