FTree query construction for virtual screening: a statistical analysis
- 99 Downloads
FTrees (FT) is a known chemoinformatic tool able to condense molecular descriptions into a graph object and to search for actives in large databases using graph similarity. The query graph is classically derived from a known active molecule, or a set of actives, for which a similar compound has to be found. Recently, FT similarity has been extended to fragment space, widening its capabilities. If a user were able to build a knowledge-based FT query from information other than a known active structure, the similarity search could be combined with other, normally separate, fields like de-novo design or pharmacophore searches. With this aim in mind, we performed a comprehensive analysis of several databases in terms of FT description and provide a basic statistical analysis of the FT spaces so far at hand. Vendors’ catalogue collections and MDDR as a source of potential or known “actives”, respectively, have been used. With the results reported herein, a set of ranges, mean values and standard deviations for several query parameters are presented in order to set a reference guide for the users. Applications on how to use this information in FT query building are also provided, using a newly built 3D-pharmacophore from 57 5HT-1F agonists and a published one which was used for virtual screening for tRNA-guanine transglycosylase (TGT) inhibitors.
KeywordsFeature trees MDDR ZINC Query building Database profiling FTrees
We are in debt to Joerg Degen and Prof. Matthias Rarey (ZBH, Hamburg) for their suggestions and to the reviewers for their constructing criticism.
- 2.Stiefl N, Zaliani A (2006) J Chem Inf Model 46:87Google Scholar
- 8.Tanimoto TT (1957) IBM Internal ReportGoogle Scholar
- 20.The MDL Drug Data Report is available from MDL Information Systems Inc., San Leandro CAGoogle Scholar
- 22.http://blaster.docking.org/zinc/subset1 (accessed Dec 2007)
- 24.http://www.chemcomp.com (accessed Dec 2007)
- 25.FTrees 1.5.1, http://www.biosolveit.de/FTrees (accessed Dec 2007) available from BioSolveIT, St. Augustin, Germany
- 26.http://www.python.org (accessed Dec 2007)
- 27.http://www.scipy.org (accessed Dec 2007)
- 31.http://www.schroedinger.com (accessed Dec 2007)
- 32.http://www.biosolveit.de/FTrees/download/ftrees_ug.pdf (accessed Dec 2007)