402
P. M. Vassiliev et al.
2. β-Adrenoblocking activity and four pharmacokinetic indices of twelve S- and
R-isomers of six known structurally similar β-adrenoblockers, 1-aryloxy-3-alkylamino-propane-2-ole derivatives (acebutolol, atenolol, carvedilol, metoprolol, pindolol, sotalol) [60]. The compounds were classified into two classes
based on each parameter; 1) the enantiomer activity was higher than that of a
racemate, or 2) the enantiomer activity was lower than that of a racemate.
3. The values for the β 1 -adrenoblocking activity of ten optical isomers of a single drug, nebivolol, which has four chiral centers [77], were divided into two
classes: highly active ones and those with a low activity.
The results of testing the best decision rules are shown in Table 12.11. The normal
strategy was the best in all of these cases, which corresponds to the characteristics
of chiral compounds; the basal activity level of chiral compounds is established by
the chemical structure, and specific effects are determined by chiral structural fragments.
In summary, IT Microcosm allows an accuracy approaching 100 % in predicting
the presence and extent of a pharmacological activity in structurally diverse and
structurally similar compounds and their stereoisomers.
12.4 IT Microcosm in the Search for Novel Drugs
The fundamental nature of the theoretical concepts behind IT Microcosm makes
it possible to use this technology to successfully solve various QSAR/QSPR optimization tasks, such as searching for novel compounds with high pharmacological
Table 12.9 General indices of prediction accuracy for high activity in structurally similar condensed azole derivatives
Strategy
ST, %
LOOCV,
%
SHCV, %
DLOOCV,
%
Min
Max
Min
Max Min
Max Min
Max
Accuracy F 0
Conservative 89
100
64
93
64
90
–
–
Normal
73
100
67
90
64
89
67
90
Risk
71
100
64
90
66
91
–
–
Sensitivity F a
Conservative 91
100
64
100
61
96
–
–
Normal
79
100
64
86
63
91
64
86
Risk
71
100
60
92
60
100
–
–
Specificity F n
Conservative 85
100
60
97
66
95
–
–
Normal
69
100
64
93
61
92
64
93
Risk
65
100
61
94
68
93
–
–
DLOOCV is not used in the conservative or risk strategy
P. M. Vassiliev et al.
2. β-Adrenoblocking activity and four pharmacokinetic indices of twelve S- and
R-isomers of six known structurally similar β-adrenoblockers, 1-aryloxy-3-alkylamino-propane-2-ole derivatives (acebutolol, atenolol, carvedilol, metoprolol, pindolol, sotalol) [60]. The compounds were classified into two classes
based on each parameter; 1) the enantiomer activity was higher than that of a
racemate, or 2) the enantiomer activity was lower than that of a racemate.
3. The values for the β 1 -adrenoblocking activity of ten optical isomers of a single drug, nebivolol, which has four chiral centers [77], were divided into two
classes: highly active ones and those with a low activity.
The results of testing the best decision rules are shown in Table 12.11. The normal
strategy was the best in all of these cases, which corresponds to the characteristics
of chiral compounds; the basal activity level of chiral compounds is established by
the chemical structure, and specific effects are determined by chiral structural fragments.
In summary, IT Microcosm allows an accuracy approaching 100 % in predicting
the presence and extent of a pharmacological activity in structurally diverse and
structurally similar compounds and their stereoisomers.
12.4 IT Microcosm in the Search for Novel Drugs
The fundamental nature of the theoretical concepts behind IT Microcosm makes
it possible to use this technology to successfully solve various QSAR/QSPR optimization tasks, such as searching for novel compounds with high pharmacological
Table 12.9 General indices of prediction accuracy for high activity in structurally similar condensed azole derivatives
Strategy
ST, %
LOOCV,
%
SHCV, %
DLOOCV,
%
Min
Max
Min
Max Min
Max Min
Max
Accuracy F 0
Conservative 89
100
64
93
64
90
–
–
Normal
73
100
67
90
64
89
67
90
Risk
71
100
64
90
66
91
–
–
Sensitivity F a
Conservative 91
100
64
100
61
96
–
–
Normal
79
100
64
86
63
91
64
86
Risk
71
100
60
92
60
100
–
–
Specificity F n
Conservative 85
100
60
97
66
95
–
–
Normal
69
100
64
93
61
92
64
93
Risk
65
100
61
94
68
93
–
–
DLOOCV is not used in the conservative or risk strategy