Related Experiment Video
Updated: May 30, 2025
07:35
A Knowledge Graph Approach to Elucidate the Role of Organellar Pathways in Disease via Biomedical Reports
Published on: October 13, 2023
1.6K
Subgraph Topology and Dynamic Graph Topology Enhanced Graph Learning and Pairwise Feature Context Relationship
Ping Xuan1,2, Xiaoying Qi1, Sentao Chen2
1School of Computer Science and Technology, Heilongjiang University, Harbin 150080, China.
Journal of Chemical Information and Modeling
|January 27, 2025
Summary
This study introduces SFPred, a novel model for predicting microRNA-disease associations by integrating local subgraph structures and contextual features. SFPred enhances disease diagnosis by accurately identifying candidate microRNAs.
Area of Science:
- Bioinformatics
- Computational Biology
- Genomics
Background:
- MicroRNAs (miRNAs) are crucial biomarkers for human diseases, making accurate prediction of disease-related miRNAs vital for early diagnosis.
- Existing prediction models often overlook subgraph structures and contextual relationships within miRNA-disease heterogeneous networks.
Purpose of the Study:
- To develop SFPred, a model that integrates local subgraph topologies, dynamic graph evolution, and pairwise feature context for improved miRNA-disease association prediction.
- To enhance the accuracy and reliability of identifying candidate disease-related miRNAs.
Main Methods:
- Formulating node importance based on neighbor subgraphs.
- Employing a multi-layer perceptron (MLP) for dynamic graph topology construction with evolving node features.
- Utilizing a context-aware transformer to integrate pairwise miRNA-disease feature relationships.
- Implementing a multiperspective residual strategy to preserve detailed features across encoding layers.
Main Results:
- SFPred significantly outperformed eight state-of-the-art methods in predicting miRNA-disease associations.
- Ablation experiments confirmed the effectiveness of the proposed model innovations.
- High recall rates for top-ranked candidate miRNAs and successful case studies demonstrated SFPred's practical utility.
Conclusions:
- SFPred offers a robust and effective approach for predicting miRNA-disease associations by leveraging advanced graph encoding and feature integration techniques.
- The model shows strong potential for screening reliable candidate miRNAs for further biological validation and advancing disease diagnostics.
