Pub Date : 2025-11-26DOI: 10.1186/s13059-025-03883-1
Ming-Ching C Wen, Joshua D Welch
Transposable elements (TEs) are often epigenetically repressed in eukaryotic cells, but still affect the molecular state of the cell in certain contexts. A flurry of recent studies have elucidated new effects of TE sequences in eukaryotic cells. We review these emerging molecular effects of TEs, including a variety of new mechanisms by which TE sequences affect the cell, including pre- and post-transcriptional regulation of gene expression; cell-to-cell transmission of genes within a multicellular organism through virus-like activity; and RNA-guided DNA insertion. Recent demonstration of TE-guided genome editing underscores the importance of these investigations for both basic and translational research. Future work is needed to continue to unravel the molecular effects of TE sequences across developmental stages, across cell types, and in various diseases.
{"title":"Molecular effects of transposable element sequences in mammalian cells.","authors":"Ming-Ching C Wen, Joshua D Welch","doi":"10.1186/s13059-025-03883-1","DOIUrl":"https://doi.org/10.1186/s13059-025-03883-1","url":null,"abstract":"<p><p>Transposable elements (TEs) are often epigenetically repressed in eukaryotic cells, but still affect the molecular state of the cell in certain contexts. A flurry of recent studies have elucidated new effects of TE sequences in eukaryotic cells. We review these emerging molecular effects of TEs, including a variety of new mechanisms by which TE sequences affect the cell, including pre- and post-transcriptional regulation of gene expression; cell-to-cell transmission of genes within a multicellular organism through virus-like activity; and RNA-guided DNA insertion. Recent demonstration of TE-guided genome editing underscores the importance of these investigations for both basic and translational research. Future work is needed to continue to unravel the molecular effects of TE sequences across developmental stages, across cell types, and in various diseases.</p>","PeriodicalId":12611,"journal":{"name":"Genome Biology","volume":"26 1","pages":"403"},"PeriodicalIF":10.1,"publicationDate":"2025-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12649098/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145632145","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-26DOI: 10.1186/s13059-025-03812-2
William Cross, Iben Lyskjær, Christopher Davies, Abigail Bunkum, Ana Maia Rocha, Tom Lesluyes, Fernanda Amary, Roberto Tirabosco, Cristina Naceur-Lombardelli, Mariam Jamal-Hanjani, Charles Swanton, Nischalan Pillay, Simone Zaccaria, Adrienne M Flanagan, Peter Van Loo
Driver mutations in IDH1 and IDH2 are initiating events in the evolution of chondrosarcoma and several other cancer types. Here, we present evidence that mutant IDH1 is recurrently lost in metastatic central chondrosarcoma. This may reflect either relaxed positive selection for the mutant IDH1 locus, or negative selection for the hypermethylation phenotype later in tumor evolution. This finding highlights the challenge for therapeutic intervention by mutant IDH1 inhibitors in chondrosarcoma.
{"title":"Loss of IDH1 and IDH2 mutations during the evolution of metastatic chondrosarcoma.","authors":"William Cross, Iben Lyskjær, Christopher Davies, Abigail Bunkum, Ana Maia Rocha, Tom Lesluyes, Fernanda Amary, Roberto Tirabosco, Cristina Naceur-Lombardelli, Mariam Jamal-Hanjani, Charles Swanton, Nischalan Pillay, Simone Zaccaria, Adrienne M Flanagan, Peter Van Loo","doi":"10.1186/s13059-025-03812-2","DOIUrl":"https://doi.org/10.1186/s13059-025-03812-2","url":null,"abstract":"<p><p>Driver mutations in IDH1 and IDH2 are initiating events in the evolution of chondrosarcoma and several other cancer types. Here, we present evidence that mutant IDH1 is recurrently lost in metastatic central chondrosarcoma. This may reflect either relaxed positive selection for the mutant IDH1 locus, or negative selection for the hypermethylation phenotype later in tumor evolution. This finding highlights the challenge for therapeutic intervention by mutant IDH1 inhibitors in chondrosarcoma.</p>","PeriodicalId":12611,"journal":{"name":"Genome Biology","volume":"26 1","pages":"404"},"PeriodicalIF":10.1,"publicationDate":"2025-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12659308/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145632119","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-24DOI: 10.1186/s13059-025-03854-6
Uberto Pozzoli, Diego Forni, Federica Arrigoni, Rachele Cagliani, Luca De Gioia, Manuela Sironi
