Scholarlink

MicroRNAs (miRNAs) are small non-coding RNAs involved in gene regulation and closely associated with various diseases, especially cancers. To resolve the issues inherent in experimental methods, we propose GCCDN, a computational model for predicting miRNA–disease associations. GCCDN integrates Chebyshev Graph Convolution and Graph Diffusion, using similarity-based graphs constructed from HMDD v2.0. Personalized PageRank enhances feature propagation, while ChebConv captures multi-hop information efficiently. Experiments on HMDD v2.0 and v3.2 show that GCCDN achieves AUCs of 94.28% and 95.04%, outperforming existing methods and demonstrating its potential for disease diagnosis and treatment discovery.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Predicting MiRNA-Disease Associations Using Chebyshev Graph Convolution and Graph

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Lihao Zhou,
  • Zhengwei Li,
  • Ru Nie,
  • Lei Wang,
  • Zhuhong You

摘要

MicroRNAs (miRNAs) are small non-coding RNAs involved in gene regulation and closely associated with various diseases, especially cancers. To resolve the issues inherent in experimental methods, we propose GCCDN, a computational model for predicting miRNA–disease associations. GCCDN integrates Chebyshev Graph Convolution and Graph Diffusion, using similarity-based graphs constructed from HMDD v2.0. Personalized PageRank enhances feature propagation, while ChebConv captures multi-hop information efficiently. Experiments on HMDD v2.0 and v3.2 show that GCCDN achieves AUCs of 94.28% and 95.04%, outperforming existing methods and demonstrating its potential for disease diagnosis and treatment discovery.