The Startup Selling 'Optimized' Babies Just Paid to Gene-Edit Human Embryos

A company that sells parents an app to rank their embryos, and whose founder says he can optimize a future child for intelligence, just paid to have the genes of human embryos rewritten. The experiment ran in Nature on September 9 under a title with no hesitation in it: “Highly efficient base editing at PCSK9 and normal human embryo development.” A Columbia team led by Dieter Egli reported editing DNA in human embryos with the kind of precision the field has chased for a decade. The check came from Nucleus Genomics. Read the science and the funding line together, because the paper looks very different depending on which one you start from.

We have been near this spot before, twice. Nearly a decade ago a lab in Oregon announced it had corrected a heart-disease mutation in human embryos, using almost exactly the words in circulation this week: safe, precise, efficient. A year later a Chinese scientist named He Jiankui stopped announcing and started implanting, and the world got its first gene-edited children, twin girls he claimed would resist HIV. He got three years in prison. The field got a lesson it promised to remember, that the ability to rewrite an embryo runs years ahead of any ability to prove the rewrite is safe, and that the gap between the two is where the harm hides.

The tool at the center of the new paper is not the CRISPR most people picture. Classic CRISPR-Cas9 works like molecular scissors, cutting both strands of the double helix and letting the cell sew the break shut, a repair that scatters small insertions, deletions, and sometimes large chromosomal wreckage across the genome. Base editing works more like a pencil with an eraser. It nicks a single strand and chemically rewrites one DNA letter into another, an A into a G, without ever snapping the helix in two. Because it does not sever both strands, the study saw far fewer of the large chromosomal deletions that CRISPR leaves behind. Far fewer, not none. The deletions still happened, only at a much lower rate, which even the authors treat as a clue about how a cell repairs the particular damage a base editor does.

The word they led with was efficient. The one further down was mosaic.

When the edit was made in the fertilized egg, before its first division, the change carried into 100 percent of the resulting embryo’s daughter cells. That figure is what reached the headlines. It holds only for that narrow case. Most of the embryos came out mosaic, the intended change reaching some cells and skipping others, so there is no single embryo you could point to and call cleanly edited. Egli said as much himself: “Mosaicism creates a range of possibilities, making it impossible to predict outcomes, and is thus preventing meaningful application for use in the clinic.”

The mosaicism is not the only asterisk. The base editor also made changes at other sites as the embryo developed, some of them near the intended target, and how many turned up depended on which guide RNA the team used. When the editing machinery was present at high levels, development failed outright, a toxicity the authors flag as its own hazard. And the assays used to certify an embryo edit as clean are not sharp enough to catch everything they miss. That last point is not this lab’s opinion. A 2023 analysis in Nature Communications warned that the standard tools for assessing embryo edits are too limited to deliver the safety certainty the field keeps claiming, which means the honest count of stray edits here is not low, it is unknown.

None of the outside scientists who read it mistook efficient for ready. Krishanu Saha of the University of Wisconsin-Madison said he “would not call it a breakthrough” and found it “hard to think about a scenario where this is medicine.” Alexis Komor, one of the chemists who invented base editing, was blunter about where it leads: the work “kind of opens the floodgates,” she said, calling it “a gateway to embryo editing to do enhancements.” Egli, to his credit, drew the line himself, telling reporters that editing embryos for a pregnancy “is currently not possible to do so safely.” Komor’s word was enhancements, not cures, and the funding line is where that word comes from.

The two genes were chosen to look like medicine.

PCSK9 and the fetal-hemoglobin genes are the most sympathetic targets in the catalog. PCSK9 governs cholesterol, and drug companies already switch it off in the livers of consenting adults to cut heart-attack risk, an edit that dies with the patient and is nobody else’s business. The HBG genes sit at the center of sickle cell disease and beta-thalassemia, both of which cause serious illness and still lack a cure for most patients. Frame the experiment as fixing those two things and it reads as mercy.

