Registration Dossier

Data platform availability banner - registered substances factsheets

Please be aware that this old REACH registration data factsheet is no longer maintained; it remains frozen as of 19th May 2023.

The new ECHA CHEM database has been released by ECHA, and it now contains all REACH registration data. There are more details on the transition of ECHA's published data to ECHA CHEM here.

Diss Factsheets

Administrative data

Key value for chemical safety assessment

Genetic toxicity in vitro

Description of key information

No genetic toxicity

Link to relevant study records
Reference
Endpoint:
in vitro gene mutation study in bacteria
Type of information:
experimental study
Adequacy of study:
key study
Study period:
08/01/2018-25/01/2018
Reliability:
1 (reliable without restriction)
Rationale for reliability incl. deficiencies:
guideline study
Qualifier:
according to guideline
Guideline:
OECD Guideline 471 (Bacterial Reverse Mutation Assay)
GLP compliance:
yes (incl. QA statement)
Type of assay:
bacterial reverse mutation assay
Specific details on test material used for the study:
TEST ITEM : 2-hydroxyethyl palmitate
OTHER NAMES / CODES : 2-HYDROXYETHYL PALMITATE LANOL P ECAILLES 35928C
IPL REGISTRATION NUMBER : 171214
BATCH NUMBER : 170301011115
APPEARANCE : white to off-white solid (flakes by naked eyes)
WATER CONTENT : <1%
PURITY / COMPOSITION : > 99% (dry matter) (Sponsor’s information)
SALT / BASE RATIO : unknown
MOLECULAR WEIGHT : unknown
DENSITY : 1.064 ± 0.013 (Sponsor’s information)
CORRECTION FACTOR : none, at the Sponsor’s request
STORAGE CONDITIONS* : room temperature (+20±5°C)
QUANTITY SUPPLIED : unknown
MANUFACTURING DATE : 27/02/2017
ANALYSIS DATE : 08/03/2017
STABILITY UNDER
STORAGE CONDITIONS : 2 years, up to 27/02/2019 for batch 170301011115
EXPIRY DATE : 27/02/2019
Target gene:
TA1535 and TA100 : his G 46
TA1537 : his C 3076
TA102 : His G 428
TA98 : his D 3052
Species / strain / cell type:
S. typhimurium TA 1535, TA 1537, TA 98, TA 100 and TA 102
Additional strain / cell type characteristics:
not specified
Metabolic activation:
with and without
Metabolic activation system:
hepatic microsomes from rat livers induced by Aroclor 1254 – Incubation period: ca. 44 h
Test concentrations with justification for top dose:
- First toxicity test : 0, 50, 150, 500, 1500 and 5000 µg/plate. Maximum dose according to OECD Guideline, i.e. 5000 μg/plate. (1 plate / dose).
An important precipitate hindering scoring was observed in all strains both without and with metabolic activation at the 2 highest tested doses of 1500 and 5000 μg/plate. Moderate and slight precipitate was also observed at 500 and 150 μg/plate, respectively. Therefore, due to the important precipitate, the maximum dose retained for the first mutagenicity assay was lowered down to 500 μg/plate in all strains both with and without metabolic activation.
- Mutagenicity test : 0, 5, 15, 50, 150, 500 µg/plate (3 plates / dose)
Vehicle / solvent:
The solvent used is Tetrahydrofuran (THF).
Prior to the implementation of the toxicity assay, trials for solubility were performed in sterile water, dimethylsulfoxide, ethanol and tetrahydrofuran (THF), at initial concentrations of 50 mg/mL, ranging from 100 to 6.25 mg/mL, ranging from 200 to 12.5 mg/mL, and at 250 mg/mL, respectively, at room temperature, demonstrating that the test is not soluble whatever the solvent. After heating the organic preparations at a temperature below the fusion point of 60-65°C (in accordance with the information provided by the Sponsor Representative), only the 250 mg/mL preparation in THF was clear. Noteworthy, when settled at room temperature, the solution in THF solidified.
For the preliminary toxicity assay, the test item 2-hydroxyethyl palmitate was dissolved in THF (Sigma, batch STBF9606V) at an initial concentration of 250 mg/mL, without following the qs (quantum satis) method (see § 11.1), in order to obtain the top dose of 5000 μg/plate when added at 20 μL/plate. Preparations for treatment were kept at ca. 37°C until treatments in order to anticipate the behavior of the test item in solution when temperature decreases.
For both mutagenicity assays, it was also prepared in THF but at an initial concentration of 25 mg/mL, without following the qs (quantum satis) method (see § 11.1), in order to obtain the top dose of 500 μg/plate when added at 20 μL/plate. As at this concentration the solution in THF did not solidified at room temperature, preparations for treatment were kept at room temperature until treatments.
Untreated negative controls:
not specified
Negative solvent / vehicle controls:
yes
True negative controls:
not specified
Positive controls:
yes
Remarks:
6 plates
Positive control substance:
9-aminoacridine
2-nitrofluorene
sodium azide
mitomycin C
Remarks:
Without S9 mix
Untreated negative controls:
not specified
Negative solvent / vehicle controls:
yes
True negative controls:
not specified
Positive controls:
yes
Remarks:
6 plates
Positive control substance:
benzo(a)pyrene
other: 2-anthramine, 2 μg/plate without pre-incubation, 1 μg/plate with pre-incubation (TA1535, TA1537, TA98, TA100)
Remarks:
With S9 Mix
Details on test system and experimental conditions:
Two test of mutagenicity were realised independently : first with metabolic activation, and the second without (with a preincubation period)

