Characterisation gaps: what current assays miss about stem cell quality

Reading time: 7 minutes

A pluripotent stem cell line can express every expected pluripotency marker, maintain a normal karyotype, and pass standard sterility tests, yet consistently fail to produce functional cardiomyocytes or mature neurons. Current assays confirm what a cell is. They are far less reliable at predicting what it will do.

This gap between identity testing and functional prediction is one of the most consequential bottlenecks in developmental and stem cell science. For the ancillary TechBio company, it is both a commercial risk and a product opportunity. This article, part of the Pillar 2 series on why ancillary technologies fail, examines what current characterisation misses and why it matters.

What the standard toolkit captures

The Pillar 1 article on characterisation and quality control described the standard assessment layers for pluripotent stem cells: identity confirmation by short tandem repeat (STR) profiling, pluripotency assessment by surface marker and transcription factor expression, genetic integrity by karyotyping, and sterility testing. The ISSCR's 2023 standards reinforced these as minimum requirements and introduced reporting checklists to improve transparency across laboratories [1].

These assays answer the question: is this cell what it claims to be? For most regulatory and banking purposes, that answer is necessary. The problem is that it is not sufficient.

Where the blind spots are

Several categories of information that matter for downstream performance are poorly served by the standard toolkit.

Epigenetic drift. Pluripotent stem cells carry a complex landscape of DNA methylation and histone modifications that determine which genes are accessible and how the cell responds to differentiation cues. These epigenetic marks can shift during prolonged culture. iPSCs can also retain residual epigenetic memory of their tissue of origin, biasing their differentiation potential in ways that surface marker expression does not reveal [2]. Standard characterisation does not routinely assess the epigenome. A 2024 longitudinal study tracking human embryonic stem cells over six years of culture found that certain genetic aberrations, particularly TP53 mutations and gains at chromosome 20q11.21, accumulated over time and altered the biological properties of the cells in ways that affected safety more than passage number alone [3]. These changes interact with epigenetic state in ways that bulk assays cannot capture.

Population heterogeneity. A flask of pluripotent stem cells is not a uniform population. It contains cells in subtly different states: naive-like, primed, transitional, and cells that have begun spontaneous differentiation. Bulk assays report an average across the population, masking subpopulations that may compromise the consistency of downstream products [4]. Single-cell transcriptomics can reveal this heterogeneity, but these methods are not yet practical as routine quality control tools. For the TechBio company building culture systems or analytical instruments, the implication is that your product may need to account for populations that look uniform by standard metrics but behave as mixtures.

Functional potency. The most important property for most applications, the ability to generate specific functional cell types, is the hardest to measure prospectively. The traditional gold standard was the teratoma assay, in which cells are injected into immunodeficient mice and assessed for their capacity to form tissues from all three germ layers. This assay is slow, expensive, ethically problematic, and highly variable [5]. Faster alternatives such as the PluriTest and ScoreCard platforms offer gene expression-based readouts, but these predict pluripotency as a state rather than as a function. They cannot reliably forecast whether a given cell line will produce mature, functional derivatives in a specific differentiation protocol [5,6].

A 2025 review of potency tests for the 31 FDA-approved cell therapy products found that over a third of potency test details were redacted from public filings, and that redactions had increased sharply in 2023 and 2024. The authors noted that no single validated approach to potency testing has emerged, and that the gap between regulatory expectation and available assay capability remains wide [7].

Metabolic profiling. Stem cells in different states have distinct metabolic signatures. Pluripotent cells rely more heavily on glycolysis; differentiated cells shift toward oxidative phosphorylation. Metabolic assays could provide rapid, non-destructive readouts of cell state, but the field has not yet standardised what metabolic profiles predict about downstream performance [8]. For technology developers, this is an area where early investment in assay development could yield significant competitive advantage.

Why this matters commercially

For the ancillary technology company, the characterisation gap creates a specific problem. If your product claims to improve cell culture, purify cell populations, or monitor cell health, its value depends on demonstrating a measurable effect on a relevant quality attribute. If the available quality attributes do not capture what matters, your product's value is difficult to prove.

This is not hypothetical. Companies building label-free analytical tools, advanced imaging platforms, or AI-driven quality prediction systems face the challenge of validating their measurements against endpoints that the field has not yet standardised. You may be measuring something genuinely informative, but if there is no accepted reference standard to benchmark against, the path to commercial adoption is longer and more expensive.

The reproducibility article in this series described how pharmaceutical partners evaluate ancillary technologies through structured assessments. Those assessments require that your product demonstrates consistent performance against defined criteria. If the criteria themselves are inadequate, even a genuinely superior product can fail evaluation because the measurement framework does not capture its advantage.

What is emerging

Several approaches offer promise for closing the characterisation gap.

