Acta Pharmacologica Sinica (2009) 30: 251-258; doi 10.1038/aps.2008.22; published online 19th January 2009

Aim: The search for molecules whose bioactivities are similar to those of given compounds or to optimize the initial lead compounds from high throughput screening has attracted increasing interest in recent years. Our goal is to provide a publically searchable database of scaffolds out from a large collection of existing chemical molecules.

Results: Although a number of in silico methods have emerged to facilitate this process, which has become known as  ”scaffold hopping?or “molecular hopping? there is an urgent need for a database system to provide such valuable data in the drug design field. Here we have systematically analyzed a collection of commercially available small molecule databases and a bioactive compound database to identify unique scaffolds and we have built apublically searchable database. The analysis of approximately 4 800 000 of these compounds identified 241 824 unique scaffolds, which are stored in a relational database (http://202.127.30.184:8080/db.html). Each entry in the database is associated with a molecular occurrence and includes its distribution of molecular properties, such as molecular weight, logP, hydrogen bond acceptor number, hydrogen bond donor number, rotatable bond number and ring number. More importantly, for scaffolds derived from the bioactive compounds database, it also contains the original compounds and their target information.

Conclusion: This Web-based database system could help researchers in the fields of medicinal and organic chemistry to design novel molecules with properties similar to the original compounds, but built on novel scaffolds.

Keywords: chemoinformatics; privileged structure; scaffold database; molecular hopping