For extensive testing of the FlexS method we established a dataset of molecules with known binding geometry based on crystallographic data. Results of these tests have been published in the Journal of Medicinal Chemistry (see below).
Our primary source of input data was the PDB but we also utilized a few complexes that have been made available by different contributors for the CASP2 competition. An early version of the dataset was established by Prof. Dr. Gerhard Klebe in 1994. Later Dr. Christian Lemmen extended the dataset to its current size for objectively assessing the performance of the superpositioning software FlexS. The authors are grateful to Dr. Markus Böhm who did a tremendous job in cleaning up the dataset and replacing ambiguities with great chemical expertise.
The dataset consists of 77 ligands from protein-ligand complexes grouped by the respective protein to which they bind. In total there are 14 such groups of varying size. All the complexes within a group have been aligned on the basis of the common protein backbone C-a atoms. This gives us the coordinates of the aligned bound ligands – information which we offer here for public scientific use.
A detailed description about how the dataset was created and the naming convention we used is available.
FlexS-77 is not public domain!
The FlexS-77 dataset can be used free of charge under the following conditions:
- BioSolveIT grants the license to download the FlexS-77 dataset for internal use. Redistribution of the dataset or integration of the dataset into other software packages or datasets is strictly prohibited.
- In any presentation or publication that contains work based around this dataset, we request that you acknowledge it as follows:
-77 dataset collected by C. Lemmen / G. Klebe / M. Böhm,
first published in
Lemmen, C; Lengauer, T; Klebe, G.
FlexS: A Method for Fast Flexible Ligand Superposition
J. Med. Chem. 1998, 41, 4502-4520