Scientists Trace a Mammal's Entire Cellular Family Tree for the First Time

For the first time, researchers have reconstructed, cell by cell, the family tree of a mammal as it grows from a single fertilized egg into a body with organs. Two independent teams, working separately but publishing within days of each other, used genetically engineered mice whose own DNA recorded every cell division along the way, producing the most detailed lineage maps of mammalian development yet assembled.

The first study, led by geneticist Jay Shendure at the University of Washington School of Medicine and Chengxiang Qiu at Dartmouth College, appeared in the journal Science on October 8. The second, from a team led by Jonathan Weissman and Zachary Smith at MIT and the Whitehead Institute, appeared as a companion paper in Cell. Both used living mouse embryos modified to carry molecular recorders that log, directly in the genome, which cell came from which, generation after generation, as the embryo develops in the womb.

A typewriter made of DNA

The Washington-Dartmouth study used a system called DNA Typewriter, a recording technology Shendure developed with Junhong Choi, now at Memorial Sloan Kettering Cancer Center. Rather than cutting DNA the way earlier genome-editing recorders did, the redesigned system writes marks onto a specific stretch of synthetic "tape" inserted into the genome of a fertilized mouse egg, adding one mark per cell division, always in the same fixed order. Because each new mark is appended after the ones that came before, and because the DNA itself is never fully severed, researchers can later read the tape out of any cell and reconstruct exactly which lineage it descended from and in what sequence those divisions happened.

The team attempted the procedure in 100 fertilized mouse eggs and obtained 10 viable embryos to analyze; one of them, labeled Embryo No. 3, carried the richest recorded history and became the focus of the full study. Researchers tracked roughly 1.3 million cells across 13.5 days of development, from the embryo's first cell division through the point at which organs had begun to take shape — out of a full mouse gestation of 19 to 21 days. A mark created at that very first division let the team distinguish the embryo's two founding cells and trace nearly every profiled cell back to one or the other. One founding cell went on to produce more descendants than the other, but both lineages generated a similar mix of specialized cell types. The data also let the researchers pinpoint when different tissues committed to their fates: blood cells and the retina specialized relatively early, while the outer layer of skin held its options open longer.

"Think of it like an actual typewriter, except it types onto a cell's own DNA instead of paper," said Haedong Kim, a postdoctoral scientist at UW Medicine and co-first author of the study, in a statement released by the university. "It writes without fully severing the DNA, keeps recording relatively steadily, and writes everything in strict order."

A second map, built for scale

The companion effort out of Cambridge, Massachusetts, took a different approach aimed at breadth rather than the finest-grained order of events. Weissman's team used a prime-editing-based recorder called PEtracer, which the group had previously demonstrated in a 2025 study tracking tumor growth, described at the time in a report from MIT. For the new embryo study, the researchers engineered stem cells carrying more than 100 heritable genetic markers, injected them into developing embryos, and used single-cell sequencing to read both the accumulated marks and each cell's gene-expression profile at the same time. That combination let them tie ancestry directly to cell identity across more than 1.4 million cells drawn from 16 mouse embryos, capturing an estimated three-quarters of all cell divisions as the embryos grew from a few hundred cells to nearly a million.

Decades of trying to see inside the womb

Biologists have wanted a complete cell lineage map of a developing animal since at least 1983, when researchers first traced every cell division in the transparent roundworm Caenorhabditis elegans by watching it under a microscope — an approach impossible to repeat in a mammal, whose embryo develops hidden inside the uterus and whose cell number runs into the millions rather than a few hundred. Later techniques tried to infer lineage retroactively from naturally occurring mutations or from engineered genetic scars, but those methods had real limits: DNA-cutting enzymes could damage or kill cells, early markings were often overwritten or erased by later edits, and the recording capacity of the DNA tape ran out well before development was complete. Weissman and Smith had taken an earlier swing at the problem in a 2019 Nature study that offered only coarse resolution. The new pair of papers is the product of years of incremental engineering on both recording methods, including work supported by the Seattle Hub for Synthetic Biology, a $70 million, five-year collaboration among the University of Washington, the Allen Institute, the Brotman Baty Institute and the Chan Zuckerberg Biohub that set out specifically to build a mouse capable of recording its own developmental history.

Why it matters beyond the lab

Both papers are being read as evidence that mammalian lineage tracing has crossed from a theoretical goal into a working, scalable tool, an advance covered widely in the science press this week. Shendure has said the approach could eventually help explain how organs form correctly, how birth defects and some cancers arise when that process goes wrong, and how stem cells might be engineered more reliably into specific tissue types. Every human body traces back to a single fertilized egg that divides into roughly 37 trillion cells, and the same kinds of developmental errors that researchers can now watch unfold in a mouse, cell by cell, underlie a long list of human conditions, from congenital malformations to neurodevelopmental disorders. The Weissman team's co-senior author Zachary Smith has described the new data, as reported by Genetic Engineering & Biotechnology News, as letting researchers see at a glance how nearly every cell relates to every other cell, where it came from and how it developed.

Neither group is claiming the mouse work translates directly to people. Inserting a synthetic recording tape into a human embryo's genome is not something any lab has proposed, and both studies stress that the findings describe mouse development specifically. What the researchers argue is more modest and, they say, more durable: a validated method for recording cell ancestry with high fidelity over nearly two weeks of embryonic life, built and tested well enough that other labs can now adapt it to their own questions about development and disease, whether in mice, other model organisms, or eventually cultured human tissue.

What comes next

Both groups describe their current embryos as proof that the recording systems work at scale, not as a finished picture of development. The Seattle-based Read Team within the synthetic biology hub is already running a large pooled screen to find tape variants with better editing fidelity, aiming to push the recording window closer to full gestation and to make the readout more reliable across different tissue types. The Weissman lab, for its part, has framed its embryo-wide lineage data as a step toward training computational models of development, with researchers describing a longer-term ambition to build predictive "virtual embryo" simulations that could flag when and how a specific genetic or environmental disruption would derail normal development. Closer to the clinic, Shendure's group has pointed to cancer as a near-term application, since tumors, like embryos, grow through repeated rounds of cell division and diversification, and a durable in-cell recorder could in principle track how a cancer's subpopulations branch and spread over time. Independent labs will need to replicate the approach in other strains and species before either recording system becomes a standard tool, but both papers mark mammalian development, long hidden inside the womb, as newly and precisely legible.