Intrinsically disordered protein regions (IDRs) are implicated in diverse cellular processes in eukaryotes and, in these organisms, they cover up to 40% of the proteome. Surprisingly little is known about IDRs in bacterial proteomes. Specifically, a number of questions remain unanswered, such as the role of these regions in host–pathogen interactions, their adaptive potential and evolutionary trajectories, as well as their biophysical properties. Here we focus on Mycobacterium tuberculosis and take advantage of the fact that, due to its extreme epidemiological relevance, several large-scale analyses are available. After benchmarking different disorder prediction tools, we integrate multiple levels of biological information to show that IDR-containing proteins are involved in virulence, in the modulation of host immune response, and in lipid metabolism. Mycobacterium tuberculosis IDRs are fast evolving and poorly antigenic, and they display specific sequence-ensemble-function relationships. Conversely, human proteins that interact with Mycobacterium tuberculosis are evolutionary constrained, widely expressed, and highly connected in the human interactome map. This indicates that the classical arms race paradigm is not universal in host–pathogen interactions. We also extend analysis to 540 human-infecting bacteria and we underscore wide variations in IDR representation and conformational properties. Our data point to a role of IDRs in contributing to bacterial virulence, interaction with the human host, and control of immune responses. Although this awaits experimental validation, we suggest that Mycobacterium tuberculosis also uses IDRs to sense and interact with its environment. Herein, we provide a database of bacterial IDRs, together with relevant parameters, for public use.
{"title":"Mycobacterium tuberculosis uses intrinsically disordered, fast evolving proteins to interact with conserved host factors","authors":"Uberto Pozzoli, Diego Forni, Federica Arrigoni, Rachele Cagliani, Luca De Gioia, Manuela Sironi","doi":"10.1186/s13059-025-03854-6","DOIUrl":"https://doi.org/10.1186/s13059-025-03854-6","url":null,"abstract":"Intrinsically disordered protein regions (IDRs) are implicated in diverse cellular processes in eukaryotes and, in these organisms, they cover up to 40% of the proteome. Surprisingly little is known about IDRs in bacterial proteomes. Specifically, a number of questions remain unanswered, such as the role of these regions in host–pathogen interactions, their adaptive potential and evolutionary trajectories, as well as their biophysical properties. Here we focus on Mycobacterium tuberculosis and take advantage of the fact that, due to its extreme epidemiological relevance, several large-scale analyses are available. After benchmarking different disorder prediction tools, we integrate multiple levels of biological information to show that IDR-containing proteins are involved in virulence, in the modulation of host immune response, and in lipid metabolism. Mycobacterium tuberculosis IDRs are fast evolving and poorly antigenic, and they display specific sequence-ensemble-function relationships. Conversely, human proteins that interact with Mycobacterium tuberculosis are evolutionary constrained, widely expressed, and highly connected in the human interactome map. This indicates that the classical arms race paradigm is not universal in host–pathogen interactions. We also extend analysis to 540 human-infecting bacteria and we underscore wide variations in IDR representation and conformational properties. Our data point to a role of IDRs in contributing to bacterial virulence, interaction with the human host, and control of immune responses. Although this awaits experimental validation, we suggest that Mycobacterium tuberculosis also uses IDRs to sense and interact with its environment. Herein, we provide a database of bacterial IDRs, together with relevant parameters, for public use.","PeriodicalId":12611,"journal":{"name":"Genome Biology","volume":"112 1","pages":"387"},"PeriodicalIF":12.3,"publicationDate":"2025-11-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145583715","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
HGMT is a database designed to analyze, explore, and visualize gut microbiomes from diverse tumor types. We process metagenomic datasets from 18,630 stool samples across 37 tumor types, including 2,207 samples from immunotherapy-treated patients across 12 tumor types. HGMT provides an interactive portal for querying taxonomic and functional profiles, visualizing cross-dataset differential abundance taxa in tumors, and identifying their pan-tumor associations. Our analysis reveals the capability of gut microbiota in diagnosing gastrointestinal tumors and predicting immunotherapy response for non-small cell lung carcinoma. HGMT represents a valuable resource for investigating the roles of gut microbiota in tumors and immunotherapy response.
