The important role of standards for the uptake of transcriptomics and metabolomics based in vitro methods in regulatory toxicology

Abstract

Driven by increasing regulatory interest in applying omics-based methods and reducing animal testing, this study reviewed existing documentary standards for transcriptomics and metabolomics, focusing on those applicable to in vitro test systems. Investigation revealed the landscape to be heterogeneous: for transcriptomics (using RNA-seq), some documentary standards have been produced by formal standardisation bodies (e.g., International Organization for Standardization), whilst for metabolomics (using MS), these were primarily driven by the work of the scientific community developing best practices. The value of reference materials is also highlighted since they enable characterisation of both (analytical) intra-laboratory repeatability and reproducibility as well as inter-laboratory reproducibility. Leveraging standards within omics-based in vitro methods that enter the OECD Test Guideline Programme will ensure their reliability, accessibility across jurisdictions, and sustainability for their long-term use. These benefits are explained with practical examples and make the case for better use of existing standards and initiating targeted development of new standards for efficient and effective development of international Test Guidelines based on in vitro transcriptomics and metabolomics methods.

Introduction

Standardisation serves to achieve uniformity of products, processes or services through development and implementation of standards (ISO 2004). Standards are essentially consensus documents developed by experts and issued by a recognised body that provide rules, guidelines or characteristics of the subject being standardised. Standards, for example those produced by the International Organization for Standardization (ISO) or the European Committee for Standardization (CEN), are not only essential for achieving efficiency across industrial sectors and facilitating international trade, they also play a critical role in protecting human health and the environment (ISO, no date a, no date b, no date c; OECD, no date a).

Currently, the Organisation for Economic Co-operation and Development (OECD) produces internationally recognised testing standards, namely Test Guidelines (TGs), for the generation of toxicological hazard data on chemicals (e.g., pesticides, biocides, industrial chemicals, and cosmetic ingredients, to name a few) to support their safe use in society (OECD, no date a). Guidelines are often complemented by OECD review papers and guidance documents that provide context on their regulatory use and practical advice on their application (OECD 2009). Guideline data are requested and used by regulatory authorities to support assessments and are often a requirement for placing chemicals and associated products on the market (OECD, no date a). When performed following the principles of Good Laboratory Practice (GLP) in a test facility inspected by an appropriate GLP compliance-monitoring programme, data from TG studies must be accepted across OECD member countries as well as non-member countries participating in the system, a process known as Mutual Acceptance of Data (MAD) (OECD, no date b). This approach has considerable benefits given that MAD removes the need for duplicate testing across jurisdictions resulting in significant saving of animals, time and money (OECD 2019, no date b). In addition, use of TGs improves the efficiency of regulatory processes, provides for higher degrees of business and legal certainty, and fosters harmonisation of hazard and risk assessment approaches across international jurisdictions.

There is significant regulatory interest in developing a new generation of OECD TGs and guidance documents to drive the uptake of New Approach Methodologies (NAMs) that utilise advanced measurement technologies such as transcriptomics, proteomics, and metabolomics (Brennan et al. 2024; Radio et al. 2024; ECHA 2025). Notable progress has been made in recent years for omics-based in vitro methods: for example, the Genomic Allergen Rapid Detection (GARD) for assessment of skin sensitisers (GARDskin) distinguishes between skin sensitisers and non-sensitisers through the measurement of gene expression in a cell-based test system and has been a part of TG 442E on in vitro skin sensitisation since 2022 (OECD 2022). This was preceded by a formal scientific peer review on the validity of the GARDskin and GARDpotency methods by ESAC, the Scientific Advisory Committee of the EU Reference Laboratory for alternatives to animal testing (EURL ECVAM) (Corsini et al. 2021). The latter method (i.e., GARDpotency) was also intended to subcategorise skin sensitisers (Category 1A vs. 1B) in accordance with the United Nations Globally Harmonized System of Classification and Labelling of Chemicals. The conclusion of the peer review was that GARDskin could progress further (i.e., to the OECD), whilst GARDpotency lacked adequate information for ESAC to support its regulatory use at the time of formulating their opinion (Corsini et al. 2021). Activities prior to the scientific peer review should be noted, such as the inter-laboratory validation study of GARDskin, published in 2019 (Johansson et al. 2019). In addition, as recorded in the Tracking System for Alternative methods towards Regulatory acceptance (TSAR), the pre-submission of the GARDskin method was received by EURL ECVAM in 2011, offering further insights on processes and timelines (European Commission, no date).

