A single AI model can now predict physical traits from DNA across four different kingdoms of life.
April 1, 2026
Original Paper
BioWorldModel: a single architecture predicts phenotype from genotype across four kingdoms of life
bioRxiv · 10.64898/2026.03.27.714912
The Takeaway
This represents a major first in genomic modeling by creating a single architecture that understands biological processes across bacteria, fungi, animals, and plants. It suggests that the 'rules' for how a genome becomes a physical organism are universal enough to be captured by a single neural network.
From the abstract
The same genome produces different phenotypes in different conditions--yet predictive models encode genotype once and treat each trait independently. Here we show that representing phenotype generation as a dynamic biological process transforms predictive accuracy across bacteria, fungi, animals and plants. BioWorldModel learns how organisms interpret their genome: frozen gene embeddings (species context) modulated by individual variation pass through four biological process layers (regulation t