6,6'-diphenylfulvene

Biodegradation in water: screening tests

Administrative data

Endpoint:

biodegradation in water: ready biodegradability

Type of information:

(Q)SAR

Adequacy of study:

key study

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

Remarks:

The value is not an experimental result, however the QSAR model is recommended by the ECHA guidance document on information requirements, is well documented with regard to validation parameters according to OECD principles. Moreover, the substance (characterised by a generic SMILES code) is fully characterised towards the applicability domain.

Justification for type of information:

1. SOFTWARE
EPIWIN software by US-EPA

2. MODEL (incl. version number)
BIOWIN v4.10

3. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL
c(ccc1C(c(ccc2)cc2)=C(C=C3)C=C3)cc1
CAS 2175-90-8

4. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL
- The model and the training and validation sets are published by US Environmental Protection Agency (USA). Details in addition to the summary provided here can be found in the program.

The Biowin program is part of the Estimation Programs Interface (EPI) Suite© and contains six models whose purpose is to provide screening-level estimates of aerobic biodegradability for organic chemicals:
Biowin1 = linear probability model;
Biowin2 = nonlinear probability model;
Biowin3 = expert survey ultimate biodegradation model;
Biowin4 = expert survey primary biodegradation model;
Biowin5 = Japanese Ministry of International Trade and Industry (MITI) linear model; and
Biowin6 = MITI nonlinear model.
Biowin1 and 2 constitute the original Biodegradation Probability Program and were the first two models developed (Howard et al., 1992). Biowin3 and 4 estimate the time required to achieve complete ultimate (Biowin3) or primary (Biowin4) biodegradation in a typical, or ‘evaluative’, aquatic environment. They were developed from estimates of biodegradability for 200 substances, provided by a panel of experts (Boethling et al., 1994). The output of the model for any given chemical is a number from one to five (continuous scale; fractional values permitted) that either can be used as it is, or if desired can be related to the words (e.g. ‘days’, ‘weeks’) used in the expert survey. In Biowin3 values N2.75 and b=3.25 are assigned the term‘weeks’. Here 2.75 was used as the criterion for ready biodegradability, so that a calculated value N2.75 is a prediction of RB and is equivalent to a prediction of ‘weeks’ or an even shorter timeframe. For Biowin4 a criterion of 3.6 (i.e., b=3.6 is NRB) yielded a reasonable separation of RB and NRB compounds. Biowin5 and 6 are linear and nonlinear probability models developed exclusively from the MITI data (MITI, 2014). Biowin5/6 classify a substance as either RB or NRB based on the MITI ready biodegradation test (OECD 301C).
Characteristics of this test, as well as details of model development and validation, are given by Tunkel et al. (2000). The model estimates the likelihood that a chemical will degrade under the test conditions. Values of the dependent variable N0.5 usually are taken as predicting that the chemical will be RB in the OECD 301C test, but as with other Biowin models the user can set the RB/NRB criterion at any value desired.

Predicting Ready Biodegradability
The following line appears at the end of the Biowin prediction results: Ready Biodegradability Prediction:  (YES or NO)
The criteria for the YES or NO prediction are as follows:  If the Biowin3 (ultimate survey model) result is "weeks" or faster (i.e. days, days to weeks, or weeks) AND the Biowin5 (MITI linear model) probability is >= 0.5, then the prediction is YES (readily biodegradable).  If this condition is not satisfied, the prediction is NO (not readily biodegradable).
This method is based on the application of Bayesian analysis to ready biodegradation data for US Premanufacture Notification (PMN) chemicals, derived collectively from all six OECD301 test methods plus OECD310.  The approach is fully described in Boethling et al. (2004).
The linear and nonlinear MITI models (Biowin5 and 6) also predict ready biodegradability, but for degradation in the OECD301C test only, and based exclusively on data from the Chemicals Evaluation and Research Institute Japan (CERIJ) database (http://www.cerij.or.jp/ceri_en/otoiawase/otoiawase_menu.html).

