Working out your hazards using read-across information

Substances with similar chemical structures may have similar physico-chemical, toxicological, ecotoxicological and/or environmental fate properties and are considered as a group of substances.

The following are our similarity criteria:

• the occurrence of common functional groups (for example, aldehyde, epoxide, ester and specific metal ion)
• the presence of common constituents or chemical classes and similar carbon range numbers
• the availability of a common mechanism of action or adverse outcome pathway
• the likelihood of common precursors and/or breakdown products via physical or biological processes that result in structurally similar chemicals (for example, the ‘metabolic pathway approach’ of examining related chemicals such as acid/ester/salt)

The read-across method uses information from 1 or more source chemicals to predict the hazard characteristics of a target chemical. If the source and target chemicals are similar enough, you can use the same information for hazard characterisation by applying our concept of substance grouping and our similarity criteria.

Source chemical/s can be from an analogue or category approach. Using data on analogues or categories means you can use the read-across method to work out the hazards of your chemical.

The analogue approach uses a single chemical as the source chemical. This single chemical is your ‘analogue chemical’. You can use its chemical information data to estimate the same hazard endpoints for your target chemical.

The category approach uses a group of chemicals as the source chemicals. With this approach, you look at the properties of the chemicals in a category as a whole. Not every chemical in the category needs to have experimental data for every endpoint. The overall data for that category should be adequate for your hazard characterisation.

For either approach, you need to consider the substance grouping similarity criteria for your read-across method. Additionally for the category approach, you should consider any steady change in properties across the category. These are often observed in physico-chemical properties (such as boiling point range). They may not be very similar but they can demonstrate a trend, which allows you to predict the hazard characteristic for your chemical.

Since both approaches can be used in identifying your chemical’s hazards, you’ll need to gather the same information and keep the same records. We’ll assess your justification on a case-by-case basis.

Follow our steps to assist you in considering the suitability of your read-across approach.

This section outlines the suitability of your read-across method using the analogue and category approaches.

You’ll need to complete up to 6 steps to assist in your read-across.

Step 1: Identify potential analogues or identify chemicals using the category/grouping approach

Using an analogue approach

If you don’t already have an analogue, the following tools can help you look for chemicals with similar structures or sub-structures.

• OECD QSAR Toolbox
• SciFinder
• ChemIDPlus
• US EPA AIM
• US EPA ChemACE

These tools often determine a quantitative similarity index. This accounts for the presence and absence of structural fragments between the target and the source chemicals.

You can also find potential analogues from these assessments:

• AICIS Chemical Information
• OECD Existing Chemicals Database
• Canadian Government Chemicals Management Plan
• US EPA
• WHO IPCS

Other sources include:

• Research Institute for Fragrance Materials (RIFM)
• CONCAWE
• American Petroleum Institute
• Hydrocarbon Solvents Producers Association
• Ecological and Toxicological Association of Dyes and Organic Pigments
• European Committee of Organic Surfactants and their Intermediates

If you find a potential analogue, go to Step 2.

Using a category approach

This approach uses a group of chemicals with properties that are likely to be similar or follow a regular pattern, usually due to:

• common functional groups
• common constituents or chemical classes
• incremental changes (for example chain-length)
• likely common precursors and/or breakdown products

Step 2: Write a rationale to explain the suitability of the read-across information

In your rationale, you should:

• provide scientific reasons by specifying the properties of the proposed analogue or category chemicals used to  predict the target chemical’s (eco)toxicological or environmental fate properties
• account for any differences in the chemical structures
• consider the analogue or category approach for each of the hazard characteristics you are trying to predict. You may have to use information on 1 or more analogue or category chemicals to read-across to the different (eco)toxicity endpoints

The specific elements below can be used to support your rationale and for the read-across to be deemed suitable. While not all of these may support your hypothesis, you will still need to consider and document them.

If you are using a category approach, you may find that the chemical’s properties may not be very similar. They may follow a regular trend, which allows you to predict the hazard characteristic for your chemical.

Similarity in structure

The main basis for using read-across information from analogue or category chemicals to a target chemical is structural similarity. The presence or absence of certain features in a chemical's structure could influence the (eco)toxicity of the chemical. Chemicals with similar structures may have similar biological and environmental activity.

You should:

• define the structural similarities between the target chemical and analogue or category chemicals.

