Registration Dossier

Ecotoxicological information

Short-term toxicity to fish

Currently viewing:

Administrative data

Link to relevant study record(s)

Reference
Endpoint:
short-term toxicity to fish
Type of information:
(Q)SAR
Adequacy of study:
key study
Study period:
April 6th 2020
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
results derived from a valid (Q)SAR model and falling into its applicability domain, with adequate and reliable documentation / justification
Justification for type of information:
1. SOFTWARE
iSafeRat® HA-QSAR toolbox v2.5

2. MODEL (incl. version number)
iSafeRat® fishLC50 (v1.9)

3. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL
C=CCCCCCCCCC(=O)NCC(=O)O

4. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL
More details in QPRF and QMRF documents attached to section "attached justification"

- Defined endpoint: Ecotoxic effects: Short-term toxicity to fish (lethality - 96 hour - LC50) (in mg/L)
- Unambiguous algorithm: See QMRF and QPRF documents attached to section "attached justification"
- Defined domain of applicability: See QMRF and QPRF documents attached to section "attached justification"

- Appropriate measures of goodness-of-fit and robustness and predictivity: 95% confidence interval (α = 0.05): 75 – 231 mg/L. Statistical characteristics of the model are given in the QMRF report KTS/QMRF/HOL/09
- Mechanistic interpretation: The toxicity of MechoA 1.1 compounds can be linked
thermodynamically to their water solubility values. The toxicity of MechoA 5.2 is also described with water solubility even if this toxicity is not well correlated with the hydrophobicity. The model still accurately predicts the toxicity because the toxicity expressed in mol/L is almost a constant over the whole training and validation sets, related to the overwhelming of the cell capacities to maintain the right pH.

5. APPLICABILITY DOMAIN
More details in QPRF and QMRF documents attached to section "attached justification"
- Descriptor domain: The Subcooled Liquid Water Solubility value (481 mg/L or -2.701 in log10 (mol/L)) given as the input to the iSafeRat® fishLC50 model falls within the descriptor domain of the model between a Subcooled Liquid Water Solubility of -3.55 to 2.00 in log10 (mol/L).
- Structural and mechanistic domains:
Structural fragment domain: All chemical groups within the molecular structure are taken into account by the model.
Mechanistic domain: The MechoA of molecules is predicted directly from the structure. The test item as an amino acid derivative is expected to exert a MechoA 5.2 (Proton release of carboxylic acids) and can be taken into account by the model (Bauer et al., 2018).
- Similarity with analogues in the training set: Kréatis Proprietary
- Other considerations (as appropriate): This model was constructed using for 7 substances with the same MechoA as the test item.

6. ADEQUACY OF THE RESULT
More details in QPRF and PMRF documents attached to section "attached justification".
The model provides 96h-LC50 predictions for fish from High Accuracy QSAR models which are based on ecotoxicologically validated studies and statistically validated regressions for a series of Mechanisms of Action described within the QMRF. As such the results from these studies are
considered accurate enough to replace an experimental study when the test substance falls within the appropriate applicability domain.
The test item falls within the applicability domain of the model and can therefore be considered a
reliably prediction for ACUTE TOXICITY TO FISH (96-HOUR LC50). Therefore, this endpoint value can be considered valid for use in risk assessment and classification and labelling.
Qualifier:
according to
Guideline:
other: OECD (2004). Principles for the validation, for regulatory purposes, of (Quantitative) Structure Activity-Relationship Models.
GLP compliance:
no
Analytical monitoring:
not required
Test type:
not specified
Water media type:
not specified
Limit test:
no
Key result
Duration:
96 h
Dose descriptor:
LC50
Effect conc.:
ca. 131 mg/L
Nominal / measured:
meas. (geom. mean)
Conc. based on:
test mat.
Basis for effect:
mortality
Details on results:
The result below is the toxicity value anticipated during a 96-hour LC50 study on fish based on measured concentrations:
96h-LC50 (mg test item.L-1) 95% confidence limits unit
-3.264 -3.508 to -3.019 in log (mol test item.L-1)
5.45E-04 3.10E-04 to 9.57E-04 in mol test item.L-1
131* 75 to 231 mg test item.L-1
*The toxicity value is greater than the limit of solubility within the exposure period of the test.

