Registration Dossier

Administrative data

Endpoint:
acute toxicity: oral
Type of information:
(Q)SAR
Adequacy of study:
weight of evidence
Reliability:
2 (reliable with restrictions)
Justification for type of information:
QSAR prediction
Cross-reference
Reason / purpose:
reference to same study

Data source

Reference
Reference Type:
other: estimation by QSAR
Title:
T.E.S.T. v 4.01
Author:
US EPA
Year:
2014
Bibliographic source:
http://www.epa.gov/nrmrl/std/qsar/qsar.html
Report Date:
2014

Materials and methods

Principles of method if other than guideline:
T.E.S.T. QSAR is developed by US EPA. The Toxicity Estimation Software Tool (T.E.S.T.) has been developed to allow users to easily estimate toxicity using a variety of QSAR methodologies. T.E.S.T allows a user to estimate toxicity without requiring any external programs.
GLP compliance:
no
Remarks:
other quality assurance

Test material

Reference
Name:
Unnamed
Type:
Constituent

Test animals

Species:
rat

Results and discussion

Effect levels
Dose descriptor:
LD50
Effect level:
>= 4 637 mg/kg bw

Any other information on results incl. tables

 

Endpoint

Predicted value – Consensus method

Predicted value -Hierarchical clustering method

Predicted value – FDA method

Predicted value – Nearest Neighbor Method

Oral rat LD50

-Log10 (mol/kg)

2.31

NA

NA

2.31

Oral rat LD50mg/kg

4636.71

NA

NA

4636.71

 Hierarchical method:The toxicity for a given query compound is estimated using the weighted average of the predictions from several different models. The different models are obtained by using Ward’s method to divide the training set into a series of structurally similar clusters. A genetic algorithm based technique is used to generate models for each cluster. The models are generated prior to runtime.

FDA method: The prediction for each test chemical is made using a new model that is fit to the chemicals that are most similar to the test compound. Each model is generated at runtime.

Nearest neighbor method: The predicted toxicity is estimated by taking an average of the 3 chemicals in the training set that are most similar to the test chemical.

Consensus method: The predicted toxicity is estimated by taking an average of the predicted toxicities from the above QSAR methods (provided the predictions are within the respective applicability domains)

Applicant's summary and conclusion

Interpretation of results:
GHS criteria not met
Remarks:
Criteria used for interpretation of results: expert judgment
Conclusions:
The T.E.S.T. QSAR estimated the oral rat LD50 to be around 4637 mg/kg. The substance is therefore not classified as acute toxic through oral route.
Executive summary:

QPRF:T.E.S.T. v 4.0.1(09092014)

 

1.

Substance

See “Test material identity”

2.

General information

 

2.1

Date of QPRF

See “Data Source (Reference)”

2.2

QPRF author and contact details

See “Data Source (Reference)”

3.

Prediction

3.1

Endpoint
(OECD Principle 1)

Endpoint

Oral rat LD50 (amount of chemical in mg/kg body weight that causes 50% of rats to die after oral ingestion)

3.2

Algorithm
(OECD Principle 2)

Model or submodel name

T.E.S.T.

Model version

v. 4.0.1

Predicted value (model result)

See “Results and discussion”

 

 

 

Input for prediction

- Chemical structure via SMILES code

Descriptor values

Different methods are used to derive the toxicity value (see result)

3.3

Applicability domain
(OECD principle 3)

Domains:

- Compounds can only contain the following element symbols: C, H, O, N, F, Cl, Br, I, S, P, Si, or As

- Compounds must represent a single pure component

- Substance contains only allowed elements

 

 

-Substance represents a single pure component

3.4

The uncertainty of the prediction
(OECD principle 4)

Statistical accuracy for training dataset:

n = 7420

Method

R2

 

K

RMSE

MAE

Coverage

Hierarchical

0.568

0.955

0.652

0.460

0.826

FDA

0.542

0.950

0.668

0.492

0.980

Nearest neighbor

0.519

0.954

0.677

0.491

0.994

Consensus

0.604

0.952

0.604

0.441

1

 

The consensus method achieved the best results in terms of both prediction accuracy and prediction coverage. The hierarchical method achieved slightly better prediction accuracy than the nearest neighbor and FDA methods but the prediction coverage was lower (83%).

3.5

The chemical mechanisms according to the model underpinning the predicted result
(OECD principle 5)

 

Hierarchical method:The toxicity for a given query compound is estimated using the weighted average of the predictions from several different models. The different models are obtained by using Ward’s method to divide the training set into a series of structurally similar clusters. A genetic algorithm based technique is used to generate models for each cluster. The models are generated prior to runtime.

FDA method: The prediction for each test chemical is made using a new model that is fit to the chemicals that are most similar to the test compound. Each model is generated at runtime.

Nearest neighbor method: The predicted toxicity is estimated by taking an average of the 3 chemicals in the training set that are most similar to the test chemical.

Consensus method: The predicted toxicity is estimated by taking an average of the predicted toxicities from the above QSAR methods (provided the predictions are within the respective applicability domains)

 

 

References

- User’s Guide for T.E.S.T. (version 4.0) (Toxicity Estimation Software Tool) – A program to estimate toxicity from molecular structure. US EPA