Molecular similarity in structure includes:

• common chemical class and subclasses
• similar molecular frameworks
• similar number of carbon atoms
• common structural fragments
• extended structural groups (side chains)

Similarity in physico-chemical properties

You should:

• define the similarities between the physico-chemical and molecular properties of the target, and analogue or category chemicals
• pay attention to functional groups or extended fragment, classes of chemicals and sub-class of the target, and analogue or category chemicals
• identify any structural differences between the target, and analogue or category chemicals and establish how differences may or may not affect toxicity

Physico-chemical properties can affect the bioavailability, persistence and partitioning behaviour in environmental and biological systems. Molecular structure and physico-chemical properties combined can affect toxicity (for example, volatility, solubility and reactivity).

There are more details on physico-chemical properties and their significance in the hazards of chemicals in the OECD Guidance on Grouping of Chemicals.

Similarity in metabolism/biodegradation

You should:

• define the similarity in potential metabolic products between the target, and analogue or category chemicals

Physical and biological processes, such as metabolic or degradation pathways, can produce precursor or breakdown products. Any similarity in toxicokinetic information between the analogue or category chemicals and target chemical is particularly useful for predicting chronic endpoints for the target chemical.

There are more details on how toxicokinetic information can influence the choice of suitable analogue or category in the OECD Guidance on Grouping of Chemicals.

Similarity in mechanisms of action (MOA) and adverse outcome pathway (AOP)

You should:

• provide the information used to determine how the target chemical and the analogue or category have similar MOA/AOP.

Certain features in a chemical structure have known modes/mechanisms of action (MOA). The concept of an adverse outcome pathway (AOP) also incorporates MOA understanding. Some (eco)toxicity endpoints have simpler modes/mechanisms while some have multiple modes/mechanisms. Mechanisms include skin sensitisation and mutagenicity.

There are more details on MOA and its significance in determining suitable read-across information in the OECD Guidance on Grouping of Chemicals.

Similarity in reactivity/stability

You should:

• provide the information used to determine how the target chemical and the analogue or category chemicals have the same reactivity or stability

Reactivity or stability between the analogue or category chemicals and target chemical is important for:

• considering the persistence of a chemical
• determining the potential for the chemical’s reaction products to be inherently toxic

Additional requirements for category approach

When using this approach, you should:

• explain the structural and mechanistic similarities between chemicals in the category/group
• make sure your grouping is clear and that you sufficiently describe the boundaries of applicability domain (you can include similarity in use profiles of the members)
• describe and justify the unimportance of any existing differences of deviations in properties that should not significantly affect their trends

There are more details on developing a category hypothesis in OECD Guidance on Grouping of Chemicals.

Step 3: Gather available data

You need to gather all the published and unpublished data you find for the potential analogue or category chemicals. You may need to consider data ownership and permissions. If possible, you should collect the following information for potential analogue or category:

• chemical name and CAS number
• molecular weight, composition and impurity profiles
• chemical structure, functional groups, carbon chain length, isomerisation and linearity/branching data
• log K_ow, water solubility, vapour pressure, particle size or molecular descriptors (for example, electrophilicity)
• (eco)toxicological data
• mode or mechanism of action (MOA) or adverse outcome pathway (AOP) properties of the analogue(s) (for example, description of the toxicological endpoints and biological events through comparison with the target organ effects)
• toxicokinetics
• environmental fate
• degradation and breakdown products information

Step 4: Use the data to scientifically justify your chemical choice using the criteria for the required hazard endpoints

You need to consider specific aspects of hazard endpoints when working out the suitability of your read-across.

We chose the following examples because there are well-defined mechanisms available for them for several classes of chemicals.

Example 1: Mutagenicity and Genotoxicity justifications

The mutagenicity potential of a chemical is linked to its ability to react with DNA in cells. If DNA is damaged and not repaired, the mutations lead to changes in the nucleotide sequence. Larger mutations result in changes to chromosomes including alterations to the number and structure.

Several databases with mutagenicity data for chemicals are available. Information from these sources enabled the mechanistic interactions and identification of potential mutagenic and non-mutagenic groups of chemicals based on DNA reactivity.

You can predict the mutagenicity potential of the target chemical using analogue(s). You compare the chemical structure (on its potential to react with DNA) to similar chemical classes based on known experimental data or mechanisms of mutagenicity.