MEDIA : agar, supplemented with 10 % of 0.5 mM biotin histidine solution

TOXICITY TEST :
The toxicity assay was carried out in all the strains to be tested under the same conditions as the first mutagenicity test with and without metabolic activation (See below).
NUMBER OF REPLICATIONS : 1 plates per dose (no positive controls included)
DETERMINATION OF THE CYTOTOXICITY : Microscopic examination

MUTAGENICITY TEST :
WIth metabolic activation:
For each strain : 0.1 mL of a bacterial suspension from a culture agitated overnight at ca. 37°C and 20 μL of the test item at the relevant initial concentrations were successively added to 2 mL of top agar, supplemented with 10 % of 0.5 mM biotin histidine solution, maintained in a state of superfusion at ca. 45°C. The content of each tube was agitated, then spread out in a Petri plate containing 20 mL of minimum agar. The plates were then incubated at ca. 37°C for approximately 44 h. At the end of the expression time, colonies of revertants were counted for each plate.
Without metabolic activation : The method was the same as the one described above except that immediately before spreading in the plates, 0.5 mL of the S9-mix metabolic activation system was added in soft agar.
NUMBER OF REPLICATIONS: 3 plates per dose (6 for the positive controls)

Evaluation criteria:
Acceptance criteria :
- Determination of the media sterility (The sterility of the test item was berified in a test where it was incubated for 44 hours at 37°C, no bacterial growth was observed, the test item was considered sterile) : ok
- Determination of the efficiency of the metabolic activation : ok
- Validation of the solvent control (The frequencies of spontaneous revertants (solvent controls) were within the limits generally observed under our experimental conditions) : ok
- Determination of the sensitivity of the strains (Statistically and biologically significant increases in the numbers of revertants were observed in the presence of positive reference test substances) : ok
- For the test, 5 doses available, as required by OECD guideline : ok (5 here)
- For the test, At least 2 plates per dose were available for the assessment of mutagenicity : ok (3 here)
Statistics:
Data were analysed by means of Dunnett's method (Mahon et al, 1989) allowing the comparison of the mean value for each dose to the mean value for the corresponding solvent control. Statistical methods may be used as an aid in evaluating the test results. However, statistical significance should not be the only determining factor for a positive response and biological relevance of the results should be considered firs
Key result
Species / strain:
S. typhimurium TA 1535
Metabolic activation:
with and without
Genotoxicity:
negative
Cytotoxicity / choice of top concentrations:
no cytotoxicity, but tested up to precipitating concentrations
Vehicle controls validity:
valid
Untreated negative controls validity:
not specified
Positive controls validity:
valid
Key result
Species / strain:
S. typhimurium TA 1537
Metabolic activation:
with and without
Genotoxicity:
negative
Cytotoxicity / choice of top concentrations:
no cytotoxicity, but tested up to precipitating concentrations
Vehicle controls validity:
valid
Untreated negative controls validity:
not specified
Positive controls validity:
valid
Key result
Species / strain:
S. typhimurium TA 98
Metabolic activation:
with and without
Genotoxicity:
negative
Cytotoxicity / choice of top concentrations:
no cytotoxicity, but tested up to precipitating concentrations
Vehicle controls validity:
valid
Untreated negative controls validity:
not specified
Positive controls validity:
valid
Key result
Species / strain:
S. typhimurium TA 102
Metabolic activation:
with and without
Genotoxicity:
negative
Cytotoxicity / choice of top concentrations:
no cytotoxicity, but tested up to precipitating concentrations
