Blood or skin? The starting cell may shape hidden DNA flaws in stem cell lines

Induced pluripotent stem cells have become one of the most powerful tools in modern biomedical research, offering scientists a way to rewind adult cells back into an embryonic-like state and then grow them into almost any tissue in the body. Yet the reprogramming process itself carries a hidden risk: it can introduce large-scale rearrangements of the genome, known as structural variants, that standard quality-control methods often miss. A new study from researchers at Stanford University School of Medicine, published in BMC Genomics, now suggests that the type of somatic cell used to create these stem cell lines may influence how much structural damage the final product carries. The team applied optical genome mapping, a high-resolution technique that reads the physical architecture of chromosomes, to dozens of early-passage induced pluripotent stem cell clones and found a clear pattern: clones derived from skin fibroblasts tended to carry a heavier burden of structural variants than clones derived from blood cells.

The research, led by first authors Leila Namvar, Kamilla Sedov and Madison James Yang under the corresponding authorship of Birgitt Schüle, examined 73 low-passage induced pluripotent stem cell clones generated from 25 parental lines. The parental cells came from two distinct sources: peripheral blood mononuclear cells, the mixed population of white blood cells that can be drawn from a simple blood sample, and fibroblasts, the connective-tissue cells typically harvested from a skin biopsy. Both cell types are widely used starting materials for making induced pluripotent stem cells, and until now there has been little systematic evidence about whether the choice between them leaves a genomic fingerprint on the resulting stem cell lines.

Optical genome mapping works in a fundamentally different way from the sequencing-based methods that dominate genomics today. Instead of breaking DNA into millions of short fragments and reassembling them computationally, the technique preserves ultra-high-molecular-weight DNA molecules, each hundreds of thousands of base pairs long. These molecules are fluorescently labeled at specific sequence motifs, creating a barcode pattern along each DNA strand. The labeled molecules are then run through a nanochannel array that stretches them out in a linear fashion, allowing cameras to image the barcode patterns directly. By aligning these observed barcodes against a reference genome, the software can detect where the actual genome deviates from the reference: deletions, duplications, inversions, insertions and translocations all leave characteristic signatures in the barcode pattern.

The practical advantage of this approach lies in its resolution and sensitivity for large, complex events. Conventional karyotyping, the century-old technique of staining and photographing chromosomes under a microscope, can only reliably detect rearrangements on the order of several million base pairs. Copy-number microarrays detect dosage changes but miss balanced events such as inversions and reciprocal translocations that preserve the total amount of DNA. Short-read whole-genome sequencing struggles with repetitive regions and with rearrangements that span large distances. Optical genome mapping, by contrast, can detect structural variants down to approximately two kilobase pairs at sufficient coverage, and because each variant call is supported by evidence at the level of individual DNA molecules, the method can also identify mosaic events present in only a fraction of the cells in a culture, a common situation in stem cell lines where a single aberrant clone can expand during passaging.

When the Stanford team applied this technology to their panel of clones, the differences between the two starting cell types were striking. Fibroblast-derived induced pluripotent stem cells showed a higher overall burden of structural variants, and this burden was enriched for duplications of at least 100 kilobase pairs. These duplications were also more likely to overlap protein-coding genes, meaning they had the potential to alter gene dosage in ways that could affect cellular function. In addition, the structural variants found in fibroblast-derived clones overlapped more frequently with fragile sites, specific regions of the genome that are prone to breakage during replication, and with recurrent chromosomal hotspot regions that have been repeatedly associated with culture-induced rearrangements in pluripotent stem cells.

Peripheral blood mononuclear cell-derived clones told a different story. These lines showed fewer structural variants overall, and a higher proportion of the clones carried no detectable clone-specific structural variants at all. In other words, when the starting material was blood, the reprogramming process more often produced genomically clean stem cell lines, at least by the standards of what optical genome mapping can see. The authors are careful to frame this as an association rather than a proven causal mechanism, but the pattern across 73 clones from 25 parental lines is difficult to dismiss as chance.