Some chemicals require metabolic activation to be considered mutagenic. Mutagenicity is usually activated in an Ames test (bacterial reverse mutation test) under metabolic activation conditions. You will need to consider if metabolic activation is an important consideration for your target chemicals and analogue or category.

You can use the OECD QSAR Toolbox as a resource.

Example 2: Skin sensitisation justification

The skin sensitisation potential of an organic chemical is linked to its ability to covalently bind and react with skin proteins to form covalently linked conjugates. The immune system can then recognise the conjugates.

Several databases containing skin sensitisation data are available. Information from these sources enabled the mechanistic interactions and identification of potential skin sensitiser and non-sensitiser groups of chemicals based on protein reactivity.

For analogue or category chemicals with test data available, you can use these data in read-across as part of a weight-of-evidence (WoE) approach. This is to confirm results of skin sensitisation and understand the mechanisms of the skin reactions.

Metabolism or abiotic transformations can also activate the ability to bind to protein.

Use the OECD QSAR Toolbox as a resource to identify potential metabolic/abiotic transformations. You’ll need to assess the reliability of the QSAR results on a case-by-case basis.

You can predict the skin sensitisation potential of the target chemical in 2 steps:

1. Inspect the chemical structure/s of the analogue or category for its potential to react with skin proteins.
2. Compare it with similar skin sensitiser or non-sensitiser chemical classes based on known experimental data or mechanisms of mutagenicity.

You can use the OECD QSAR Toolbox as a resource.

Step 5: Maintain your documentation

You must:

• maintain documentation to provide clear evidence of the approach you used to identify suitable analogue or category
• support your hypothesis with data to justify why the analogue or category is similar enough to the target chemical

If we conduct an audit, you’ll need to show us your documentation setting out how your information meets our criteria for identifying suitable analogue or category.

For example, if you choose an analogue or category because it has a similar metabolic pathway as the target chemical, you need to have toxicokinetics or metabolism studies to support this.

How to set out your information

• Hypothesis: describe the structural and mechanistic similarities between the target chemical and analogue or category. Highlight uncertainties and any actions to reduce them.
• Analogue or category information: provide CAS numbers, chemical names, chemical structure, functional groups, composition and uses of the analogue or category.
• Purity/impurity: define the purity/impurity profile of the target chemical and analogue or category.
• Justification: provide data and/or reasons to support your analogue or category hypothesis. This can be in vitro and in vivo studies, MOA/AOP information, computational and non-computational theoretic models, and bioavailability and reactivity profiles.
• Applicability domain: identify the structural requirements and ranges of physico-chemical, environmental fate or (eco)toxicological properties within which reliable estimations can be made. You should define structural alerts and functional group(s) of the analogue or category where possible to set the boundaries that are used as inclusion/exclusion criteria. This can include:
  • carbon number range
  • branching and position of branching
  • aromatic content
  • cyclicity
  • the presence of stereoisomers
  • position and frequency of double bonds
  • steric hindrance
• Reasoning: describe the process you used in reading across information to the target chemical.

There are more reporting formats in the  OECD Guidance on Grouping of Chemicals.

Step 6: Optional data matrix for category/grouping process

You may build a data matrix with the available data for each category member. The data matrix should show the similarities/trends or the data gaps within the category.

You can check Case study 2 for an example or see the OECD Guidance on Grouping of Chemicals.

Specific issues for certain chemical types

You may need to consider specific issues of certain types of chemicals when identifying their suitable analogue or category. In this section, we discuss specific issues for UVCBs and polymers.

More examples of chemical types with specific issues are available in the OECD Guidance on Grouping of Chemicals.

Specific issues for UVCBs

There are many types of UVCBs with diverse properties. There is no common approach to identifying suitable analogue or category for UVCBs.

You should consider the following elements when you are developing your category hypothesis or gathering information to determine suitable analogues:

• UVCBs are made of many chemicals, so a simple chemical structure or a specific molecular formula cannot represent them.
• The way the CAS number is defined may vary.
• Many UVCBs are of natural origin and can’t be fully separated into their constituent chemicals (for example, crude oil, coal, plant extracts and reaction products).
• You need to state the composition of the UVCB and the relationship between the various category members. You may present this in a table of the category members indicating changing elements and those staying constant within the category, providing some indication of structure.