{"title":"HGMT: a database of human gut microbiota for tumors and immunotherapy response","authors":"Jinxin Liu, Mingyu Wang, Chentao Xu, Longhao Jia, Senying Lai, Zi-Chao Zhang, Jinglong Zhang, Wei-Hua Chen, Yucheng T. Yang, Xing-Ming Zhao","doi":"10.1186/s13059-025-03865-3","DOIUrl":"https://doi.org/10.1186/s13059-025-03865-3","url":null,"abstract":"HGMT is a database designed to analyze, explore, and visualize gut microbiomes from diverse tumor types. We process metagenomic datasets from 18,630 stool samples across 37 tumor types, including 2,207 samples from immunotherapy-treated patients across 12 tumor types. HGMT provides an interactive portal for querying taxonomic and functional profiles, visualizing cross-dataset differential abundance taxa in tumors, and identifying their pan-tumor associations. Our analysis reveals the capability of gut microbiota in diagnosing gastrointestinal tumors and predicting immunotherapy response for non-small cell lung carcinoma. HGMT represents a valuable resource for investigating the roles of gut microbiota in tumors and immunotherapy response.","PeriodicalId":12611,"journal":{"name":"Genome Biology","volume":"223 1","pages":""},"PeriodicalIF":12.3,"publicationDate":"2025-11-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145583713","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-24DOI: 10.1186/s13059-025-03862-6
Yang Li, Guanyu Qiao, Hongli Du, Xin Gao, Guohua Wang
Single-cell transcriptomics enables precise characterization of cellular heterogeneity, but current pre-trained models relying solely on expression data fail to capture gene associations. We present scKGBERT, a knowledge-enhanced foundation model integrating 41 M single-cell RNA-seq profiles and 8.9 M protein–protein interactions to jointly learn gene and cell representations. scKGBERT employs Gaussian attention to emphasize key genes and improve biomarker identification, achieving superior performance across gene annotation, drug response, and disease prediction tasks. scKGBERT enhances biological interpretability and offers a powerful resource for precision medicine and disease mechanism discovery.
{"title":"scKGBERT: a knowledge-enhanced foundation model for single-cell transcriptomics","authors":"Yang Li, Guanyu Qiao, Hongli Du, Xin Gao, Guohua Wang","doi":"10.1186/s13059-025-03862-6","DOIUrl":"https://doi.org/10.1186/s13059-025-03862-6","url":null,"abstract":"Single-cell transcriptomics enables precise characterization of cellular heterogeneity, but current pre-trained models relying solely on expression data fail to capture gene associations. We present scKGBERT, a knowledge-enhanced foundation model integrating 41 M single-cell RNA-seq profiles and 8.9 M protein–protein interactions to jointly learn gene and cell representations. scKGBERT employs Gaussian attention to emphasize key genes and improve biomarker identification, achieving superior performance across gene annotation, drug response, and disease prediction tasks. scKGBERT enhances biological interpretability and offers a powerful resource for precision medicine and disease mechanism discovery.","PeriodicalId":12611,"journal":{"name":"Genome Biology","volume":"100 1","pages":""},"PeriodicalIF":12.3,"publicationDate":"2025-11-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145583712","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The rise of genomic sequencing raises privacy concerns due to the identifiable nature of genomic data. The GA4GH Beacon Project enables privacy-preserving data sharing but is vulnerable to membership inference attacks that reveal individual participation. Existing defenses, such as noise addition and query restrictions, rely on static policies that attackers can bypass. We introduce the first reinforcement learning (RL)-based dynamic defense for the beacon protocol, training defender and attacker agents in a multiplayer setting. Our approach adapts responses in real time, distinguishing users from adversaries and balancing privacy with utility against evolving threats.