Remarks:
Toxicity test : statistically significant decreases were noted in the first assay in strain TA102 at 500 and 150 μg/plate without and with metabolic activation, respectively. Nevertheless, these effects had no meaning in terms of mutagenicity hazard
Vehicle controls validity:
valid
Untreated negative controls validity:
not specified
Positive controls validity:
valid
Key result
Species / strain:
S. typhimurium TA 100
Metabolic activation:
without
Genotoxicity:
negative
Cytotoxicity / choice of top concentrations:
no cytotoxicity, but tested up to precipitating concentrations
Vehicle controls validity:
valid
Untreated negative controls validity:
not specified
Positive controls validity:
valid
Key result
Species / strain:
S. typhimurium TA 100
Metabolic activation:
with
Genotoxicity:
ambiguous
Remarks:
a statistically but not biologically significant increase in the number of revertants was observed at the intermediate dose of 15 μg/plate.
Cytotoxicity / choice of top concentrations:
no cytotoxicity, but tested up to precipitating concentrations
Vehicle controls validity:
valid
Untreated negative controls validity:
not specified
Positive controls validity:
valid
Additional information on results:
Mutagenicity test :
In two independent assays performed both with and without metabolic activation (the second assay with S9-mix was performed according to the pre-incubation protocol), no biologically significant increases in the mean number of revertants were noted in the five Salmonella typhimurium strains TA1535, TA1537, TA98, TA100 and TA102 tested, in the presence of 2-hydroxyethyl palmitate.
The test item 2-hydroxyethyl palmitate is thus not mutagenic in these conditions.
In the first assay in strain TA100 in presence of metabolic activation without pre-incubation, a statistically but not biologically significant increase in the number of revertants was observed at the intermediate dose of 15 μg/plate. The mean value for revertants was very slightly higher than the upper bound of the intervals of historical data for negative controls, i.e. 162 vs. 161. However, the threshold ratio for a biologically significant effect (set at 2 in this strain) was not reached, with an induction ratio of 1.3, and no dose-effect relationship was observed. Furthermore, the second assay performed under the more sensitive pre-incubation method was clearly negative. The test item 2-hydroxyethyl palmitate was thus considered as not mutagenic in this experimental condition.
Conclusions:
The mutagenic activity of the test item 2-hydroxyethyl palmitate (Batch 170301011115) sponsored by SEPPIC was assessed by means of the Ames’ test in the five Salmonella typhimurium strains TA1535, TA1537, TA98, TA100 and TA102 tested both in presence and in absence of metabolic activation, in two independent assays according to OECD guideline (OECD 471, 1997), using the maximum dose compatible with the solubility of the test item in the test system, i.e. 500 μg/plate. The validity criteria for the assay were fulfilled. The study is thus valid. Under these experimental conditions, no mutagenic activity was revealed.
Endpoint conclusion
Endpoint conclusion:
no adverse effect observed (negative)

Genetic toxicity in vivo

Description of key information

No genetic toxicity

Endpoint conclusion
Endpoint conclusion:
no adverse effect observed (negative)

Additional information

Genetic toxicity in vitro

Justification for grouping of substances and read-across

The Glycol ester category covers esters of an aliphatic diol (ethylene glycol (EG), propylene glycol (PG) or 1,3-butyleneglycol (1,3-BG)) and one or two carboxylic fatty acid chains. The fatty acid chains comprise carbon chain lengths ranging from C6 to C18, mainly saturated but also mono unsaturated C16 and C18, branched C18 and epoxidized C18.

