Previously available under the MolecularNeuropathology.org website, our flagship CNS tumor methylation classifier is now live on our portal. It currently supports both Illumina Human Methylation EPIC v1 and EPIC v2 BeadChip arrays, as well as the older 450k platform.

METHOD 

Our classifier is trained on a well-annotated reference set constituting 7,495 samples covering 184 subclasses of CNS tumors. For swift, reproducible classification, we employ a random forest algorithm, providing a confidence level for each class determination. This ensemble method merges several independent component classifiers to enhance overall prediction accuracy. 

Our classifier algorithm rapidly compares each new case with thousands from our reference cohort, outputting a classification score reflecting resemblance to a particular tumor class. Each class yields a confidence score between 0 and 1. For a positive identification of a tumor type, the methylation subclass score must surpass a 0.9 threshold.

DATA 

The classifier operates on DNA methylation data. The cancer methylome combines somatically acquired DNA methylation alterations and attributes indicative of the cell of origin. Proven to be highly robust and reproducible, DNA methylation profiling can even be successful with small or low-quality samples. This technique has been extensively employed to meaningfully subclassify CNS tumors previously viewed as homogeneous diseases. 

USE 

To leverage our methylation profiling classifier, simply upload unprocessed IDAT-files of your samples from either Illumina Human Methylation 450K BeadChip or EPIC v1/v2 BeadChip arrays. Your data will automatically be classified with our classifier resulting in predictions for 4 hierarchy levels- methylation subclass, class, family and superfamily. You will also receive a copy number profile and the MGMT promoter methylation status prediction. A detailed report of your methylation profiling will be available in the portal. You can also download it as a PDF file. 

GET STARTED

The classifier is available at https://app.epignostix.com as the ‘CNS Tumor Methylation Classifier’. You will need to register as a new user here. Once your registration is approved, you can login to the portal using your email credentials. For step-by-step instructions on how to use our portal, please check our instructions video. If you have any questions, please check our FAQs or reach out to us through our Contact Form.