← Back to Home

Writing Cell Histories into DNA: Scientists Reconstruct a Family Tree of 1.28 Million Mouse Cells

From the fertilized egg to late organ formation, molecular recording allows cells to leave traceable clues to their ancestry. This lineage map, spanning 138 cell types, also reveals that different cells take on specialized functions at different times.

By SURL BioNews

How does a single cell give rise to blood, skin, and eyes? Understanding embryonic development requires knowing not only what cells ultimately become, but also where they come from and when their paths diverge. A team at the Seattle Hub for Synthetic Biology has now enabled mouse embryos to write these histories into DNA, then used the records to reconstruct a family tree of approximately 1.28 million cells, providing a new tool for tracing how organs form and how development deviates from its normal course.

The study, published in *Science* on October 8, traces development from the fertilized egg’s first division to embryonic day 13.5, a late stage of organ formation. According to data on the public platform NextCell, the lineage tree includes 1,281,141 fully analyzed cells spanning 138 cell types, accounting for approximately one-tenth of the embryo’s total cells at that stage. Despite its scale, this remains a map reconstructed from sampled cells.

The key technology is called DNA Typewriter. Researchers placed artificial “recording tapes” in the cells’ genomes and used a prime editor to write short DNA tags sequentially. Existing tags are passed to descendants as cells divide, while new records accumulate over time. Comparing the tags shared by cells and their order allows researchers to infer common ancestors and branching times. In this experiment, each cell carried 11 recording tapes, each with six writing positions. Combining these records with gene expression data linked ancestry to cell identity.

The family tree shows that the first two cells produced by division left unequal numbers of descendants, yet both contributed to multiple cell types in similar proportions. According to the University of Washington School of Medicine, the study also identified differences in when cells committed to particular fates: blood and retinal cells established their direction earlier, while cells forming the outer layer of the skin did so later. These findings are consistent with years of developmental research and can now be examined together within the lineage of a single mouse.

However, a million cells do not amount to a million independent pieces of embryonic evidence. The team attempted recording in 100 fertilized eggs, obtained 10 embryos for examination, and ultimately focused the full analysis on embryo No. 3, which had the richest records. These results therefore primarily demonstrate the feasibility of large-scale recording and reconstruction. Whether different embryos exhibit the same branching patterns still requires comparisons across more samples.

The reconstruction itself also has different layers. NextCell explains that the researchers first built a reference framework using 655,701 cells for which at least seven recording tapes had been read, then placed the remaining cells on the tree. The identities of ancestral cells at the time were estimated by a model, and the platform explicitly labels them as inferred results. When reading this map, it is necessary to distinguish lineage relationships supported by DNA records from ancestral states filled in by analytical methods.

Through NextCell, the team has made available an interactive browser, lineage trees with timing information, analysis code, and links to sequencing data. The Allen Institute notes that tools of this kind could eventually be used to study developmental disorders and cancer cell expansion, or to track where cells go during stem cell therapies. This experiment has not yet validated these disease and treatment applications. Its most immediate value is to let researchers trace backward along cellular branches, turning “when did their paths diverge, and where did they come from?” into questions that can be tested concretely.

References

  1. Allen Institute
  2. UW Medicine
  3. NextCell — Shendure lab, University of Washington and Seattle Hub for Synthetic Biology