Non-destructive, label-free methods based on Raman spectroscopy, dielectric properties, or cell mechanical measurements can assess cell state in real time without consuming the cells being measured. Our own collaborative research demonstrated that biophysical techniques including dielectrophoresis can discriminate between pluripotent stem cells and their differentiating derivatives based on physical properties alone [9]. These approaches could enable continuous monitoring during culture rather than assessment only at fixed checkpoints.

Machine learning applied to imaging, process, or molecular data is increasingly able to predict differentiation outcomes and detect quality drift earlier than conventional assays. The constraint is data: these models require large, well-annotated training sets, and the stem cell field has not yet produced these at the scale needed for robust generalisation [10].

Single-cell multi-omics, combining transcriptomic, epigenomic, and proteomic measurement at single-cell resolution, offers the most comprehensive view of cell state but remains too costly and labour-intensive for routine quality control. As costs fall and automation improves, elements of this approach may become practical for in-process monitoring.

What this means for TechBio founders

The characterisation gap is not just an academic concern. It defines the measurement infrastructure on which the entire ancillary technology sector depends. If you cannot measure the thing that matters, you cannot prove your product improves it. If the field lacks potency assays that predict clinical function, then every product that claims to improve manufacturing quality is building its evidence base on incomplete foundations.

The companies that will do well in this space are those that invest in understanding which measurements predict real-world performance, that build validation data against those measurements, and that contribute to the development of standards the field can adopt. The GMP article in Pillar 1 described the Quality by Design framework that regulators expect. That framework requires Critical Quality Attributes, and quality attributes are only as good as the assays that measure them.

References

[1] International Society for Stem Cell Research. Standards for Human Stem Cell Use in Research. 2023. Available from: https://www.isscr.org/basic-research-standards

[2] Kim K, Doi A, Wen B, et al. Epigenetic memory in induced pluripotent stem cells. Nature. 2010;467(7313):285-290. DOI: 10.1038/nature09342

[3] Park JW, Lee JH, Kim SW, et al. Longitudinal analysis of genetic and epigenetic changes in human pluripotent stem cells in the landscape of culture-induced abnormality. Exp Mol Med. 2024;56:2400-2413. DOI: 10.1038/s12276-024-01334-8

[4] Hough SR, Thornton M, Mason E, Mar JC, Wells CA, Pera MF. Single-cell gene expression profiles define self-renewing, pluripotent, and lineage primed states of human pluripotent stem cells. Stem Cell Reports. 2014;2(6):881-895. DOI: 10.1016/j.stemcr.2014.04.014

[5] Smith L, Quelch-Cliffe R, Liu F, Aguilar AH, Przyborski S. Evaluating Strategies to Assess the Differentiation Potential of Human Pluripotent Stem Cells: A Review, Analysis and Call for Innovation. Stem Cell Rev Rep. 2025 Jan;21(1):107-125. doi: 10.1007/s12015-024-10793-5

[6] Bock C, Kiskinis E, Verstappen G, et al. Reference maps of human ES and iPS cell variation enable high-throughput characterization of pluripotent cell lines. Cell. 2011;144(3):439-452. DOI: 10.1016/j.cell.2010.12.032

[7] Simon, C.G., Bozenhardt, E.H., Celluzzi, C.M. et al. Analysis of the measurements used as potency tests for the 31 US FDA-approved cell therapy products. J Transl Med 23, 259 (2025). https://doi.org/10.1186/s12967-025-06253-4

[8] Zhang J, Nuebel E, Daley GQ, Koehler CM, Bhatt AP. Metabolic regulation in pluripotent stem cells during reprogramming and self-renewal. Cell Stem Cell. 2012;11(5):589-595. DOI: 10.1016/j.stem.2012.10.005

[9] Pethig R, Menachery A, Pells S, De Sousa P. Dielectrophoresis: a review of applications for stem cell research. J Biomed Biotechnol. 2010;2010:182581. DOI: 10.1155/2010/182581

[10] Kusumoto D, Lachmann M, Kunber T, et al. Automated deep learning-based system to identify endothelial cells derived from induced pluripotent stem cells. Stem Cell Reports. 2018;10(6):1687-1695. DOI: 10.1016/j.stemcr.2018.04.007

About StemCells.Help

StemCells.Help is an advisory consultancy that aids innovation and real-world impact of life science applications built on developmental and stem cell biology. Founded by Dr Paul De Sousa, it draws on over four decades of experience spanning early embryo development, animal cloning, pluripotent stem cell manufacturing, and technology commercialisation. If you build tools for these domains or work in an emerging application where the biology is the enabling technology, StemCells.Help can provide experienced scientific counsel to ground your decisions. To discuss your needs, talk to Paul.

ORCID: 0000-0003-0745-2504

Web: stemcells.help