GPN-Star model uses evolutionary trees to predict genetic variant effects
A genomic language model called GPN-Star uses species trees and whole-genome alignments to learn functional constraints directly from DNA. Built by C. Ye and colleagues, the phylogeny-aware architecture evaluates evolutionary timescales across vertebrates, mammals and primates. In tests, the model beat existing methods at prioritizing pathogenic variants and fine-mapped genome-wide association study variants while improving rare variant association testing. The authors also developed versions for Mus musculus, Gallus gallus, Drosophila melanogaster, Caenorhabditis elegans and Arabidopsis thaliana.
The framework provides a scalable tool for human genome interpretation that also aids laboratory model research.
Written by the Genomes desk from the primary source cited and linked above and checked against it. Research use only; not medical advice. Corrections: [email protected].