Rapid brain tumour classification from sparse epigenomic data
This study showcases a live brain tumour classification method combining Oxford Nanopore sequencing with MethyLYZR, an epigenomic analysis tool. By leveraging real-time genetic and epigenetic data from Oxford Nanopore sequencing, MethyLYZR achieved 94.5% accuracy within 15 minutes and required minimal computational power — highlighting its potential for future intraoperative diagnostics.
Key points:
MethyLYZR is a naïve Bayesian framework that runs in parallel with Oxford Nanopore sequencing for live methylation-based classification
Over 200 brain tumour samples were analysed, including 10 intraoperatively
MethyLYZR classified tumours with 94.52% accuracy in under 15 minutes, with a 1-hour biopsy-to-result turnaround time
It requires under 1 minute of processing time and less than 3GB of RAM, making MethyLYZR an attractive alternative to computationally intensive machine learning models
Tumours were also classified from cerebrospinal fluid cell-free DNA, highlighting potential for non-invasive liquid biopsy diagnostics in the future
Watch Franz-Josef Müller and Helene Kretzmer discuss their research at NCM 2021
Sample type: human brain tissue