Chemistry Reference
In-Depth Information
compounds from fragments may help identify hits that possess enough structural
novelty to warrant patentability.
It is of interest for workers to have available chemical information of compounds
that can be synthetized from commercial building blocks as well as purchasable
compounds (including websites of academia and industrial libraries). In order to
perform virtual screening most of collections require some level of 'curation',
Depending on the virtual screening routine, collections with 2D sdf files for
example, need to undergo 'curation' by the users. We note that raw formats
permit rapid handling of large chemical data. Most commercial vendors update
regularly their chemical collections.
Applications of computational procedures to process chemical data and include
required information for virtual screening entails database 'curation' steps. This
provides standardization protocols for same chemical information/record. It is,
however, a formidable task since collections can involve millions of compounds
and substantial information with differing formats involved. Hardware/software
may be required to manage/store/share the large data influencing setups/potential
usage and molecular properties (including molecular descriptors, annotations).
3D generators, not conformation sampling, should be standardized. It may also be
necessary to remove salts as well as co-crystalized small molecules that are not a
part of the inhibition mechanism and may mislead computational results.
Another process is filtering to obtain selected range of pharmacokinetic and
physicochemical properties and reduce clinic attrition of drug candidates, using
for example Lipinsky's Rule of Five. On the other hand, unfiltered databases may
yield useful scaffold information on families of ligands that interact well with
specific targets and may be useful for improvement of pharmacokinetic profiles
and binding affinities.
Another problem is format standardization and redundancy in merging databases
from different chemical sources. Non-redundant repository and non-ambiguous
identification of chemical entities is important (using elimination of superfluous
information and compound recognition during the storage process). We can use
