使用深度学习和图傅里叶变换识别复杂和3D组织的空间转录组学中的空间域

IF 10.1 1区生物学 Q1 BIOTECHNOLOGY & APPLIED MICROBIOLOGY

Genome Biology Pub Date : 2025-06-03 DOI:10.1186/s13059-025-03631-5

Shuli Sun, Jixin Liu, Guojun Li, Bingqiang Liu

{"title":"使用深度学习和图傅里叶变换识别复杂和3D组织的空间转录组学中的空间域","authors":"Shuli Sun, Jixin Liu, Guojun Li, Bingqiang Liu","doi":"10.1186/s13059-025-03631-5","DOIUrl":null,"url":null,"abstract":"The rapid advancements in spatially resolved transcriptomics (SRT) enable the characterization of gene expressions while preserving spatial information. However, high dropout rates and noise hinder accurate spatial domain identification for understanding tissue architecture. We present DeepGFT, a method that simultaneously models spot-wise and gene-wise relationships by integrating deep learning with graph Fourier transform for spatial domain identification. Benchmarking results demonstrate the superiority of DeepGFT over existing methods. DeepGFT detects tumor substructures with immune-related differences in human breast cancer, identifies the complex germinal centers accurately in human lymph node, and accurately reveals the developmental changes in 3D Drosophila data.","PeriodicalId":12611,"journal":{"name":"Genome Biology","volume":"43 1","pages":""},"PeriodicalIF":10.1000,"publicationDate":"2025-06-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"DeepGFT: identifying spatial domains in spatial transcriptomics of complex and 3D tissue using deep learning and graph Fourier transform\",\"authors\":\"Shuli Sun, Jixin Liu, Guojun Li, Bingqiang Liu\",\"doi\":\"10.1186/s13059-025-03631-5\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"The rapid advancements in spatially resolved transcriptomics (SRT) enable the characterization of gene expressions while preserving spatial information. However, high dropout rates and noise hinder accurate spatial domain identification for understanding tissue architecture. We present DeepGFT, a method that simultaneously models spot-wise and gene-wise relationships by integrating deep learning with graph Fourier transform for spatial domain identification. Benchmarking results demonstrate the superiority of DeepGFT over existing methods. DeepGFT detects tumor substructures with immune-related differences in human breast cancer, identifies the complex germinal centers accurately in human lymph node, and accurately reveals the developmental changes in 3D Drosophila data.\",\"PeriodicalId\":12611,\"journal\":{\"name\":\"Genome Biology\",\"volume\":\"43 1\",\"pages\":\"\"},\"PeriodicalIF\":10.1000,\"publicationDate\":\"2025-06-03\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Genome Biology\",\"FirstCategoryId\":\"99\",\"ListUrlMain\":\"https://doi.org/10.1186/s13059-025-03631-5\",\"RegionNum\":1,\"RegionCategory\":\"生物学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"BIOTECHNOLOGY & APPLIED MICROBIOLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Genome Biology","FirstCategoryId":"99","ListUrlMain":"https://doi.org/10.1186/s13059-025-03631-5","RegionNum":1,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"BIOTECHNOLOGY & APPLIED MICROBIOLOGY","Score":null,"Total":0}

引用次数: 0

摘要

空间解析转录组学（SRT）的快速发展使基因表达的表征，同时保留空间信息。然而，高辍学率和噪声阻碍了准确的空间域识别以理解组织结构。我们提出了DeepGFT，这是一种通过将深度学习与用于空间域识别的图傅立叶变换相结合，同时对点和基因关系进行建模的方法。基准测试结果证明了DeepGFT优于现有方法。DeepGFT检测人类乳腺癌中具有免疫相关差异的肿瘤亚结构，准确识别人类淋巴结中复杂的生发中心，准确揭示三维果蝇数据中的发育变化。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

查看原文本刊更多论文

DeepGFT: identifying spatial domains in spatial transcriptomics of complex and 3D tissue using deep learning and graph Fourier transform

The rapid advancements in spatially resolved transcriptomics (SRT) enable the characterization of gene expressions while preserving spatial information. However, high dropout rates and noise hinder accurate spatial domain identification for understanding tissue architecture. We present DeepGFT, a method that simultaneously models spot-wise and gene-wise relationships by integrating deep learning with graph Fourier transform for spatial domain identification. Benchmarking results demonstrate the superiority of DeepGFT over existing methods. DeepGFT detects tumor substructures with immune-related differences in human breast cancer, identifies the complex germinal centers accurately in human lymph node, and accurately reveals the developmental changes in 3D Drosophila data.

求助全文

通过发布文献求助，成功后即可免费获取论文全文。去求助

来源期刊

Genome Biology Biochemistry, Genetics and Molecular Biology-Genetics

CiteScore

21.00

自引率

3.30%

发文量

241

审稿时长

2 months

期刊介绍： Genome Biology stands as a premier platform for exceptional research across all domains of biology and biomedicine, explored through a genomic and post-genomic lens. With an impressive impact factor of 12.3 (2022),* the journal secures its position as the 3rd-ranked research journal in the Genetics and Heredity category and the 2nd-ranked research journal in the Biotechnology and Applied Microbiology category by Thomson Reuters. Notably, Genome Biology holds the distinction of being the highest-ranked open-access journal in this category. Our dedicated team of highly trained in-house Editors collaborates closely with our esteemed Editorial Board of international experts, ensuring the journal remains on the forefront of scientific advances and community standards. Regular engagement with researchers at conferences and institute visits underscores our commitment to staying abreast of the latest developments in the field.