You may need to provide more justifications and assessments for the read-across method for UVCBs to account for the:

• increasing complexity of the substance’s composition
• composition’s impacts on the predictions

Specific issues for polymers

A polymer can differ in molecular weight, monomer composition and percentage of monomers with low molecular weight (<1000g/mol). You should consider the following elements when determining the suitability of an analogue or category for a polymer:

• Similarity in structure
• Similarity in functional groups
• Similarity in molecular weight range
• Use of monomer data

Similarity in structure

An analogue or category is suitable if it contains most of the monomers present in the target polymer. We allow differences in monomer composition if the:

• functional groups of concern are still present
• polymer class is similar to that of the target polymer

Any changes in monomer composition should not significantly change the physico-chemical properties compared with that of the target polymer. For example, water solubility of analogue or category and target polymers should be in the range of 50-200%.

Similarity in functional groups

An analogue or category is suitable if it contains all the functional groups of concern of the target polymer. This includes anionic, cationic and potentially cationic groups. The density of the functional groups (functional group equivalent weight or FGEW) should also be within close range of the target polymer. If the FGEW is significantly greater than that of the target polymer, it won’t be a suitable analogue or category.

Similarity in molecular weight range

An analogue or category is suitable if the number average molecular weight (NAMW) is within close range of the target polymer. The percentage of low molecular weight species less than 500 g/mol and less than1000 g/mol (if known) should also be within close range of the target polymer.

Use of monomer data

If suitable analogue or category with data are not available, you may estimate the hazard profile of the target polymer using (eco)toxicological data on its monomers, as long as the polymer doesn’t have any additional functional groups of concern.

These case studies show:

• similarities between the target chemical and analogue chemical(s) (based on available information)
• category hypothesis (for category approach)
• justification for analogue suitability
• the assessment outcome

Case study 1: analogue approach

The target chemical

Carbamic acid, [(butylthio)thioxomethyl]-, butyl ester. The applicant proposed introduction volume up to 200 tonnes and use as a flotation agent in mineral processing (gold) up to 70% concentration (solution).

Information available for the target chemical and proposed analogue chemical

Chemical information	Target chemical	Analogue chemical
CAS number	1001320-38-2	39142-36-4
Structure
Chemical name	Carbamic acid, [(butylthio)thioxomethyl]-, butyl ester	Thioimidocarbonic acid ((HO)C(O)NHC(S)(OH), O1,O3-dibutyl ester
Molecular weight	249	233
Physico-chemical properties
Melting point	30.85 ± 0.5°C (solid at RT)	-7.15°C (liquid at RT)
Boiling point	211.85 ± 0.5°C	238°C
Vapour pressure	0.98 x 10-5 kPa at 25°C	2.8 x 10-5 kPa at 25°C
Water solubility	2.58 x 10-3g/L at 20 ± 0.5°C	6.48 x 10-2 g/L at 20°C
Partition co-efficient	logPow = 4.09	logPow at 20°C = 3.20
Absorption/Desorption	logKoc = 3.72	logKoc = 3.13
Acute toxicity
Oral	LD50>2000 mg/kg bw	LD50 ~ 500 mg/kg bw
Dermal	None provided	LD50 > 2000 mg/kg bw
Corrosion/Irritation
Skin irritation	None provided	Slightly irritating
Eye irritation	None provided	Slightly irritating
Sensitisation
Skin sensitisation	None provided	Limited evidence of sensitisation (not classified)
Repeat dose toxicity
Oral	None provided	No NOEL/NOAEL was determined, as toxicologically significant effects were observed at all dose levels (LOAEL = 15 mg/kg/day)
Genotoxicity
In vitro Ames	Negative	Negative
In vivo Chromosomal Aberration	None provided	Negative
In vivo Micronucleus	None provided	Weakly genotoxic at maximum tolerated dose (320 mg/kg bw/day)

The certificate applicant provided the following justification

We consider the structural difference between the target and analogue chemical to be insignificant for the oral repeated dose toxicity endpoint. The reason for this is as follows.

We do not expect the substitution of oxygen for a sulphur atom is to markedly alter the results of the physico-chemical properties because:

1. The functionality of the substance remains unchanged.
2. The difference in the molecular weights of the target and analogue chemicals (249.39 and 233.33 g/mole) is small.
3. We do not expect the insertion of a sulphur group instead of oxygen to markedly reduce the conformational flexibility of the molecule as a whole.

