Rice University researchers evolved compact CRISPR enzymes to recognize previously inaccessible DNA stretches by addressing weak PAM engagement. Published in Nature Communications, the approach used an AI-supported platform called Sequence Display to generate sequence–activity datasets rapidly, enabling SlugCas9 variants to edit target sites in genes APOC3 and PCSK9 that earlier versions could not reliably access. For genome-editing developers, the work is a practical demonstration that PAM-bottleneck solutions can be engineered without altering the genomic target itself, potentially expanding clinical editing utility for cardiovascular risk genes. —