Evaluating NGS variant callers in a challenging genomic context with a focus on the PRSS1-PRSS2 locus for hereditary pancreatitis

Hereditary pancreatitis is frequently associated with pathogenic variants in PRSS1.1 However, the PRSS1 and PRSS2 loci reside within a genomic region with extensive sequence homology to multiple pseudogenes (ie, PRSS3P2 and TRY7) with reads misaligning to PRSS1/2, causing erroneous variant calling in conventional pipelines.1 2 Therefore, accurate identification of PRSS1/2 variants with high sensitivity and specificity is essential for reliable clinical diagnosis and genetic risk assessment. To address these challenges, Lou et al recently developed NGS.PRSS1-2caller,1 which improves variant calling accuracy by using an alternative contig of chromosome 7 from GRCh38 with all relevant pseudogenes.1 2 Concurrently, the DRAGEN platform has demonstrated high accuracy in genome analysis due to its algorithmic optimisation and hardware acceleration.3 This study aims to systematically evaluate the performance in clinically relevant settings of multiple next-generation sequencing (NGS) variant calling workflows (online supplemental methods; figure 1A) including popular approach (GATK HaplotypeCaller4), the custom NGS.PRSS1-2caller1 and DRAGEN3 using our Australian discovery5 (n=44) and validation cohorts (n=24) to detect PRSS1/PRSS2 variants.SP110.1136/gutjnl-2025-336988.supp1Supplementary data