Why Do I Build Models Membrane protein models

Why Do I Build Models? • Membrane protein models – what is the structure of proteins currently intractable to experimental characterization and what do the structures suggest about function? • Docking – how do small molecules fit into proteins and how does this help rational design/selection of new candidates? • Virtual Screening – what additional small molecules will bind to the target proteins?

Atomic Scale • Virtual screening tools – example: QSAR • Statistical uncertainty quantitated for training/test sets • Models convincing to broader community when BROADLY effective

Atomic Scale • Virtual screening tools – example: pharmacophore • Uncertainty can be reflected by – Hit rate – Enrichment factor • Models convincing to broader community when effective

Atomic Scale • Protein/Peptide Design • Uncertainty quantitation mechanism not as clear • Models convincing to broader community when design produces desired result

Where can we go next? • Model-driven experimentation – Currently collaborative (theory/experiment) – Data/Model sharing might broaden model use • Challenge: common model formats not available • Challenge: conveying model scope to novice • Challenge: conveying model limitations to novice – Scale integration broadens use/acceptance