Ziqi Xia
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Ziqi Xia has a total of 11 co-authors. Here are the top co-authors by number of shared publications:
- Lei Wang
School of Computer Science and Technology/School of Artificial Intelligence, China University of Mining and Technology, Xuzhou 221116, China
3 shared publications - Meng-Meng Wei
School of Computer Science and Technology/School of Artificial Intelligence, China University of Mining and Technology, Xuzhou 221116, China
3 shared publications - Xin-Fei Wang
College of Computer Science and Technology, Jilin University, Changchun, China
2 shared publications - Yue-Chao Li
Northwestern Polytechnical University
2 shared publications - Zhu-Hong You
School of Computer Science, Northwestern Polytechnical University, Xi’an 710129, China
2 shared publications - Shengli Li
Clinical Research Institute, The Affiliated Hospital of Xuzhou Medical University, Xuzhou, China
1 shared publication
Latest Publications
- 16 July 2026
- Published by Springer Nature
- p. 79-89
A Hybrid Transformer-GCN Framework for CircRNA-Disease Association Prediction
- 16 July 2026
- Published by Springer Nature
- p. 79-89
A Hybrid Transformer-GCN Framework for CircRNA-Disease Association Prediction
- Open Access
- 13 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 231-236
scGraphPilot: Functionally Guided Graph Transfer for Robust Cross-Patient Cell Type Annotation
Cross-patient cell type annotation in single-cell RNA sequencing (scRNA-seq) remains challenging due to strong biological heterogeneity and distribution shifts between patient datasets. Many existing annotation methods rely on transcriptional similarity or marker gene information, which may lead to unstable predictions when transferring knowledge across patients. In this work, we propose scGraphPilot, a graph-based framework designed for robust cross-patient cell type annotation. The proposed method introduces KEGG functional-guided graph learning, where pathway activity signals derived from KEGG categories are incorporated to refine cell–cell graph structures. This mechanism encourages biologically coherent neighborhoods and reduces noisy edges caused by purely expression-based similarity. In addition, a cross-patient representation alignment strategy is employed to enable knowledge transfer from labeled reference datasets to unlabeled query datasets while preserving patient-specific graph topology. Experiments on the leukemia dataset GSE132509 demonstrate that scGraphPilot achieves stable annotation performance across multiple cross-patient transfer tasks and consistently outperforms several representative annotation models in terms of accuracy and robustness. Furthermore, biological validation using known immune-related gene interactions confirms that the learned graph connectivity reflects meaningful cellular communication patterns. These results indicate that integrating functional signals into graph learning provides an effective and interpretable solution for cross-patient cell type annotation in heterogeneous single-cell datasets.
- 13 March 2026
- Published by Association for Computing Machinery (ACM)
- p. 231-236
Open AccessscGraphPilot: Functionally Guided Graph Transfer for Robust Cross-Patient Cell Type Annotation
Cross-patient cell type annotation in single-cell RNA sequencing (scRNA-seq) remains challenging due to strong biological heterogeneity and distribution shifts between patient datasets. Many existing annotation methods rely on transcriptional similarity or marker gene information, which may lead to unstable predictions when transferring knowledge across patients. In this work, we propose scGraphPilot, a graph-based framework designed for robust cross-patient cell type annotation. The proposed method introduces KEGG functional-guided graph learning, where pathway activity signals derived from KEGG categories are incorporated to refine cell–cell graph structures. This mechanism encourages biologically coherent neighborhoods and reduces noisy edges caused by purely expression-based similarity. In addition, a cross-patient representation alignment strategy is employed to enable knowledge transfer from labeled reference datasets to unlabeled query datasets while preserving patient-specific graph topology. Experiments on the leukemia dataset GSE132509 demonstrate that scGraphPilot achieves stable annotation performance across multiple cross-patient transfer tasks and consistently outperforms several representative annotation models in terms of accuracy and robustness. Furthermore, biological validation using known immune-related gene interactions confirms that the learned graph connectivity reflects meaningful cellular communication patterns. These results indicate that integrating functional signals into graph learning provides an effective and interpretable solution for cross-patient cell type annotation in heterogeneous single-cell datasets.
- Open Access
- 1 January 2026
- Published by Science Exploration Press in Computational Biomedicine
- Vol. 1 (4) , 202614
Identification of potential associations between cirRNAs and diseases based on meta relation aware
Aims: Circular RNAs (circRNAs) have been shown to be closely associated with the occurrence and progression of various diseases. However, most existing circRNA-disease association prediction methods are limited to homogeneous networks and are unable to effectively capture deep semantic associations through high-order meta-paths. This study aims to develop an efficient computational method for accurately predicting potential circRNA-disease associations. Methods: We propose a meta-relation-aware heterogeneous graph learning framework for circRNA-disease association prediction. Specifically, known circRNA-disease associations are first used to compute Gaussian interaction profile kernel similarity and extract node attribute features, based on which a heterogeneous graph network is constructed. A graph neural network is then employed to perform multi-layer message passing on the heterogeneous graph, aggregating neighborhood information to achieve deep fusion of multi-source features and generate node embeddings that encode both local and global structural information. Finally, the learned embeddings are fed into a gradient boosting decision tree classifier, and an ensemble strategy is adopted to improve prediction accuracy. Five-fold cross-validation is used for performance evaluation. Results: Experimental results on three benchmark datasets, CircR2Disease V2.0, circAtlas 3.0, and circRNADisease V2.0, show that the proposed model achieves area under the receiver operating characteristic curve (AUC) values of 92.17%, 91.83%, and 91.73%, respectively. The model outperforms traditional methods in terms of accuracy, precision, and recall. Furthermore, ablation studies validate the effectiveness of the meta-relation-aware strategy. Conclusions: Overall, this work provides an efficient and reliable computational framework for molecular association prediction and biomarker discovery in the biomedical domain.
- 1 January 2026
- Published by Science Exploration Press in Computational Biomedicine
- Vol. 1 (4) , 202614
Open AccessIdentification of potential associations between cirRNAs and diseases based on meta relation aware
Aims: Circular RNAs (circRNAs) have been shown to be closely associated with the occurrence and progression of various diseases. However, most existing circRNA-disease association prediction methods are limited to homogeneous networks and are unable to effectively capture deep semantic associations through high-order meta-paths. This study aims to develop an efficient computational method for accurately predicting potential circRNA-disease associations. Methods: We propose a meta-relation-aware heterogeneous graph learning framework for circRNA-disease association prediction. Specifically, known circRNA-disease associations are first used to compute Gaussian interaction profile kernel similarity and extract node attribute features, based on which a heterogeneous graph network is constructed. A graph neural network is then employed to perform multi-layer message passing on the heterogeneous graph, aggregating neighborhood information to achieve deep fusion of multi-source features and generate node embeddings that encode both local and global structural information. Finally, the learned embeddings are fed into a gradient boosting decision tree classifier, and an ensemble strategy is adopted to improve prediction accuracy. Five-fold cross-validation is used for performance evaluation. Results: Experimental results on three benchmark datasets, CircR2Disease V2.0, circAtlas 3.0, and circRNADisease V2.0, show that the proposed model achieves area under the receiver operating characteristic curve (AUC) values of 92.17%, 91.83%, and 91.73%, respectively. The model outperforms traditional methods in terms of accuracy, precision, and recall. Furthermore, ablation studies validate the effectiveness of the meta-relation-aware strategy. Conclusions: Overall, this work provides an efficient and reliable computational framework for molecular association prediction and biomarker discovery in the biomedical domain.