1,1'-(benzene-1,3-diyldimethanediyl)bis(1H-pyrrole-2,5-dione)

Partition coefficient

Link to relevant study record(s)

Reference

Endpoint:

partition coefficient

Type of information:

(Q)SAR

Adequacy of study:

key study

Study period:

2017-12-05

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
Estimation Programs Interface Suite™ for Microsoft® Windows v4.11. US EPA, United States Environmental Protection Agency, Washington, DC, USA.

2. MODEL (incl. version number)
KOWWIN v1.68 (September 2010)

3. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL
SMILES: O=C3C=CC(=O)N3Cc1cc(CN2C(=O)C=CC2(=O))ccc1
CAS-no.: 13676-53-4

4. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL

- Defined endpoint: log Kow; (log P)- ogarithmic octanol/water partition coefficient.
- Unambiguous algorithm: KOWWIN uses a "fragment constant" methodology to predict log P. KOWWIN’s methodology is known as an Atom/Fragment Contribution (AFC) method. Coefficients for individual fragments and groups were derived by multiple regression of 2447 reliably measured log P values.
To estimate log P, KOWWIN initially separates a molecule into distinct atom/fragments. In general, each non-hydrogen atom in a structure is a "core" for a fragment; the exact fragment is determined by what is connected to the atom. Several functional groups are treated as core atoms. Connections to each core "atom" are either general or specific; specific connections take precedence over general connections.
Log P estimates made from atom/fragment values alone could or need to be improved by inclusion of substructures larger or more complex than "atoms"; hence, correction factors were added to the AFC method. They are either factors involving aromatic ring substituent positions, or miscellaneous factors. Correction factors are values for various steric interactions, hydrogen-bondings, and effects from polar functional substructures. Individual correction factors were selected through a tedious process of correlating the differences (between log P estimates from atom/fragments alone and measured log P values) with common substructures.
The general regression equation has the following form:

log P = ∑(f(i)*n(i)) + ∑(c(j)*n(j)) + b

where f(i) is the coefficient of atom/fragment i, n(i) – the number of times the fragment i occurs in the molecule, c(j) is the coefficient for the correction factor j, and n(j) the number of times the factor j occurs (or is applied) in the molecule. b is the linear equation constant; b = 0.229.
Values of f and c coefficients are available.

- Defined domain of applicability:
Currently there is no universally accepted definition of model domain. However, it should be considered that log P estimates may be less accurate for compounds outside the molecular weight range of the training set compounds, and/or that have more instances of a given fragment than the maximum for all training set compounds. Although the training set of the model contains a large number of diverse molecules and can be considered abundant, it is also possible that a compound may be characterised by structural features (e.g. functional groups) not represented in the training set, with no respective fragment/correction coefficient developed. These points should be taken into consideration when interpreting model results.
Training set molecular weights:
Minimum MW: 18.02
Maximum MW: 719.82 (in the validation set: 991.15)
Average in the training set: 199.98.

- Appropriate measures of goodness-of-fit and robustness and predictivity:
Training set statistics:
N = 2447 compounds
correlation coefficient R2= 0.982
standard deviation = 0.217
absolute deviation = 0.159

Training set estimation error:
within ≤ 0.10 – 45.0%
within ≤ 0.20 – 72.5%
within ≤ 0.40 – 92.4%
within ≤ 0.50 – 96.4%
within ≤ 0.60 – 98.2%

To be effective an estimation method must be capable of making accurate predictions for chemicals not included in the training set. Currently, KOWWIN has been tested on an external validation dataset of 10,946 compounds (compounds not included in the training set). The validation set includes a diverse selection of chemical structures that rigorously test the predictive accuracy of any model. It contains many chemicals that are similar in structure to chemicals in the training set, but also many chemicals that are different from and structurally more complex than chemicals in the training set. The average molecular weight of compounds in the validation set is 258.98 versus 199.98 for the training set.

