IEEE transactions on computational biology and bioinformatics最新文献

{"title":"TMESurv: A Tumor Microenvironment-Informed Interpretable Neural Network for Survival Analysis Across Multiple Cancer Types.","authors":"Xiaobao Ding, Lin Zhang, Ming Fan, Lihua Li","doi":"10.1109/TCBBIO.2026.3718442","DOIUrl":"10.1109/TCBBIO.2026.3718442","url":null,"abstract":"<p><p>Cancer survival analysis is critical for prognosis; however, integrative analyses across multiple cancer types remain limited, particularly in capturing tumor microenvironment (TME) heterogeneity. To address this, we developed TMESurv, a biologically informed neural network that integrates hierarchical insights from the TME to predict survival outcomes across cancers. TMESurv employs a structured framework linking genes, cellular components, cells, cell roles, and clinical outcomes, with key connections defined using OpenAI API prompts and established biological knowledge. Predictive performance was assessed using the concordance index (C-index), the area under the time-dependent ROC curve (AUC), and the hazard ratio (HR). When benchmarked against five survival models across eight TCGA cohorts and validated on an independent SKCM immunotherapy cohort, TMESurv demonstrated superior performance, achieving a C-index of 0.68 for KIRC and an AUC of 0.62 for SKCM, outperforming its closest competitors by 2-3 points. It effectively stratified patients into high- and low-risk groups (HR > 1) and improved predictive accuracy by integrating immunostimulatory and immunosuppressive cell roles, boosting the C-index and AUC for SKCM by 6 points when incorporated into the loss function. Validation on the SKCM immunotherapy cohort further highlighted the distinct roles of these cells in differentiating responders from non-responders. Guided by TME principles, TMESurv offers robust survival predictions and interpretable insights, emphasizing the foundational role of the TME in tumor biology and showcasing the benefits of integrating biological constraints to enhance predictive accuracy and interpretability.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":"2103-2113"},"PeriodicalIF":0.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148623862","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Sparse Logistic Regression on Genomic Data for Prediction of Tumour Pathological Subtype.","authors":"Ozlem Kaymaz, Dodi Vionanda, Fatma Z Dogru, Youngjo Lee, Henry M Wood, Arief Gusnanto","doi":"10.1109/TCBBIO.2026.3732668","DOIUrl":"10.1109/TCBBIO.2026.3732668","url":null,"abstract":"<p><p>The correct prediction of tumour subtype is critical for the treatment of cancer patients to maximise the chance of survival. The patients' genomic information, such as copy number alterations (CNA) profile, has increasingly become an important factor in the prediction to supplement the traditional pathological subtyping. The incorporation of the CNA information in a prediction model, such as logistic regression, faces two major statistical challenges: first, how to estimate the model parameters in the thousands and, second, how to deal with the correlation of CNA between genomic regions. To address them, we propose a sparse logistic regression model with random effects where some of its parameters are estimated to zero while the other parameters are non-zero. In effect, a variable selection is embedded in the modelling. To deal with the correlation of CNA across genomic regions, we extend further the model to incorporate an additional penalty in the corresponding likelihood function in the logistic regression. The results show that we can identify selected genomic regions that are informative to distinguish different tumour subtypes, while giving a good prediction ability. We illustrate the methodology using CNA dataset from a lung cancer cohort.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":"2334-2345"},"PeriodicalIF":0.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148922356","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"EpiCRISPR: Improving CRISPR/Cas9 On- and Off-Target Prediction by Flanking Sequences, Epigenetic Marks, and High-Throughput Datasets.","authors":"Alon Luboshitz, Michal Rahimi, Ayal Hendel, Yaron Orenstein","doi":"10.1109/TCBBIO.2026.3727752","DOIUrl":"10.1109/TCBBIO.2026.3727752","url":null,"abstract":"<p><p>CRISPR/Cas9 technology has revolutionized gene-editing technologies. However, the endogenous gene-editing efficiency at the target site, unintended off-target sites (OTSs), and the DNA repair outcomes are affected by various factors, among them are the sequence of the site, its flanking sequences, and the epigenetic marks harboring it. Researchers have developed computational methods to predict endogenous efficiency, OTSs, and repair outcomes. However, previous studies have failed to successfully incorporate the factors that affect on- and off-target activity, and repair outcomes, and the methods were trained and evaluated on small datasets. A recent dataset of endogenous on-target efficiencies and repair outcomes of nearly 1,600 measurements in T cells, and a recently preprocessed dataset of 78 OTS experiments in T cells provide a