The story of the GARDskin test method illustrates that although the benefits of incorporating such methods into standard TGs are clear, this particular path to regulatory acceptance and use of NAMs is a multistage, technically complex, resource-intensive and time-consuming endeavour. This is unsurprising given the rigour of the process to meet the requirements of MAD and expectations of international regulators, policy makers, industrial users and the broader stakeholder community. There is an opportunity here, as the scientific communities across areas relevant to NAMs have already been working towards greater standardisation – both, to enhance excellent and collaborative research, and to aid the translation of results into commercial applications. In addition, the importance of standardisation in applying omics technologies in chemical risk assessment has already been recognised by others over the years (Sauer et al. 2017; Pain et al. 2020). Thus, a logical step next step is to increase awareness of already existing standards (Hollmann et al. 2021), which could be adapted and applied in the development of TGs that are based on NAMs. However, this can be somewhat challenging since not all relevant standards are explicitly called as such, often operating under pseudonyms such as “guidance”, “best practice”, “framework” and so on. Acknowledging this fact upfront is important to 1) ensure that potentially important resources are not missed and 2) consider what steps might be required to incorporate them into formal standards recognised by relevant authorities. It is worth noting that a standard does not have to be document-based: some are in fact, physical, for example, reference materials. The value of these should not be underestimated as they constitute useful tools in demonstrating, for example, analytical repeatability and reproducibility of a method within and across laboratories.

To address the above, this study reviewed available documentary standards that are considered of potential value to support incorporating omics-based in vitro methods into TGs. In particular, a set of existing documentary standards is identified and presented for transcriptomics- and metabolomics-based measurements, namely, (i) RNA sequencing (RNA-seq) and (ii) mass spectrometry (MS, with primary focus on liquid chromatography – mass spectrometry (LC–MS)), respectively. These documentary standards include approaches that measure either a pre-defined set or a broad range of biomolecules (i.e., targeted or untargeted assays). Also reviewed here are some of the currently available reference materials for transcriptomics and metabolomics to highlight their value in the assessment of reliability of methods. Lastly, the manuscript provides insights into incorporating standards into TGs that feature omics-based methods, and the value of doing so, for the reliability, accessibility and sustainability of TGs.

Existing documentary standards for transcriptomics- and metabolomics-based in vitro methods

Sources of documentary standards

Sources of documentary standards identified for this review are outlined below. Although some documentary standards included here were developed for application in clinical settings, they hold relevance for use in regulatory toxicology and could be repurposed for this context. The sources are:

1. ISO Standards and Technical Specifications,

2. Guidance documents from the OECD,

3. Guidelines from the International Council for Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH),

4. Guidance from the Next Generation Sequencing (NGS) Quality Initiative,

5. Best practices from the scientific community, including outputs from transcriptomics and metabolomics societies or consortia. Although these are not formal documentary standards, they may become de facto standards on the topic they address.

Some documentary standards, if relevant to omics-based methods but not directly applicable to in vitro test systems, have been excluded. For example, the ISO document “Specifications for pre-examination processes in metabolomics in urine, venous blood serum and plasma” (ISO 2021c) was considered out of scope. Each documentary standard from resources above was mapped onto an omics-based workflow for testing of chemicals in vitro (Fig. 1) that has been divided into the following steps:

1. Experimental design – decisions taken prior to the experiment to ensure data will be of satisfactory quality to provide correct scientific conclusions for a given purpose, e.g., the number of biological and technical replicates or the choice of an analytical platform.

2. Exposure of the test system to a chemical – not addressed here as it does not directly relate to conducting omics measurements; however, it is recommended to consult the OECD guidance on Good In Vitro Method Practices (GIVIMP) to demonstrate sufficient quality of samples (OECD 2018).

3. Sample collection – activities to preserve the biomolecular profiles of samples for subsequent omics analysis, e.g., washing of cells (to remove remnants of cell media) or freezing of samples.

