Open-Source Bioinformatic Pipeline to Improve PMS2 Genetic Testing Using Short-Read NGS Data

The molecular diagnosis of mismatch repair–deficient cancer syndromes is hampered by difficulties in sequencing the PMS2 gene, mainly owing to the PMS2CL pseudogene. Next-generation sequencing short reads cannot be mapped unambiguously by standard pipelines, compromising variant calling accuracy. This study aimed to provide a refined bioinformatic pipeline for PMS2 mutational analysis and explore PMS2 germline pathogenic variant prevalence in an unselected hereditary cancer (HC) cohort. PMS2 mutational analysis was optimized using two cohorts: 192 unselected HC patients for assessing the allelic ratio of paralogous sequence variants, and 13 samples enriched with PMS2 (likely) pathogenic variants screened previously by long-range genomic DNA PCR amplification. Reads were forced to align with the PMS2 reference sequence, except those corresponding to exon 11, where only those intersecting gene-specific invariant positions were considered. Afterward, the refined pipeline's accuracy was validated in a cohort of 40 patients and used to screen 5619 HC patients. Compared with our routine diagnostic pipeline, the PMS2_vaR pipeline showed increased technical sensitivity (0.853 to 0.956, respectively) in the validation cohort, identifying all previously PMS2 pathogenic variants found by long-range genomic DNA PCR amplification. Fifteen HC cohort samples carried a pathogenic PMS2 variant (15 of 5619; 0.285%), doubling the estimated prevalence in the general population. The refined open-source approach improved PMS2 mutational analysis accuracy, allowing its inclusion in the routine next-generation sequencing pipeline streamlining PMS2 screening.