unique opportunity to develop a method to predict endogenous on-target efficiency, OTSs, and repair outcomes. Here, we developed EpiCRISPROn and EpiCRISPROff to improve on-target efficiency and OTSs prediction, respectively, by combining various inputs: the site and its flanking sequences, multiple epigenetic marks, and a high-throughput-based prediction. In EpiCRISPROn and EpiCRISPROff, the additional non-sequence features improved prediction performance from an average Spearman correlation of 0.31 to 0.51 in 5-fold cross-validation and from an average AUPRC of 0.436 to 0.441 on a held-out test set, respectively. Moreover, EpiCRISPROn and EpiCRISPROff were trained on one cell type and successfully generalized to other cell types. In contrast, the inclusion of flanking sequences and epigenetic features did not improve repair-outcome prediction. Furthermore, by interrogating EpiCRISPROn and EpiCRISPROff trained models we shed light on on-target and OTS editing preferences. We expect EpiCRISPROn and EpiCRISPROff to advance the field of gene editing by providing improved cell-type-aware prediction of endogenous CRISPR/Cas9 on- and off-target activity.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":"2281-2294"},"PeriodicalIF":0.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148836141","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Multi-Objective Evolutionary Algorithm Integrating Topological and Gene Ontology Information for Overlapping Protein Complex Detection.","authors":"Mustafa N Abbas, David Broneske, Gunter Saake","doi":"10.1109/TCBBIO.2026.3710850","DOIUrl":"10.1109/TCBBIO.2026.3710850","url":null,"abstract":"<p><p>Accurate identification of protein complexes is essential for understanding the structural and functional organization of cells. Since many proteins participate in multiple complexes, protein complex detection should account for overlapping modular structures rather than assuming disjoint communities. However, many existing computational methods still rely mainly on the topological properties of protein-protein interaction (PPI) networks, which may limit their biological interpretability and robustness in noisy networks. To address this challenge, we propose MOEGO-OPCD, a multi-objective evolutionary framework for overlapping protein complex detection that jointly exploits network topology and Gene Ontology (GO)-based functional information. The proposed framework first represents candidate solutions using an edge-based encoding scheme, then decodes them into overlapping node complexes, and subsequently optimizes them within a decomposition-based multi-objective evolutionary process. Three complementary objectives are considered to promote intra-complex topological compactness, penalize inter-complex topological dispersion, and enhance GO-based functional coherence. To improve the search process, the framework incorporates an objective-guided uniform crossover operator and a self-adaptive heuristic mutation operator that uses both structural and functional cues to balance exploration and exploitation. Extensive experiments on three benchmark yeast PPI networks, Yeast-D1, Yeast-D2, and Collins, evaluated against the MIPS and CYC2008 reference complexes, show that the proposed method achieves consistently strong and stable performance and outperforms a range of existing baselines in terms of Recall, Precision, and F-score. Additional sensitivity and runtime analyses further demonstrate that the framework is robust across a broad range of parameter settings and scales reasonably across the tested datasets. These results indicate that biologically informed multi-objective evolutionary optimization provides an effective and interpretable strategy for detecting overlapping protein complexes in PPI networks.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":"1967-1980"},"PeriodicalIF":0.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148407896","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"BAMPDA: A Matrix Refactoring Framework With Heterogeneous Inference for Potential PTM-Disease Association Identification.","authors":"Xinting Zhang, Jie Ni, Zihao Song, Shan Lu, Shenqi Jing, Xin Zhang, Zhuoying Xie, Jie Sun, Yun Liu","doi":"10.1109/TCBBIO.2026.3723432","DOIUrl":"10.1109/TCBBIO.2026.3723432","url":null,"abstract":"<p><p>Post-translational modifications (PTMs) are modifications of proteins that occur after translation. They exert their influence on human health by altering the properties of proteins. Recent progress in biomedical research has yielded extensive heterogeneous datasets detailing protein PTMs and their disease relevance. These datasets provide a substantial foundation for the development of methods predicting PTM-disease associations. Thus we develop BAMPDA, a matrix refactoring framework with heterogeneous inference designed to enhance the inference of PTM-disease associations. Our approach integrates multi-source data encompassing both protein and disease characteristics, leveraging two successively applied matrix-based algorithms for association prediction. Furthermore, rigorous 5-fold cross-validation demonstrates that BAMPDA significantly outperforms five recent baseline methods. Finally, we applied BAMPDA to four representative diseases to predict associated proteins and validate the role of PTMs in mediating these disease relationships. The results underscore BAMPDA's robust predictive capability in practical scenarios.