Assessment outcome

The analogue chemical is acceptable for oral repeated dose toxicity endpoint.

The applicant’s justification is reasonable. They discussed structure and physico-chemical properties.

We do not expect the difference in potential metabolites to be significant for risk assessment following repeated oral exposure to the chemical. This is because the analogue chemical is more toxic than the target chemical via the oral route.

The analogue chemical is classified as ‘Specific target organ toxicity (repeated exposure) – Category 1’ (H372: Causes damage to organs through prolonged or repeated exposure). Thus, as a worst-case scenario, we recommend the target chemical for the same hazard classification.

We made recommendations and secondary notification conditions (for highly controlled uses) for the target chemical.

Case study 2: category approach

The chemicals

We used the category approach to fill data gaps for a group of chemicals.

The group comprised 6 isomers of xylidene:

• benzenamine, 2,3-dimethyl- (2,3-DMA) (CAS number 87-59-2)
• benzenamine, 2,4-dimethyl- (2,4-DMA) (CAS number 95-68-1)
• benzenamine, 2,5-dimethyl- (2,5-DMA) (CAS number 95-78-3)
• benzenamine, 2,6-dimethyl- (2,6-DMA) (CAS number 87-62-7)
• benzenamine, 3,4-dimethyl- (3,4-DMA) (CAS number 95-64-7)
• benzenamine, 3,5-dimethyl- (3,5-DMA) (CAS number 108-69-0)

Chemical information	Category chemicals
CAS number	87-59-2	87-62-7	95-64-7
Structure
Chemical name	Benzenamine, 2,3-dimethyl-	Benzenamine, 2,6-dimethyl-	Benzenamine, 3,4-dimethyl-
Molecular formula	C8H11N	C8H11N	C8H11N
Acute toxicity
Oral	No data available	Harmful if swallowed (Xn; R22) (TG 423) Median lethal dose – LD50 – 300-2000 mg/kg bw, OECD 2012.	No data available
Dermal	No data available	Harmful in contact with skin (Xn; R21)	No data available
Inhalation	No data available	Harmful by inhalation (Xn;R20)	No data available
Observation in humans	No data available	Observed methaemoglobinaemia	No data available
Corrosion/Inhalation
Respiratory irritation	No data available	Irritating to respiratory system (Xi; R37)	No data available
Skin irritation	No data available	Irritating to skin (Xi;R38) (TG404) erythema, severe oedema, scale formation	No data available
Eye irritation	No data available	(TG405) slightly irritate the eyes	No data available
Sensitisation
Skin sensitisation	No data available	(TG 429) No dermal skin sensitisation	No data available
Repeat dose toxicity
Oral	(TG407) NOAEL 2 - 12 mg/kg bw/day based on kidneys and liver	(TG422) NOAEL 10 mg/kg bw/day based on kidneys and liver	(TG407) NOAEL 2 - 12 mg/kg bw/day based on kidneys and liver
Inhalation	No data available	No data available	No data available
Genotoxicity
In vitro – positive results indicated	(TG471) positive metabolic activation seen in TA100 strain. (TG473) positive result for chromosomal aberration with and without metabolic activation. A positive result for sister chromatid exchange assay in CHL cells.	(TG471) positive metabolic activation seen in TA100 and TA1535 strains. A positive result for sister chromatid exchange assay in CHL cells. A positive result for sister chromatid exchange assay in CHO cells.	A positive result for a bacterial reverse mutation assay of S. typhimurium strains with and without metabolic activation.
In vivo	Micronucleus assay – negative result in bone marrow. Single cell gel electrophoresis – induced DNA damage in lungs, kidneys and liver.	Gene mutation assay – with MutaTM mice gave a positive result in nasal tissue. Micronucleus assay – negative result in bone marrow and peripheral blood Single cell gel electrophoresis – induced DNA damage in lungs, kidneys and liver.	Micronucleus assay – negative result in bone marrow. Single cell gel electrophoresis – induced DNA damage in lungs, kidneys, liver and bone marrow.
Carcinogenicity	No data available	Test study similar to (TG453) an increase incidence of carcinomas of the nasal cavity, papillary adenomas and rhabdomyosarcoma seen in both sexes Autoradiography study in female Sprague Dawley rats showed metabolites bound in tissues in the nasal olfactory mucosa, upper alimentary and respiratory tracts. The carcinogenic effect in these tissues was correlated with the capacity to bioactivate the compound in vivo.	No data available
Reproductive and Developmental Toxicity	No data available	(TG 422) observed a decrease in the number of implantations and maternal toxicity.	No data available

