01
Why It Matters
It turns cellular modeling into a falsifiable response-prediction problem.
02
What Changed
Shared perturbation datasets now support comparative evaluation.
03
Institution
Mapping in progress
04
Researchers
05
Related Problems
06
Sources
Deep-learning perturbation prediction does not yet outperform simple linear baselines
Peer-reviewed benchmark and important negative result.
Nature MethodsOpen source ↗Predicting transcriptional outcomes of novel multigene perturbations
Peer-reviewed GEARS perturbation-prediction study.
Nature BiotechnologyOpen source ↗