 

The available data allows for an accurate hazard and risk assessment of the category and the category concept is applied for the assessment of environmental fate and environmental and human health hazards. Thus, where applicable, environmental and human health effects are predicted from adequate and reliable data for source substance(s) within the group by interpolation to the target substances in the group (read-across approach) applying the group concept in accordance with Annex XI, Item 1.5, of Regulation (EC) No 1907/2006. In particular, for each specific endpoint the source substance(s) structurally closest to the target substance is/are chosen for read-across, with due regard to the requirements of adequacy and reliability of the available data. Structural similarities and similarities in properties and/or activities of the source and target substance are the basis of read-across.

A detailed justification for the grouping of chemicals and read-across is provided in the technical dossier (see IUCLID Section 13).


Data Matrix

CAS #

627-83-8 (a)

68583-51-7

84988-75-0

624-03-3

91031-31-1 (b)

151661-88-0

853947-59-8

1323-39-3

4219 -49 -2 

Genetic toxicity (mutagenicity) in bacteria in-vitro

Negative

Negative

Negative

RA: CAS 91031-31-1

RA: CAS 627-83-8

Negative

Negative

Negative

Negative (c)

Negative 

Genetic toxicity (cytogenicity) in mammalian cells in-vitro

RA: CAS 853947-59-8

RA: CAS 853947-59-8

RA: CAS 853947-59-8

RA: CAS 853947-59-8

RA: 853947-59-8

--

Negative

--

  

RA: CAS 853947-59-8

Genetic toxicity (mutagenicity) in mammalian cells in-vitro

RA: CAS 91031-31-1

RA: CAS 91031-31-1

RA: CAS 91031-31-1

RA: CAS 91031-31-1

Negative

--

--

--

 

RA: CAS 91031-31 -1

Genetic toxicity in vivo

RA: CAS 151661-88-0

RA: CAS 151661-88-0

RA: CAS 151661-88-0

RA: CAS 151661-88-0

--

Negative

--

--

  

--

(a) Category members subject to the REACh Phase-in registration deadline of 31 May 2013 are indicated in bold font.

(b) Substances that are either already registered under REACh or not subject to the REACh Phase-in registration deadline of 31 May 2013 are indicated in normal font.

(c) Original data is not available.

For all category members registered under REACh a full data set for each endpoint is provided. For substances not subject to the current REACh Phase-in registration, lack of data for a given endpoint is indicated by "--".

 


Genetic toxicity (mutagenicity) in bacteria in-vitro

CAS 627-83-8

A bacterial gene mutation assay (Ames test) was performed with ethylene distearate following OECD guideline 471 and in compliance with GLP (Grötsch, 1996). The plate incorporation procedure was performed with Salmonella typhimurium strains TA 100, TA 1535, TA 1537 and TA 98 in the absence and presence of metabolic activation (Aroclor 1254-induced rat liver S9-mix). Two independent experiments were conducted each in triplicates at concentrations from 8 to 5000 µg/plate (vehicle: DMSO). No increase in the number of revertant colonies was noted in any of the bacterial strains, with and without metabolic activation system. No cytotoxicity was observed up to the highest dose tested. The included positive and negative controls showed the expected results, however instead of the solvent control DMSO, water was included as negative control substance. Under the conditions of the study, the test substance did not induce mutations in the bacterial mutation assay in the absence and presence of a metabolic activation system in any of the strains tested.

CAS 68583-51-7

Two bacterial gene mutation assays were performed with Decanoic acid, mixed diesters with octanoic acid and propylene glycol following OECD guideline 471 and in compliance with GLP (Banduhn, 1991a; Ebert, 1995). The strains Salmonella typhimurium TA 98, TA 100, TA 1535, TA 1537 and TA 1538 were tested in two independent experiments according to the plate incorporation procedure at concentrations from 8 to 5000 µg/plate (vehicle: suspension medium Tween 80/bidest. water) with and without a metabolic activation system (Aroclor 1254-induced rat liver S9-mix). No increase in the number of revertant colonies was noted in any strain of bacteria tested, with and without a metabolic activation system. No cytotoxicity was observed up to the highest dose tested. The included positive and negative controls showed the expected results and were therefore considered as valid. Under the conditions of this study, the test substance did not induce mutations in the bacterial mutation tests in the absence and presence of a metabolic activation system in any of the strains tested (Banduhn, 1991a). A further experiment in the tester strains Salmonella typhimurium TA 1535, TA 1537, TA 98 and TA 100 confirmed the above described results (Ebert, 1995). In concentrations from 50 to 5000 µg/plate with and without a metabolic activation system, no cytotoxicity was apparent and no increase in the number of revertant colonies was observed in any of the strains tested.

