On October 14, 2026, Colossal Biosciences announced GenoMorph, an AI platform for de-extinction research. Using comparative genomics, 3D morphology, and developmental biology, it identified approximately 2,300 DNA-editing targets associated with dodo cranial and facial morphology. The model was trained on developmental data from chickens, quail, ducks, and pigeons, plus genetic information and 3D skull measurements from more than 70 bird species. Colossal is working from only one dodo genome, compared with approximately 100 genomes available for its woolly mammoth project. [1]
GenoMorph does not edit genes; it ranks genetic changes for laboratory investigation by learning which DNA regions regulate beak and facial development across living birds. Colossal is also developing lipid nanoparticles to deliver gene-editing tools into developing bird embryos. The key technical value is inference under severe archival scarcity: cross-species biological data can identify plausible restoration targets when direct reference genomes are incomplete. [1]
The Ark team should monitor GenoMorph’s laboratory validation, prediction-to-phenotype accuracy, off-target effects, embryo-delivery performance, and reproducibility across independent datasets. Ark biological archives should preserve not only reference genomes but also developmental time-series data, 3D anatomical scans, regulatory-region annotations, and living proxy-species genomes; these combined datasets may enable post-collapse reconstruction of lost traits and ecosystem functions.