[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-seo-377908-105":3,"detail-sidebar-cat-0-en-105":81,"doc-detail-377908-en":130},{"code":4,"msg":5,"data":6},0,"ok",{"site_id":7,"language":8,"slug":9,"title":10,"keywords":11,"description":12,"schema_data":13,"social_meta":74,"head_meta":76,"extra_data":78,"updated_unix":80},105,"en","deep-learning-model-accurately-classifies-metastatic-tumors-from-primary-tumors-based-on-mutational-signatures","Deep learning model accurately classifies metastatic tumors from primary tumors based on mutational signatures","","Metastatic propagation drives cancer morbidity and mortality, and identifying metastatic tumors and their origins is essential for effective clinical treatment. This study develops MetaWise, a deep neural network that uses mutational signatures computed from whole-exome sequencing data as input features. MetaWise accurately classifies metastatic tumors from primary tumors, outperforming traditional machine learning and the deep learning baseline DiaDeL. Non-coding mutation signatures substantially affect performance, while SHAP and LIME reveal signatures correlated with metastatic spread.",{"@graph":14,"@context":73},[15,34,56],{"@type":16,"itemListElement":17},"BreadcrumbList",[18,23,27,31],{"item":19,"name":20,"@type":21,"position":22},"https://docshare.wps.com","Home","ListItem",1,{"item":24,"name":25,"@type":21,"position":26},"https://docshare.wps.com/document/","Document",2,{"item":28,"name":29,"@type":21,"position":30},"https://docshare.wps.com/document/healthcare/","Healthcare",3,{"item":32,"name":10,"@type":21,"position":33},"https://docshare.wps.com/document/deep-learning-model-accurately-classifies-metastatic-tumors-from-primary-tumors-based-on-mutational-signatures/377908/",4,{"url":32,"name":10,"@type":35,"image":36,"author":41,"headline":10,"publisher":44,"fileFormat":47,"inLanguage":8,"description":12,"dateModified":48,"datePublished":49,"encodingFormat":47,"isAccessibleForFree":50,"interactionStatistic":51},"DigitalDocument",{"url":37,"@type":38,"width":39,"height":40},"https://docshare.wps.com/thumbnails/deep-learning-model-accurately-classifies-metastatic-tumors-from-primary-tumors-based-on-mutational-signatures/377908.png","ImageObject",300,407,{"name":42,"@type":43},"Aurelia","Person",{"url":19,"name":45,"@type":46},"DocShare","Organization","application/pdf","2026-10-22","2026-09-24",true,{"@type":52,"interactionType":53,"userInteractionCount":55},"InteractionCounter",{"@type":54},"ViewAction",5,{"@type":57,"mainEntity":58},"FAQPage",[59,65,69],{"name":60,"@type":61,"acceptedAnswer":62},"What is MetaWise and what data does it use?","Question",{"text":63,"@type":64},"MetaWise is a deep neural network that uses mutational signatures as input features. These signatures are calculated from whole-exome sequencing (WES) data from TCGA and other metastatic cohorts.","Answer",{"name":66,"@type":61,"acceptedAnswer":67},"How does MetaWise perform compared with other models?",{"text":68,"@type":64},"MetaWise accurately classifies metastatic tumors from primary tumors and outperforms traditional machine learning models and the deep learning model DiaDeL.",{"name":70,"@type":61,"acceptedAnswer":71},"Which analyses and signatures explain the model’s predictions?",{"text":72,"@type":64},"SHAP and LIME analyses identify mutational signatures correlated with metastatic spread, including APOBEC mutagenesis, UV-induced signatures, and DNA damage response deficiency signatures. Non-coding mutation signatures also meaningfully influence performance.","https://schema.org",{"og:url":32,"og:type":75,"og:title":10,"og:site_name":45,"og:description":12},"article",{"robots":77,"canonical":32},"index,follow",{"doc_id":79,"site_id":7},377908,1791099926,{"code":4,"msg":82,"data":83},"success",[84,88,92,96,100,105,109,114,119,122,126],{"id":22,"doc_module":4,"doc_module_name":25,"category_name":85,"show_sort_weight":86,"slug":87},"Story & Novel",90,"story-novel",{"id":26,"doc_module":4,"doc_module_name":25,"category_name":89,"show_sort_weight":90,"slug":91},"Literature",80,"literature",{"id":33,"doc_module":4,"doc_module_name":25,"category_name":93,"show_sort_weight":94,"slug":95},"Exam",70,"exam",{"id":55,"doc_module":4,"doc_module_name":25,"category_name":97,"show_sort_weight":98,"slug":99},"Comic",60,"comic",{"id":101,"doc_module":4,"doc_module_name":25,"category_name":102,"show_sort_weight":103,"slug":104},6,"Technology",50,"technology",{"id":106,"doc_module":4,"doc_module_name":25,"category_name":29,"show_sort_weight":107,"slug":108},7,40,"healthcare",{"id":110,"doc_module":4,"doc_module_name":25,"category_name":111,"show_sort_weight":112,"slug":113},8,"Research & Report",30,"research-report",{"id":115,"doc_module":4,"doc_module_name":25,"category_name":116,"show_sort_weight":117,"slug":118},9,"Religion & Spirituality",20,"religion-spirituality",{"id":117,"doc_module":4,"doc_module_name":25,"category_name":120,"show_sort_weight":117,"slug":121},"World Cup","world-cup",{"id":123,"doc_module":4,"doc_module_name":25,"category_name":124,"show_sort_weight":123,"slug":125},10,"Lifestyle","lifestyle",{"id":127,"doc_module":4,"doc_module_name":25,"category_name":128,"show_sort_weight":55,"slug":129},19,"General","general",{"code":4,"msg":82,"data":131},{"doc_id":79,"user_id":132,"nickname":42,"user_avatar":133,"doc_module":4,"category_id":106,"category_name":29,"doc_title":10,"doc_description":12,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":55,"is_deleted":4,"is_public":22,"is_downloadable":22,"audit_status":22,"page_count":139,"language":140,"language_code":8,"site_id":7,"html_lang":8,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":12,"update_tm":144,"read_time":145},1099514068365,"https://ap-avatar.wpscdn.com/avatar/10000253d8d9f28188e?