5′-O-(4,4′-dimethoxytrityl)-2′-deoxythymidine-3′-O-[O-(2-cyanoethyl)-N,N′-diisopropylphosphoramidite]

Skin sensitisation

Administrative data

Endpoint:

skin sensitisation, other

Remarks:

in silico

Type of information:

(Q)SAR

Adequacy of study:

supporting study

Study period:

2019-07-01

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

results derived from a valid (Q)SAR model, but not (completely) falling into its applicability domain, with adequate and reliable documentation / justification

Justification for type of information:

1a. SOFTWARE
VEGA QSAR Models

2a. MODEL (incl. version number)
Skin Sensitization model (CAESAR) 2.1.6
Core version: 1.2.8

3a. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL
Compound SMILES: N#CCCOP(OC4CC(OC4(COC(c1ccccc1)(c2ccc(OC)cc2)c3ccc(OC)cc3))N5C=C(C(=O)NC5(=O))C)N(C(C)C)C(C)C

4a. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL
[Explain how the model fulfils the OECD principles for (Q)SAR model validation. Consider attaching the QMRF or providing a link]
- Defined endpoint: skin sensitization activity

5a. APPLICABILITY DOMAIN
Global AD Index
AD index = 0
Explanation: the predicted compound is outside the Applicability Domain of the model.

Similar molecules with known experimental value
Similarity index = 0.644
Explanation: only moderately similar compounds with known experimental value in the training set have been found.

Accuracy of prediction for similar molecules
Accuracy index = 1
Explanation: accuracy of prediction for similar molecules found in the training set is good.

Concordance for similar molecules
Concordance index = 0.492
Explanation: similar molecules found in the training set have experimental values that disagree with the predicted value.

Model's descriptors range check
Descriptors range check = False
Explanation: 3 descriptors for this compound have values outside the descriptor range of the compounds of the training set.

Atom Centered Fragments similarity check
ACF index= 0.28
Explanation: a prominent number of atom centered fragments of the compound have not been found in the compounds of the training set or are rare fragments (7 unknown fragments and 4 infrequent fragments found).

1b. SOFTWARE
VEGA QSAR Models (Core Version 1.2.8)

2b. MODEL (incl. version number)
Skin Sensitization model IRFMN/JRC 1.0.0

3b. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL
Compound SMILES: N#CCCOP(OC4CC(OC4(COC(c1ccccc1)(c2ccc(OC)cc2)c3ccc(OC)cc3))N5C=C(C(=O)NC5(=O))C)N(C(C)C)C(C)C

4b. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL
[Explain how the model fulfils the OECD principles for (Q)SAR model validation. Consider attaching the QMRF or providing a link]
- Defined endpoint: skin sensitization activity

5b. APPLICABILITY DOMAIN
Global AD Index
AD index = 0.24
Explanation: the predicted compound is outside the Applicability Domain of the model.

Similar molecules with known experimental value
Similarity index = 0.68
Explanation: only moderately similar compounds with known experimental value in the training set have been found.

Accuracy of prediction for similar molecules
Accuracy index = 1
Explanation: accuracy of prediction for similar molecules found in the training set is good.

Concordance for similar molecules
Concordance index = 0.537
Explanation: similar molecules found in the training set have experimental values that disagree with the predicted value.

Model's descriptors range check
Descriptors range check = True
Explanation: descriptors for this compound have values inside the descriptor range of the compounds of the training set.

Atom Centered Fragments similarity check
ACF index= 0.34
Explanation: a prominent number of atom centered fragments of the compound have not been found in the compounds of the training set or are rare fragments (5 unknown fragments and 1 infrequent fragments found).

Data source

Reference

Reference Type:

other: QSAR prediction

Title:

Report Prediction and Applicability Domain analysis for models: Skin Sensitization model (CAESAR) 2.1.6 Core version: 1.2.8; Skin Sensitization model (IRFMN/JRC) 1.0.0

Author:

Rücker T.

Year:

2019

Report date:

2019

Applicant's summary and conclusion

Interpretation of results:

other: equivocal results

Conclusions:

The test item is predicted on the one hand to be a non-sensitizer in the VEGA QSAR Model CAESAR 2.6.1 and on the other hand predicted to be a sensitizer in the VEGA QSAR Model IRFMN/JRC 1.0.0. Based on these results, no clear prediction can be derived from the QSAR modelling and the results need to be evaluated as equivocal.

Executive summary:

The test item was predicted as NON-Sensitizer in the VEGA QSAR model CAESAR 2.6.1. Even though there were only moderately similar compounds with known experimental value in the training set (Similarity index = 0.644), and similar molecules found in the training set had experimental values that disagree with the predicted value, and the descriptors for this compound had values outside the descriptor range of the compounds of the training set, the accuracy of prediction for similar molecules found in the training set was considered good.

In the VEGA QSAR model IRFMN/JRC 1.0.0 the test item was predicted as Sensitizer. Even though there were only moderately similar compounds with known experimental value in the training set (Similarity index = 0.68) and similar molecules were found in the training set which had experimental values that disagree with the predicted value, descriptors for this compound had values inside the descriptor range of the compounds of the training set and the accuracy of prediction for similar molecules found in the training set was considered good.

Overall, the QSAR prediction does not allow to draw a conclusion on the sensitizing potential of the test item and the prediction from the both models need to be evaluated as equivocal.