Reaction product of Fatty acids, C18-unstaturated, dimer with 4,7,10-trioxa-1,13-tridecane-diamine

Bioaccumulation: aquatic / sediment

Link to relevant study record(s)

Reference

Reference 1

Endpoint:

bioaccumulation in aquatic species: fish

Type of information:

(Q)SAR

Adequacy of study:

key study

Study period:

25 June 2020

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:

1. SOFTWARE : Episuite 4.11
2. MODEL (incl. version number) : BCFBAF v.3.01
3. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL : see below
4. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL
- Defined endpoint: Bioconcentration factor (BCF)
- Unambiguous algorithm:
BCFBAF contains four methods for determining BCF depending on log Kow and substance ionization. The predicted value of the log Kow for neutral substances with log Kow >7 is calculated through the following equation:
log BCF = -0.49 x log Kow + 7.554 + C
where log Kow is the log of the octanol-water partition coefficient and C is sum of any structure-specific correction factors
- Defined domain of applicability:
Training set parameters (n = 35 for subset of neutral substances with log Kow > 7)
molecular weight range 226.45 D to 959.17 D
log Kow range 7 to 11.59
No amides are present in the subset of the training set comprising neutral substances having log Kow >7, although several hydrocarbons and aromatic esters are represented. A correction for eight or more contiguous methylene groups is applicable only between log Kow seven and twelve. No specific factor is applied for ester groups. While amides are present in the subset of neutral substances with log Kow less than seven, this also did not result in a specific correction factor. This suggests that carbonyl moieties are little different from other substructures as regards the relationship between log Kow and BCF, although a conclusive statement cannot be made given the limited training/validation data available. The log Kow >7 validation subset consists mostly of halogenated aromatic hydrocarbons, although trioctyl amine (CAS# 1116-76-3) is present.
- Appropriate measures of goodness-of-fit and robustness and predictivity:
From the plot of predicted and experimental log BCF values for the training set (527 compounds), the correlation coefficient is 0.833, the standard deviation is 0.502, and the absolute mean error is 0.382. For the subset of training set chemicals with log Kow >7 (n = 35), the correlation coefficient is 0.627, the standard deviation is 0.616, and the absolute mean error is 0.460. From the plot of predicted and experimental log BCF values for the validation set (158 compounds), the correlation coefficient is 0.82, the standard deviation is 0.59, and the absolute mean error is 0.46.
- Mechanistic interpretation:
The model is parameterized empirically. The basic mechanistic assumption is that bioconcentration is driven by hydrophobicity as represented using log Kow. A limited number of structural features were identified manually based on common occurrence in outlying data points, and the quality of the fit improved by deriving correction factors for these structural features by multiple regression. For non-ionic substancess, bioconcentration dependence on hydrophobicity is normally distributed with a maximum at log Kow = 7, while bioconcentration of ionic materials has less dependence on log Kow and can be adequate described by a grouping approach.

5. APPLICABILITY DOMAIN
With respect to the modelled dimers, a subset of 35 of the training set substances and 11 validation set substances are neutral molecules with log Kow greater than seven.
The parametric domain of this subset includes chemicals with molecular weight of 226.45 D to 959.17 D and log Kow of 7 to 11.59. The molecular weight and predicted log Kow for the modelled dimers is slightly outside the parametric range of the model. However, because the difference is not great and materials of this size and hydrophobicity are not expected to bioaccumulate, the result is considered acceptable.
No amides are present in this log Kow >7 subset of the training set, although several hydrocarbons and aromatic esters are represented. A correction for eight or more contiguous methylene groups is applicable only between log Kow values from seven and twelve. No specific factor is applied for ester groups. While amides are present in the subset of neutral substances with log Kow less than seven, this also did not result in a specific correction factor. This suggests that carbonyl moieties are little different from other substructures as regards the relationship between log Kow and BCF, although a conclusive statement cannot be made given the limited training/validation data available. The log Kow >7 validation subset consists mostly of halogenated aromatic hydrocarbons, although trioctyl amine (CAS# 1116-76-3) is present. A set of nine related structures (ester, hydrocarbon, and amine) was hand selected from the training and validation set data provided within the software supporting information. Statistical variation of the 9 pairs of experimental and predicted log BCF values indicate that there is a 0.871 correlation coefficient (r-squared), a standard deviation of 0.468 in residuals, and an absolute mean error of 0.339. It should be noted that all but one of these estimated log BCFs are based on estimated log Kow values. When an estimated log Kow is used in all cases, model performance improves slightly to R² value of 0.896, standard deviation of 0.436, and absolute mean error of 0.297. In either case, the model performs better when the training set as a whole is considered.

