Population-scale genome-wide prioritization of targets for allele-specific antisense oligonucleotide therapy design

Background: Allele-specific antisense oligonucleotides (ASOs) are increasingly used to selectively suppress transcripts containing pathogenic gain-of-function or dominant negative (GOF/DN) variants. However, most current approaches rely on patient-specific target selection of SNVs in cis with the disease-causing variant. This individualized approach may limit broader ASO applicability and could exacerbate inequities in access. Population-level selection of ASO-targetable SNVs is an alternative approach not previously explored at scale. Methods: For a curated set of >400 disease genes with putative toxic GOF/DN mechanisms, we applied a greedy algorithm to identify ASO-targetable SNVs that maximize population coverage in the 1000 Genomes Project dataset. Population coverage was evaluated across ″superpopulation″ ancestry groups and assessed in multiple diverse replication datasets. For each prioritized SNV we attempted to design allele-specific 15-mer and 20-mer ASOs meeting standard chemistry and sequence criteria. The presence of prioritized SNVs was assessed in two local cohorts with GOF/DN disorders: infants with epilepsy diagnosed by rapid short-read genome sequencing, and children with long-read genome sequencing data generated for bespoke ASO development. Results: We identified 7,927 SNVs (median 17.5 per gene) achieving a median cumulative heterozygous population coverage of 88.9% across 408 disease genes. More than two-thirds of genes (70.6%) reached ≥80% coverage using a median of only four prioritized SNVs. Coverage was consistently high for all ″superpopulation″ ancestry groups, with median coverage proportions of ≥85%. Prioritized SNVs generalized to independent ancestrally diverse replication datasets, achieving median coverage of 86.4% in the Human Genome Diversity Project and 90.3% in the All of Us Research Program . Many prioritized targets were compatible with established ASO design principles, enabling generation of candidate allele-specific 15-mer and 20-mer ASOs at scale for future experimental validation. The presence of prioritized targets in the rapid short-read sequencing cohort, and the overlap between prioritized targets and manually curated long-read genome sequencing targets, supports the clinical relevance of population-informed target selection. All SNV targets and candidate allele-specific ASO sequences are available at hawk-eye.research.sickkids.ca/knockdown. Conclusions: Prioritization of common SNV targets may facilitate the development of ASO therapies with broader population coverage, helping move the field beyond single-patient applications and toward more equitable opportunities for access.