On November 24, 2025, researchers from Harvard Medical School in the United States and the Center for Genomic Regulation in Barcelona, Spain, unveiled their groundbreaking findings in the prestigious journal Nature Genetics. They introduced an innovative AI model dubbed popEVE. This model leverages evolutionary data spanning hundreds of thousands of species and genetic variation information from the entire human population. As a result, it is capable of precisely pinpointing the most disease-associated mutations in human proteins and ranking the extent of harm these mutations can inflict on the human body. This breakthrough holds the promise of revolutionizing the diagnostic landscape for genetic diseases.

During the validation process, popEVE demonstrated remarkable accuracy. In 98% of the cases, it successfully flagged co-mutations as the most detrimental variants, outperforming comparable tools. Moreover, the model uncovered 123 previously unidentified genes linked to developmental disorders, opening up new avenues for research and potential treatments.