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":"2167-2180"},"PeriodicalIF":0.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148724173","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Enhancing Cancer Driver Genes Prediction via Multi-view Hypergraph and Dynamic Reward-Penalty Model.","authors":"Xiaoyan Kui, Zhipeng Hu, Qinsong Li, Shen Jiang, Canwei Liu, Ziwei Zou, Chengtao Liu, Beiji Zou","doi":"10.1109/TCBBIO.2026.3713986","DOIUrl":"10.1109/TCBBIO.2026.3713986","url":null,"abstract":"<p><p>Cancer driver genes can grant cancer cells the ability to proliferate continuously and resist apoptosis, thereby promoting tumor evolution toward malignancy. Identifying cancer driver genes not only helps to reveal the molecular pathological network of cancer, but also provides important targets for the development of precise targeted drugs. Although the existing methods can partially compensate for the limitations of single-omics data to a certain extent, they are still inadequate in mining deep relationships among diverse omics features and gene synergy, thus impacting the accuracy of driver gene identification. In this paper, a cancer driver gene identification method integrating the multi-view hypergraph with the dynamic reward-penalty model is proposed and termed MHDP. Firstly, the functional module and spatial co-localization hypergraphs are constructed independently in functional and spatial dimensions, respectively, to capture the synergistic information among genes, as well as the functional and spatial information of genes. Secondly, the Bhattacharyya distance is employed to evaluate gene expression differences between normal and tumor samples. Then, miRNA importance scores are calculated based on the number of miRNA connections for each gene within the mRNA-miRNA bipartite graph. Finally, through analysis of correlations among different features, we propose a Dynamic Reward-Penalty model that adaptively adjusts weights of omics features to achieve efficient multi-omics integration. Experimental evaluation using breast, lung adenocarcinoma, and prostate cancer datasets demonstrates that MHDP markedly outperforms 8 state-of-the-art methods, including HWC, DriverRWH, and WMDS.netP, exhibiting significant improvements in accuracy, functional consistency, and partial area under the ROC curve (PAUC).</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":"1995-2005"},"PeriodicalIF":0.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148475789","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Prediction of GO Terms Based on Partitioning PPI Networks Into Highly Connected Components.","authors":"Milana Grbic, Branislava Gemovic, Savka Vracevic, Radoslav Davidovic, Dragan Matic, Aleksandar Kartelj","doi":"10.1109/TCBBIO.2026.3706474","DOIUrl":"10.1109/TCBBIO.2026.3706474","url":null,"abstract":"<p><p>This paper proposes a novel two-phase method for Gene Ontology (GO) term prediction, which integrates a Variable Neighborhood Search (VNS)-based strategy for partitioning biological networks with a dedicated procedure for functional annotation. In the first phase, we introduce a VNS metaheuristic for detecting highly connected components in protein-protein interaction (PPI) network, enabling the grouping of biologically correlated elements into cohesive subnetworks. In the second phase, we utilize the obtained network partitions for functional enrichment analysis, employing the DINGO annotation tool to predict GO terms. Experimental results on real-world and artificial datasets confirm the effectiveness of the proposed approach in producing high-quality network partitions and its potential to support accurate GO term annotation. Although demonstrated here on biological networks, the proposed VNS framework is general and applicable to any network where the identification of strongly connected modules is relevant.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":"1916-1927"},"PeriodicalIF":0.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148311340","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exploring Complex Genetic Mechanisms in Brain Imaging Genetics via a New Multi-Task Learning Method.","authors":"Yiming Wang, Cui-Na Jiao, Ying-Lian Gao, Xin-Chun Cui, Chun-Hou Zheng, Jin-Xing Liu","doi":"10.1109/TCBBIO.2026.3709159","DOIUrl":"10.1109/TCBBIO.2026.3709159","url":null,"abstract":"<p><p>Brain imaging genetics generally combines genotype data with brain structure and functional measures to investigate the genetic basis of neurological disorders. Multimodal brain imaging data carry different but complementary information, which can clearly measure the same brain. Multi-task sparse canonical correlation analysis (MTSCCA) is often used to analyze the bi-multivariate correlation between genotype data and multimodal imaging features; however, it lacks the capacity to identify class-specific biomarkers