MetDecode: methylation-based deconvolution of cell-free DNA for non-invasive multi-cancer typing

Antoine Passemiers; Stefania Tuveri; Dhanya Sudhakaran; Tatjana Jatsenko; Tina Laga; Kevin Punie; Sigrid Hatse; Sabine Tejpar; An Coosemans; Els Van Nieuwenhuysen; Dirk Timmerman; Giuseppe Floris; Anne‐Sophie Van Rompuy; Xavier Sagaert; Antonia Carla Testa; Daniela Ficherova; Daniele Raimondi; Frédéric C. Amant; Liesbeth Lenaerts; Yves Moreau; Joris Robert Vermeesch

doi:10.1093/bioinformatics/btae522

Genetic & Genomic MedicineBioinformaticsAug 22, 2024

Peer-reviewed journal article

MetDecode: methylation-based deconvolution of cell-free DNA for non-invasive multi-cancer typing

Antoine Passemiers iD, Stefania Tuveri iD, Dhanya Sudhakaran, Tatjana Jatsenko iD, Tina Laga iD, Kevin Punie iD, Sigrid Hatse iD, Sabine Tejpar iD, An Coosemans iD, Els Van Nieuwenhuysen iD, Dirk Timmerman iD, Giuseppe Floris iD, Anne‐Sophie Van Rompuy iD, Xavier Sagaert iD, Antonia Carla Testa iD, Daniela Ficherova, Daniele Raimondi iD, Frédéric C. Amant iD, Liesbeth Lenaerts iD, Yves Moreau iD, Joris Robert Vermeesch iD

PDF DOI Preprint

Post to X

Threads on X Post to Bluesky Threads on Bluesky

Abstract

Cell-free DNA (cfDNA) mediated early cancer detection is based on detecting alterations in the cfDNA components. However, the underlying pathology can usually not be readily identified. We built a reference atlas based on the methylome of multiple cancer and blood-cell types and developed MetDecode, an epigenetic signature-based deconvolution algorithm. MetDecode accurately estimates the tumour proportion in in-silico mixtures and identifies the tissue of origin in 81.25% cfDNA samples from cancer patients. This method will complement cancer screening programs and guide clinical follow-up.