{"title":"A reinforcement learning-based approach for dynamic privacy protection in genomic data sharing beacons","authors":"Masoud Poorghaffar Aghdam, Sobhan Shukueian Tabrizi, Kerem Ayöz, Erman Ayday, Sinem Sav, A. Ercüment Çiçek","doi":"10.1186/s13059-025-03871-5","DOIUrl":"https://doi.org/10.1186/s13059-025-03871-5","url":null,"abstract":"The rise of genomic sequencing raises privacy concerns due to the identifiable nature of genomic data. The GA4GH Beacon Project enables privacy-preserving data sharing but is vulnerable to membership inference attacks that reveal individual participation. Existing defenses, such as noise addition and query restrictions, rely on static policies that attackers can bypass. We introduce the first reinforcement learning (RL)-based dynamic defense for the beacon protocol, training defender and attacker agents in a multiplayer setting. Our approach adapts responses in real time, distinguishing users from adversaries and balancing privacy with utility against evolving threats.","PeriodicalId":12611,"journal":{"name":"Genome Biology","volume":"143 1","pages":""},"PeriodicalIF":12.3,"publicationDate":"2025-11-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145583714","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-21DOI: 10.1186/s13059-025-03873-3
Karina Ray, Christina Mulch, Samuel M. Peterson, Sebastian Benjamin, Nathan Gullicksrud, Adam J. Ericsen, Eric J. Vallender, Betsy M. Ferguson, Jeffrey D. Wall, Benjamin N. Bimber
Due to their close evolutionary relationship with humans, rhesus macaques are an important pre-clinical model. While genetic diversity driven by short nucleotide variation has long been studied in rhesus macaques, there is comparatively little known about structural variation, with most published studies focused on cross-species comparative analyses. Understanding the degree and implications of intraspecies structural variation is essential to all biomedical research using rhesus macaques as a model. Here we present long-read sequencing of 59 rhesus macaques, identifying a catalog of 339,334 structural variants (SVs), which we subsequently genotype in a cohort of 2,645 individuals with short read whole genome sequencing data to create the largest public dataset of rhesus macaque SVs. These data reveal population structure within rhesus macaque SVs based on both geographic ancestry and to a lesser degree, breeding center. While there is evidence of strong purifying selection against SVs within exons, 0.7% of SVs overlap exons, with an average of 16.9 rare SVs per subject predicted to have a high impact on protein coding sequences. Notably, rhesus macaque SVs are dominated by Alu retrotransposition events, which comprise 55.7% of SVs and suggest significantly different modes of SV formation relative to humans and great apes. This dataset represents the largest study of structural variation in rhesus macaques to date and demonstrates use of both long and short-read datasets to generate SV genotype data. These data enable the consideration of structural variation impact in rhesus macaque-based research and will also aid the development of primate pangenomes.
{"title":"Long-read structural variant discovery and targeted short read genotyping enables population scale characterization of structural variation in rhesus macaques","authors":"Karina Ray, Christina Mulch, Samuel M. Peterson, Sebastian Benjamin, Nathan Gullicksrud, Adam J. Ericsen, Eric J. Vallender, Betsy M. Ferguson, Jeffrey D. Wall, Benjamin N. Bimber","doi":"10.1186/s13059-025-03873-3","DOIUrl":"https://doi.org/10.1186/s13059-025-03873-3","url":null,"abstract":"Due to their close evolutionary relationship with humans, rhesus macaques are an important pre-clinical model. While genetic diversity driven by short nucleotide variation has long been studied in rhesus macaques, there is comparatively little known about structural variation, with most published studies focused on cross-species comparative analyses. Understanding the degree and implications of intraspecies structural variation is essential to all biomedical research using rhesus macaques as a model. Here we present long-read sequencing of 59 rhesus macaques, identifying a catalog of 339,334 structural variants (SVs), which we subsequently genotype in a cohort of 2,645 individuals with short read whole genome sequencing data to create the largest public dataset of rhesus macaque SVs. These data reveal population structure within rhesus macaque SVs based on both geographic ancestry and to a lesser degree, breeding center. While there is evidence of strong purifying selection against SVs within exons, 0.7% of SVs overlap exons, with an average of 16.9 rare SVs per subject predicted to have a high impact on protein coding sequences. Notably, rhesus macaque SVs are dominated by Alu retrotransposition events, which comprise 55.7% of SVs and suggest significantly different modes of SV formation relative to humans and great apes. This dataset represents the largest study of structural variation in rhesus macaques to date and demonstrates use of both long and short-read datasets to generate SV genotype data. These data enable the consideration of structural variation impact in rhesus macaque-based research and will also aid the development of primate pangenomes.","PeriodicalId":12611,"journal":{"name":"Genome Biology","volume":"11 1","pages":""},"PeriodicalIF":12.3,"publicationDate":"2025-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145559410","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Advancements in single-cell RNA sequencing have enabled the analysis of millions of cells, but integrating such data across samples and methods while mitigating batch effects remains challenging. Deep learning approaches address this by learning biologically conserved gene expression representations, yet systematic benchmarking of loss functions and integration performance is lacking. We evaluate 16 integration methods using a unified variational autoencoder framework, incorporating batch and cell-type information. Results reveal limitations in the single-cell integration benchmarking index (scIB) for preserving intra-cell-type information. To address this, we introduce a correlation-based loss function and enhance benchmarking metrics to better capture biological conservation. Using cell annotations from lung and breast atlases, our approach improves biological signal preservation. We propose a refined integration framework, scIB-E, and metrics that provide deeper insights into the integration process and offer guidance for advanced developments in integrating increasingly complex single-cell data. This benchmark highlights the potential of deep learning-based approaches for single-cell data integration, emphasizing the importance of biologically informed metrics and improved benchmarking strategies.