Thus, the test substance did not induce mutations in the bacterial mutation assay in the absence and presence of a metabolic activation system in any of the strains tested.

CAS 84988-75-0

A study investigating the genetic toxicity in vitro of Fatty acids, C14-18 and C16-18-unsatd., esters with propylene glycol is available. The study was conducted according to OECD guideline 471 under GLP conditions (Banduhn, 1991b). In two independent experiments, the tester strains Salmonella typhimurium TA 98, TA 100, TA 1535, TA 1537 and TA 1538 were tested according to the plate incorporation procedure. Concentrations from 8 to 5000 µg/plate were investigated with and without a metabolic activation system (Aroclor 1254-induced rat liver S9-mix). No increase in the number of revertant colonies was noted in any of the bacterial strains, with and without metabolic activation system. No cytotoxicity was observed up to the highest dose tested. The included positive and negative controls showed the expected results and were therefore considered as valid. Thus, under the conditions of this study, the test substance did not induce mutations in the bacterial mutation tests in the absence and presence of a metabolic activation system in any of the strains tested.

CAS 91031-31-1

The gene mutation properties of the category member Fatty acids, C16-18, esters with ethylene glycol were determined according to OECD guideline 471 under GLP conditions (Banduhn, 1991c). The tester strains, Salmonella typhimurium TA 98, TA 100, TA 1535, TA 1537 and TA 1538 were used. The main study was performed in triplicates each in two independent experiments according to the plate incorporation procedure at concentrations from 8 to 5000 µg/plate (vehicle: Tween 80; 1:1 (w/w) dilution with water) with and without a metabolic activation system (Aroclor 1254-induced rat liver S9-mix). No increase in the number of revertant colonies was noted in any of the bacterial strains, with and without metabolic activation system. No cytotoxicity was observed up to the highest dose tested. Precipitation was noted at the highest test concentration of 5000 µg/plate. The positive and negative controls showed the expected results and were therefore considered as valid. Under the conditions of this study, the test substance did not induce mutations in the bacterial mutation tests in the absence and presence of a metabolic activation system in any of the strains tested.

Genetic toxicity (cytogenicity) in mammalian cells in-vitro

CAS 853947-59-8

An in vitro mammalian chromosome aberration test was conducted with C8-10, 1,3-Butandiolester in accordance with OECD guideline 473 under GLP conditions (Dechert, 1997). The induction of structural chromosome aberrations was evaluated in vitro in Chinese hamster lung fibroblasts (V79) cells, incubated for 18 and 28 h with and without a metabolic activation system (S9-mix from rats treated with Aroclor 1245). Concentrations of 10-100 µg/mL (18 h incubation) and 80 and 100 µg/mL (28 h incubation) of the test substance in the vehicle ethanol were applied. The solubility limit of the test substance in the vehicle ethanol in the culture medium was determined to be 100 µg/mL. In the first experiment without metabolic activation, the negative controls exhibited a mitotic index of 2.0% only and the experiment was therefore repeated. Thereafter, the negative as well as the positive controls showed the expected results and were within the range of historical control data. The frequency of polyploidy cells with and without metabolic activation was within the expected range (< 10%). In the experiments both with and without metabolic activation, a systematic influence of the test substance was observed, which led to a reduction in the mitotic index. No statistically or biologically significant increase in the incidence of chromosome aberrations was observed.

Therefore, under the conditions of the study, the test substance did not show clastogenic activity in this chromosomal aberration test with and without metabolic activation performed in Chinese hamster lung fibroblasts in vitro.

 

