Dongsheng Yuan, Robin Jugas, Petra Pokorna, Jaroslav Sterba, Ondrej Slaby, Simone Schmid, Christin Siewert, Brendan Osberg, David Capper, Skarphedinn Halldorsson, Einar O Vik-Mo, Pia S Zeiner, Katharina J Weber, Patrick N Harter, Christian Thomas, Anne Albers, Markus Rechsteiner, Regina Reimann, Anton Appelt, Ulrich Schüller, Nabil Jabareen, Sebastian Mackowiak, Naveed Ishaque, Roland Eils, Sören Lukassen, Philipp Euskirchen
{"title":"crossNN是跨平台的基于DNA甲基化的肿瘤分类的可解释框架。","authors":"Dongsheng Yuan, Robin Jugas, Petra Pokorna, Jaroslav Sterba, Ondrej Slaby, Simone Schmid, Christin Siewert, Brendan Osberg, David Capper, Skarphedinn Halldorsson, Einar O Vik-Mo, Pia S Zeiner, Katharina J Weber, Patrick N Harter, Christian Thomas, Anne Albers, Markus Rechsteiner, Regina Reimann, Anton Appelt, Ulrich Schüller, Nabil Jabareen, Sebastian Mackowiak, Naveed Ishaque, Roland Eils, Sören Lukassen, Philipp Euskirchen","doi":"10.1038/s43018-025-00976-5","DOIUrl":null,"url":null,"abstract":"<p><p>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% and 97.8% precision for the brain tumor and pan-cancer models, respectively.</p>","PeriodicalId":18885,"journal":{"name":"Nature cancer","volume":" ","pages":""},"PeriodicalIF":23.5000,"publicationDate":"2025-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"crossNN is an explainable framework for cross-platform DNA methylation-based classification of tumors.\",\"authors\":\"Dongsheng Yuan, Robin Jugas, Petra Pokorna, Jaroslav Sterba, Ondrej Slaby, Simone Schmid, Christin Siewert, Brendan Osberg, David Capper, Skarphedinn Halldorsson, Einar O Vik-Mo, Pia S Zeiner, Katharina J Weber, Patrick N Harter, Christian Thomas, Anne Albers, Markus Rechsteiner, Regina Reimann, Anton Appelt, Ulrich Schüller, Nabil Jabareen, Sebastian Mackowiak, Naveed Ishaque, Roland Eils, Sören Lukassen, Philipp Euskirchen\",\"doi\":\"10.1038/s43018-025-00976-5\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><p>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% and 97.8% precision for the brain tumor and pan-cancer models, respectively.</p>\",\"PeriodicalId\":18885,\"journal\":{\"name\":\"Nature cancer\",\"volume\":\" \",\"pages\":\"\"},\"PeriodicalIF\":23.5000,\"publicationDate\":\"2025-06-06\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Nature cancer\",\"FirstCategoryId\":\"3\",\"ListUrlMain\":\"https://doi.org/10.1038/s43018-025-00976-5\",\"RegionNum\":1,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"ONCOLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Nature cancer","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1038/s43018-025-00976-5","RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ONCOLOGY","Score":null,"Total":0}
crossNN is an explainable framework for cross-platform DNA methylation-based classification of tumors.
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% and 97.8% precision for the brain tumor and pan-cancer models, respectively.
期刊介绍:
Cancer is a devastating disease responsible for millions of deaths worldwide. However, many of these deaths could be prevented with improved prevention and treatment strategies. To achieve this, it is crucial to focus on accurate diagnosis, effective treatment methods, and understanding the socioeconomic factors that influence cancer rates.
Nature Cancer aims to serve as a unique platform for sharing the latest advancements in cancer research across various scientific fields, encompassing life sciences, physical sciences, applied sciences, and social sciences. The journal is particularly interested in fundamental research that enhances our understanding of tumor development and progression, as well as research that translates this knowledge into clinical applications through innovative diagnostic and therapeutic approaches. Additionally, Nature Cancer welcomes clinical studies that inform cancer diagnosis, treatment, and prevention, along with contributions exploring the societal impact of cancer on a global scale.
In addition to publishing original research, Nature Cancer will feature Comments, Reviews, News & Views, Features, and Correspondence that hold significant value for the diverse field of cancer research.