Appendix D — Glossary
ADMET
Absorption, distribution, metabolism, excretion, and toxicity. These properties shape whether a compound can become a drug.
AI-ready data
Data with sufficient metadata, provenance, structure, quality control, and governance for reliable machine learning use.
Backbone generation
Protein design task that proposes a protein fold or structural scaffold before sequence selection.
Benchmark leakage
Evaluation failure where training data or related examples contaminate a test set.
Cell Painting
High-content imaging assay that profiles cellular morphology using multiplexed fluorescent stains.
Closed-loop experimentation
Workflow where measurements update a model that selects future experiments.
Foundation model
Model pretrained on broad data so its representations transfer to downstream tasks.
Inverse folding
Protein design task that proposes a sequence for a target structure.
Perturbation prediction
Forecasting how a biological system changes after a genetic, chemical, or environmental intervention.
Virtual cell
Computational model intended to predict cell behavior, usually for defined outputs rather than complete cellular simulation.