hexyl nitrite

Biodegradation in water: screening tests

Administrative data

Endpoint:

biodegradation in water: ready biodegradability

Type of information:

(Q)SAR

Adequacy of study:

key study

Study period:

15/10/2020

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

accepted calculation method

Justification for type of information:

1. SOFTWARE QSAR

2. MODEL (incl. version number) : Model 4.0.4 Molcode Ltd., Turu 2, Tartu, 51014, Estonia; http://www.molcode.com
Endpoint (OECD Principle 1): QSAR 2.3.a
Algorythm (OECD Principle 2): QSAR ANN model for Persistence: Biotic degradation in water, Model version 22.05.2012


3. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL
SMILES: CCCCCCON=O, not used for prediction
3D Mol file used for prediction

4. 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: QSAR 2.3.a; Ready Biodegradability (Aerobic Mineralisation in Surface Water - Simulation Biodegradation Test); OECD 309
- Unambiguous algorithm: QSAR ANN model for Persistence: Biotic degradation in water (Reference to QMRF - attached separately)
- Defined domain of applicability:
a) descriptor domain
b) structural fragment domain
c) mechanims domain
- Uncertainty of the prediction (OECD Principle 4):
The source experimental data for the model originate from different labs and different experiment series, adding to uncertainty, however, previous (and present) successful modeling add to the consi stence of the data. The significant size of the dataset; the diversity of the structures covering a large parts of the chemical space, and the statistical quality (RMS, correlation coefficients etc.) of the model supports reliable predictions within the margins of the experimental error. The similarity of the analogues together with the correct estimates supports potential prediction consistency.
Considering the dataset size, model statistical quality and prediction reliability, a reliability score (Klimisch score) “2” could be assigned to the present prediction.
The prediction reliability in terms of the persistence category is estimated as 84 %
- Mechanistic interpretation:
Chemical degradation half-life has been defined as a quasi-intensive property since it is independent of quantity and is a characteristic of a molecule within a defined environmental medium. The struc tural features of a chemical determine its intrinsic ability or tendency to persist in the environment, independent of its partitioning properties and environmental conditions. Particular features accepted as indicators of high persistence include (per)halogenated aliphatic and aromatic carbon systems, the molecular size being also important.
Because of the nonlinear nature of the ANNs, deeper analysis of the descriptors is difficult compared to the normal multilinear analysis. The present ANN model descriptors are mainly related to the c harge distributions of the compounds e.g. FPSA3 Fractional PPSA (PPSA-3/TMSA) (Zefirov), Negatively Charged Surface Area (Zefirov), Partial Charged (Zefirov) Surface Area of H atoms. In additio n the halogens reactivity plays also important role for the water degradation. Generally the halogens have larger LogT values, the tendency that well agrees with the general understanding of the pro perty.
Artificial Neural networks (ANN) are used in many areas, such as pattern recognition, process analysis and non-linear modelling. An advantage of neural nets is that the neural net model is very flexi ble in contrast to the classical statistical models. A significant disadvantage is the amount of data needed and the causal ambiguity of the network. The neural net ‘learns’ from examples by one of tw o different approaches, supervised or unsupervised learning. During supervised learning, the system is forced to assign each object in the training set to a specific class, while during unsupervised l earning, the clusters are formed without any prior information. One approach commonly used is multi-layer feed-forward (MLF) networks consisting of three or more layers: one input layer, one outp ut layer and one or more intermediate (hidden) layer (Smiths et al., 1994; Xu et al., 1994; De Saint Laumer et al., 1991).
In this model report: nonlinear regression QSAR artificial neural networks model with architecture 6-5-1 trained with back propagation of the error.
ANN is mentioned many times in REACH related official documents ANN is considered as an acceptable algorithm for non-linear correlations.

