Registration Dossier

Administrative data

in vitro gene mutation study in bacteria
Type of genotoxicity: gene mutation
Type of information:
Adequacy of study:
supporting study
2 (reliable with restrictions)
Justification for type of information:
QSAR prediction: migrated from IUCLID 5.6

Data source

Reference Type:
study report
Report Date:

Materials and methods

Principles of method if other than guideline:
Gene mutation as microbial in vitro Salmonella was estimated by using four predictors: Leadscope, ACD/Percepta, Vega and Toxtree decision rule system.
GLP compliance:
Type of assay:
bacterial reverse mutation assay

Test material


Results and discussion

Any other information on results incl. tables

 Leadscope  ACD/Percepta  Vega  Toxtree  Consensus prediction


(moderate reliable)

 Out of the domain


(not reliable)


Not reliable


Moderate reliable

Leadscope FDA Model Applier prediction formicrobialin vitroSalmonellaresulted to be NEGATIVE, since the positive prediction probability was equal to 0.19. Since 4 features were found, it was concluded that 2-chloro-4-(1H-pyrazol-5-yl)benzonitrile is well represented by the model. Additionally, all the identified features are mainly represented in negative training compounds.The robustness of the prediction was further evaluated by examining compounds similar to the 2-chloro-4-(1H-pyrazol-5-yl)benzonitrile from the training set.While this information don’t take part to the prediction, it provides complementary means to see how similar compounds are predicted and what the experimental values of similar compounds are. Fourteen structures were identified in the training set as analogues to 2-chloro-4-(1H-pyrazol-5-yl)benzonitrile (similarity > 30%) and one of them, Dichlobenil, is characterized by similarity indices greater than 0.5.This training compound exhibits moderate similarity with respect to 2-chloro-4-(1H-pyrazol-5-yl)benzonitrile (similarity index equal to 0.54), meaning that the 2-chloro-4-(1H-pyrazol-5-yl)benzonitrile compound is moderately represented in the training set of the model, and experimental consistent negative Ames test results. Based on these considerations, Leadscopeprediction was assessed as moderate reliable.

Vega predicted 2-chloro-4-(1H-pyrazol-5-yl)benzonitrile as positive, however the prediction was assessed as not reliable since the applicability domain index was equal to 0.67. In factthe similar molecules found in the training set have experimental values that disagree with the predicted value. Thus, the prediction was considered not reliable.

Toxtree predicted the 2-chloro-4-(1H-pyrazol-5-yl)benzonitrile as NOT-mutagen since it didn’t identify any structural alert for genotoxicity. The prediction was assessed as not reliablesince theADI (global Applicability Domain Index) was equal to 0.0. In fact, only moderately similar compounds with known experimental value in the training set have been found; the similar molecules found in the training set have experimental values that agree/disagree with the predicted value; the accuracy of prediction for similar molecules found in the training set is not optimal, and some atom centered fragments of the compound have not been found in the compounds of the training set or are rare fragments.

Applicant's summary and conclusion

Interpretation of results (migrated information):
negative Moderate reliability

In the consensus assessment only reliable predictions are to be taken into account. In the case of 2-chloro-4-(1H-pyrazol-5-yl)benzonitrile, ACD/Percepta did not provide any prediction since the target was out of the model domain, while both Vega and Toxtree predictions resulted to be not reliable. Thus, based on Leadscope prediction, it was concluded that 2-chloro-4-(1H-pyrazol-5-yl)benzonitrile is predicted with a moderate level of confidence as NEGATIVE for microbial in vitro Salmonella.