designed by this method which may be believed to carry higher importance for
discovering novel therapeutic candidates.
3.2 Rooted Substructure Searching for Drug Discovery
In the previous section, we showed how the exact matching algorithm can help find
structures of active compounds which could be obtained from the trees generated
from the topological distance distribution information of activity-related vertices
obtained from other active compounds. In this section, we describe the use of two
other matching algorithms—strong matching and weak matching—along with
exact matching algorithm for searching active compounds in a data set in the form
of tree and sub-tree matching. As given in the method section, these sub-trees are
obtained by means of applying node deviation and node migration in the actual tree
obtained from the distance distribution associated with an activity-related vertex.
The presence of such trees and sub-trees are then searched for in the compounds
present in a data set to identify potential drug candidates. In doing that, we have
considered two known TB drugs—Isoniazid and Streptomycin—to describe the
usefulness of the present method in finding potential antitubercular compounds
from a data set (named GTB data set) of 3779 compounds [22, 23] for which MIC
values against H37Rv strain of Mtb have been measured. The authors have made
MIC = 5.0 as the cut-off point and the MIC value of any compound which is higher
than 5.0 give an inactive compound in the data set. It therefore seems reasonable to
consider the same cut-off value for the present purpose. We will first furnish the
results obtained for Isoniazid which will be followed by those obtained for
Streptomycin. It may be noted that the activity-related vertices for both Isoniazid
and Streptomycin have been taken from the literature information and not by using
rule-based method in the ordering of vertex indices which has been done for the
barbiturate and NA series of compounds. In fact, it shows that the method can be
used successfully in identifying potential drug candidates by picking
activity-related vertices by other means than by the rule-based method.
3.2.1 Studies with Isoniazid
Isoniazid is a known first line drug for the treatment of tuberculosis. However, it
may become resistant in situations, and therefore, this leads researchers look for
novel drug candidates to overcome drug resistance problem for the treatment of
tuberculosis . We have described in this subsection how structures generated from
activity-related vertex information of Isoniazid using the present method can help
search for potential TB drugs from a data set of 3779 compounds [22, 23]. It is
known that the chemical/biochemical reaction takes place at the point of the first
98
Md.I. H. Rizvi et al.
discovering novel therapeutic candidates.
3.2 Rooted Substructure Searching for Drug Discovery
In the previous section, we showed how the exact matching algorithm can help find
structures of active compounds which could be obtained from the trees generated
from the topological distance distribution information of activity-related vertices
obtained from other active compounds. In this section, we describe the use of two
other matching algorithms—strong matching and weak matching—along with
exact matching algorithm for searching active compounds in a data set in the form
of tree and sub-tree matching. As given in the method section, these sub-trees are
obtained by means of applying node deviation and node migration in the actual tree
obtained from the distance distribution associated with an activity-related vertex.
The presence of such trees and sub-trees are then searched for in the compounds
present in a data set to identify potential drug candidates. In doing that, we have
considered two known TB drugs—Isoniazid and Streptomycin—to describe the
usefulness of the present method in finding potential antitubercular compounds
from a data set (named GTB data set) of 3779 compounds [22, 23] for which MIC
values against H37Rv strain of Mtb have been measured. The authors have made
MIC = 5.0 as the cut-off point and the MIC value of any compound which is higher
than 5.0 give an inactive compound in the data set. It therefore seems reasonable to
consider the same cut-off value for the present purpose. We will first furnish the
results obtained for Isoniazid which will be followed by those obtained for
Streptomycin. It may be noted that the activity-related vertices for both Isoniazid
and Streptomycin have been taken from the literature information and not by using
rule-based method in the ordering of vertex indices which has been done for the
barbiturate and NA series of compounds. In fact, it shows that the method can be
used successfully in identifying potential drug candidates by picking
activity-related vertices by other means than by the rule-based method.
3.2.1 Studies with Isoniazid
Isoniazid is a known first line drug for the treatment of tuberculosis. However, it
may become resistant in situations, and therefore, this leads researchers look for
novel drug candidates to overcome drug resistance problem for the treatment of
tuberculosis . We have described in this subsection how structures generated from
activity-related vertex information of Isoniazid using the present method can help
search for potential TB drugs from a data set of 3779 compounds [22, 23]. It is
known that the chemical/biochemical reaction takes place at the point of the first
98
Md.I. H. Rizvi et al.