Molecular database characterisation

Silicos' proprietary Spectrophore™ descriptors are ideally suited to characterise, compare, or cluster large number of molecular databases in terms of their molecular diversity or similarity. Important hereby is the fact that the Spectrophore™ technology provides an accurate representation of the molecular properties that are important for ligand-receptor interactions, which guarantees that the clustering or characterisation results are more related to drug-receptor interactions than when the molecular database characterisation calculations have been performed with standard chemical connectivity descriptors such as chemical fragment bitstrings.

The use of Spectrophores™ for database mapping is illustrated in the following example. A 40-by-40 Self Organising Map was trained based on different conformation of 2.5 million lead-like, commercially available compounds. Next, two distinct libraries, the Asinex Building Blocks and the Comprehensive Medicinal Chemistry, were mapped onto this SOM and the enrichment in each cell was computed. The resulting enrichment map is shown in the two figures below. This type of analysis allows to quickly assess how a new library relates to the current collection.
Asinex Building Blocks Comprehensive Medicinal Chemistry

Contact
© 2005-2008 Silicos NV. All rights reserved.
Latest update: 2008-07-05.