IEEE transactions on computational biology and bioinformatics最新文献

{"title":"Protein FT-Transformer: Species Classification via Amino Acid-Encoded DNA Sequences.","authors":"Meetesh Nevendra, Deepak Singh","doi":"10.1109/TCBBIO.2026.3737459","DOIUrl":"10.1109/TCBBIO.2026.3737459","url":null,"abstract":"<p><p>DNA barcoding is widely used for species identification, yet many computational classifiers continue to operate directly on raw nucleotide sequences. For large and heterogeneous barcode repositories, this representation can introduce avoidable computational cost because long nucleotide strings often contain redundant codon-level information and highly uneven class distributions. This study investigates whether a biologically grounded representation can improve supervised barcode classification while reducing the effective input length. Each DNA barcode is translated codon by codon into its corresponding amino acid sequence, preserving protein-level constraints that remain informative for species discrimination. The translated sequence provides a compact and interpretable input for learning, without requiring large-scale pretraining or an additional decompression stage. On this representation, we propose Protein FT-Transformer, an amino-acid-based transformer architecture for species classification. The model employs a dual-stream feature tokenizer to combine residue-level information with broader contextual features, followed by a multi-head self-attention encoder that models dependencies across distant sequence positions. Training further incorporates class-weighted focal loss to address class imbalance, class-wise temperature scaling to improve probability calibration, and lightweight ensemble classification heads to improve prediction stability. The proposed framework is evaluated on six publicly available DNA barcode benchmarks spanning invertebrate, vertebrate, and plant taxa. Protein FT-Transformer outperforms classical machine learning classifiers and nucleotide-based deep learning baselines, achieving an average classification accuracy of 98.20% with improved F1-score and AUC values under modest computational and memory requirements. t-SNE visualizations of the learned embeddings also show more coherent species-level grouping. The results suggest that codon translation coupled with a task-specific transformer provides an effective supervised strategy for DNA barcode-based species identification.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149067293","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":"MoTRANet: Safety-Aware Drug Recommendation via Molecular Representation Learning and Dynamic Patient Modeling.","authors":"Pengtao Jia, Junhao Gao, Jingtao Sha, Wenqian Dong","doi":"10.1109/TCBBIO.2026.3736819","DOIUrl":"10.1109/TCBBIO.2026.3736819","url":null,"abstract":"<p><p>With the rapid development of artificial intelligence in healthcare, medication recommendation systems have become important tools for achieving personalized treatment. However, existing methods still have limitations in modeling drug molecular functions and perceiving dynamic changes in patient conditions. To address these issues, this paper proposes MoTRANet, a safety-aware combinatorial medication recommendation model. The model designs a dynamic patient modeling module that integrates GRU, Transformer, and diagnosis-procedure cross attention to effectively capture the dynamic evolution of patient health status. On the drug side, an ECFP-enhanced graph attention network and LightGCN-based motif propagation are employed to construct drug representations with molecular functional information. Furthermore, a DDI-aware encoder and a curriculum learning strategy are introduced to dynamically balance recommendation accuracy and medication safety during training. Experimental results on two public datasets, MIMIC III and MIMIC-IV, demonstrate that MoTRANet outperforms existing mainstream baseline methods in terms of Jaccard, F1, and PRAUC, while achieving a relatively low DDI rate, validating the model's accuracy and safety. Future work will focus on multi center real-world data validation, dose sensitivity analysis, and model interpretability.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149027462","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":"AI-Driven Bio-Constrained Fusion Modeling for Biologically Consistent Multiomics Analysis.","authors":"Jothi Prakash V, Arul Antran Vijay S, Ganeshkumar Pugalendhi","doi":"10.1109/TCBBIO.2026.3731395","DOIUrl":"10.1109/TCBBIO.2026.3731395","url":null,"abstract":"<p><p>Multiomics learning is central to computational biology, yet existing models often prioritise predictive accuracy over biological admissibility, yielding brittle representations and limited cross-cohort generalisation. We address the problem of learning transferable multiomics representations that remain consistent with known biological structure under missing modalities and distribution shifts. We propose Bio-Constrained Fusion Modeling (BCFM), a constraint-aware framework that integrates pathway coherence and interaction sparsity directly into representation learning via an augmented Lagrangian