Long‐Read Sequencing Enhances Pharmacogenomic Profiling by Resolving Complex Haplotypes, Novel Star Alleles, and Structural Variants

Pharmacogenomics has shifted drug therapy from trial-and-error to personalized approaches by leveraging genetic biomarkers. However, traditional genotyping assays and short-read sequencing often fail to resolve structural complexity in pharmacogenes, leading to ambiguous haplotypes and incorrect phenotype predictions. To evaluate the potential of emerging long-read sequencing technologies in overcoming these limitations, we analyzed diplotypes and drug response phenotypes across 20 clinically actionable pharmacogenes using Oxford Nanopore or PacBio data from 100 healthy individuals in the 1,000 Genomes Project and 159 participants from the Genomics England cancer and rare disease cohorts, alongside Illumina short-read sequencing data. Long reads achieved phasing accuracies of >98% (Oxford Nanopore) and 96.5% (PacBio), with most genes covered by a single phased haploblock. Variant detection metrics were high (i.e., precision, recall, and F1 > 0.92). Genotype and phenotype concordance between short- and long-read predictions exceeded 99%. Short-read predictions contributed nearly twice as much to overall discordant cases compared to those of long reads. Notably, long reads identified 19 novel star (*) alleles and 106 novel suballeles in nine pharmacogenes, including CYP2D6, CYP2B6, CYP2C9, CYP2C19, CYP4F2, and SLCO1B1, which have been submitted to PharmVar. Additionally, long reads resolved 13 ambiguous CYP2D6 structural variants, including a potentially novel structure. Also, a homozygous TA repeat (UGT1A1*80 + *28) was identified that is associated with a poor metabolizer phenotype. Overall, we reassigned 77 genotypes across nine genes for 58 of the 100 investigated 1,000 Genomes Project subjects. These results demonstrate the superiority of long-read sequencing in phasing and resolving complex genomic regions enabling more precise pharmacogenomic profiling. As sequencing costs decline with rapid technological advances, long-read sequencing may become the method of choice in clinical pharmacogenomics, enhancing therapeutic safety and efficacy.