408
P. M. Vassiliev et al.
The results of the computational-accuracy testing for the prediction of the extent of
the antiarrhythmic activity of condensed azole derivatives are shown in Table 12.13.
Not a single adequate decision rule was obtained for the moderate activity class.
According to the predicted results, 25 substances were selected out of 752 novel,
untested condensed azole derivatives for an experimental study. The selected substances showed 3A for a high activity level in all three strategies, a conformity coefficient of prediction estimates spectrum (21) K High = 1.0 and a membership function
in the conservative strategy (17) Fb High > 0.65.
Experimentally, all of the 25 compounds (100 %) showed an expressed antiarrhythmic activity MEC < 2.8·10
-4
M; 20 of them (80.0 %) showed a high activity
level MEC < 7.3·10
-5
M. Of 20 highly active substances, 11 (55.0 %) are comparable with Moricizine in the extent of the antiarrhythmic activity; nine substances
(45.0 %) showed a higher extent of activity.
Previously, the primary screening of condensed azole derivatives with a high
antiarrhythmic activity employed, apart from intuitive prediction, the Hansch methTable 12.13  Indices of the prediction accuracy for the antiarrhythmic activity level of condensed
azole derivatives
Activity
level
ST, %
LOOCV, %
SHCV, %
F 0
F a
F n F u F 0
F a
F n
F u F 0
F a
F n
F u
CS
High
98
97
98 12 87
72
91
14 89
78
92
14
Moderate
98
97
99 18 72
40
81
21 74
49
81
18
High or
moderate
94
94
94 13 75
77
73
15 78
81
75
16
NS
High,
BA
81
82
80 –
78
79
78
–
79
82
78
–
Moderate, BA
66
91
58 –
56
67
52
–
59
61
58
–
High or
moderate,
LDM
90
90
91 –
74
74
75
–
75
76
75
–
RS
High,
DM11
83
82
84 –
81
80
81
–
80
80
80
–
Moderate,
BA11
73
91
68 –
59
59
59
–
66
66
66
–
High or
moderate,
DM11
79
78
81 –
72
71
73
–
72
72
72
–
The designations are the same as in Table 12.12