50 Comparative evaluation of long- and short-read methylation-aware sequencing for CNS tumour diagnostics

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Abstract Introduction DNA methylation–based classification has transformed the diagnosis of central nervous system (CNS) tumours. However, methylation profiling alone is insufficient for definitive diagnosis and must be integrated with genomic alterations such as single nucleotide variants, structural variants, and gene fusions. Molecular diagnostics are therefore moving towards unified sequencing approaches capable of simultaneously capturing epigenomic and genomic information within a single assay. Rapid nanopore sequencing has demonstrated the feasibility of combined methylation and genomic profiling within an intraoperative timeframe. Illumina 5-base whole-genome sequencing is a recently released short-read technology that distinguishes methylated from unmethylated cytosines, enabling integrated epigenomic and genomic analysis. Here, we compare the diagnostic performance of these long- and short-read methylation-aware sequencing platforms in CNS tumours. Method We analysed a matched cohort of 30 CNS tumour samples previously characterised using a rapid nanopore diagnostic workflow. Illumina 5-base whole-genome sequencing was generated for each case, alongside conventional methylation array profiling. Tumour classification was performed using Random Forest, NanoDx, Sturgeon, and mnp-Flex, referencing Heidelberg classifier versions 11 and 12.8. Concordance between nanopore, Illumina 5-base, and standard-of-care methylation array classifications was assessed. Detection of diagnostically relevant genomic alterations, including single nucleotide variants, structural variants, and gene fusions, was compared across platforms. Results Illumina 5-base sequencing demonstrated strong concordance with rapid nanopore classifications and established methylation array diagnoses. Agreement was consistent across classifiers and reference versions. Both approaches identified key diagnostic genomic alterations, supporting comprehensive molecular characterisation from a single assay. Conclusion Integrated methylation-aware sequencing platforms enabling simultaneous epigenomic and genomic profiling show robust cross-platform concordance and clinical feasibility for CNS tumour diagnostics.