Why might fibroblasts be more prone to acquiring structural variants during reprogramming? The study does not definitively answer this question, but several biological considerations are consistent with the findings. Fibroblasts are adherent cells that typically require more extensive expansion in culture before and during reprogramming, and prolonged cell division is a well-known driver of copy-number changes, particularly duplications that arise through replication errors and DNA breakage at fragile sites. Blood cells, by contrast, are naturally short-lived and are isolated from a sample that can be processed relatively quickly. It is also possible that the epigenetic and replicative state of the two cell types at the moment of reprogramming influences how faithfully the genome is copied and reorganized as the cells pass through the pluripotent state. The enrichment of variants at fragile sites and known culture-associated hotspots in the fibroblast-derived lines points toward replication stress and cell culture itself as contributing factors.

The clinical and research implications of this work are substantial. Induced pluripotent stem cells are used to model diseases ranging from Parkinson’s disease to Alzheimer’s disease, and the Stanford study drew its donor samples from the Stanford Alzheimer’s Disease Research Center, with funding from the National Institutes of Health, the Michael J. Fox Foundation for Parkinson’s Research and the California Institute for Regenerative Medicine. In disease modeling, a structural variant that disrupts a gene could be mistaken for a disease-relevant finding or could mask the true cellular phenotype of interest. In cell therapy, where reprogrammed cells may eventually be transplanted into patients, a duplication affecting an oncogene or a deletion affecting a tumor suppressor could raise safety concerns. The finding that structural variants frequently overlapped protein-coding genes, fragile sites and recurrent chromosomal hotspots underscores why clone-level genomic assessment matters before any downstream application.

What makes the study particularly actionable is its argument for optical genome mapping as a routine quality-control step during stem cell line generation and selection. The authors position the technology as a post-reprogramming screening tool capable of catching clone-specific events below the resolution of conventional cytogenetic and SNP-array-based assays. Because the method detects variants at approximately two kilobase pairs and provides molecule-level evidence for mosaic events, it occupies a sweet spot between the coarse resolution of karyotyping and the computational complexity of long-read sequencing. For laboratories producing banks of induced pluripotent stem cell lines, the practical message is that screening each clone with a genome-wide structural variant assay before committing it to experiments or therapies could prevent costly surprises downstream.

The study also carries a subtler message about experimental design. When researchers plan a new panel of induced pluripotent stem cell lines, the choice between blood and skin as the source tissue is often made for practical reasons, such as patient accessibility or biopsy logistics. This work suggests that the choice may have genomic consequences that persist into the pluripotent state even at early passages, before extended culture has had time to accumulate additional changes. Fibroblast-derived lines may require closer scrutiny, particularly for large duplications affecting genes, while blood-derived lines may more often yield genomically clean clones. As induced pluripotent stem cell technology moves from the research bench toward clinical applications, understanding how the origin of a stem cell line shapes its genome is becoming an essential part of ensuring that the cells scientists study and eventually transplant are as faithful to the original genome as possible. Optical genome mapping, this study suggests, offers a practical and high-resolution way to make that assessment.

Subject of Research: Structural variant differences between blood- and fibroblast-derived human induced pluripotent stem cells detected by optical genome mapping

Article Title: Optical genome mapping identifies source-associated structural variant differences across early-passage human iPSCs

Article References: Namvar, L., Sedov, K., Yang, M. J., Hermosillo, R., Zafar, F., & Schüle, B. (2026). Optical genome mapping identifies source-associated structural variant differences across early-passage human iPSCs. BMC Genomics. https://doi.org/10.1186/s12864-026-13395-2

Image Credits: AI Generated

DOI: 10.1186/s12864-026-13395-2

Keywords: induced pluripotent stem cells, optical genome mapping, structural variants, fibroblasts, peripheral blood mononuclear cells, nuclear reprogramming, genomic quality control, copy number variation, fragile sites, chromosomal hotspots, disease modeling, BMC Genomics