6. ADEQUACY OF THE RESULT
The simulation of BCF is relevant to the environmental fates and effects (specifically bioconcentration in aquatic organisms) when experimental bioconcentration data are not available. The model gives better performance with structures more closely related to the modeled dimers than with the training set overall. The molecular weight and predicted log Kow for the modelled dimers is partially within the parametric range of the model. Based on these data, this study is considered as acceptable as a QSAR simulation of BCF for the target substances. Its use in evaluating transport and fate in the environment is considered adequate for regulatory purposes.

Qualifier:

according to guideline

Guideline:

other: Guidance on information requirements and chemical safety assessment: Chapter R.6: QSARs and grouping of chemicals

Deviations:

no

Principles of method if other than guideline:

- Software tool(s) used including version: Episuite 4.11
- Model(s) used: BCFBAF v.3.01
- Model description: see field 'Attached justification'
- Justification of QSAR prediction: 'Attached justification'

GLP compliance:

no

Remarks:

QSAR model

Specific details on test material used for the study:

Reaction products of fatty acids, C18-unsaturated, dimers with 3,3'-[oxybis(ethane-2,1-diyloxy)]dipropan-1-amine:
1) Dimer 1: Molecular formula: C56H110N4O8 (molecular weight, 967.51), Smile: NCCCOCCOCCOCCCNC(=O)CCCCCCCCC1C=CC(CCCCC)C(CCCCCCC)C1CCCCCCCCC(=O)NCCCOCCOCCOCCCN
2) Dimer 2: Molecular formula: C56H112N4O8 (molecular weight, 969.54), Smile: NCCCOCCOCCOCCCNC(=O)CCCCCCCCC(=CCCCCCCC)C(CCCCCCCC)CCCCCCCCC(=O)NCCCOCCOCCOCCCN

Key result

Type:

BCF

Value:

16 L/kg

Remarks on result:

other: QSAR result for Dimer 1

Key result

Type:

BCF

Value:

42 L/kg

Remarks on result:

other: QSAR result for Dimer 2

Table 1. The predicted and experimental log BCF for related structures

CAS No.	Chemical name	Experimental Log BCF	Estimated log BCF	Residual
117-81-7	1,2-Benzenedicarboxylic acid, bis(2-ethylhexyl) ester	2.77	3.23 (2.85)¹	0.46 (0.08)¹
544-76-3	Hexadecane	3.7	2.91	-0.79
1460-02-2	1,3,5-Tri-tert-butylbenzene	4.36	3.77	-0.59
3319-31-1	1,2,4-Benzenetricarboxylic acid, tris(2-ethylhexyl) ester	1.17	1.28	0.11
4390-04-9	Nonane, 2,2,4,4,6,8,8-heptamethyl-	3.82	3.74	-0.08
26603-23-6	Benzenamine, ar-octyl-N-(octylphenyl)-	1.54	1.44	-0.1
26761-40-0	1,2-Benzenedicarboxylic acid, diisodecyl ester	1.16	1.88	0.72
35860-37-8	Naphthalene, tris (1-methylethyl)-	3.7	3.86	0.16
1116-76-3	1-Octanamine, N,N-dioctyl-(validation set substance)	1.93	1.89	-0.04

1, values in parentheses are calculated values using KOWWIN estimate for log Kow.  Log Kow inputs for all other substances were calculated using KOWWIN during model development.   

Validity criteria fulfilled:

not applicable

Conclusions:

Reaction products of fatty acids, C18-unsaturated, dimers and trimers with 3,3'-[oxybis(ethane-2,1-diyloxy)]dipropan-1-amine have multiple components. The estimated BCF values are 16 and 42 for two representative dimers, using BCFBAF v.3.01 as implemented through EPISuite v4.11.