4. Sample preparation – activities performed to obtain biomolecules of interest for subsequent data generation: RNA extraction and purification for transcriptomics; metabolite extraction for metabolomics.

5. Data generation (and elements pertaining to the analytical technique) – activities performed to measure biomolecules of interest: library preparation and subsequent sequencing for transcriptomics; mass spectrometry-based analysis for metabolomics.

6. Data processing and analysis – activities performed after data acquisition, from initial assessment of raw data to statistical analysis.

7. Data interpretation – activities performed to draw toxicological conclusions from the data generated using an omics-based in vitro method.

8. Reporting – activities performed to transparently describe one or more element of the omics-based workflow explained above.

Fig. 1.

Existing documentary standards were mapped onto an omics-based workflow for testing of chemicals in vitro. The greyed-out box represents that the “exposure of the test system to a chemical” is not addressed here

Key documentary standards for transcriptomics- and metabolomics-based in vitro methods

Key documentary standards for commonly used technologies in transcriptomics (RNA-seq) and metabolomics (MS) are presented in Tables 1 and 2 (a non-exhaustive list), focusing on those applicable to in vitro methods and separating journal articles or book chapters from other guidance documents. Some of the documentary standards encompass several steps of an omics-based workflow, whilst others focus on a single step (indicated by a tick in a relevant box).

Table 1.

Key documentary standards for transcriptomics studies

Step of the workflow
1	2	3	4	5	6	7	8	Description	Type	Reference
✓								Considerations on applying high-throughput gene expression measurements for testing of chemicals covering study design and the use of QC samples, reference materials, and reference chemicals. Focus on using the data for concentration-response modelling and predicting mechanisms of action of chemicals.	Journal article	Harrill et al., (2019)
✓					✓			Survey on best practice in RNA-seq data processing and analysis including quality metrics, alignment, quantification, and identification of differentially expressed genes. The publication also mentions aspects of experimental design.	Journal article	Conesa et al., (2016)
✓					✓			Guidance on designing RNA-seq experiments which includes considerations and protocols regarding RNA extraction, assessing sample quality, design and preparation of RNA-seq libraries as well as data processing and analysis.	Book chapter	Chatterjee et al., (2018)
		✓						Guideline on genomic sampling (including RNA) and data management for clinical studies. Although the application area differs and its focus is on collecting clinical specimens, many principles pertaining the topic are universal (e.g., stability, degradation, transport, and storage of samples).	Guidance document	ICH, (2017)
			✓					Guidance on QC for Illumina-based sequencing providing checkpoints and quality metrics for sample preparation.	Guidance document	The Next Generation Sequencing Quality Initiative, (2023)
			✓	✓				Guidance on evaluating quality of nucleic acids prior to sequencing, preparation of libraries as well as the use of reference materials and controls. Annexes include checklists for assessing quality of samples (before preparing a library), quality metrics criteria for common sequencing platforms, and recommended sample amounts and concentration per application area as well as examples of reference materials to assess sequencing methods.	Guidance document	ISO, (2022)
			✓		✓			Criteria for assessment of high-throughput gene expression data (covers only short-read RNA-seq). Quality metrics and considerations are provided (where appropriate) for RNA sample quality (i.e., concentration, purity, integrity) as well as data processing and analysis covering the assessment of sequencing reads, alignment, gene expression, and differentially expressed genes (including a brief acknowledgment of interpretation). Spike-in controls, proficiency testing, and process management are also briefly discussed.	Guidance document	ISO, (2021d)
					✓			Requirements for assessing quality of raw (sequencing) data as well as alignment and mapping. Some considerations for validation of bioinformatics pipelines and quality metrics as well as documentation are provided. Annexes include recommendations for coverage and read per application area as well as some examples of software for alignment and mapping.	Guidance document	ISO, (2021a)
					✓			Guidance on assessing the quality of sequencing data using FASTQC with a description on the metrics.	Guidance document	The Next Generation Sequencing Quality Initiative, (no date)
					✓			R-ODAF is a bioinformatics pipeline used to obtain differentially expressed genes (the final output of this pipeline) and may serve as a baseline for assessing future bioinformatics tools.	Journal article	Verheijen et al., (2022)
							✓	Guidance document on reporting toxicological studies, which incorporate omics technologies (e.g., RNA-seq and targeted RNA-seq). The guidance has the following reporting modules: 1) Study Summary Reporting Module, 2) Toxicology Experiment Reporting Module, 3) Data Acquisition and Processing Reporting Module, and 4) Data Analysis Reporting Module.	Guidance document	OECD, (2023a)

