nitrogen (N) atom (underlined) of the fragment (–NH–NH2) in isoniazid molecule
to convert this pro-drug into its metabolite that works as the effector molecule.
Therefore, this vertex (N atom) may be regarded as an activity-related vertex for
Isoniazid. Accordingly, the distance distribution associated with the vertex representing this nitrogen (N) atom has been considered for generating structures. In
order to screen out potential antitubercular compounds having high activities, the
exact, strong and weak matching algorithms (method section) have been applied on
the GTB data set of 3779 compounds considered for the present study. A number of
highly active compounds have been obtained in the process and the information for
some of them obtained applying different node deviation and node migration on the
tree obtained from the distance distribution associated with the root vertex are
shown in Table 6 along with the structures of Isoniazid (with root vertex specified)
and the screened compounds. As said earlier, in their studies [22], the researchers
have considered a compound having MIC value less than 5.0 to be active. In this
way, data set is composed of almost equal number of active and inactive compounds implying no bias for active or inactive compounds in forming the data set.
Accordingly, compound nos. 1–1890 are active compounds and the other compounds are inactive. Considering the same cut-off value, one can see that only
compound no. 3296 has MIC value higher than 5.0 and the rest of the compounds
may be screened out as potential active compounds. In particular, compound no.
180 which is obtained by two types of node deviation and node migration in
generating structures from the root vertex has quite low MIC value which identifies
it as a highly active compound. Therefore, the result clearly shows that the method
may be used to successfully screen potentially highly active antitubercular compounds from this data set starting from Isoniazid.
3.2.2 Studies with Streptomycin
Streptomycin is another antitubercular drug in use, an antibiotic. For this compound, the removal of even one of the two guanidino groups present in the structure
reduces the activity of the compound. Considering that, we have taken the vertex
representing the nitrogen (N) atom in one of the guanidino groups as the root vertex
to start generating/designing novel structures. Out of a number of structures
designed using the present method, i.e., using exact matching as well as strong
matching and weak matching algorithms in relation to node deviation and node
migration on the trees obtained from the distance distribution associated with the
root vertex, information about some of these compounds are given in Table 7 along
with the structures of Streptomycin having root vertex indicated and the matched/
searched compounds from GTB data set. It is found from this table that all the
compounds shown here are active according to the adopted criterion (MIC
5.0 is
active) with compound no. 183 being the most active among them. Therefore, it
appears from this finding that the method may be used successfully to screen
potentially highly active antitubercular compounds from the data set of 3779
compounds starting from Streptomycin.
Combinatorial Drug Discovery from Activity-Related Substructure …
99
to convert this pro-drug into its metabolite that works as the effector molecule.
Therefore, this vertex (N atom) may be regarded as an activity-related vertex for
Isoniazid. Accordingly, the distance distribution associated with the vertex representing this nitrogen (N) atom has been considered for generating structures. In
order to screen out potential antitubercular compounds having high activities, the
exact, strong and weak matching algorithms (method section) have been applied on
the GTB data set of 3779 compounds considered for the present study. A number of
highly active compounds have been obtained in the process and the information for
some of them obtained applying different node deviation and node migration on the
tree obtained from the distance distribution associated with the root vertex are
shown in Table 6 along with the structures of Isoniazid (with root vertex specified)
and the screened compounds. As said earlier, in their studies [22], the researchers
have considered a compound having MIC value less than 5.0 to be active. In this
way, data set is composed of almost equal number of active and inactive compounds implying no bias for active or inactive compounds in forming the data set.
Accordingly, compound nos. 1–1890 are active compounds and the other compounds are inactive. Considering the same cut-off value, one can see that only
compound no. 3296 has MIC value higher than 5.0 and the rest of the compounds
may be screened out as potential active compounds. In particular, compound no.
180 which is obtained by two types of node deviation and node migration in
generating structures from the root vertex has quite low MIC value which identifies
it as a highly active compound. Therefore, the result clearly shows that the method
may be used to successfully screen potentially highly active antitubercular compounds from this data set starting from Isoniazid.
3.2.2 Studies with Streptomycin
Streptomycin is another antitubercular drug in use, an antibiotic. For this compound, the removal of even one of the two guanidino groups present in the structure
reduces the activity of the compound. Considering that, we have taken the vertex
representing the nitrogen (N) atom in one of the guanidino groups as the root vertex
to start generating/designing novel structures. Out of a number of structures
designed using the present method, i.e., using exact matching as well as strong
matching and weak matching algorithms in relation to node deviation and node
migration on the trees obtained from the distance distribution associated with the
root vertex, information about some of these compounds are given in Table 7 along
with the structures of Streptomycin having root vertex indicated and the matched/
searched compounds from GTB data set. It is found from this table that all the
compounds shown here are active according to the adopted criterion (MIC
5.0 is
active) with compound no. 183 being the most active among them. Therefore, it
appears from this finding that the method may be used successfully to screen
potentially highly active antitubercular compounds from the data set of 3779
compounds starting from Streptomycin.
Combinatorial Drug Discovery from Activity-Related Substructure …
99