CHEMSPIDER: A PLATFORM FOR CROWDSOURCED COLLABORATION TO CURATE DATA DERIVED FROM PUBLIC COMPOUND DATABASES - Collaborative Computational Technologies for Biomedical Research - page 370

4. Chemical structures associated with a particular asserted label can have
timelines. An originally reported chemical structure for some newly
extracted material may be assigned a particular label and enter the his-
torical literature archive. Some period of time later the same compound
may be freshly elucidated with newer experimental data and new struc-
tural details identifi ed. The same asserted chemical name will now be
associated with a new structure. This process can occur many times with
the result that a single chemical may have a multitude of associated
structures. One particular example is hexacyclinol, a natural product that
generated a signifi cant amount of blog discussion and resulted in two
structures forever being associated with that chemical name [43 - 46] .
As an example of the challenges of locating the “correct” chemical structure
for what should be a well-known and easily locatable chemical compound, we
will initiate a search for a well-known vitamin, vitamin K 1 , commonly known
as phylloquinone. A Google search will direct us to a number of resources and
databases utilized by life scientists, and these include Wikipedia [18], PubChem
[3], DrugBank [14], Chemical Entities of Biological Interest (ChEBI) [47],
DailyMed [48], and ChemSpider [31], to name just a few. A review of the data
concisely demonstrates the confusion that can exist online and the quality of
available data. Figure 22.1 shows the images of the structures of vitamin K 1
extracted from a number of these databases. PubChem alone lists 10 different
structures under the name vitamin K 1 . It should be noted that there are dif-
ferences in the structures shown, specifi cally in the stereochemistry and the
E / Z orientation of the alkene bond in the phytyl side chain.
Some observations from Figure 22.1 include (1) the Wikipedia article and
the Kyoto Encyclopedia of Genes and Genome (KEGG) database record
contain no explicit stereochemistry, (2) DrugBank has ambiguous orientation
around the alkene bond, and (3) ChEBI and ChemSpider are consistent with
Figure 22.1 Images of chemical structures of vitamin K 1 extracted from series of
databases labeled with name of associated database. The asserted structures of vitamin
K 1 are surrounded by a bolded box and are consistent with those from the Merck Index
and Common Chemistry .