DNA Typewriter recorded a mouse embryo’s cell history by having a prime editor write heritable, ordered marks into an engineered DNA target that was introduced into zygotes. At embryonic day 13.5 (E13.5), a stage within organogenesis, the team read those marks alongside each cell’s transcriptome and reconstructed a time-calibrated lineage tree with 1,340,794 annotated cells from a single embryo. The study is a proof of concept for one embryo, not a complete census of mouse development.
How the recorder works
DNA Typewriter has three working parts: a prime editor, engineered prime-editing guide RNAs (epegRNAs), and an engineered genomic target called TAPE. Together they turn a stretch of DNA into a running log. Each edit writes an information-bearing insertion into the target, and the act of writing exposes the next target in the sequence for the following edit.
As an Amazon Associate I earn from qualifying purchases.
Writing marks in order
Because each edit prepares the next, insertions accumulate in temporal order along the array. The order of marks therefore reflects the order in which they were written, provided the array has room left. Capacity is finite: once the writable targets are used up, later events cannot be recorded. The cited sources do not give a target count for the TAPE array in this experiment.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Clear out junk files and repair common Windows errors3Fix the driver behind crashes, sound loss and screen glitchesWhat the method can and cannot show
DNA Typewriter is a lineage recorder, not a live microscope. It does not image each division as it happens. The measurement is destructive: the embryo is collected at the end of the experiment, and the accumulated marks and transcriptional profiles are used afterward to infer which cells descended from which. Branches and ancestral states in the resulting tree are therefore reconstructions, not direct observations of the ancestral cells.
How the mouse experiment was run
- Introduce the recorder components into wildtype zygotes by pronuclear injection. The construct is carried into the genome by piggyBac integration.
- Allow the embryo to develop to E13.5 while the editing system writes marks into TAPE as cells divide.
- Dissociate the embryo and profile single nuclei transcriptionally. The paper reports approximately 1.75 million single-nucleus transcriptomes.
- Recover circularized TAPE RNA (circTAPE) from the same nuclei, so each cell carries both its recorded marks and its cell-state profile.
- Build the tree in two parts: a parsimony-supported backbone of 640,012 cells, then distance-based placement of the remaining cells to reach 1,340,794 annotated cells.
- Integrate a separate mouse developmental single-cell time series to assign states and annotations to most internal nodes of the tree.
The numbers, and which version of each to cite
Several counts appear in connection with this study, and they come from two different places. The table separates them so each figure keeps its origin.
| Figure | Value | What it counts | Source and qualification |
|---|---|---|---|
| Single-nucleus transcriptomes | Approximately 1.75 million | Profiled nuclei from the E13.5 embryo | Yu et al., 2026 (paper) |
| Annotated cells in the expanded phylogeny | 1,340,794 | Cells placed in the time-calibrated tree from one embryo | Yu et al., 2026 (paper) |
| Parsimony-supported backbone | 640,012 cells | Cells in the tree before distance-based placement | Yu et al., 2026 (paper) |
| Profiled cells | 1,281,141 | Current figure on the project page | NextCell project page (next-cell.org), not the paper |
| Backbone cells | 655,701 | Current figure on the project page | NextCell project page (next-cell.org), not the paper |
The website and the paper report different totals. The sources reviewed do not explain the gap, so use the paper’s figures when describing the study’s analysis, and attribute the website figures to the project page with the date you checked them. Do not combine the two sets.
What the lineage tree shows
The authors used the tree to ask three questions: where the embryo’s cells come from, how sibling cells relate in type, and when lineages diverge. Their reported findings are below, along with the qualifications the paper attaches to them.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Early founders contribute unevenly
The paper reports that a modest group of founder cells, present before gastrulation, contributes disproportionately to the embryo while remaining broadly multipotent. During organogenesis, a second phase of clonal dominance appears in particular lineages. A small number of early founders, in other words, does not mean that only a few cells build the embryo; it means that a few founders expand more than others.
