Clinical Impact of Pharmacogenomics in Pediatric Care: Insights Extracted from Clinical Exome Sequencing

Simran D S Maggo^1*Yachen Pan²Dejerianne Ostrow²Jenny Nguyen²Jaclyn A Biegel²Matthew A Deardorff²Xiaowu Gai³

• ¹School of Pharmacy, Shenandoah University, Winchester, United States
• ²Children's Hospital of Los Angeles, Los Angeles, California, United States
• ³Medical College of Wisconsin, Milwaukee, Wisconsin, United States

Pharmacogenomic (PGx) testing improves drug efficacy and reduces risk of toxicity for commonly prescribed medications, with most pharmacogenomic studies largely focused on individuals of European descent to date. The impact of pharmacogenomic testing in a racially diverse population is still emerging, especially for Admixed American patients. In this study, we assessed the frequency of actionable PGx variants by analyzing anonymized exome sequencing data of a racially diverse cohort of 1777 pediatric patients, collected for routine clinical genetic diagnosis at Children’s Hospital Los Angeles (CHLA). Utilizing exome data, we used the Illumina DRAGEN germline pipeline v4.2, to determine the predicted phenotypes of 25 pharmacogenes including HLA-A and HLA-B, including CPIC Level A genes and genes recommended for PGx testing by the U.S. Food and Drug Administation. To evaluate cross-platform consistency, we compared results against that of a modified version of the Stargazer pipeline, PyPGx. As the distribution of PGx alleles is ancestry specific, we estimated genetic ancestry bioinformatically using the Somalier tool. Genetic ancestry analysis demonstrated that 60% of our cohort was Admixed American, followed by 23% European, 8% East Asian, 5% African American, and 2% South East Asian. Actionability analysis showed that: 1) 93% of all exome cases had at least one actionable PGx phenotype, 2) one in five cases (22%) had at least three actionable PGx phenotypes, and 3) CYP2B6 (54%) and CYP2D6 (33%) had the highest number of actionable phenotypes. Further analysis revealed notable differences, including higher rates of poor metabolizers for CYP2B6 and variations in CYP2D6 metabolizer statuses, in PGx phenotypes compared to previously collated frequencies in the PharmGKB database, especially within the Admixed American population. In conclusion, our study reinforces the importance of PGx testing, underscores the diversity of PGx variation in ancestral backgrounds, and supports the clinical utility of preemptive PGx testing using exome or genome sequencing approaches.