341
11 Computational Toxicology in Drug Discovery: Opportunities and Limitations
11.4 Development of New Medicines
Availability of side effects in drugs is one of the major problems in clinical practice.
Serious pathologies that affect the number of patients due to taking of some drugs
are the cause for rejection of using these drugs in the clinic. Approximately 35–40 %
failures of the drug application fall on these cases [37, 38, 39]. This leads to withdrawal of the drug from the market and increasing costs, since the usage of these
drugs results in severe injuries in various organs and tissues, which may cause the
patient death. The central nervous system, cardiovascular system and liver (9–26 %
of all cases of stopping treatment) are affected in the most commonly cases [40, 41].
Table 11.6 Comparison of software for the toxicity prediction
ACD/
Percepta
ADMET
predictor
CASE
Ultra
Meteor/
Derek
GUSAR PASS
StarDrop TOPKAT
Modification
−
−
+
+
+
+
+
−
Availability of training set
+
a
−
+
−
−
−
−
+
Method
description
+
+
+
+
+
+
+
+
Internal
validation
+
+
+
+
+
+
+
+
External
validation
+
+
−
+
a
+
+
+
+
a
Applicability domain
+
−
−
+
+
+
+
+
Number of
end-points
(ADME/T)
30/26
37/30
6/471
dozens/40 0/44
282/638 11/40
7/16
Source
ACD/
Labs
Simulations
Plus, Inc.
Multicase Inc.
Lhasa
Ltd.
GeneXplain
GeneXplain
Optibrium,
Ltd.
Accelerys
a
for some models
Table 11.7 Characteristics of GUSAR models for the rat acute toxicity predictions
Administration N train
N test
N models R
2
Q
2
R
2
test
RMSE test Coverage, %
Oral
6280
2692 40
0.61 0.57
0.59 0.57
97.5
Intraperitoneal 2480
1065 68
0.66 0.56
0.57 0.57
96.1
Intravenous
920
394 50
0.73 0.66
0.63 0.62
99.2
Subcutaneous
759
325
7
0.69 0.59
0.50 0.69
92.0
N train number of compounds in the training set, N test number of compounds in the test set, R
2
average R
2
of the models calculated for the appropriate training set, Q
2
average Q
2
of the models
calculated for the appropriate training set, Coverage percent of the compounds from the test set
fell in the Applicability Domain
11 Computational Toxicology in Drug Discovery: Opportunities and Limitations
11.4 Development of New Medicines
Availability of side effects in drugs is one of the major problems in clinical practice.
Serious pathologies that affect the number of patients due to taking of some drugs
are the cause for rejection of using these drugs in the clinic. Approximately 35–40 %
failures of the drug application fall on these cases [37, 38, 39]. This leads to withdrawal of the drug from the market and increasing costs, since the usage of these
drugs results in severe injuries in various organs and tissues, which may cause the
patient death. The central nervous system, cardiovascular system and liver (9–26 %
of all cases of stopping treatment) are affected in the most commonly cases [40, 41].
Table 11.6 Comparison of software for the toxicity prediction
ACD/
Percepta
ADMET
predictor
CASE
Ultra
Meteor/
Derek
GUSAR PASS
StarDrop TOPKAT
Modification
−
−
+
+
+
+
+
−
Availability of training set
+
a
−
+
−
−
−
−
+
Method
description
+
+
+
+
+
+
+
+
Internal
validation
+
+
+
+
+
+
+
+
External
validation
+
+
−
+
a
+
+
+
+
a
Applicability domain
+
−
−
+
+
+
+
+
Number of
end-points
(ADME/T)
30/26
37/30
6/471
dozens/40 0/44
282/638 11/40
7/16
Source
ACD/
Labs
Simulations
Plus, Inc.
Multicase Inc.
Lhasa
Ltd.
GeneXplain
GeneXplain
Optibrium,
Ltd.
Accelerys
a
for some models
Table 11.7 Characteristics of GUSAR models for the rat acute toxicity predictions
Administration N train
N test
N models R
2
Q
2
R
2
test
RMSE test Coverage, %
Oral
6280
2692 40
0.61 0.57
0.59 0.57
97.5
Intraperitoneal 2480
1065 68
0.66 0.56
0.57 0.57
96.1
Intravenous
920
394 50
0.73 0.66
0.63 0.62
99.2
Subcutaneous
759
325
7
0.69 0.59
0.50 0.69
92.0
N train number of compounds in the training set, N test number of compounds in the test set, R
2
average R
2
of the models calculated for the appropriate training set, Q
2
average Q
2
of the models
calculated for the appropriate training set, Coverage percent of the compounds from the test set
fell in the Applicability Domain