Transport across contexts
Unseen perturbation, donor, state, dataset and calibrated abstention. Primary outputs are pseudobulk response and, when cell-level truth exists, a predicted cell population.
Scientific program · released contract
The core object is now perturbation × cell state × disease context × time × phenotype. A gene alone is not the unit of biological validation.
Two tracks · separate gates
Unseen perturbation, donor, state, dataset and calibrated abstention. Primary outputs are pseudobulk response and, when cell-level truth exists, a predicted cell population.
DMD and matched control, functional primary endpoint, target engagement, toxicity, reagent concordance, donor heterogeneity and prospective outcome registration.
Mechanism chain
These windows are planning defaults and must be calibrated to the selected cell system before registration.
Flagship studies
Which perturbations produce reproducible DMD-relevant functional effects?
DRAFT_PROTOCOL_NO_EXPERIMENT_STARTEDHow do perturbation effects change across myogenic state, time and microenvironment?
REGISTERED_DATA_MODEL_DATA_MISSINGDoes frozen model selection improve future experimental hit rate over random, expert and simple baselines?
REGISTERED_FRAMEWORK_NO_PREDICTIONS_OR_OUTCOMESGoverned evidence transition