_k=1776742907772140068","[www. nature.com/scientificreports](www. nature.com/scientificreports)  \nOPEN  \nDeep learning model accurately classifies metastatic tumors from primary tumors based on mutational signatures  \nWeisheng Zheng1,5, Mengchen Pu1,5, Xiaorong Li1,2, Zhaolan Du1,3, Sutong Jin1,4, Xingshuai Li1, Jielong Zhou1 & Yingsheng Zhang1*  \nMetastatic propagation is the leading cause of death for most cancers. Prediction and elucidation of metastatic process is crucial for the treatment of cancer. Eventhough somatic mutations have been linked to tumorigenesis and metastasis, it is less explored whether metastatic events can be identified through genomic mutational signatures, which are concise descriptions ofthe mutational processes. Here, we developed MetaWise, a Deep Neural Network (DNN) model, by applying mutational signatures as input features calculated from Whole-Exome Sequencing (WES) data of TCGA and other metastatic cohorts. This model can accurately classify metastatic tumors from primary tumorsand outperform traditional machine learning (ML) models and a deep learning (DL) model, DiaDeL. Signatures of non-coding mutations also have a major impact on the model’s performance. SHapley Additive exPlanations (SHAP) and Local Surrogate (LIME) analyses identify several mutational signatures which are directly correlated to metastatic spread in cancers, including APOBECmutagenesis, UV-induced signatures, and DNA damage response deficiency signatures.  \nFor most cancers, the metastatic spread is the main cause of cancer morbidity and mortality1. Studies have shown that nearly 90% of cancer deaths are the result of metastasis2. Once detaching from the primary site, metastatic cancer cells can travel through the circulatory system and establish themselves in new sites, potentially impacting multiple organs and systems3,4. The malignant cells can invade the local blood or lymphatic vessels to access to the systemic circulation. The lymphatic vessels transport immune cells throughout the body. Cancer cells can exploit these vessels to migrate to nearby lymph nodes and regional lymph nodes, evade the self-destruction meditated by immune cells, and eventually to distant organs and initiate tumorigenesis in those metastatic sites5. The distribution of metastatic sites for a given primary tumor is not arbitrary, but rather dictated by several factors such as the anatomical location, the origin of the cell, and molecular subtypes6. As a result, the identification of metastatic tumors and their origins is pivotal for choosing appropriate and effective therapeutic treatment in clinical practice.  \nIndividuals accumulate somatic mutations over their lifetime as a result of various genetic and environmental factors7,8. These mutations can occur in any part of the genome and may have different effects on gene function and regulation. While the majority of these somatic mutations are neutral and accumulated in a passive manner, some genomic changes on the DNA sequences alter the regulation and function of genes, leading to the abnormal phenotypes of cells7. Accumulation of mutations in critical regulatory genes can lead to the development of diseases, such as cancers. Therefore, mutations serve as valuable features to decipher the developmental stages of tumorigenesis, including primary and metastatic tumors. By analyzing the pattern of mutations, researchers can gain insights into the genetic changes that occur during tumorigenesis.  \nAberrant somatic mutations in DNA take many different forms, including single base substitutions (SBS), doublet base substitutions (DBS), small insertions and deletions (ID), and others. These genomic alterations are caused by a variety of mechanisms, including replication infidelity, spontaneous or enzymatic conversionsand exposure to exogenous or endogenous mutagens9. Each of these mechanisms results in a distinctive pattern of mutations, known as mutational signatures10. Mathematical extraction of mutational signatures in ","cbCaivKo5NLOpfGw","https://ap.wps.com/l/cbCaivKo5NLOpfGw","pdf",2032201,11,"English","# Introduction\n## Clinical importance of metastasis\n## Somatic mutations and mutational signatures\n# Methods and model performance\n## MetaWise framework and inputs\n## Comparison with machine learning and DiaDeL\n## Feature attribution with SHAP and LIME\n# Prior research context\n## Pan-cancer signature catalogs\n## Primary-to-metastatic signature transformation\n## Signature-based diagnosis using neural networks\n# Study contribution\n## Combining SBS, DBS, and ID signatures","[{\"question\":\"What is MetaWise and what data does it use?\",\"answer\":\"MetaWise is a deep neural network that uses mutational signatures as input features. These signatures are calculated from whole-exome sequencing (WES) data from TCGA and other metastatic cohorts.\"},{\"question\":\"How does MetaWise perform compared with other models?\",\"answer\":\"MetaWise accurately classifies metastatic tumors from primary tumors and outperforms traditional machine learning models and the deep learning model DiaDeL.\"},{\"question\":\"Which analyses and signatures explain the model’s predictions?\",\"answer\":\"SHAP and LIME analyses identify mutational signatures correlated with metastatic spread, including APOBEC mutagenesis, UV-induced signatures, and DNA damage response deficiency signatures. Non-coding mutation signatures also meaningfully influence performance.\"}]","Deep learning model accurately classifies metastatic tumors from primary tumors based on mutational signatures | PDF",1790227590,28]