Solving Lynch syndrome missing heritability by integrating long-read DNA sequencing and short-read RNA sequencing

ASHG 2026

Oct 20, 2026
Oncology
Presentation

Christopher J. Hatch, Sami Belhadj, John Harting, Cassidy Carraway, Hoda Mirsafian, Stuti Joshi, Kenny Chen, Stephen Tanner, Bojan Losic, Marcy E. Richardson, Seth I. Berger, Rachid Karam

Abstract
Lynch Syndrome (LS) is an inherited genetic condition that increases the risk of developing numerous cancers, including colorectal and endometrial cancer, and affects approximately 1:300 individuals. It is caused by pathogenic germline variants in the DNA mismatch repair (MMR) genes MLH1, MSH2, MSH6 and PMS2; and select deletions in EPCAM. Constitutional MLH1 promoter methylation (hereinafter epimutation), can also cause LS. Many individuals suspected to have LS receive negative clinical testing results potentially due to inherent limitations of conventional testing in detecting structural variants (SVs), regions complicated by pseudogenes, and epimutations. Advancements in long-read sequencing (LRS) and targeted RNA sequencing analysis may hold the potential to close these gaps.

 

Two cohorts of unexplained MMR-deficient patients were analyzed. Cohort 1 included 96 patients diagnosed with at least 1 LS-associated cancer that underwent targeted LRS (T-LRS). Cohort 2 included 250 individuals, unselected for personal or family cancer history, that underwent allele-specific expression (ASE) analysis to prioritize cases for Long-Read WGS (LR-WGS). For both cohorts, DNA isolated from whole blood, buccal swabs or saliva specimens was available for additional testing.

 

In Cohort 1, T-LRS did not identify additional causative germline variants affecting MMR genes. Single nucleotide variants (SNV) with a deleterious in-silico prediction were already classified as (likely) benign. Intronic SVs showed no significant impact on splicing with short-read RNA sequencing. One sample was identified as a possible epimutant based on ASE of MLH1. 54% of samples could not undergo ASE because of a lack of exonic SNPs, but two showed decreased normalized expression levels of MLH1 suggestive of an epimutation.

 

Both T-LRS and clinical tests are PCR-based approaches that lose methylation signals. However, LR-WGS can endogenously detect 5-mC methylation directly from native DNA. With this in mind, ASE analysis ran on cohort 2 identified 3/250 cases (1.2%) with complete MLH1 ASE. Subsequent LR-WGS confirmed MLH1 epimutations in the silenced allele.

 

In these cohorts, where most individuals already received paired DNA and RNA clinical testing, the ability to detect epimutants had a greater impact on undiagnosed LS cases than searching for additional SNVs and SVs, suggesting that further investigation is warranted to assess the contribution of epimutations to hereditary cancer risk.