Registration Dossier

Diss Factsheets

Physical & Chemical properties

Partition coefficient

Currently viewing:

Administrative data

Link to relevant study record(s)

Reference
Endpoint:
partition coefficient
Type of information:
(Q)SAR
Adequacy of study:
key study
Study period:
June 18, 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) : KOWWIN v.1.68
3. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL : see below
4. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL
- Defined endpoint: logarithm of the n-octanol:water partition coefficient
- Unambiguous algorithm:The predicted value of the log Kow is calculated through the following equation: log KOW = Σfj*nj + Σcj*nj + 0.229
where fj is the coefficient for each atom/fragment), cj is the coefficient for each correction factor, and ni is the number of times the atom/fragment (and correction factor) occurs in the test species.
- Defined domain of applicability:
Training set parameters (n = 2447):
Molecular weight range: 18.02 D – 719.92 D
Average molecular weight: 199.98
Log Kow range (measured): -4.22 – 8.18
Validation set parameters (n = 10,946):
Molecular weight range: 27.03 D – 991.15 D
Average molecular weight: 258.98
Log Kow range (measured): -5.08 - 11.29
KOWWIN v1.68 predicts log Kow of a given test chemical based on the functional groups which are part of that given test chemical’s structure. The log Kow is calculated as the summation of the atom fragment constants of individual chemical functionalities (fj). The contribution of individual chemical functionality to the overall log Kow is corrected according to a corresponding coefficient correction (cj) for a given atom type or chemical functionality. The summation of these atom fragment constants and associated correction factors is representative of the overall log Kow (i.e. the ratio of n-octanol solubility to water solubility). The predictive capacity of KOWWIN is limited to chemical functionalities which are contained in the reference data set and therefore only test chemicals which fall in that domain are accurately estimated. KOWWIN v1.68 accurately estimates test chemical which contain most commonly occurring aliphatic and aromatic chemical functional groups. Inorganics, organometallics, and boron containing species are generally outside the predictive domain of KOWWIN v1.68
- Appropriate measures of goodness-of-fit and robustness and predictivity:
From the log-log plot of predicted and experimental n-octanol/water partition coefficients (Kow’s) of the training set (2,447 compounds), the correlation coefficient (r2) is 0.982, the standard deviation (sd) is 0.217, and the absolute mean error (me) is 0.159. From the log-log plot of predicted and experimental n-octanol/water partition coefficients (Kow’s) of the validation set (10,946 compounds), the correlation coefficient (r2) is 0.943, the standard deviation (sd) is 0.479, and the absolute mean error (me) is 0.356.
- Mechanistic interpretation:
The mechanistic basis for this QSAR is the specific chemical functional groups behave a certain way in n-octanol and water (the ratio of which is Kow). These fragments impart solubility characteristics on the given chemical and this is a constant. These constants have been calculated for a wide variety of functional groups and provide a basis by which to predict the Kow may be predicted for a chemical which contains functional groups from the training set. The individual mechanisms of this dissolution will vary between different chemical species and functional groups.

5. APPLICABILITY DOMAIN
The molecular weights of target compounds are 967.52 and 969.54 for dimer 1 and dimer 2, respectively. Although they are outside of the MW range of the training set, they are within the validation set MW range of the model. Effectively, the inclusion of the higher MWs compounds in the validation set extends the applicability domain of the model to encompass the MW range of the target compounds.
The target compounds’ fragments are not overrepresented in the structure, except the -CH2 [aliphatic carbon] fragment. The number of –CH2 aliphatic carbon in both dimer 1 and dimer 2 exceeded the maximum 18 within a structure in training set and maximum 28 within a structure in validation set. The exact list of chemicals used to determine fragment coefficients and structure corrections are not available for each fragment type.
Structural analogues of the target substances were hand selected from the data set provided within the software supporting information. No structural analogue that contains all fragments in the target chemicals was identified. However, very hydrophobic (log Kow > 7) fatty acid and fatty acid ester, amide and amine compounds are included, and these are used to determine prediction uncertainty. The experimental and predicted log Kow values of 27 analogous compounds indicate that there is a 0.9857 correlation coefficient (r-squared), a standard deviation of 0.351 and an absolute mean error of 0.280.