An omics-based workflow has been divided to the following steps: 1) experimental design, 2) exposure of the test system to a chemical (outside the scope of this work), 3) sample collection, 4) sample preparation, 5) data generation, 6) data processing and analysis, 7) data interpretation and 8) reporting. A check mark indicates which step of the workflow the standard addresses

Table 2.

Key documentary standards for metabolomics studies

Step of the workflow
1	2	3	4	5	6	7	8	Description	Type	Reference
✓		✓	✓	✓	✓		✓	Best practices and reporting standards for untargeted and targeted metabolomics studies. The document covers aspects of experimental design, types of QC samples, sampling, metabolite extraction, data acquisition, data processing and analysis, metabolite annotation and identification as well as data management and reporting.	Journal article	Viant et al., (2019)
✓		✓	✓	✓	✓		✓	Guide for metabolomics studies including aspects of experimental design, sampling, metabolite extraction, data acquisition, metabolite annotation and identification, and reporting.	Journal article	Alseekh et al., (2021)
			✓	✓				Framework for targeted metabolomics assays proposing a tiered strategy for the assessment of their reliability. The parameters and acceptance criteria for analytical method validation and qualification (e.g., analytical precision and accuracy) are provided.	Journal article	Sarmad et al., (2023)
			✓	✓	✓			Guidelines and considerations for the use of system suitability and QC samples. Although the focus is on clinical studies, many principles pertaining the use of system suitability and QC samples are universal.	Journal article	Broadhurst et al., (2018)
							✓	General recommendations from the mQACC on usage and reporting of intra-study QC samples following literature survey.	Journal article	Broeckling et al., (2023)
							✓	Requirements for reporting metabolite annotation and identification covering four levels of annotation and identification including requirements to achieve them. Developed by Metabolomics Standards Initiative (MSI) in 2007.	Journal article	Sumner et al., (2007)
							✓	Requirements for reporting metabolite annotation and identification covering five levels of annotation and identification including requirements to achieve them. Developed by Schymanski et al., 2014 that provides an alternative reporting format to the one initially proposed by MSI.	Journal article	Schymanski et al., (2014)
							✓	Framework from the mQACC for describing QA/QC procedures and results. Supplementary Information contains templates for reporting the information.	Journal article	Kirwan et al., (2022)
							✓	Requirements for reporting metabolomics data processing and analysis split into six steps, which describe data handling, i.e., pre-processing, pre-treatment, processing, post-processing, validation and interpretation.	Journal article	Goodacre et al., (2007)
							✓	Guidance document on reporting toxicological studies, which incorporate omics technologies (e.g., MS-based metabolomics). The guidance has the following reporting modules: 1) Study Summary Reporting Module, 2) Toxicology Experiment Reporting Module, 3) Data Acquisition and Processing Reporting Module, and 4) Data Analysis Reporting Module.	Guidance document	OECD, (2023a)

An omics-based workflow has been divided to the following steps: 1) experimental design, 2) exposure of the test system to a chemical (outside the scope of this work), 3) sample collection, 4) sample preparation, 5) data generation, 6) data processing and analysis, 7) data interpretation and 8) reporting. A check mark indicates which step of the workflow the standard addresses

Not all available standards fit into the workflow above, however, some of these “outliers” provide useful considerations in areas such as terminology in pharmacogenomic and pharmacogenetic studies (ICH 2007), regulatory submissions for qualification of genomic biomarkers (ICH 2010) or the format of data exchange for omics (ISO 2021b). Similarly, a framework for applying omics under GLP (a key element of MAD) was previously discussed by others and it should be consulted when developing TGs that incorporate omics-based methods (of note, the authors of this framework acknowledged challenges regarding full GLP compliance of omics studies) (Kauffmann et al. 2017). Others commented on how confidence in metabolite identification dictates the use of metabolomics under different regulatory scenarios, an important consideration for method developers working in this area (Malinowska and Viant 2019). Some documents excluded from Tables 1 and 2 were considered too general to include but could serve as a foundation for development of new standards and are mentioned below. It is also important to emphasise that standardisation of transcriptomics- and metabolomics-based methods is an evolving topic, and new documentary standards are under development including an OECD guidance document to standardise collection of samples for subsequent omics analysis and a module on reporting omics data when used for chemical grouping (OECD 2024b; Viant, Barnett, et al. 2024).

