On April 15 the Food and Drug Administration released a draft guidance titled "Safety Assessment of Genome Editing in Human Gene Therapy Products Using Next-Generation Sequencing" — the first standardized safety framework the agency has written for CRISPR, base-editing, and prime-editing therapies. [1] The document specifies which off-target detection methods the FDA will accept in an Investigational New Drug submission, at what depth of sequencing coverage, and with what orthogonal confirmation. [2]

It is paired, on the same Wednesday, with a bespoke rare-disease pathway modeled on the Baby KJ case — the infant whose individualized base-editing therapy was designed, manufactured, and administered in under six months in 2025. [3] The two documents read as a single regulatory on-ramp: a standard the field can cite, and a pathway that accepts population-of-one trials when the disease is lethal and the editing is surgical. Jennifer Doudna, David Liu, and Fyodor Urnov each posted within the day that this was the most consequential FDA guidance for gene editing since 2017's CAR-T framework. [4]

The economics matter as much as the biology. Until Wednesday, the blocker every biotech named — the reason Andrew Hessel and others were lobbying in public — was the absence of safe-harbor language: without a standard, every off-target finding became litigation exposure, and every sponsor priced that risk at zero clinical trials. [5] The draft does not eliminate the risk. It names what "safe enough to proceed" looks like.

Comment period runs sixty days. The finalized guidance will arrive after a Congress that has spent April fighting about AI infrastructure and oil. It will arrive anyway. The field now knows what the FDA will ask for, and — for the first time — what it will accept.

-- KENJI NAKAMURA, Tokyo