Evaluation and statistics
The parameters of the model used in this in silico study are as following:
· Local model used: proton release of carboxylic acids (MechoA 5.2)
· N (training set) = 11 chemicals
· R² = 0.4829
· RMSE = 0.2270
· N (validation set) = 5 chemicals
· Q² = 0.2674
· RMSEP = 0.2650


Further statistical evidence of high prediction accuracy are provided in the QMRF (KREATiS, 2020).


The correlation coefficient R2 and Q² values for the QSAR model were respectively 0.4829 and 0.2674. In this case the correlation coefficients R2 and Q² are not relevant since the aquatic toxicity of carboxylic acids is almost constant.
However the other parameters relative to error like Root Mean Standard Errors (RMSE) and Root Mean Standard Errors of Predictivity (RMSEP), respectively 0.2270 and 0.2650, show the good accuracy of the predictions.
Therefore, the model was still considered as reliable for providing accurate predictions falling within its applicability domain.


 


 

Validity criteria fulfilled:
not applicable
Conclusions:
The QSAR model used to achieve the study has been fully validated following the OECD recommendations (OECD,2004). The test item falls within the applicability domain of the model and was therefore reliably predicted for its ACUTE TOXICITY TO FISH (96-HOUR LC50). Therefore, this endpoint value can be considered valid for use in risk assessment and classification and labelling.

The ACUTE TOXICITY TO FISH (96-HOUR LC50) of the test item was predicted as 131 mg.L-1. Therefore, the toxicity value is greater than the limit of solubility within the exposure period of the test.

95% confidence interval (α = 0.05): 75 – 231 mg.L-1.
Executive summary:

Introduction. A Quantitative Structure Activity Realtionship (QSAR) model was used to calculate the ACUTE TOXICITY TO FISH (96-HOUR LC50) for the test item. This QSAR model has been validated to be compliant with the OECD recommendations for QSAR modeling (OECD, 2004) and predicts the endpoint value which would be expected when testing the substance under experimental conditions in a laboratory following the Guideline for Testing of Chemicals No. 203, "Fish Acute Toxicity Test" (OECD, 2019), referenced as Method C.1 of Commission Regulation No. 440/2008 (European Commission, 2008). The criterion predicted was the LC50 (Median Lethal Concentration), a statistically derived concentration which is expected to cause mortality in 50% of test animals within a period of 96 hours.



Methods. The ACUTE TOXICITY TO FISH (96-HOUR LC50) was determined using the iSafeRat® fishLC50, a validated QSAR model for the Mechanism of Action (MechoA) in question (MechoA 5.2, i.e. proton release of carboxylic acids) (Bauer et al., 2018). The QSAR is based on validated data for a training set of 7 chemicals derived from 96-hour test on fish, for which the concentrations of the test item had been determined by chemical analyses over the test period.



Results. The result below is the toxicity value anticipated during a 96-hour LC50 study on fish based on measured concentrations.















Time (h) LC50 (mg test item.L-1) 95% confidence limits (mg test item.L-1)
96  131*75 – 231

*The toxicity value is greater than the limit of solubility within the exposure period of the test.

Description of key information

A preliminary acute toxicity to fish (OECD 203) study was performed with the registered substance showing a LC50 higher than 79 mg/L (based on initial concentrations). Due to technical difficulties encountered (the variation of the test item concentration levels was higher than 20%) the main test was not carried out and a QSAR study was conducted.


The quantitative Structure Activity Relationship (QSAR) model was used to calculate the acute toxicity to fish (96h LC50) for the registered substance. The acute toxicity to fish (96h LC50) was determined using the iSafeRat® fishLC50, a validated QSAR model for the Mechanism of Action (MechoA) in question (MechoA 5.2, i.e. proton release of carboxylic acids) (Bauer et al., 2018). The QSAR is based on validated data for a training set of 7 chemicals derived from 96-hour test on fish, for which the concentrations of the test item had been determined by chemical analyses over the test period. The result is the toxicity value anticipated during a 96h LC50 study on fish based on measured concentrations: LC50 ) 131 mg/L.

Key value for chemical safety assessment

LC50 for freshwater fish:
131 mg/L

Additional information