Genetic toxicity (mutagenicity) in mammalian cells in-vitro

CAS 91031-31-1

The in vitro mammalian cell gene mutation study of Fatty acids, C16-18, esters with ethylene glycol was carried out according to OECD guideline 476 under GLP conditions (Verspeek-Rip, 2010). Gene mutations in the thymidine kinase locus were investigated in L5178Y mouse lymphoma cells in the presence and absence of a metabolic activation system (Phenobarbital/β-naphtoflavone-induced rat liver S9). In the first experiment, cells were exposed for 3 h to test substance at concentrations of 0.1-333 µg/mL (in DMSO) with and without metabolic activation. Concentrations of the second experiment without metabolic activation for an exposure time of 24 h ranged from 3-175 µg/mL and with metabolic activation (3 h; 12% S9-mix) from 0.1-333 µg/mL. The vehicle and positive controls in the study showed the expected results and were within the range of historical control data. No cytotoxicity was observed up to the precipitating concentration of 100 µg/mL and up to 333 µg/mL, respectively. There was no significant increase in the number of forward mutations at the thymidine kinase locus of L5178Y mouse lymphoma cells treated with the test material, neither in the presence nor in the absence of a metabolic activation system. Under the conditions of the study, Fatty acids, C16-18, esters with ethylene glycol did not show gene mutation activity in this test performed in L5178Y mouse lymphoma cells in vitro.

 

Genetic toxicity in vivo

CAS 151661-88-0

The in vivo micronucleus assay of Fatty acids, C18 and C18 unsatd., epoxidized, ester with ethylene glycol was carried out according to OECD guideline 474 under GLP conditions (Banduhn, 1990). Based on the results of a preliminary dose range finding study, the test substance diluted in arachis oil was administered at 3000, 4000 and 5000 mg/kg bw as single oral doses to groups of 6 male and female CFW1 mice, observed for 24, 48 and 72 h post-dose. A concurrent negative control with the vehicle alone and a positive control group given cyclophosphamide was included in the study. No mortality and no signs at clinical examinations were reported. The test substance did not induce a statistically significant increase in the number of micronucleated polychromatic erythrocytes in the bone marrow of the animals. The negative and positive controls showed the expected results. Therefore, under the conditions of the study, Fatty acids, C18 and C18 unsatd., epoxidized, ester with ethylene glycol did not induce chromosomal mutations in the bone marrow of mice.

 

Conclusion for genetic toxicity

In summary, several studies investigating the genetic mutation in bacteria in-vitro are available within the Glycol Ester category for Ethylene distearate (CAS No. 627-83-8), Decanoic acid, mixed diesters with octanoic acid and propylene glycol (CAS No. 68583-51-7), Fatty acids, C14-18 and C16-18-unsatd., esters with propylene glycol Fatty acids (CAS No. 84988-75-0) and C16-18, esters with ethylene glycol (CAS No. 91031-31-1), all providing negative results. Furthermore, no cytogenicity in mammalian cells in-vitro (CAS No. 853947-59-8), no mutagenicity in mammalian cells in-vitro (CAS No. 151661-88-0) and no genetic toxicity in-vivo (CAS No. 91031-31-1) was observed with members of the Glycol Ester category.

Therefore, no properties for genetic toxicity were observed within the Glycol Ester group for any member.

For a detailed reference list please refer to the CSR or IUCLID section 13.


Short description of key information:
Negative results in Salmonella typhimurium TA 98, TA 100, TA 1535 and TA 1537, with and without metabolic activation (OECD 471, GLP).
Negative results in mammalian chromosomal aberration test with Chinese hamster lung cells (OECD 473, GLP, category approach).
Negative results in mammalian cell gene mutation tests using Chinese hamster ovary cells, with and without metabolic activation (OECD 476, GLP).
Negative results in mammalian erythrocyte micronucleus test in vivo (OECD 474, GLP, category approach).

Endpoint Conclusion: No adverse effect observed (negative)

Justification for classification or non-classification

According to Article 13 of Regulation (EC) No. 1907/2006 "General Requirements for Generation of Information on Intrinsic Properties of substances", information on intrinsic properties of substances may be generated by means other than tests e.g. from information from structurally related substances (grouping or read-across), provided that conditions set out in Annex XI are met. Annex XI, "General rules for adaptation of this standard testing regime set out in Annexes VII to X” states that “substances whose physicochemical, toxicological and ecotoxicological properties are likely to be similar or follow a regular pattern as a result of structural similarity may be considered as a group, or ‘category’ of substances. This avoids the need to test every substance for every endpoint". Since the group concept is applied to the members of the Glycol Ester Category, data will be generated from representative reference substance(s) within the category to avoid unnecessary animal testing. Additionally, once the group concept is applied, substances will be classified and labeled on this basis.

Therefore, based on the group concept, all available data on genetic toxicity do not meet the classification criteria according to Regulation (EC) 1272/2008 or Directive 67/548/EEC, and are therefore conclusive but not sufficient for classification.