Executive summary:

Reaction products of fatty acids, C18-unsaturated, dimers and trimers with 3,3'-[oxybis(ethane-2,1-diyloxy)]dipropan-1-amine have multiple components. Approximately 62.7% of compounds are identified as dimers, and 31.5% of compounds are identified as trimers. Among the identified dimers and trimers, two dimers have ≥ 10% of composition and no trimers has ≥ 10% composition. Additionally, the trimer is of such high molecular weight that bioconcentration is not a relevant property, and cannot be meaningfully estimated. Therefore, the two dimers (dimer 1 and dimer 2) with ≥ 10% of composition are modeled for the bioconcentration factor (BCF) using BCFBAF v.3.01 as implemented through EPISuite v4.11.

The QSAR for non-ionic materials having log Kow greater than seven is:

log BCF = -0.49 * log Kow + 7.554 + sum of correction factors

The software is an accepted, valid model for estimation of bioconcentration factor. The molecular weights and estimated log Kow values for dimer 1 and dimer 2 are slightly outside ranges in the training and validation sets. However, because the difference is not great and materials of this size and hydrophobicity are not expected to bioaccumulate, the result is considered acceptable.
The structural domain contains no amides having log Kow between one and seven but does a number of esters, hydrocarbons and amines (related structures). The sole identified structural correction factor is for eight contiguous -CH2- groups applied to chemicals with log Kow between seven and twelve. When the related structures are considered, performance statistics have a correlation coefficient (r²) of 0.871, a standard deviation of 0.468 and an absolute mean error of 0.339. This performance improves slightly through consistent use of modeled log Kow values. The use of this QSAR to predict bioconcentration factor for the two representative dimers was determined to be applicable and reliable with restriction.

This study is classified as an acceptable QSAR and satisfies the requirements for bioconcentration factor estimation. Its use in evaluating transport and fate in the environment is considered adequate for regulatory purposes.

Reference 2

Endpoint:

bioaccumulation in aquatic species: fish

Type of information:

(Q)SAR

Adequacy of study:

supporting study

Study period:

19 - 23 June 2020

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

results derived from a (Q)SAR model, with limited documentation / justification, but validity of model and reliability of prediction considered adequate based on a generally acknowledged source

Justification for type of information:

1. SOFTWARE : CATALOGIC 5.11.17

2. MODEL (incl. version number) : BCF base-line model v.02.09

3. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL: see below

4. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL
- Defined endpoint: Bioconcentration factor (BCF): The BCF base-line model consists of two major components: a model for predicting the maximum potential for bioaccumulation based solely on chemicals’ lipophilicity (based on multi-compartment diffusion), and a set of mitigating factors that account for the reduction of the bioaccumulation potential of chemicals based on chemical (e.g., molecular size, ionization, etc.) and organism-dependent factors (e.g., metabolism). In the BCF base-line model the tissue metabolism simulator is used to account for the effect of metabolism.
- Unambiguous algorithm:
- Defined domain of applicability: The applicability domain of the BCF base-line model contains three layers:
General properties requirements
Structural domain
Mechanistic domain (discriminates between modes of bioaccumulation - passive (partitioning in lipid phase) or active (based on protein binding). Only chemicals with expected passive diffusion driven bioaccumulation are considered to be in the mechanistic domain of the model)
- Appropriate measures of goodness-of-fit and robustness and predictivity:
Residual Sum of Squares = 176
Coefficient of Correlation, R = 0.92
Root mean square error, 0.52
Distribution of residual error: ca. 88% of residuals of the fitted data are <0.75 log units
- Mechanistic interpretation: The BCF base-line model reflects the current understanding of the process by which lipophilic organic chemicals are bioaccumulated in fish through the respiratory organs only. Chemicals, bioaccumulating by other mechanisms (e.g., binding to proteins) are considered out of the mechanistic domain of the model. The BCFmax model is a theoretical model based on the assumption that the only driving force of bioconcentration is lipophilicity and the effect of any other factors are insignificant. Its mathematical formalism is derived considering multi-compartment diffusion. The bioconcentration predicted by BCFmax model could be limited by variety of mitigating factors that account for the reduction of the bioaccumulation potential of chemicals based on chemical and organism-dependent factors. The effect of mitigating factors mathematically is quantified by probabilities: to penetrate through the cell membrane, to be ionized, to be metabolised, etc. In the BCF baseline model the tissue metabolism simulator is used to account for the effect of metabolism. It consists of a sequence of spontaneous abiotic and enzyme controlled steps. Probabilities of these molecular transformations are assessed by fitting the training set data.