Then read the funder. The study was paid for by Nucleus Genomics. Kian Sadeghi, the founder and CEO of Nucleus Genomics, has spent the past year selling a $30,000 program called IVF+ that scans both parents and up to 20 embryos and screens for more than 2,000 traits and conditions, from disease risk to eye color, and he has said publicly that he partnered with a real couple to “optimize their embryos based on intelligence.” A senior co-author, Nathan Treff, the chief clinical officer at Nucleus, put his name on the paper. Dieter Egli, the Columbia biologist who led the study, ran the bench. This is not a cancer center that happened to accept a grant. It is a company built to sell genetic selection, paying to demonstrate that genetic writing works, and picking the two diseases least likely to draw an objection to demonstrate it with.

The Center for Genetics and Society, which tracks this industry, read the collaboration the same way. “Claims of ‘safe’ and ‘precise’ changes to the DNA of human embryos are reminiscent of similar claims nearly ten years ago,” said Katie Hasson, the group’s associate director, before naming what the disease framing covers: “It’s extremely troubling to see this research collaboration with a company that is explicit about its plans to bring designer baby technology to the market.” Heritable editing, she noted, is already prohibited in the United States and in at least 70 countries. The purpose of proving it works here, she argued, is not therapy but a market: “The risk of a market-based techno-eugenics is much too large.”

A somatic edit dies with the patient; a germline edit gets inherited.

Here is the biology that separates this from the cholesterol edit a cardiologist might give an adult. An edit made in a one-cell embryo does not stay in that embryo. It copies into every cell of the person who grows from it, including the eggs or sperm that person will later make, which means it flows into their children, and their children’s children, down a line that never signed a consent form. A somatic edit is a treatment you choose. A germline edit is a permanent experiment run on descendants who cannot choose it and cannot undo it. The off-target changes nobody could fully count in a six-day-old embryo would ride down that same family tree.

This is not, as this publication reads it, a two-sided scientific debate. It is a question of who holds the pen over human inheritance, and the answer on offer is a venture-backed startup selling optimization to people who can afford it. Hasson put the stakes plainly: the decision to edit the human germline “belongs to all of us, and certainly should not be driven by Silicon Valley’s transhumanist fantasies,” and without “stronger prohibitions in the US and internationally” the endpoint is a world of genetic “haves” and “have nots.” Even Nature’s own news coverage filed the result under “praise and alarm.” The alarm is the honest half.

None of this is territory the field has avoided. In 2022, Chinese teams reported using a related base editor to rewrite mitochondrial DNA in abnormally fertilized human embryos, and the same off-target ghosts showed up then too. What is new in 2026 is not the pipette. It is the business model bolted to it.

The line that has kept the germline closed in America is not a law of physics. It is a rider that Congress renews every year, one that forbids the FDA from so much as reviewing an application to start a pregnancy from an edited embryo, propped up beside a patchwork of national bans that a well-funded company can shop around by moving offshore. So the number to watch next is not the next efficiency figure out of Egli’s lab. It is whether that rider survives its next renewal, whether the National Academies reconvene on heritable editing, and whether Nucleus Genomics, having now paid to show the writing works, goes looking for a jurisdiction that will let it sell the result. The tool arrived before the guardrails did. It usually does. The open question is who gets to decide what happens next, and right now the loudest voice in that argument belongs to the people with the most to sell.

Sources

1. Nature – Egli et al., “Highly efficient base editing at PCSK9 and normal human embryo development” (Sept 9, 2026)
2. News-Medical – summary of the Columbia base-editing study, Egli quotes and risks
3. Center for Genetics and Society – Katie Hasson statement on the study and Nucleus Genomics
4. Scientific American – expert reaction (Komor, Saha) and He Jiankui context
5. Precision Medicine Online – Nucleus Genomics funding of the Egli lab and Nathan Treff’s role
6. Inc. – Kian Sadeghi on “optimizing” embryos for intelligence
7. The San Francisco Standard – inside Nucleus Genomics’ embryo-selection product
8. Nature Communications – “Limitations of gene editing assessments in human preimplantation embryos” (2023)
9. Nature news – “Precise genome editing of human embryos triggers praise and alarm”
10. Cell Discovery – mitochondrial base editing in human embryos (2022)