Colossal's new AI hunts for dodo gene-editing targets

There isn't enough dodo DNA left to study the normal way. Scientists usually compare many genomes from one species to find the genes behind a trait. The dodo doesn't have enough surviving genetic material for that. So on Oct. 14, Colossal Biosciences unveiled an AI platform called GenoMorph to work around that gap. It's already flagged about 2,300 candidate genes tied to the dodo's beak and skull shape.

Beth Shapiro, Colossal's chief science officer, explained why this tool matters. Normal gene-hunting needs lots of genomes from the same species to compare. Extinct species like the dodo don't have that luxury. GenoMorph instead compares three things: body shape in 3D, genomes from related living birds, and how those birds develop. From that, it guesses which genes likely shape a trait like beak size. That's a real scientific shortcut. But it's still a guess about which genes matter, not proof those genes will work once edited. And the actual editing isn't happening in dodo cells at all. It's happening in a living relative, the Nicobar pigeon, the closest thing to a dodo still alive.

Keep reading — what to watch

So when does an actual dodo-like bird show up? Not soon. GenoMorph only narrows down which genes are worth editing next — it doesn't build an animal. Colossal still has to test those edits in Nicobar pigeon cells. Then it needs to confirm they actually produce dodo-like traits. Then carry that through an entire breeding and development process. Every one of those steps can fail or take years. The company has made bold timelines before. Whether this tool actually speeds things up, or just better organizes the guesswork, is the real test ahead.