formulation. Modality-specific encoders project heterogeneous omics data into a shared latent space, fused through adaptive gating and aligned across modalities, while biological constraints are enforced as feasibility conditions rather than heuristic regularisers. Extensive experiments on TCGA pan-cancer data and external GEO cohorts show consistent gains over nine representative baselines: up to 3.4% higher classification accuracy, 3.1% higher macro-F1, and a 2.5% absolute gain in survival C-index, alongside substantially reduced biological constraint violations, mitigated degradation under missing modalities and cross-cancer transfer, and stable constraint convergence during training. Overall, embedding biological admissibility into optimisation yields robust, generalisable, and biologically plausible multiomics representations that transfer across cohorts and cancer types for bioinformatics applications in biomedical research.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-09-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149032723","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":"GLIMP: An Integrated Graph Neural Network-Large Language Model for Promoter Recognition.","authors":"Haitao Ma, He Jia, Chenning Yuan, Changyong Yu, Yuhai Zhao, Keming Mao","doi":"10.1109/TCBBIO.2026.3735721","DOIUrl":"10.1109/TCBBIO.2026.3735721","url":null,"abstract":"<p><p>Accurate identification of DNA promoter regions is crucial for understanding gene regulation, yet remains challenging. Existing models often struggle to generalize across species or promoter types due to their inability to model functional degeneracy and capture subtle regulatory signals. To address this, we propose GLIMP, a novel promoter classification framework that integrates a large-scale genomic language model with Graph Neural Networks (GNNs). The core innovation of GLIMP lies in its multimodal fusion of a Shapelet-guided Weighted Context-Aware Interaction Network (WC-IDNG) graph representation-which topologically connects functionally similar but sequence-divergent motifs-and deep semantic features extracted by DNABERT. Specifically, we construct a WC-IDNG graph from $k$-mers to capture local proximity and structural similarity, refined by biologically meaningful shapelets. These graph-aggregated signals are then fused with global contextual semantics from DNABERT via a Transformer-based architecture to capture long-range dependencies. Extensive experiments on multi-species datasets demonstrate that GLIMP achieves state-of-the-art performance, with an Accuracy of 96.3% and an F1-score of 96.3% (on A. thaliana), substantially improving recognition robustness and generalization across diverse promoter types.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-09-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149032734","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":"RaFT-DM: A Residue-Aware Fusion Transformer With Domain-Wise Memory for Accurate Multi-Label Protein Function Prediction.","authors":"Jiakai Zhang, Zihang Zhang, Jiaru Yang, Weiping Ding, Zhenyu Lei, Shangce Gao","doi":"10.1109/TCBBIO.2026.3734966","DOIUrl":"10.1109/TCBBIO.2026.3734966","url":null,"abstract":"<p><p>Accurate protein function prediction via modeling protein sequences as long structured symbolic sequences is essential for understanding cellular mechanisms and guiding drug discovery. However, conventional deep learning models often use global pooling, which weakens domain-specific signals and fails to capture contextual dependencies across domains. Here, we propose Residue-aware Fusion Transformer with Domain-wise Memory (RaFT-DM), a domain-conditioned information routing framework for multi-label protein function prediction. Rather than treating domain annotations as auxiliary inputs, RaFT-DM uses functional domains as structured intermediate memory units between residue-level multimodal evidence and protein-level GO labels. It first aligns sequence and structural signals at the residue level, then routes the fused evidence through domain-wise bottlenecks that preserve local functional cues and model contextual dependencies among domains. Experiments on standard benchmarks show that RaFT-DM consistently outperforms state-of-the-art baselines. By replacing global pooling with domain-conditioned routing, RaFT-DM reduces false negatives and misclassifications, enabling more accurate and interpretable predictions. The implementation of RaFT-DM is available at https://github.com/AI4DS-wk/RaFT-DM.