Screening is typically performed with a combination of carefully selected diverse sets and cherry picked sets, to screen both broadly and focused for specific targets. This is possible due to the highly automated sample storage and retrieval system.

A typical screening set contains 100,000 compounds, 80,000 picked for diversity, using computational filters and hierarchical clustering, and 20,000 picked based on target criteria. The focused set can be designed by structure-based virtual screening, 3D pharmacophore filtering, chemical similarity to known compounds etc., or to increase the sampling in selected parts of the chemical space represented by the compound collection. We can of course also screen the full 300,000 compound collection.

For fragment based screening we have a set of ca 1,000 fragments in a carefully designed screening library.