5. APPLICABILITY DOMAIN
The EPI Suite data files are files containing the experimental data sets used to derive and validate program methodologies or test program accuracy.  In EPI Suite v4.00, the data files were available in either the appendices of the individual program help files or via Internet download from: http://esc.syrres.com/interkow/EpiSuiteData.htm

6. ADEQUACY OF THE RESULT
The accuracy of the models is described in detail in the program itself and e.g. by Boethling (Science of the Total Environment 497–498 (2014) 60–67) and suffices to give suitable predictions for substances falling into the applicability domain of the model.

Two independent training sets were used to develop four mathematical models for predicting aerobic biodegradability from chemical structure.  All four of the models are based on multiple regressions against counts of 36 preselected chemical substructures plus molecular weight.  Two of the models, based on linear and nonlinear regressions, calculate the probability of rapid biodegradation and can be used to classify compounds as rapidly or not rapidly biodegradable.  The training set for these models consisted of qualitative summary evaluations of all available experimental data on biodegradability for 295 compounds.  The other two models allow semi-quantitative prediction of primary and ultimate biodegradation rates using multiple linear regression.  The training set for these models consisted of estimates of primary and ultimate biodegradation rates for 200 compounds, gathered in a survey of 17 biodegradation experts.  The two probability models correctly classified 90% of the compounds in their training set, whereas the two survey models calculated biodegradation rates for the survey compounds with R2 = 0.7.  These four models are intended for use in chemical screening and in setting priorities for further review.

Data source

Reference

Title:

EPI Suite Version 4.10

Year:

2012

Bibliographic source:

BIOWIN v4.10

Materials and methods

Test guideline

Guideline:

other: REACH guidance on QSARs Chapter R.6

Version / remarks:

May 2008

GLP compliance:

no

Test material

Test material information

Constituent 1

Reference substance name:

6,6'-diphenylfulvene

EC Number:

218-533-4

EC Name:

6,6'-diphenylfulvene

Cas Number:

2175-90-8

Molecular formula:

C18H14

IUPAC Name:

1,1'-(cyclopenta-2,4-dien-1-ylidenemethylene)dibenzene

Test material form:

solid

Results and discussion

% Degradation

Key result

Parameter:

probability of ready biodegradability (QSAR/QSPR)

Remarks on result:

not readily biodegradable based on QSAR/QSPR prediction

Applicant's summary and conclusion

Validity criteria fulfilled:

not applicable

Interpretation of results:

not readily biodegradable

Conclusions:

The study report describes a scientifically accepted calculation method for the ready biodegradability using the US-EPA software BIOWIN v1.40. No GLP criteria are applicable for the usage of this tool and the QSAR estimation is easily repeatable. The substance falls within the applicability domain of the QSAR model, hence, the result can be considered as reliable. The result is adequate for the regulatory purpose.

Executive summary:

The ready biodegradability of the substance was determined by the QSAR program BIOWIN v1.40 (EPIWIN software) by US-EPA.

BIOWIN estimates the probability of rapid aerobic and anaerobic biodegradation of an organic compound in the presence of mixed populations of environmental microorganisms. Two independent training sets were used to develop four mathematical models for predicting aerobic biodegradability from chemical structure.  All four of the models are based on multiple regressions against counts of 36 preselected chemical substructures plus molecular weight.  Two of the models, based on linear and nonlinear regressions, calculate the probability of rapid biodegradation and can be used to classify compounds as rapidly or not rapidly biodegradable.  The training set for these models consisted of qualitative summary evaluations of all available experimental data on biodegradability for 295 compounds.  The other two models allow semi-quantitative prediction of primary and ultimate biodegradation rates using multiple linear regression.  The training set for these models consisted of estimates of primary and ultimate biodegradation rates for 200 compounds, gathered in a survey of 17 biodegradation experts. The two probability models correctly classified 90% of the compounds in their training set, whereas the two survey models calculated biodegradation rates for the survey compounds with R²= 0.7. In this case, the substance was estimated to be not readily biodegradable.

Adequacy of the QSAR:

- QSAR model is scientifically valid.

- The substance falls within the applicability domain of the QSAR model.

- The prediction is fit for regulatory purpose.