6. ADEQUACY OF THE RESULT
The molecular weights of target compounds are 967.52 and 969.54 for dimer 1 and dimer 2, respectively. Although they are outside of the MW range of the training set, they are within the validation set MW range of the model. The predicted log Kow values are 11.72 and 12.11 for the dimer 1 and dimer 2, respectively, outside of the training and validation set log Kow range of the model. However, the target structures do not contain any unidentified fragments, and there is complete coverage of the structure by the molecule descriptors. The estimated and experimental data for the analogous compounds is in good agreement. Based on these data, this study is considered as acceptable as a QSAR simulation of log Kow 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: KOWWIN v.1.68
- Model description: see field 'Attached justification'
- Justification of QSAR prediction: 'Attached justification'
GLP compliance:
no
Remarks:
QSAR model
Type of method:
calculation method (fragments)
Remarks:
KOWWIN v1.68 as implemented through EPI Suite v4.11
Partition coefficient type:
octanol-water
Key result
Type:
log Pow
Partition coefficient:
11.7
Temp.:
25 °C
pH:
7
Remarks on result:
other: for Dimer 1
Key result
Type:
log Pow
Partition coefficient:
12.1
Temp.:
25 °C
pH:
7
Remarks on result:
other: for Dimer 2

Table 1. The predicted and experiment log Kow

CAS Number

Exp LogKow

Kowwin

Est

Residual

Analogous

 

 

 

929-77-1

10.20

10.20

0.00

1120-34-9

9.32

9.98

0.66

1120-28-1

9.30

9.21

-0.09

506-30-9

9.29

8.93

-0.36

20292-08-4

8.03

8.65

0.62

112-61-8

8.35

8.23

-0.12

112-62-9

7.45

8.02

0.57

112-63-0

6.82

7.80

0.98

57-11-4

8.23

7.94

-0.29

2027-47-6

7.64

7.73

0.09

60-33-3

7.05

7.51

0.46

463-40-1

6.46

7.30

0.84

106-18-3

6.51

6.76

0.25

3681-78-5

6.11

6.27

0.16

124-10-7

6.41

6.27

-0.14

106-33-2

5.71

5.78

0.07

95715-61-0

3.12

3.26

0.14

5830-33-1

2.19

2.28

0.09

00-06-1

2.28

2.56

0.28

82700-02-5

3.70

4.01

0.31

54704-34-6

2.96

3.03

0.07

111-86-4

2.90

2.80

-0.10

111-92-2

2.83

2.77

-0.06

111-68-2

2.57

2.31

-0.26

20193-21-9

2.12

2.28

0.16

111-26-2

2.06

1.82

-0.24

110-58-7

1.49

1.33

-0.16

Target compounds

 

 

 

dimer 1

 

11.7

 

Dimer 2

 

12.1

 

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 log Kow values are 11.7 and 12.1 for two representative dimers, using KOWWIN v1.68 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 partitioning behavior 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 partition coefficients n-octanol/water (log Kow) using KOWWIN v1.68 as implemented through EPISuite v4.11.

The software is an accepted, valid model for estimation of partition coefficient. The molecular weights for dimer 1 and dimer 2 are within the validation set MW range of the model but are outside of the MW range of the training set. The dimers have a fragment exceeded the maximum fragment numbers within a structure in both training set and validation set. However, the KOWWIN software has estimation accuracy for exceed fragment instance domain with a correlation coef (r2)of 0.939 and a standard deviation of 0.731.

The predicted log Kow values are 11.72 and 12.11 for the dimer 1 and dimer 2, respectively, outside of the training and validation set log Kow range of the model. However, the dimers’ structures do not contain any unidentified fragments, and there is complete coverage of the structure by the molecule descriptors. Based on representation of analogous substances within the data set and performance statistics (a correlation coefficient (r^2) of 0.9857, a standard deviation of 0.351 and an absolute mean error of 0.280) derived from a comparison of experimental and estimated log Kow values for 27 representative analogous substances, the use of this QSAR to predict log Kow 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 partition coefficient study. Its use in evaluating transport and fate in the environment is considered adequate for regulatory purposes.

Description of key information

The predicted log Kow values are 11.72 and 12.11 for two representative dimers using KOWWIN v1.68 as implemented through EPISuite v4.11.
.

Key value for chemical safety assessment

Log Kow (Log Pow):
11.7
at the temperature of:
25 °C

Additional information

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 components are identified as dimers, and 31.5% of components 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 partitioning behavior is not a relevant property or cannot be meaningfully estimated. Therefore, the two dimers (dimer 1 and dimer 2) with ≥ 10% of composition are modeled for the partition coefficients n-octanol/water (log Kow) using KOWWIN v1.68 as implemented through EPISuite v4.11.

The predicted log Kow values are 11.72 and 12.11 for the dimer 1 and dimer 2, respectively, outside of the training and validation set log Kow range of the model. However, the dimers’ structures do not contain any unidentified fragments, and there is complete coverage of the structure by the molecule descriptors. The molecular weights for dimer 1 and dimer 2 are within the validation set MW range of the model but are outside of the MW range of the training set. The dimers have a fragment exceeded the maximum fragment numbers within a structure in both training set and validation set. However, estimation outside of domain is accepted based on good estimation accuracy for exceed fragment instance domain with a correlation coef (r2)of 0.939 and a standard deviation of 0.731.

 

The software is an accepted, valid model for estimation of partition coefficient. The use of this QSAR to predict log Kow for the two representative dimers was determined to be reliable with restriction, based on representation of analogous substances within the data set and performance statistics derived from a comparison of experimental and estimated log Kow values for representative analogous substances.


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