Fig. 1

Overview of the TAS-LRS clinical workflow and validation framework. (a) Schematic of the end-to-end pharmacogenomics testing workflow based on Targeted Adaptive Sampling–Long Read Sequencing (TAS-LRS), comprising four main phases: optional pre-test consultation and sample collection; wet lab processing (DNA extraction, shearing, and Oxford Nanopore sequencing); bioinformatics analysis (including alignment, variant and star-allele calling, HLA typing, and genome-wide imputation); and post-test clinical consultation. Adaptive sampling enriches for 35 target pharmacogenes, including all PharmGKB Very Important Pharmacogenes (VIPs) and HLA-A, while generating low-depth off-target data for broader genome-wide applications. Blue, green, orange, and red indicate distinct workflow stages. (b) Validation framework for the TAS-LRS workflow, consisting of four core studies: (1) Limit of Detection (LOD), evaluating variant calling performance (SNVs, indels, and SVs) across different DNA input amounts (800 ng vs. 1000 ng) and basecalling models (High-accuracy (HAC) vs. Super-accuracy (SUP)); (2) Accuracy, assessed by diplotype and phenotype concordance across reference cell lines (GIAB, GeT-RM, 1KGP), external quality assessment (CAP EQA), and clinical samples from Tan Tock Seng Hospital (TTSH); (3) Precision, evaluated via reproducibility across sequencing runs and repeatability across independent bioinformatics analyses; and (4) Specificity, assessed through interference testing using triglyceride-spiked blood samples and cross-contamination testing using VerifyBamID2.