TY  - JOUR
AU  - Yuan, Dongsheng
AU  - Jugas, Robin
AU  - Pokorna, Petra
AU  - Sterba, Jaroslav
AU  - Slaby, Ondrej
AU  - Schmid, Simone
AU  - Siewert, Christin
AU  - Osberg, Brendan
AU  - Capper, David
AU  - Halldorsson, Skarphedinn
AU  - Vik-Mo, Einar O
AU  - Zeiner, Pia S
AU  - Weber, Katharina
AU  - Harter, Patrick N
AU  - Thomas, Christian
AU  - Albers, Anne
AU  - Rechsteiner, Markus
AU  - Reimann, Regina
AU  - Appelt, Anton
AU  - Schüller, Ulrich
AU  - Jabareen, Nabil
AU  - Mackowiak, Sebastian
AU  - Ishaque, Naveed
AU  - Eils, Roland
AU  - Lukassen, Sören
AU  - Euskirchen, Philipp
TI  - crossNN is an explainable framework for cross-platform DNA methylation-based classification of tumors.
JO  - Nature cancer
VL  - 6
IS  - 7
SN  - 2662-1347
CY  - London
PB  - Nature Research
M1  - DKFZ-2025-01181
SP  - 1283-1294
PY  - 2025
N1  - 2025 Jul;6(7):1283-1294
AB  - DNA methylation-based classification of (brain) tumors has emerged as a powerful and indispensable diagnostic technique. Initial implementations used methylation microarrays for data generation, while most current classifiers rely on a fixed methylation feature space. This makes them incompatible with other platforms, especially different flavors of DNA sequencing. Here, we describe crossNN, a neural network-based machine learning framework that can accurately classify tumors using sparse methylomes obtained on different platforms and with different epigenome coverage and sequencing depth. It outperforms other deep and conventional machine learning models regarding accuracy and computational requirements while still being explainable. We use crossNN to train a pan-cancer classifier that can discriminate more than 170 tumor types across all organ sites. Validation in more than 5,000 tumors profiled on different platforms, including nanopore and targeted bisulfite sequencing, demonstrates its robustness and scalability with 99.1
LB  - PUB:(DE-HGF)16
C6  - pmid:40481322
DO  - DOI:10.1038/s43018-025-00976-5
UR  - https://inrepo02.dkfz.de/record/301911
ER  -