In transcriptomics, documentary standards have been mainly focused on a particular step of an omics-based workflow (e.g., sample preparation), with several of them produced by formal standardisation organisations such as ICH or ISO (Table 1). Other documentary standards were proposed by the community, e.g., the omics data analysis framework for regulatory application (R-ODAF) pipeline (Verheijen et al. 2022). R-ODAF was developed to act as a benchmark for processing and analysis of microarray and RNA-seq data. As such, it consists of 1) quality control, 2) processing/normalisation, and 3) differential expression analysis. It is also worth drawing the attention to the NGS Quality Initiative launched by the Centers for Disease Control and Prevention and the Association of Public Health Laboratories to develop a quality management system for NGS in public health laboratories (U.S. Centers for Disease Control and Prevention 2024). As a result, the NGS Quality Initiative offers ample of resources and tools related to NGS (e.g., guidance documents, standard operating procedures), some of which have been included in Table 1.

In metabolomics, standardisation efforts were mostly driven by the work of the scientific community with some documentary standards addressing multiple steps of the workflow (Table 2). Many of the examples provided in Table 2 are not formal standards, rather de facto standards, with a notable example of the requirements for metabolite annotation and identification from Metabolomics Standards Initiative (MSI) (Sumner et al. 2007). More recently, the MEtabolomics standaRds Initiative in Toxicology (MERIT) project proposed applications of untargeted and targeted metabolomics to regulatory toxicology as well as best practices and reporting standards (Viant et al. 2019). The community of metabolomics practitioners continues to work on development and harmonisation of quality assurance and quality control in untargeted metabolomics through the metabolomics Quality Assurance & Quality Control Consortium (mQACC). The consortium has published on topics such as current uses of pooled quality control (QC) samples for untargeted LC–MS metabolomics by surveying the literature (Broeckling et al. 2023), reporting of quality assurance (QA) and QC procedures and results (Kirwan et al. 2022), and a review on available reference materials for metabolomics and lipidomics (Lippa et al. 2022).

For reporting of laboratory-based toxicological studies that employ omics technologies, the OECD developed the OECD Omics Reporting Framework (OORF), which has been a turning point in promoting their regulatory uptake at the international level (OECD 2023a). Currently, the framework focuses on reporting transcriptomics- and metabolomics-based studies, but it is being extended to reporting of proteomics-based studies (OECD 2023a, 2024a). Technologies currently addressed by OORF include microarray, RNA-seq (including targeted RNA-seq), and quantitative reverse transcription polymerase chain reaction (qRT-PCR) for transcriptomics as well as MS and nuclear magnetic resonance (NMR) spectroscopy for metabolomics. The framework has a modular structure consisting of 1) Study Summary Reporting Module, 2) Toxicology Experiment Reporting Module, 3) Data Acquisition and Processing Reporting Module, and 4) Data Analysis Reporting Module. The latter covers reporting of univariate and multivariate methods (with the former used for obtaining differentially abundant molecules) as well as benchmark dose analysis. Further efforts are also ongoing to extend OORF to report enrichment analysis (OECD 2024a).

