A Science paper describes an AI-designed synthetic TnpB enzyme, SynTnpBs, designed to expand the range of programmable gene-editing options beyond natural references. Researchers report that the synthetic enzyme outperformed a natural reference, demonstrating that computational design can generate editing proteins with improved performance characteristics. For gene therapy and genome engineering teams, the work signals continued momentum to diversify editing components—potentially affecting guide choice, delivery constraints, and compatibility with different target contexts.