External validation set statistics:
N = 10946 compounds
correlation coefficient R2= 0.943
standard deviation = 0.479
absolute deviation = 0.356

Validation set estimation error:
within ≤ 0.20 – 39.6%
within ≤ 0.40 – 66.0%
within ≤ 0.50 – 75.6%
within ≤ 0.60 – 82.5%
within ≤ 0.80 – 91.6%
within ≤ 1.00 – 95.6%
The external validation set of 10946 compounds contains 372 compounds that exceed the domain of instances of a given fragment or correction factor maximum for all training set compounds. The estimation accuracy for these compounds is:

Exceed Fragment Instance Domain - Accuracy Statistics:
number in dataset = 372
correlation coef (r2) = 0.939
standard deviation = 0.731
absolute deviation = 0.564
avg Molecular Weight = 460.0

Exceed Molecular Weight Domain - Accuracy Statistics:
number in dataset = 103
correlation coef (r2) = 0.879
standard deviation = 0.815
absolute deviation = 0.619
avg Molecular Weight = 802.16

Exceed BOTH Fragment & MW Domain - Accuracy Statistics:
number in dataset = 75
correlation coef (r2) = 0.879
standard deviation = 0.905
absolute deviation = 0.706
avg Molecular Weight = 812.70

A list of the 75 compounds that exceed both the fragment instance domain and molecular limit domain can be found in the help section of EPISuite.

- Mechanistic interpretation:
KOWWIN’s "reductionist" fragment constant methodology (i.e. derivation via multiple regression) differs from the "constructionist" fragment constant methodology of Hansch and Leo (Hansch, C. and Leo, A.J., Substituent Constants for Correlation Analysis in Chemistry and Biology, Wiley, New York, 1979). More complete description of KOWWIN methodology is described in: Meylan, W.M., and Howard, P.H., Atom/Fragment Contribution Method for Estimating Octanol-Water Partition Coefficients, J. Pharm. Sci 84: 83-92, 1995.

5. APPLICABILITY DOMAIN

- Descriptor domain:
molecular weight, structure fragments

- Similarity with analogues in the training set:
Since KOWWIN uses a "fragment constant" methodology to predict log P, the functional groups within m-Xylylenebismaleimide, i.e. the aliphatic carbons and the attached groups (olefinc , aromatic and Pyrrole-2,5-dione ring) can be found in the training set provided by the KOWWIN help section. The similar substance 1H-Pyrrole-2,5-dione, 1-phenyl- (CAS 941-69-5) is in the validation set of the tool. The substance has an estimated logKow of 0.88 and an experimental determined logKow of 1.09 (Hansch et al 1995). Also, experimental data for another very similar substance (1,1'-(methylenedi-p-phenylene)bismaleimide CAS 13676-54-5) support the use of the estimation tool /QSAR. The estimated logKow for CAS 13676-54-5 with the tool is 1.8 whereas the experimental logKow is 1.5 (see dissiminated dossier at the ECHA website ).


6. ADEQUACY OF THE RESULT
m-Xylylenebismaleimide contains a structure with chemical groups present in the training/validation data sets used by KOWWIN. Thus, m-Xylylenebismaleimide falls into the applicability domain and the results are considered reliable.

For more informations on the model and the results please refer to the attached jusitification.

Qualifier:

no guideline followed

Principles of method if other than guideline:

QSAR prediction using EPISuite software and KOWWIN v1.68 (Spetember 2010).

GLP compliance:

no

Type of method:

other: QSAR prediction

Key result

Type:

log Pow

Partition coefficient:

1.6

Temp.:

25 °C

Remarks on result:

other: QSAR prediction, pH was not reported

Conclusions:

The partition coefficient was estimated using EPISuite software and the KOWWIN model (v1.68, September 2010). The log Kow is 1.60. The results are considered to be reliable because the structural fragments contained in m-Xylylenebismaleimide are part of the training/validation set data and the substance has a molecular weight which is in the range of the training set molecular weights.

Description of key information

- QSAR estimation using EPISUite software and KOWWIN model, log Kow: 1.60

Key value for chemical safety assessment

Log Kow (Log Pow):

1.6

at the temperature of:

25 °C

Additional information

The partition coefficient was estimated using EPISuite software and the KOWWIN model (v1.68, September 2010). The log Kow is 1.60. The results are considered to be reliable because the structural fragments contained in m-Xylylenebismaleimide are part of the training/validation set data and the substance has a molecular weight which is in the range of the training set molecular weights.