5. APPLICABILITY DOMAIN
- Descriptor domain: General properties requirements (log Kow, MW, WS) are in domain.
- Structural domain:
Dimer 1:
Correct fragments = 100%
Incorrect fragments = 0%
Unknown fragments = 0%
Dimer 2:
Correct fragments = 98.53%
Incorrect fragments = 0%
Unknown fragments = 1.47%
"Unknown" structural features are atom centered fragments which do not have a prediction.
- Mechanistic domain: Active bioaccumulation not expected.
- Similarity with analogues in the training set: not reported

6. ADEQUACY OF THE RESULT
The simulation of BCF is relevant to the environmental fates and effects (specifically bioconcentration in aquatic organisms) when experimental bioconcentration data are not available. The model provides important supplementary information including calculated molecular size and predicted metabolic pathways. The general parameters are within the parametric range of the model and no incorrectly predicted atom centered fragments are present. Based on these data, this study is considered as acceptable as a QSAR simulation of BCF for the target substances. Its use in evaluating transport and fate in the environment is considered adequate for regulatory purposes.

Qualifier:

according to guideline

Guideline:

other: Guidance on information requirements and chemical safety assessment: Chapter R.6: QSARs and grouping of chemicals

Deviations:

no

Principles of method if other than guideline:

Principles of method if other than guideline
- Software tool(s) used including version: Catalogic v.5.11.17
- Model(s) used: BCF base-line model v.02.09
- Model description: see field 'Attached justification'
- Justification of QSAR prediction: 'Attached justification'

GLP compliance:

no

Remarks:

QSAR model

Specific details on test material used for the study:

Reaction products of fatty acids, C18-unsaturated, dimers with 3,3'-[oxybis(ethane-2,1-diyloxy)]dipropan-1-amine:
1) Dimer 1: Molecular formula: C56H110N4O8 (molecular weight, 967.51), Smile: NCCCOCC
OCCOCCCNC(=O)CCCCCCCCC1C=CC(CCCCC)C(CCCCCCC)C1CCCCCCCCC(=O)NCCCOCCOCCOCCCN
2) Dimer 2: Molecular formula: C56H112N4O8 (molecular weight,
969.54), Smile: NCCCOCCOCCOCCCNC(=O)CCCCCCCCC(=CCCCCCCC)C(CCCCCCCC)CCCCCCCCC(=O)NCCCOCCOCCOCCCN

Key result

Type:

BCF

Value:

7.4 L/kg

Remarks on result:

other: QSAR result for Dimers 1 and 2

Details on results:

The model uses a baseline prediction of log BCF calculated using log Kow, with a predicted maximum log BCF occurring at log Kow = 6. The model then adjusts baseline BCF according to several model factors including metabolism, water solubility, and molecular size. The two modeled dimers differ slightly in log Kow but have an identical predicted log BCF. The other model factors had different weighting factors in arriving at this result. In order of weight the factors are:

Molecular size:
Dimer 1 and Dimer 2 have calculated average values for Maximum diameter of 36.4 angstrom and 38.3 angstroms, respectively.

Metabolism:
Dimer 1 has a probability of 80% to metabolize, while Dimer 2 has a probability of 44%. In both cases epoxidation of olefin groups, terminal hydroxylation of alkyl groups, and amide hydrolysis are the predicted pathways.

Water solubility:
Dimer 1 and Dimer 2 each have an estimated water solubility of 9.7E-07 mg/L

Validity criteria fulfilled:

not applicable

Conclusions:

Reaction products of fatty acids, C18-unsaturated, dimers and trimers with 3,3'-[oxybis(ethane-2,1-diyloxy)]dipropan-1-amine have multiple components. The estimated BCF value is 7.4 for two representative dimers, using BCF baseline model v.02.09 as implemented through Catalogic v.5.11.17. In addition, Dimer 1 and Dimer 2 have calculated average values for Maximum diameter of 36.4 angstrom and 38.3 angstroms, respectively.

