Guide reselection expands canonical PAM candidates in an HGD editing resource for alkaptonuria

Rare monogenic diseases offer defined molecular targets, but sequence-level editing options require explicit selection and verification. Using artificial intelligence assisted analysis and public tools within a community-college research cohort, we assembled an editing resource for the homogentisate 1,2-dioxygenase gene (HGD) in alkaptonuria. The inventory contained 278 variants and represented 268 of 271 eligible records in a defined ClinVar census. We compared filtering fixed prime-editing designs with re-selection requiring canonical NGG protospacer-adjacent motifs at both guide targets. Re-selection preserved 93 compatible configurations and added 140, all using a second nick without predicted edited-allele discrimination (PE3), while retaining 92 predicted allele-discriminating configurations (PE3b). Under a shared genomic count screen including NGG, NAG and NGA sites, 70 preserved and 79 added configurations remained. With an unchanged base-editing arm, combined representation increased from 110 to 188 variants, or 106 to 180 of 271 eligible ClinVar records. Preserving fixed configurations makes nonnegative gain structural; a 33-setting threshold grid quantified additional representation, with gains of 2–121. Conservative base-window exclusions and a prime-only control retained the increase. Sequence reconstruction and independent PrimeDesign comparisons checked component conventions. The result identifies nicking-guide selection as a source of additional canonical candidates and motivates planned tests of editing activity, product purity and restoration of HGD function.