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148993606","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":"Allosteric Pocket Prediction Based on Multi-modal Protein Language Models and Ensemble Learning.","authors":"Huiling Zhang, Shuwen Wen, Ziqi Xu, Xiaochuan Chen, Shaozhen Cai, Guanshang Huang, Haowen Zhao, Ling Yin, Yanjie Wei","doi":"10.1109/TCBBIO.2026.3734697","DOIUrl":"10.1109/TCBBIO.2026.3734697","url":null,"abstract":"<p><p>Allosteric regulation plays a crucial role in modulating protein function, and allosteric pockets represent promising targets for drug discovery. However, identifying allosteric pockets remains challenging due to their structural and evolutionary diversity. In this study, we propose AlloPoc, a novel framework that integrates multi-modal protein language model (PLM) representations with biochemical features for accurate allosteric pocket prediction. Unlike existing PLM-based methods that rely exclusively on sequence-level features, AlloPoc leverages complementary sequence- and structure-based information to capture a more comprehensive view of allosteric pockets. To our knowledge, this represents the first integration of dual-modality PLMs in this field. AlloPoc comprises three core modules: feature extraction, feature processing (including selection and enhancement), and ensemble prediction. By integrating sequence- and structure-based PLM embeddings with physicochemical properties extracted by Fpocket, AlloPoc leverages a hierarchical ensemble strategy to mitigate prediction biases inherent to individual classifiers. Evaluated on an independent test set, AlloPoc achieves state-of-the-art performance with an AUC of 0.927 and an MCC of 0.524, surpassing existing methods including AlloFusion, DeepAllo, and AlloPED. These results demonstrate the effectiveness of integrating multi-modal PLMs with ensemble learning for allosteric pocket identification.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148993613","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":"Transferring Cancer Drug Response Prediction Through Multi-Label Heuristic Domain Adaptation.","authors":"Dan Luo, Jian Zhang, Lei You, Jie Yang, Xuan Liu","doi":"10.1109/TCBBIO.2026.3734533","DOIUrl":"10.1109/TCBBIO.2026.3734533","url":null,"abstract":"<p><p>Accurate prediction of cancer drug response (CDR) is critical for oncology drug screening. Given the scarcity of labeled patient data, domain adaptation has emerged as a promising method to transfer supervision signals from the label-rich cell line data (source domain) into the patient prediction (target domain). However, existing efforts usually treat each drug's response prediction as an independent task, requiring model retraining when drugs change, and struggle to capture reliable domain-invariant information between cell lines and patients through regular alignment strategies. Herein, we propose AdaptCDR, a multi-label heuristic domain adaptation model for transferring CDR prediction from cell lines to patients. Specifically, we pre-train an autoencoder to extract shared features across two domains, and a multi-label association classifier to map these features to response outcomes for various drugs simultaneously in the source domain. Then, we fine-tune the autoencoder to capture domain-invariant representations by distinguishing domain-specific counterparts as heuristics, thereby bridging the domain gap and enabling robust transfer of multi-label classifier to the target domain. Extensive experiments conducted on the TCGA and PDX datasets demonstrate that AdaptCDR boosts transfer accuracy and efficiency compared to previous methods. Subsequent case studies involving unmeasured patient samples further highlight its potential for reliable screening of clinical anti-cancer drugs.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-09-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148984590","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":"SSMGCN: Multi-View Graph Clustering with Shared-Specific Information Modelling for Spatially Resolved Transcriptomics.","authors":"Wei Zhang, Danyang Dong, Bangyi Zhang, Ziqi Zhang, Jun Zhou, Yun Zuo, Zhaohong Deng, Weiping Ding, Xiaoyong Pan, Jian Liu","doi":"10.1109/TCBBIO.2026.3733661","DOIUrl":"10.1109/TCBBIO.2026.3733661","url":null,"abstract":"<p><p>The rapid development of Spatial Transcriptomics (ST) enables simultaneous acquisition of gene expression and spatial locations, offering new avenues to explore tissue organization. However, effectively integrating spatial and transcriptional information for spatial domain identification remains challenging in existing methods. Specifically, most existing methods construct graphs using a single spatial similarity metric, making results sensitive to metric choice. Even methods that build dual graphs from spatial and expression views often fail to disentangle shared and view specific information systematically, which limits their robustness in complex biological contexts. To this end, a novel Shared-Specific Multi-view Graph Convolutional Network (SSMGCN) is proposed. SSMGCN first constructs spatial and gene-expression graphs. Subsequently, to fully exploit the shared and specific information between the two types of graphs, a shared-specific information decoupling autoencoder framework is proposed based on the graph convolutional network. In this framework, we first employ a shared encoder and view specific encoders to capture common and unique knowledge across views. In the decoding phase, a zero-inflated negative binomial (ZINB) decoder is applied to the expression data to model zero inflation and over-dispersion, while two structure decoders are assigned to the spatial and expression graphs to simultaneously reconstruct their adjacency relationships in the latent space, thereby preserving local topological continuity and long-range functional connectivity. Finally, a Student's t-distribution-based clustering is adopted in the embedding space, which leverages soft assignments and KL divergence-based self-training to enhance intra-cluster compactness and inter-cluster separation, thereby yielding discriminative representations for spatial domain partition. Extensive experiments on diverse ST datasets demonstrate that SSMGCN consistently outperforms existing methods in spatial domain identification. Moreover, its unified embedding supports downstream tasks such as cell-type annotation, spatial localization, and functional analysis, providing a robust foundation for mechanistic exploration. The code and dataset of this study are available at https://github.com/ddddoreen/SSMGCN.