In JAMA Network Open, a team from the Fred Hutchinson Cancer Research Center, the University of Washington School of Medicine, and elsewhere outlines a strategy for finding SARS-CoV-2 mutations and variant strains with the help of a haplotype-based artificial intelligence approach. After training and validating the model with sequences from more than 10 million viral isolates collected from sites around the world, the researchers showed that it could pick up new SARS-CoV-2 mutations, variant strains, or mutation mixtures in a set of more than 344,900 prospectively collected isolates, identifying patterns that tracked closely with those reported in the "Global initiative on sharing avian influenza data" (GISAID). Along with variant mixtures involving Omicron, Delta, or Alpha viruses found in combination with other strains, for example, the model pointed to new mutations that appeared to be on the rise since May 2022. "The successful application of [haplotype-based artificial intelligence] in this study suggests its utility in identifying novel emerging SARS-CoV-2 variants even if such variants have not been observed previously," the authors write, adding that the approach "learned from the large collection of viral sequences in GISAID to identify core polymutants that were specific to viral variants."