Researchers Develop Model Predicting Avian Flu Spillover

A research team led by Liam Brierley developed a machine learning model, detailed in a preprint under peer review, to predict which avian influenza strains can jump to humans. Trained on roughly 19,000 viral sequences (including 618 human samples), the model achieved 91.9% identification accuracy and flagged short motifs in polymerase, HA, NP, and NS1 genes. The approach could guide surveillance prioritization and targeted vaccine design.