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148979290","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":"Trustworthy Drug-Drug Interaction Prediction Model via Weighted Aggregating Conformal Prediction.","authors":"Kaiyuan Zhang, Ting Wang, Zhixin Zhou, Wen Jung Li, Rui Luo","doi":"10.1109/TCBBIO.2026.3731205","DOIUrl":"10.1109/TCBBIO.2026.3731205","url":null,"abstract":"<p><p>Drug-Drug Interaction (DDI) prediction is crucial for ensuring the safety and efficacy of combination drug therapies. However, most existing methods lack uncertainty estimation, limiting their reliability and practical adoption. While conformal prediction (CP) offers guaranteed coverage in classification tasks, its direct application to DDI prediction faces challenges in achieving valid marginal coverage and efficient prediction set sizes. To overcome these limitations, this work introduces WeightedCPDDI, the first conformal prediction framework tailored for DDI tasks. WeightedCPDDI incorporates a novel weighting mechanism to achieve user-specific coverage and enhance prediction efficiency. Evaluation across transductive and inductive settings demonstrates WeightedCPDDI's strong performance. In the transductive setting, it achieves target coverage with smaller or comparable prediction sets, indicating improved calibration and efficiency. In the inductive setting, WeightedCPDDI remains competitive with or better than state-of-the-art methods in both seen-unseen and unseen-unseen scenarios, with the largest efficiency gains in the challenging unseen-unseen drug-pair setting, where it reduces prediction set sizes relative to the strongest baselines while maintaining the target coverage rate. Its strong performance across various backbone models highlights its generalizability and practical reliability for DDI prediction under uncertainty.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-09-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148979308","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":"SPFuseRanker: A Multi-Importance Score Fusion Framework for Core Microbiome Identification in Metagenomic Data.","authors":"Sirong Chen, Qi Guan, Da Zhou, Lihua Duan, Jie Hu","doi":"10.1109/TCBBIO.2026.3733474","DOIUrl":"10.1109/TCBBIO.2026.3733474","url":null,"abstract":"<p><p>Clinical metagenomic data are typically high-dimensional, sparse, and zero-inflated, and are often characterized by limited sample sizes and measurement noise. In addition, different feature-importance methods may produce inconsistent taxon rankings, which limits the stability of single-method feature selection and complicates the identification of candidate core microbiome members. To address this issue, we propose SPFuseRanker, a score-fusion-based ranking framework for integrating multiple microbial importance measures in metagenomic data. The method constructs a consensus score vector by combining heterogeneous importance signals, including statistical tests, correlation analysis, univariate classification performance, and tree-based feature importance. A top-weighted distance function based on Softmax normalization is introduced to emphasize highly ranked taxa during the fusion process. The optimization of the fused score vector is formulated as a minimum-distance problem and solved using a genetic algorithm (GA). We evaluate the proposed method using both synthetic simulations and a real-world systemic lupus erythematosus (SLE) gut microbiome dataset. Experimental results show that SPFuseRanker achieves more stable ranking performance compared with several representative rank aggregation and score fusion methods, particularly in terms of ranking consistency and robustness under noise. In addition, the selected candidate microbial taxa demonstrate improved predictive performance in disease classification tasks, suggesting their potential relevance to SLE-associated microbial signatures. Overall, SPFuseRanker provides a practical framework for integrating multi-source importance information and may serve as a useful tool for candidate core microbiome identification in metagenomic studies.</p>","PeriodicalId":520987,"journal":{"name":"IEEE transactions on computational biology and bioinformatics","volume":"PP ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2026-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148946500","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}