5. APPLICABILITY DOMAIN
- Descriptor domain: All descriptor values for Hexyl nitrite fall in the applicability domain (training set value ±30%).
- Structural domain: Hexyl nitrite is structurally similar to the model compounds, the model contains compounds featuring short alkyl chains and ether functionalities The training set contains comp ounds of si milar size to the studied molecule.
- Mechanistic domain: Hexyl nitrite is considered to be in the same mechanistic domain as the molecules in the training set as it is structurally similar to the model compounds.
- Similarity with analogues in the training set: The experimental acute oral toxicity values for compounds of similar functionalities (saturated alkyl ethers) all fall to the “low persistence” category, The structural analogues are relatively similar to the studied compound. The descriptor values of the analogues are close to those of the studied compound. The analogues are considered to be within the same mechanistic domain. All the analogues are relativ ely well estimated within the model. The following aspects have been considered for the selection and analysis of structural analogues:
Presence and number of common functional groups;
Presence and relevance of non-common functional groups;
Similarity of the ‘core structure’ apart from the (non-)common functional groups;
Potential differences due to reactivity;
Potential differences due to steric hindrance;
Presence of structural alerts;
Position of the double bonds;
Presence of stereoisomers.

6. ADEQUACY OF THE RESULT
6.1 Regulatory purpose:
The present prediction may be used for preparing the REACH Joint Registration Dossier on the Substance(s) for submission to the European Chemicals Agency (“ECHA”) as required by Regulation (EC) N° 1907/2006 of the European Parliament and of the Council of 18 December 2006 concerning the Registration, Evaluation, Authorisation and Restriction of Chemicals ("REAC H") and as required by Biocide Product Directive 98/8/EC ("98/8/EC")
6.2 Approach for regulatory interpretation of the model result
The predicted result has been presented in the formats directly usable for the intended regulatory purposes. The predicted numeric value in physicochemical units has been presented. The possible classification of the values according to classification systems has also been discussed.
6.3 Outcome
See section 3.2(e) for the classification of the prediction in light of the regulatory purpose described in 4.1.
6.4 Conclusion
Considering the above, the predicted result can be considered adequate for the regulatory conclusion described in 4.1.

Data source

Reference

Reference Type:

study report

Title:

Unnamed

Year:

2020

Report date:

2020

Materials and methods

Test guideline

Test guideline 1

Qualifier:

according to guideline

Guideline:

other: OECD Guideline 309: Ready Biodegradability (Aerobic Mineralisation in Surface Water - Simulation Biodegradation Test)

Deviations:

no

Test guideline 2

Qualifier:

according to guideline

Guideline:

OECD Guideline 301 A (Ready Biodegradability: DOC Die Away Test)

Principles of method if other than guideline:

3.3. Comment on endpoint:
The half-life is the time required for the concentration of a substance to halve its original value in a particular environmental medium. The half-lives of organic compounds are among the most commonly used criteria for studying persistence [1]. The semiquantitative data based on expert judgment and actual experimental values have already been suggested by Webster et al. [2] as preferable for half life identification, and are commonly used to develop the widely applied multimedia models [3,4]. In addition, a simple QSPR regression model has been demonstrated to be an useful tool for the identification and prioritization of existing or not yet synthesized potential persistent organic pollutants
3.4.Endpoint units:
The half-life values (in h) were transformed into logarhitmic form
for modelling
3.5.Dependent variable:
log T(0.5)
3.6.Experimental protocol:
The dataset of structurally heterogeneous and highly
representative of many classes of already defined problematic chemicals
includes 206 organic compounds of known half-lives for transformation
into air [6].
3.7.Endpoint data quality and variability:
min value training set: 1.23
max value training set: 4.74
avrg training set: 2.63

GLP compliance:

no

Test material

Reference

Name:

Unnamed

Type:

Constituent

Test material form:

liquid

Study design

Oxygen conditions:

aerobic

Inoculum or test system:

natural water: marine

Parameter followed for biodegradation estimation

Parameter followed for biodegradation estimation:

not specified

Results and discussion

% Degradation

Key result

Parameter:

half-life in days (QSAR/QSPR)

Value:

> 60

Remarks on result:

not readily biodegradable based on QSAR/QSPR prediction

Details on results:

A substance fulfils the persistence/vP criterion (P/vP) if the half-life in marine water is higher than 60 days, OR the half-life in fresh- or estuarine water is higher than 40/60 days.

Applicant's summary and conclusion

Validity criteria fulfilled:

not applicable

Interpretation of results:

under test conditions no biodegradation observed

Conclusions:

Hexyl nitrite falls into the "Low persistence" category.

Executive summary:

Hexyl nitrite falls into the "Low persistence" category.