Executive summary:

Reaction products of fatty acids, C18-unsaturated, dimers and trimers with 3,3'-[oxybis(ethane-2,1 -diyloxy)]dipropan-1-amine have multiple components. Approximately 62.7% of compounds are identified as dimers, and 31.5% of compounds are identified as trimers. Among the identified dimers and trimers, two dimers have ≥ 10% of composition and no trimers has ≥ 10% composition. Additionally, the trimer is of such high molecular weight that bioconcentration is not a relevant property, and cannot be meaningfully estimated. Therefore, the two dimers (dimer 1 and dimer 2) with ≥ 10% of composition are modeled for the bioconcentration factor (BCF) using BCF baseline model v.02.09 as implemented through Catalogic v.5.11.17.

This model is a validated QSAR with extensive parameterization to take molecular size, metabolism, and other parameters into account. The molecular weights, estimated water solubility and estimated log Kow values for dimer 1 and dimer 2 are within the models parametric range. All atom centered fragments in Dimer 1 and 98.5% of fragment in Dimer 2 are in the set of fragments with correct prediction, with the remaining fragment of Dimer 2 not recognized. No incorrectly predicted atom centered fragments are present. Based on these data, the use of this QSAR to predict bioconcentration factor for the two representative dimers was determined to be applicable and reliable with restriction.

This study is classified as an acceptable QSAR and satisfies the requirements for bioconcentration factor estimation. Its use in evaluating transport and fate in the environment is considered adequate for regulatory purposes.

Description of key information

The largest BCF modeled for Dimeric constituents of MTDID 18990 was 42. This is based on an estimated log Kow of 12.1. The smallest calculated average of maximum diameters was 36.4 angstroms.

Key value for chemical safety assessment

BCF (aquatic species):

42 L/kg ww

Additional information

Bioaccumulation endpoints are not required at this annual volume. QSAR modeling of bioconcentration factors (BCFs) was done in support of the PBT assessment. Reaction products of fatty acids, C18-unsaturated, dimers and trimers with 3,3'-[oxybis(ethane-2,1-diyloxy)]dipropan-1 -amine have multiple components. Approximately 62.7% of compounds are identified as dimers, and 31.5% of compounds are identified as trimers. Among the identified dimers and trimers, two dimers have ≥ 10% of composition and no trimers has ≥ 10% composition. Additionally, the trimer is of such high molecular weight that cannot be meaningfully estimated. Therefore, the two dimers (dimer 1 and dimer 2) with ≥ 10% of composition are modeled for the BCF. Two models were used: BCFBAF v.3.01 (within EPISUITE 4.11) and BCF baseline model v.02.09 (Catalogic v.5.11.17). Both models are accepted and validated for estimation of BCF.

BCFBAF applicability was determined as follows. The molecular weights and estimated log Kow values for dimer 1 and dimer 2 are slightly outside ranges in the training and validation sets. However, because the difference is not great and materials of this size and hydrophobicity are not expected to bioaccumulate, the result is considered acceptable. The structural domain contains no amides but does a number of esters, hydrocarbons and amines (related structures). The sole structural identifier is for eight contiguous -CH2- groups applied to chemicals with log Kow between seven and twelve. When the related structures are considered, performance statistics are a correlation coefficient (r²) of 0.871, a standard deviation of 0.468 and an absolute mean error of 0.339. This performance improves slightly through consistent use of modeled log Kow values.

BCF baseline model was determined as follows. BCF baseline model is a validated QSAR with extensive parameterization to take molecular size, metabolism, and other parameters into account. The molecular weights, estimated water solubility and estimated log Kow values for dimer 1 and dimer 2 are within the models parametric range. All atom centered fragments in Dimer 1 and 98.5% of fragment in Dimer 2 are in the set of fragments with correct prediction, with the remaining fragment of Dimer 2 not recognized. No incorrectly predicted atom centered fragments are present.

The use of these QSAR to predict bioconcentration factor for the two representative dimers was therefore determined to be applicable. The estimated BCF values are 16 and 42 for the two representative dimers, using BCFBAF v.3.01. BCF baseline model v. 02.09 gave an estimated BCF of 7.4 for both representative dimers, and also provided calculated average values for Maximum diameter of 36.4 angstrom and 38.3 angstroms, respectively.

These in silico studies were deemed reliable with restriction and satisfy the requirements for bioconcentration factor estimation. Their use in evaluating transport and fate in the environment is considered adequate for regulatory purposes such as PBT assessment.