{"title":"Benchmarking deep learning methods for biologically conserved single-cell integration","authors":"Chenxin Yi, Jinyu Cheng, Jiajun Chen, Wanquan Liu, Junwei Liu, Yixue Li","doi":"10.1186/s13059-025-03869-z","DOIUrl":"https://doi.org/10.1186/s13059-025-03869-z","url":null,"abstract":"Advancements in single-cell RNA sequencing have enabled the analysis of millions of cells, but integrating such data across samples and methods while mitigating batch effects remains challenging. Deep learning approaches address this by learning biologically conserved gene expression representations, yet systematic benchmarking of loss functions and integration performance is lacking. We evaluate 16 integration methods using a unified variational autoencoder framework, incorporating batch and cell-type information. Results reveal limitations in the single-cell integration benchmarking index (scIB) for preserving intra-cell-type information. To address this, we introduce a correlation-based loss function and enhance benchmarking metrics to better capture biological conservation. Using cell annotations from lung and breast atlases, our approach improves biological signal preservation. We propose a refined integration framework, scIB-E, and metrics that provide deeper insights into the integration process and offer guidance for advanced developments in integrating increasingly complex single-cell data. This benchmark highlights the potential of deep learning-based approaches for single-cell data integration, emphasizing the importance of biologically informed metrics and improved benchmarking strategies.","PeriodicalId":12611,"journal":{"name":"Genome Biology","volume":"177 1","pages":""},"PeriodicalIF":12.3,"publicationDate":"2025-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145554403","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-20DOI: 10.1186/s13059-025-03860-8
Weiming Yu, Zhuobin Chen, Yaohua Hu, Jing Qin, Le Ou-Yang
Inferring gene regulatory networks (GRNs) is essential for understanding biological regulation. Although numerous deep learning approaches have been developed for GRN inference, most require large amounts of labeled data. We present Meta-TGLink, a structure-enhanced graph meta-learning model for few-shot GRN inference. By formulating GRN inference as a link prediction task, Meta-TGLink captures transferable regulatory patterns while reducing dependence on extensive labeled datasets. The model combines graph neural networks with Transformer architectures to integrate relational and positional information, thereby improving predictive performance under data-scarce conditions. Experiments on real datasets demonstrate its superiority over state-of-the-art baselines, particularly in cross-domain few-shot scenarios.
{"title":"Structure-enhanced graph meta learning for few-shot gene regulatory network inference","authors":"Weiming Yu, Zhuobin Chen, Yaohua Hu, Jing Qin, Le Ou-Yang","doi":"10.1186/s13059-025-03860-8","DOIUrl":"https://doi.org/10.1186/s13059-025-03860-8","url":null,"abstract":"Inferring gene regulatory networks (GRNs) is essential for understanding biological regulation. Although numerous deep learning approaches have been developed for GRN inference, most require large amounts of labeled data. We present Meta-TGLink, a structure-enhanced graph meta-learning model for few-shot GRN inference. By formulating GRN inference as a link prediction task, Meta-TGLink captures transferable regulatory patterns while reducing dependence on extensive labeled datasets. The model combines graph neural networks with Transformer architectures to integrate relational and positional information, thereby improving predictive performance under data-scarce conditions. Experiments on real datasets demonstrate its superiority over state-of-the-art baselines, particularly in cross-domain few-shot scenarios.","PeriodicalId":12611,"journal":{"name":"Genome Biology","volume":"135 1","pages":""},"PeriodicalIF":12.3,"publicationDate":"2025-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145554407","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号