and to deeply capture complex genetic mechanisms. In this paper, a linear discrimination and decomposition method based on MTSCCA (LDMTSCCA) is proposed to solve above problems. Specifically, LDMTSCCA first extracts disease-related genetic information using sparse linear discriminant analysis, and then employs parameter decomposition to learn multi-level expression patterns of genetic loci. To better explore intricate genetic mechanisms, LDMTSCCA jointly incorporates disease states, parameter decomposition, and network connectivity constraints into the association analysis between multilocus genetic information and multimodal neuroimaging phenotypes. The proposed method is compared with the traditional single task, multi-task, multi-view, and deep learning-based canonical correlation analysis methods in Alzheimer's Disease Neuroimaging Initiative, which shows that LDMTSCC not only achieves the highest canonical correlation coefficients but also identifies multi-level biomarkers.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":"1893-1903"},"PeriodicalIF":0.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148371819","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MAGNN-DTA: Cross-Modal Mutual Attention Graph Neural Networks for Drug-Target Affinity Prediction.","authors":"Yun Zhou, Yibo Zhang, Dong Liu, Wei Wang","doi":"10.1109/TCBBIO.2026.3732146","DOIUrl":"10.1109/TCBBIO.2026.3732146","url":null,"abstract":"<p><p>Drug-Target Affinity (DTA) prediction is a crucial step in developing new pharmaceuticals. However, existing approaches remain limited in capturing cross-modal interactions, often failing to fully characterize complementary information of drugs and proteins at the sequence, structural, and topological levels. To address these limitations, we propose MAGNN-DTA, a cross-modal mutual attention graph neural network framework for DTA prediction. First, MAGNN-DTA adopts a dual-path feature extraction module to extract sequence features and graph features of drug and protein. For drug representation, a Conformer block is used to capture long-range dependencies and local functional patterns, while a Graph Isomorphism Network (GIN) and multi-head attention are used to extract local and global features from molecular graphs. The drug features are further refined via a Kolmogorov-Arnold Network (KAN) to improve representational capacity and interpretability. For protein representation, sequence and structural information are integrated through a multi-scale deep convolutional network and a gated graph attention network (Gated-GAT) derived from residue distance maps. Subsequently, sequence feature and graph feature are concatenated by a gated pooling mechanism to form the representations of the drug and the protein. Finally, the model employs a cross-modal mutual attention module to capture bidirectional interactions between drug and protein modalities, enabling precise identification of key functional groups in drugs and critical binding residues in proteins. Extensive experiments on the Davis and KIBA benchmark datasets demonstrate that MAGNN-DTA achieves highly competitive performance compared with recent advanced methods across standard evaluation metrics.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":"2360-2371"},"PeriodicalIF":0.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148915060","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"GraphGDel: Constructing and Learning Graph Representations of Genome-Scale Metabolic Models for Growth-Coupled Gene Deletion Prediction.","authors":"Ziwei Yang, Takeyuki Tamura","doi":"10.1109/TCBBIO.2026.3714738","DOIUrl":"10.1109/TCBBIO.2026.3714738","url":null,"abstract":"<p><p>In genome-scale constraint-based metabolic models, gene deletion strategies are essential for achieving growth-coupled production, where cell growth and target metabolite synthesis occur simultaneously. Despite the inherently networked nature of genome-scale metabolic models, existing computational approaches rely primarily on sequential data and lack graph representations that capture their complex relationships, as both well-defined graph constructions and learning frameworks capable of exploiting them remain largely unexplored. To address this gap, we present a twofold solution. First, we introduce a systematic pipeline for constructing graph representations from constraint-based metabolic models. Second, we develop a deep learning framework that integrates these graph representations with gene and metabolite sequence data to predict growth-coupled gene deletion strategies. Across three metabolic models, our approach consistently outperforms established baselines, with improvements in overall accuracy of 14.04%, 16.26%, and 13.18% over a deep feedforward neural network baseline, 6.17%, 4.96%, and 5.31% over a sequence-learning baseline, and 5.10%, 4.36%, and 4.70% over a topology-aware graph aggregation baseline on the same metabolite graph, respectively.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":"2052-2065"},"PeriodicalIF":0.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148475782","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}