Rank #3
- Excellent science series aligned to current state standards
- Helps build understanding of physical, life, and earth science
- Engaging activities from songs, rhymes and hands-on projects motivate and inspire
- Lessons focus on one science concept at a time for focused learning
- Also aligned to Next Generation Science
The first division and the two blastomere lineages
The accumulated edits mark the first cell division. The paper reports that the two lineages arising from that division contribute asymmetrically to the embryo proper. The authors also found those blastomeres to be fate-neutral in their analyses, so the asymmetry in contribution does not, on their reading, reflect a difference in the fates the two lineages were committed to.
Sibling cells and shared cell types
The paper reports that sibling cells share a cell type 9-fold more often than chance would predict. For cell types arising from spatially restricted founder pools, the reported excess rises to 68- to 107-fold. These are the authors’ comparisons against chance, not absolute probabilities, and they should be described that way.
Germ layers and dated branch points
From clade co-occurrence in the tree, the authors recover the organization of the three germ layers and a dated hierarchy of cell-type couplings. The branch points they date begin at E8.5.
Free tools Windows power users keep installed
One-click scans. No signup required.
Observed, reconstructed, and imputed
The study mixes three kinds of information, and each carries a different level of confidence.
Best Value
- Students build unmatched deductive-reasoning skills as they become crime-solving stars
- Most scenarios have more than one plausible outcome, allowing individuals or groups to broadly interpret evidence
- Includes interpretive handwriting, body language, fingerprinting, and many more activities
- Observed: the edited records and transcriptional profiles read from cells of the E13.5 embryo.
- Reconstructed: the branching structure of the tree, built from a parsimony backbone with distance-based placements added.
- Imputed: the cell states and annotations assigned to most internal nodes, which depend on integrating the external developmental time series. The ancestral cells at those nodes were not sampled, so their states are inferences.
Limits of the study
- The reconstruction covers one E13.5 embryo. It is a proof of concept, and the paper does not establish how the findings vary across embryos or mice.
- The tree ends at E13.5. Development after that stage, and the adult animal, lie outside the reconstructed interval.
- The endpoint is destructive, so the same embryo cannot be followed live.
- Internal-node states rely on an external atlas, so their accuracy is bounded by how well that atlas matches the embryo studied.
- Distance-based placements sit outside the parsimony backbone and carry more uncertainty than the backbone itself.
Version status and where to find the data
The study’s PMC record labels the included version as a bioRxiv preprint dated July 30, 2026, and states that it has not been peer reviewed. The NextCell project page cites the work as a 2026 Science paper, DOI 10.1126/science.ael0508. The description in this article follows the preprint text. Wording and figures in the published journal version may differ, so cite the Science article for formal references and for any close quotation. The PMC record is at https://pmc.ncbi.nlm.nih.gov/articles/PMC13484211/.
The NextCell project page at https://next-cell.org/ links to the following:
- Full and backbone dated lineage trees, and an interactive browser for exploring the tree and its annotations.
- Analysis code, released under GPL-3.0.
- A Zenodo archive, and processed cell metadata.
- Sequencing data deposited at GEO under accession GSE341627.
- Data, stated as released under CC BY 4.0.
How to compare lineage-recording approaches
DNA Typewriter is a single method demonstration, not a head-to-head comparison with other tools. The paper treats the following as design considerations, and they are a useful checklist for comparing any lineage recorder.
| Question | How this study handles it |
|---|---|
| Is event order encoded directly, or reconstructed afterward? | Encoded directly: each edit exposes the next target, so insertions follow temporal order within the array’s capacity. |
| How many writable targets, and how much recording capacity? | Finite, set by the engineered TAPE array. A specific target count is not stated in the cited sources. |
| Can marks be recovered alongside cell state? | Yes: circTAPE is recovered alongside single-nucleus transcriptomes. |
| What are the delivery, expression, and information-loss constraints? | Delivery by pronuclear injection into zygotes with piggyBac integration. The paper discusses expression and information loss as challenges, not as solved problems. |
| What scale of cells and developmental interval is reconstructed? | About 1.34 million annotated cells from one embryo, spanning zygote to E13.5. |
The table shows what the study does and does not establish. A reader comparing it with other recorders should check each row against the other method’s published data rather than against this summary alone.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
