Oxford Nanopore at the Annual Cancer Research Meeting

Retinoblastoma (RB) is the most common pediatric ocular cancer due to biallelic inactivation of the retinoblastoma gene (RB1). Those that carry a germline variant in RB1 have increased risk of additional malignant neoplasms that are associated with significant morbidity and mortality. There is mounting evidence that parent-of-origin of the germline variant has a significant role in disease severity for both ocular and extraocular tumors. Parent-of-origin of germline RB cannot be deciphered in the vast majority of cases with familial testing, limiting potential proper prognostic and therapeutic considerations. In this proposal we leverage the ability of targeted long-read sequencing to simultaneously assay genetic and epigenetic base level changes to provide haplotype-phased long-read genome and CpG methylome in a single experimental run without the need for tiered sequencing. We benchmarked this multi-omic approach in cases with known familial inheritance of retinoblastoma to show that we can correctly identify the parent-of-origin of the germline variant in each case by phasing it with an imprinted methylation signal in intron 2 of RB1. We then applied this approach to heritable retinoblastoma cases due sporadic RB1 with a diverse set of mutations, some of which had evaded clinical genetic characterization, and resolved parent-of-origin alleles in each case and demonstrated that it correlated with extent of disease. Finally, we showed that our targeted long-read sequencing approach can provide germline, parent-of-origin diagnosis in hours. This demonstrates that there is tremendous potential for this multi-omic approach to be developed into a clinical test to reveal the molecular basis of disease in RB in a single sequencing run and allow for a more personalized approach for diagnosis and treatment in this vulnerable pediatric population to improve disease outcomes.

Acute leukemia (AL) is an aggressive blood cancer that requires precise molecular classification and urgent treatment. However, standard-of-care diagnostic tests are time and resource intensive and may not resolve the full spectrum of AL disease heterogeneity. In this presentation, Dr. Griffin will discuss recent efforts to use long-read (nanopore) sequencing and machine learning to rapidly classify AL solely on the basis of epigenomic (DNA methylation) information. The promise and limitations of these approaches to complement standard-of-care diagnostics for AL in the clinic will be discussed, as will opportunities to further develop methylation-based rapid diagnostics for cancer.