(888) 440-0954


Early detection and diagnosis of cancer with interpretable machine learning to uncover cancer-specific DNA methylation patterns

Biol Methods Protoc, Vol. 9, Issue 1, Article Number: bpae028, 2024

PMID: 38903861

In this study, the authors present an interpretable machine learning model that can classify 13 cancer types and non-cancer tissue samples using only DNA methylome data, achieving 98.2% accuracy. The features identified by this model are utilized to develop EMethylNET, a robust framework consisting of an XGBoost model that informs a deep neural network, capable of generalizing to independent data sets. The authors also demonstrate that the methylation-associated genomic loci detected by the classifier are linked to genes, pathways, and networks involved in cancer, providing insights into the epigenomic regulation of carcinogenesis.

Read More


Considering Genomic Research?

    Fields marked in * is mandatory to fill

    2024 © MedGenome • All Rights Reserved
    Request a Quote