database from the National Cancer Institute (NCI),
is composed of 1990 compounds. Each compound represents
a structural chemical family. For this reason Diversity is an interesting
database for development studies. In order to select only "Drug-Like"
compounds and simulate a real High-Throughput Docking study we have
filtered out the database using derived-Lipinski rules.
|• SlogP > 7 (Wildman
and Crippen, 1999)
• MW < 100 et > 800.
• H donnors > 5
• H acceptors >10
• Number of rotatable bonds > 15
• Contain reactive groups (Oprea, 2000)
• No N or O atom
• Contain linear chains > -(CH2)6CH3
• Contain perfluorinated chains
• Ring numbers > 6
• Big rings> à 7 atoms
• Contain non organic compounds
• Wrong structures
• Number of halogens > 7
|Our final Diversity
"Drug-Like" database is composed of 1420 compounds.
This database is provided by the NCI in three dimensions (3D) structures
stored in a Structural Data File (SDF) which was then transformed
into a Sybyl database for High-Throughput Docking with FlexX.
the 1420 filtered diversity library :