Chemistry Reference
In-Depth Information
conversion tools,
i.e.
algorithms that encode molecular structure information on
string codes such as molecular hash codes. Taking into consideration duplicate
information, using compound recognition, the files can be cleaned and
standardized. A common chemical format such as multiple sdf files can be used.
For the development of novel bioactive compounds, stereoisomers can play an
important role. Different enantiomeric forms can influence different roles. The
treatment using commercial compounds can be delicate. The information is not
often provided. This can lead to
in silico
recalculation of stereoisomeric forms
creating mirror images and identification of stereogenic centers.
Working with molecules that have at least three stereocenters can double the size
of database collections. The increased database size could be very dramatic if
more than eight centers are considered and most collections consider only four
centers. Some virtual screening programs can calculate these large numbers of
centers on the fly. Calculations of all forms are necessary for pharmacophere
screenings (when say, four sites are involved, stereomeric forms and chiralties
may be recognized).
The ligand in the binding site may show different protonation states than at
physiological pH. This information may be crucial in pharmacophore and virtual
screening/docking methods particularly when the information is poor regarding
the target at issue. Some databases give a range of pH at which the ligands are
often protonated.
It is well known that tautomeric forms can influence the binding of a ligand to a
protein. Different forms can yield different pharmacophere hypothesis in virtual
screening. The absence of this information can result in incorrect information on
compound recognition/binding.
Since the bioactive conformation is not necessarily the most stable conformation,
when using shape-based or pharmacophore screening, compound sampling in
databases can be important. This can be particularly true in virtual screening when
we can determine very closely the conformation that the ligand will assume in a
ligand-protein complex. It is not necessary to pre-calculate the conformations
