3 Results and Discussion
We furnish in this section the results obtained using the method, described in the
previous section, that can generate chemical structures combinatorially using
activity-related substructural topological information, predict activity for the biological endpoints under consideration, prioritize compounds and screen them to
help discover novel therapeutic candidates. The results given here are for a series of
19 convulsant–anticonvulsant barbiturates [18], a series of 20 nucleoside analogues
(NA) having anti-HIV activities [20, 21] and a data set of 3779 compounds [22, 23]
for which minimum inhibitory concentration (MIC) values have been measured
against H37Rv strain of Mycobacterium tuberculosis (Mtb).
3.1 Activity Prediction–Compound Prioritization–Molecular
Design
We describe in this section the results obtained for combinatorial structure generation from the substructural information of activity-related vertices (atoms), activity
prediction using a rule-based system [18, 19] and prioritization and screening of
potential drug candidates using a newly defined Molecular Priority Score
(MPS) [15]. The application of different algorithms incorporated in the computer
program developed using the method, and the results obtained therefrom are given
here and discussed accordingly. In particular, the method has been used for activity
prediction, compound prioritization using MPS and structure generation considering
barbiturates and the NA series of compounds. On the other hand, structure matching
algorithm based on distance distribution has been used for searching potential antitubercular compounds from the data set of 3779 compounds mentioned above.
3.1.1 Studies with Barbiturates
The activity prediction for the series of barbiturates [18] considered for the present
study is reported here using the rule-based method [18, 19] considering
hydrogen-filled (H-filled) graphs of the compounds. Along with activity prediction
considering H-suppressed graphs, the method also supports activity prediction
using H-filled graphs and that option available in the computer program has been
used for the activity prediction studies with the barbiturates. The R-groups of the
barbiturates considered here and built on the core structure shown in Fig. 6 are
given in Table 1.
Activity prediction for this series of compounds has already been reported [18]
by considering information theoretical vertex indices V
d (vertex distance complexity) and V
d
n (normalized V
d ), which are also available in this software for use.
Although V
d
n has produced very high percentage of correct predictions [18], we
90
Md.I. H. Rizvi et al.
We furnish in this section the results obtained using the method, described in the
previous section, that can generate chemical structures combinatorially using
activity-related substructural topological information, predict activity for the biological endpoints under consideration, prioritize compounds and screen them to
help discover novel therapeutic candidates. The results given here are for a series of
19 convulsant–anticonvulsant barbiturates [18], a series of 20 nucleoside analogues
(NA) having anti-HIV activities [20, 21] and a data set of 3779 compounds [22, 23]
for which minimum inhibitory concentration (MIC) values have been measured
against H37Rv strain of Mycobacterium tuberculosis (Mtb).
3.1 Activity Prediction–Compound Prioritization–Molecular
Design
We describe in this section the results obtained for combinatorial structure generation from the substructural information of activity-related vertices (atoms), activity
prediction using a rule-based system [18, 19] and prioritization and screening of
potential drug candidates using a newly defined Molecular Priority Score
(MPS) [15]. The application of different algorithms incorporated in the computer
program developed using the method, and the results obtained therefrom are given
here and discussed accordingly. In particular, the method has been used for activity
prediction, compound prioritization using MPS and structure generation considering
barbiturates and the NA series of compounds. On the other hand, structure matching
algorithm based on distance distribution has been used for searching potential antitubercular compounds from the data set of 3779 compounds mentioned above.
3.1.1 Studies with Barbiturates
The activity prediction for the series of barbiturates [18] considered for the present
study is reported here using the rule-based method [18, 19] considering
hydrogen-filled (H-filled) graphs of the compounds. Along with activity prediction
considering H-suppressed graphs, the method also supports activity prediction
using H-filled graphs and that option available in the computer program has been
used for the activity prediction studies with the barbiturates. The R-groups of the
barbiturates considered here and built on the core structure shown in Fig. 6 are
given in Table 1.
Activity prediction for this series of compounds has already been reported [18]
by considering information theoretical vertex indices V
d (vertex distance complexity) and V
d
n (normalized V
d ), which are also available in this software for use.
Although V
d
n has produced very high percentage of correct predictions [18], we
90
Md.I. H. Rizvi et al.