Searching Large Sets of Virtual Libraries Using Modal Fingerprints describes the use of Modal Fingerprints as library descriptors. Using a single descriptor for a library instead of the complete ensemble of compounds allows searching huge selections of libraries in a very small timeframe. Background information is given about Fingerprints, the generation of modal fingerprints and the comparison of fingerprints using Tanimoto similarity. The quality of hits resulting from Modal Fingerprints for libraries are compared with a conventional approach being used on several hundred Pfizer libraries. The modal fingerprint is ~2000 times faster identifying the nearest library in chemical space with only a ~10% degradation in quality for the 1st hit.

Presentation slides:

• Link to Presentation