Although some standards were identified in Tables 1 and 2 under “data processing and analysis”, this is a broad area with a variety of methods and tools available – as such, existence of some standards should not give the illusion of standardisation being achieved and are context specific. For example, although R-ODAF (Verheijen et al. 2022) exists, its purpose is to provide a set of differentially expressed genes and does not address other steps of data analysis (e.g., derivation of points of departure from concentration–response experiments). However, awareness of already existing standards should guide prioritisation and development of future ones. From even the briefest review of Tables 1 and 2, it is evident that data interpretation is currently lacking comprehensive documentary standards for both transcriptomics and metabolomics. Although there are publications that address this topic (Geistlinger et al. 2021; Wieder et al. 2021, 2022; Chicco and Agapito 2022), and some of them contain general recommendations, broader scientific consensus is still needed. As shown by Wijesooriya et al. (2022), open-access scientific publications often inappropriately use or report elements pertaining data interpretation: for over-representation analysis, for example, only 8 out of 197 cases reported an appropriate background list (of genes). Others have also recently drawn the attention to the challenges of using pathway analysis tools in metabolomics studies (Lee et al. 2025). These examples highlight the need to establish minimum good practices in applying and reporting methods and tools to interpret transcriptomics and metabolomics data.

Reference materials

A reference material is a generic term used to describe a material that is well-characterised (qualitatively or quantitatively) with respect to a given property and intended use (ISO 2015). Reference materials play an important role in standardising omics-based methods by demonstrating their accuracy or reproducibility, since they provide known “truths” qualitatively (e.g., indicating presence or absence of an analyte) or quantitatively (e.g., by comparing measured and expected values for an analyte) (Hardwick, Deveson and Mercer 2017; Lippa et al. 2022). Others have extensively discussed currently available reference materials for transcriptomics and metabolomics alongside their potential uses offering a more in-depth examination than covered here (Hardwick, Deveson and Mercer 2017; Lippa et al. 2022). However, a snapshot of available reference materials for both transcriptomics and metabolomics is provided below.

Generally, physical reference materials are either biological or synthetic – the former better recapitulates the complexity of the transcriptome and metabolome, while the latter can be easily spiked into study samples (Hardwick, Deveson and Mercer 2017; Lippa et al. 2022). Perhaps the two most well-known reference materials for transcriptomics are the Universal Human Reference RNA (UHRR, from ten cancer cell lines) and Human Brain Reference RNA (HBRR, from donors’ brain regions) extensively used within the MAQC/SEQC project (MAQC Consortium 2006; SEQC/MAQC-III Consortium 2014). These reference materials allowed to characterise RNA-seq across laboratories, sequencing platforms and bioinformatics pipelines (SEQC/MAQC-III Consortium 2014). With the advent of targeted RNA-seq such as TempO-seq (Yeakley et al. 2017), which are compatible with crude cell lysates, there is now the need for reference materials that mimic study samples for such analysis. Harrill et al. (2019, 2021) previously described generation of bulk cell lysates that acted as reference samples – the cells were treated with either solvent control or a chemical, which was followed by their lysis, pooling, aliquoting and freezing. This strategy allowed the investigators to assess the performance of the TempO-seq assay on samples that better represented the actual study samples. Synthetic spike-ins for RNA-seq are also available, e.g., the ERCC Spike-In Control Mix composed of 92 polyadenylated transcripts that are added to the study samples before or after selection or depletion procedures (Ambion (Life Technologies), 2012). The ERCC Spike-In Control Mix is currently available in the form of two products: either as “ERCC RNA Spike-In Mix” (to determine dynamic range and lower limit of detection) or “ERCC ExFold RNA Spike-In Mixes” (to determine dynamic range, lower limit of detection, and fold changes).

Regarding metabolomics, the mQACC previously published a comprehensive review paper on the use of reference materials for MS-based untargeted metabolomics and lipidomics (Lippa et al. 2022). The authors provided examples of available biological reference materials (e.g., Standard Reference Material 1950 – Metabolites in Frozen Human Plasma from the National Institute of Standards and Technology, NIST), synthetic chemical standard mixtures and commercially available reference libraries. This is an ongoing area of research at NIST that is currently developing new reference materials for future metabolomics use, including the Human Liver Suite for Proteomics and Metabolomics (Candidate Reference Material 8462) (NIST 2024). As noted by others, reference materials can also take the form of digital datasets to test bioinformatics pipelines (Hardwick, Deveson and Mercer 2017). For example, a dataset for direct infusion mass spectrometry-based metabolomics was previously proposed for such a purpose (Kirwan et al. 2014).