Chemical information	Category chemicals
CAS number	95-68-1	95-78-3	108-69-0
Structure
Chemical name	Benzenamine, 2,4-dimethyl-	Benzenamine, 2,5-dimethyl-	Benzenamine, 3,5-dimethyl-
Molecular formula	C8H11N	C8H11N	C8H11N
Acute toxicity
Oral	(TG 401) moderate acute toxicity - LD50 of 1259 mg/kg bw, OECD 2012.	No data available	No data available
Dermal	No data available	No data available	No data available
Inhalation	(TG403) LC50 + 1.53 mg/L	No data available	No data available
Observation in humans	No data available	No data available	No data available
Corrosion/Inhalation
Respiratory irritation	No data available	No data available	No data available
Skin irritation	Slightly irritating to skin	No data available	No skin irritation
Eye irritation	(TG405) slightly irritate the eyes	No data available	No data available
Sensitisation
Skin sensitisation	(TG 429) No dermal skin sensitisation	No data available	No data available
Repeat dose toxicity
Oral	(TG407) NOAEL 2 - 12 mg/kg bw/day based on kidneys and liver	(TG407) NOAEL 2 - 12 mg/kg bw/day based on kidneys and liver	(TG407) NOAEL 2 - 12 mg/kg bw/day based on kidneys and liver
Inhalation	(TG412) NOAEC 0.033 mg/L air	No data available	No data available
Genotoxicity
In vitro – positive results indicated	(TG471) positive metabolic activation seen in TA100 strain. (TG473) positive chromosomal aberration with and without metabolic activation A positive result for sister chromatid exchange assay in CHL cells.	A positive result for a bacterial reverse mutation assay of S. typhimurium strains with and without metabolic activation.	(TG473) positive chromosomal aberration with and without metabolic activation. A positive result for sister chromatid exchange assay in CHL cells.
In vivo	Micronucleus assay – negative result in bone marrow. Single cell gel electrophoresis – induced DNA damage in lungs, kidneys and liver.	Gene mutation assay – with MutaTM mice gave a positive result in nasal tissue and bone marrow. Micronucleus assay – negative result in bone marrow and peripheral blood Single cell gel electrophoresis – induced DNA damage in lungs, kidneys and liver.	Gene mutation assay – with MutaTM mice gave a negative result in nasal tissue and bone marrow. Micronucleus assay – negative result in bone marrow and peripheral blood Single cell gel electrophoresis – induced DNA damage in lungs, kidneys, liver and bone marrow.
Carcinogenicity	An increased incidence of pulmonary tumours was seen in female HaM/ICR mice at high doses only.	An increase in subcutaneous fibromas and fibrosarcomas in male Charles River CD rats and vascular tumours in male HaM/ICR mice.	No data available
Reproductive and Developmental Toxicity	No data available	No data available	No data available

Category justification

We justified the inclusion of these chemicals into a group due to the:

• functional group/structural similarity including 2 methyl and 1 amino group attached directly to the benzene ring
• similarity of the physico-chemical properties including melting points, boiling points, water solubility, log Kow and dissociation constants in water
• similarity in the toxicologically relevant human health effects (where available), including acute toxicity, repeated dose toxicity and genotoxicity
• similarity in the ecotoxicological effects on aquatic organisms, including vertebrates, invertebrates and plants

We used the available data to fill data gaps required to make decisions about classification regarding the following endpoints:

• acute oral/dermal/inhalation toxicity
• respiratory and skin irritation
• genotoxicity and carcinogenicity
• acute toxicity to fish, invertebrates and algae

Assessment outcome

We made the following recommendations:

• Extend classifications for 1 of the chemicals under the Hazardous Chemicals Information System (HCIS) and the Globally Harmonized System of Classification and Labelling (GHS) to the rest of the chemicals in the group for these endpoints despite data gaps.
• Classify all the chemicals for genotoxicity.

We used the category approach to draw inferences about the hazard endpoints for data-poor chemicals with acceptable levels of uncertainty given the regulatory context.