We sort the ligands by molecular weight as a proxy for complexity. We then use the algorithm of Bienfait to incrementally select compounds that differ from all previous by the given Tanimoto cutoff. This is a very cheap clustering technique that scales well with set size and gives some indication of the amount of chemical redundancy in the dataset at various Tanimoto levels. N/A indicates that clustering is pending.
Molecules are available in four formats: isomeric SMILES, mol2,
SDF and flexibase. Molecules are represented as a single pH=7 form.
Additional representations (protonation variants and tautomers) are
available in three incremental subsets to augment the single
representative: medium pH (5.75 to 8.25), high pH 7.0-9.5
(e.g. for docking to metals), and low pH 4.5-7.0 (e.g. for docking
to a positively charged binding site).
Larger files are broken up into slices to faciliate downloading.
You may download individual slices or use c-shell scripts to download
a single representation (pH 7.0),
all Usual ligands (pH 5.75 to 8.25), ligands for metals (5.75-9.5) or All.
Note that these sets are overlapping so you do not want to download
both Metals and All.
We expect dockers will want to just download the "Usual" subset.
Chemical informaticists who require only a single form of each molecule
may want just the "Single" representation. If files appear to be missing
or incomplete, please try again tomorrow as the export may still be in
progress. If problems persist for 48 hours please complain to
comments at docking dot org
A product of
BCIRC, the
Bioinfomatics and Chemical Informatics Research Center @ UCSF.
Last updated Aug 6, 2009.
questions and discussion to zinc-fans at docking.org;
bug reports to support at docking.org;
any other correspondence to comments at docking.org.
Terms of use.
Privacy policy.