Recently, a multi-omics reference material has been proposed by the Quartet Project (Yu et al. 2024; Zhang et al. 2024; Zheng et al. 2024). The reference material (of DNA, RNA, proteins and metabolites) was derived from immortalised B lymphoblastoid cell lines from four family members and is complemented by reference datasets as well as quality metrics to assist data assessment. In addition, the authors proposed the use of ratio-based quantitative profiling where the absolute values of study samples are scaled to those measured in a selected reference sample – this approach supports integration of datasets horizontally and vertically, i.e., within and across omics modalities, respectively (Zheng et al. 2024). The Quartet Project constitutes an important addition to the family of reference materials given the multi-omics focus. Nonetheless, these are project-specific reference materials and as such, their uptake by the broader community is yet to be seen.

Reliability, accessibility and sustainability of omics-based in vitro methods within TGs

It is imperative that the uptake of new in vitro test methods, such as those incorporating omics-based measurements, adheres to the fundamental requirements of international TGs, namely – Reliability: the transferability and reproducibility of a test method within and across independent laboratories over time (OECD 2005); Accessibility: the ability of independent laboratories in different OECD countries to use a test method without unreasonable technical or economic barriers; and Sustainability: the long-term usability of a test method irrespective of changes in technology platforms or vendors of essential test method components. More extensive use of documentary standards and reference materials may further facilitate the peer review of omics-based in vitro methods, as it is an important yet often challenging step towards their incorporation into OECD TGs. These fundamental requirements are discussed below alongside illustrative examples.

Reliability refers to transferability and reproducibility of a test method within and across independent laboratories over time (OECD 2005). Reliability is a key characteristic evaluated in validation studies and one of the prerequisites for international acceptance of new or updated test methods. According to EURL ECVAM’s modular approach to validation, the “reliability assessment” entails assessing 1) within-laboratory variability, 2) transferability, and 3) between-laboratory variability of a method (Hartung et al. 2004). Documentary standards in Tables 1 and 2 will support reliability by providing considerations for laboratories that wish to establish and implement TGs that incorporate omics-based in vitro methods. Such laboratories include Contract Research Organisations (CROs) that provide toxicological studies as a service. For this reason, inclusion of standards within omics-based in vitro methods that enter the TG Programme will provide a crucial support for their correct execution across a range of laboratories with different technical infrastructures. For example, some documentary standards (e.g., ISO 20397–1:2022 (ISO 2022)) provide quality metrics criteria (e.g., sample amount and concentration, fragmentation size range) for several sequencing platforms. Others, like R-ODAF (Verheijen et al. 2022), have been developed to benchmark bioinformatics pipelines. Likewise, physical reference materials have been used to assess and compare the performance of platforms for a given purpose, as shown in the SEQC/MAQC-III (SEQC/MAQC-III Consortium 2014). Reference materials could also be leveraged to monitor the analytical performance of a method over time (Broadhurst et al. 2018; Lippa et al. 2022). Practically, given that omics measurements require both technical expertise as well as specialised (and often expensive) instrumentation, CROs will need to weigh establishing new internal capabilities to perform all the steps of an omics-based workflow for testing of chemicals (Fig. 1) against outsourcing certain steps to specialised providers (e.g., sample preparation and data generation). As this outsourcing model develops, heightening the likelihood of inter-laboratory collaboration, it adds strength to the argument that future TGs (including the use of reagents, instrumentation and bioinformatics pipelines) must be designed to support transferability and reproducibility of omics-based in vitro methods across different steps of the workflow (Fig. 1). This could be achieved by incorporating appropriate quality controls and associated metrics, which demonstrate that facilities with different technical infrastructure produce comparable results.

Accessibility is the ability of independent laboratories (e.g., in different OECD countries) to use a test method without unreasonable technical or economic barriers. An example of a barrier could be access to a reagent that is not readily available in another country, but one that is required by a test method. This is an important consideration as the accessibility of reagents and instrumentation could differ across the many countries adhering to MAD. Greater use of standards within TGs should support their adaptation without compromising overall performance. This concept is similar to OECD Performance Standards supporting validation of analogous test methods, e.g., see (OECD 2024c). Performance standards define essential test method components (amongst other elements) allowing for adaptation of a method where appropriate. There are important endeavours demonstrating that omics-based methods may be technically heterogeneous yet have equivalent performance: for example, a metabolomics ring trial on plasma samples from rats exposed to eight test substances established that reliable grouping of chemicals was feasible in five (out of six) participating laboratories that applied different LC–MS metabolomics workflows (Viant, Amstalden, et al. 2024). As noted by the authors, further guidance is required to establish quality control metrics; this should be prioritised to improve accessibility of future TGs that feature omics-based methods to allow for their use across laboratories of different technical infrastructures as well as across geographical locations. Furthermore, reference materials could further support characterisation of different instrumentation to promote accessibility of a given method.

