Population frequency of Predicted pathogenic MisMatch Repair (MMR) gene variants in Lynch syndrome from bioinformatic analyses of the general population

Lynch syndrome is characterised by heterozygous germline mutations in the MisMatch Repair (MMR) genes and an increased risk of cancer. Previous population estimates based on cohorts with colorectal cancer suggest one in 300 people have a disease-causing variant. This study calculated the population frequency of Lynch syndrome from Predicted pathogenic variants in MMR genes in the general population. MLH1, MSH2, MSH6 and PMS2 variants were downloaded from gnomAD v.2.1, and annotated in ANNOVAR. Our population frequencies of heterozygous Predicted pathogenic variants were calculated from the sum of structural, null, rare computationally-damaging missense changes, and founder variants. Population frequencies were also derived from the proposed ClinGen variant specifications, and from pathogenic variants in the ClinVar, LOVD or InSiGHT websites. Predicted pathogenic variants were found in one in 94 people in gnomAD v.2.1.1 using our strategy, and one in 122, 203, 199 or 594 using the proposed ClinGen specifications, and the ClinVar, LOVD or InSiGHT databases, respectively. The frequencies derived from ClinVar (one in 203) and LOVD (one in 199) were based on accurate assessments of penetrant variants since they were largely derived from patient testing, but were underestimates because not all gnomAD variants had been assessed. Our strategy and that of the proposed ClinGen specifications examined each variant in gnomAD and resulted in more common population frequencies but some assessments may have been inaccurate, and variants incompletely penetrant. The number of Predicted pathogenic MMR gene variants in the general population suggests that Lynch syndrome is more common than reported previously.