Screening policy

Predicted binding-affinity threshold: 0.7

This case study models a virtual screen of 50,000 candidate molecules against a target protein. A model predicts each candidate's binding affinity score. Only candidates above the threshold advance to expensive, slow wet-lab testing, where the true hit rate is finally confirmed.

Screening outcome

Candidates advanced to wet lab—

True hits among advanced—

Estimated wet-lab cost—

Candidates by predicted affinity (cyan = advanced, gold ring = true hit)