With respect to sustainability, we define it as the long-term usability of a test method, irrespective of changes in technology platforms or vendors of essential test method components. For example, a discontinuation of the commercial production of a particular piece of equipment required by a test method must not be a threat to its future use. The Bovine Corneal Opacity and Permeability (BCOP) test method illustrates the importance of method sustainability: it is currently featured in TG 437 to determine whether exposure to a chemical results in opacity and greater permeability in bovine cornea (OECD 2023b). The test method requires measurement of opacity and permeability, with the former originally performed using the OP-KIT opacitometer (Van Rompay, Adriaens and Verstraelen 2020; OECD 2023b). However, it became increasingly challenging to commercially source a reliable device, which severely restricted the use of the TG. This eventually prompted development of an alternative opacitometer to provide a sustainable solution to measure opacity (Schrage et al. 2011; Van Rompay, Adriaens and Verstraelen 2020; RL, 2/10). It is not difficult to imagine similar issues occurring for omics-based methods if they are reliant on a specific reagent, instrument, or software application. Omics-based methods should therefore incorporate documentary standards to outline the desired characteristics and performance metrics of essential reagents or equipment, rather than the specific items themselves, to support long-term sustainability. There are already examples of this included in TGs, e.g., the GARDskin test method provides a generic description of the RNA isolation step (OECD 2022). Whilst it gives the name and model of the kit, this is offered as an example. Similarly, this is the case for the instrument used to measure quantity and quality of RNA for this test method. As such, other instruments that have equivalent specifications and performance could be used contributing to the flexibility and sustainability of the overall test method (OECD 2022). However, this same test method specifies measuring the endpoint using the NanoString nCounter system, despite the fact that other targeted technologies could be potentially suitable for this purpose. Whilst each technology has its advantages and limitations, an ideal solution would be to develop a method not limited to implementation on a specific technology or commercial platform.

Although several elements discussed here are already being addressed to some extent within TGs for new or updated methods, it is rather ad hoc and case-dependant. For omics-based in vitro methods, more deliberate and systematic incorporation of documentary standards and reference materials is not only feasible, but also offers significant advantages for developers, users, regulators, and for the OECD TG Programme itself that underpins MAD. Simultaneously, development of new international standards for omics-based in vitro methods must continue to help support this endeavour since there are still gaps to be filled and technology is continuously evolving. Although this is often seen as a daunting task, many standards (e.g., in the form of best practices) already exist offering a baseline for the future efforts to build upon.

Conclusions

A standard is not a collection of “rules”, rather a practically applicable knowledge product that represents a consensus of an expert community at the time of writing. Use of standards is fundamental in increasing the quality of omics measurements and bolstering the confidence in the results obtained. As shown in Tables 1 and 2, standards have many names including “best practices”, “guidelines”, “frameworks” etc. Thus, acknowledging these pseudonyms as standards is important to accelerate their use and visibility as well as providing a basis for development of formal standards that are recognised by appropriate authorities. The use of physical standards in the form of reference materials can assist with characterising analytical repeatability and reproducibility of a method, again increasing confidence in results obtained.

Incorporation of documentary standards and reference materials into omics-based in vitro methods that enter the TG Programme will support their reliability, accessibility and sustainability. Formal standardisation activities should leverage existing efforts by the community to propel use and acceptance of new non-animal approaches, such as omics-based in vitro methods, in regulatory toxicology. Overall responsibility for ensuring reliability, accessibility and sustainability must be shared: developers of NAMs should incorporate standards, whilst reviewers and users of such methods should require their use, where feasible. Only then, omics-based in vitro methods that enter a TG Programme will provide a long-lasting impact.

Declarations

Conflict of interest

The authors have no relevant financial or non-financial interests to disclose.