Long-read sequencing for detecting methylation dysregulation

Abstract DNA methylation is a critical epigenetic signature that plays a role in human development and disease and is especially recognized for its role in imprinted diseases, such as Angelman, Prader-Willi, or Russel-Silver syndrome. Further, it may serve as a molecular marker of disease severity or progression even in diseases not primarily driven by methylation. We single-molecule, whole-genome sequenced 55 pediatric patients with severe, acute disease using the Oxford Nanopore platform. Using these data, we developed a novel algorithm for detecting regions across the genome with aberrant methylation by using our samples to train an algorithm on ‘normal’ methylation. Using additional short-read parental variant data we can parent-of-origin phase across the proband methylome. Further, our method can be used to detect imprinting errors even without parental samples by taking advantage of the phased nature of the sequencing data. We successfully re-identified 58 known regions of imprinting and detected an additional 45 shared regions that appeared to have differential methylation based on parent of origin. Additionally, we identify broad methylation dysregulation in known imprinting diseases, as well as in immune patients and a patient with trisomy 21. Together these data show the value of long-read, single-molecule sequencing to elucidate methylation patterns ‘for free’ when whole-genome sequencing. These data can identify known imprinted regions, potentially diagnose diseases of methylation, and can even identify molecular markers of pathology in diseases that are not driven by methylation. This further punctuates the multifaceted role of long-read sequencing to potentially diagnose a plethora of diseases. Biography Matthew Bainbridge has worked with next-generation sequencing since 2006. He is currently the head of research at Rady Children's Institute of Genomic Medicine.