The preprint followed every existing guardrail and still opened a debate the U.S. has been deferring since the CRISPR babies case, because the country never wrote those guardrails into law.
In May 2026, Dieter Egli's lab at Columbia University posted a preprint on bioRxiv describing base editing of human embryos at two well-studied genomic sites, a piece of work that followed every U.S. rule on the books and still managed to reopen the question American science has been deferring since He Jiankui announced the first CRISPR babies in 2018.
The embryos were zygotes, single cells created for research and not intended for pregnancy. The targets were PCSK9, a cholesterol-regulating gene, and HBG1 and HBG2, which govern fetal hemoglobin. According to the authors' preprint, those genes were picked because they are well-characterized, not because the work was aimed at any specific patient. The editing tool was base editing, a single-letter chemical change made with a guide RNA, distinct from the double-strand DNA cuts that defined earlier CRISPR work and the He Jiankui case. The Columbia team delivered the editor as a protein because, per the preprint, RNA delivery halted embryo development.
Technically, the result is ordinary by 2026 standards. Chinese groups have used base editing in embryos since 2017, and none of those embryos were ever implanted. What made the Columbia preprint land differently is the context. It was done in the United States, funded entirely outside the federal system, by a team that voluntarily cleared an Embryonic Stem Cell Research Oversight committee, an institutional review board, and an independent-ethicist donor-consent review, all of which are checkpoints the U.S. requires for federal grants and asks of private work only if the institution agrees.
Independent scientists read the work the same way. Alexis Komor, a base-editing researcher at UCSD's Sanford Stem Cell Innovation Center, told Scientific American that the project is "a gateway to embryo editing to do enhancements" and that "the cat's out of the bag." Krishanu Saha, a biomedical engineer at the University of Wisconsin-Madison, was blunter. "I would not call it a breakthrough," he said. "I find it hard to think about a scenario where this is medicine." His reasoning: embryos should serve a therapeutic end, and the Columbia embryos did not carry a pathogenic mutation the editing was meant to fix.
Megan Allyse, a bioethicist at Case Western Reserve University, focused on the governance gap. There is no U.S. statute requiring IRB or ESCRO review for privately funded embryo research, and the long-standing informal "gentleman's agreement" against embryo editing rests on the assumption that no one will simply publish. The Egli preprint is what that assumption looks like when it fails.
The 2020 international commission convened by the National Academies, the National Academy of Medicine, and the Royal Society set a narrow set of conditions for heritable human genome editing to be considered acceptable, limited to serious monogenic diseases with no alternatives, and paired with long-term follow-up and public engagement. By those criteria, the Columbia work is not even on the clinical pathway. The targets are not pathogenic, the embryos are not headed for transfer, and the manuscript is a preprint that has not been peer reviewed. The commission's framework is the most developed U.S.-facing governance document on the subject, and it is not law.
What changed this month is that the informal pause that has held since 2018 just produced a publication, not a scandal. The New York Times first reported on the preprint on June 4, 2026. Egli says the manuscript has been submitted to a peer-reviewed journal. Funding sources listed on the preprint include the Institute of Organic Chemistry and Biochemistry in Prague, the New York Stem Cell Foundation, Genomic Prediction, a commercial preimplantation genetic testing company, and the Korean Fund for Regenerative Medicine.
The mix is the point. The work cleared every U.S. rule that exists, was funded outside the federal system, and used embryos that no current statute was written to stop. The preprint reports editing efficiencies, mosaicism that could be reduced by editing five to twelve hours after fertilization, and no large chromosomal abnormalities in the embryos that were sequenced. The He Jiankui case ended with a prison sentence in China. The 2020 commission spelled out the conditions for a future that did not yet exist. The Columbia preprint sits in the gap those events left: legal in the U.S., governed only by voluntary review, and now public.
The question the preprint forces is not whether the science was done well. It is who, in the United States, is supposed to decide when research like this becomes acceptable, and on what evidence. That is the conversation the next twelve months will turn on, and it is one the country has not yet had at the level of statute.