Population-scale long-read methylation signatures for rare disease classification | LC26

Abstract
There is growing evidence that long-read sequencing (LRS) is superior to standard genetic testing workflows in the clinical environment. As a single assay, LRS can replace many stepwise tests performed today while also providing improved resolution of complex genomic regions, including repetitive elements, structural variants, and phased haplotypes. While an increasing number of clinical laboratories are validating LRS-based workflows, there remains a critical need for population-scale reference datasets derived from both affected and unaffected individuals to support variant filtering, prioritization, and clinical validation. A key and underexploited advantage of LRS is its ability to directly detect native DNA methylation without additional library preparation or bisulfite conversion, enabling the simultaneous assessment of genetic and epigenetic variation from a single experiment. In the context of rare disease diagnosis — particularly for disorders driven by epigenetic dysregulation — this capability enables the identification of disease-associated episignatures that are often undetectable using sequence-only approaches. However, publicly available, population-scale LRS methylation reference datasets remain limited. To address this gap, we developed EpiSignaLR, a tool for classifying rare disease-associated episignatures from LRS-derived methylation data. EpiSignaLR currently supports classification of more than 30 genetic conditions and includes refined and improved LRS-derived episignatures for Coffin–Siris syndrome that are superior to array-derived models. By integrating sequence, structural variation, and epigenetic signatures into a unified analytic framework, EpiSignaLR enables more comprehensive interpretation of LRS data in both research and clinical settings. We conclude with a call for broad community collaboration to expand this framework through the generation, validation, and public dissemination of LRS-derived episignatures for additional conditions. The development of openly accessible reference datasets will be essential to realizing the full diagnostic potential of LRS as a single, integrated assay for rare disease evaluation.

Biography
Danny Miller is an Assistant Professor in the Department of Pediatrics, Division of Genetic Medicine, and the Department of Laboratory Medicine and Pathology at the University of Washington and is an attending physician at Seattle Children’s Hospital. His laboratory is developing long-read sequencing-based clinical genetic tests with a goal of increasing the rate of genetic diagnoses, reducing the time required to make a genetic diagnosis, and lowering barriers to obtaining comprehensive clinical testing. Clinically, he cares for patients in both general genetics and skeletal dysplasia clinics.