A cross-modal graph structure learning prediction model for CircRNA-MiRNA interactions

doi:10.1016/j.bspc.2026.110082

A cross-modal graph structure learning prediction model for CircRNA-MiRNA interactions

2026-03-20

0

PRE

AI

DOI：10.1016/j.bspc.2026.110082

Original

Original request for help

Save

Abstract

En 中文

• MAGPI integrates sequence, structure, and expression similarities into a compact multimodal representation. • A hybrid Transformer-GCN with KNN hypergraphs and a learnable global node captures global semantics and high-order structure. • Layer attention and a stacked ensemble classifier improve discrimination and robustness across circRNA-miRNA tasks. • Experiments on three benchmarks show MAGPI surpasses state-of-the-art methods in AUC/AUPR with strong generalization.

Keywords:

MAGPI

circRNA-miRNA interactions

multimodal representation

hybrid Transformer-GCN

graph structure learning

A cross-modal graph structure learning prediction model for CircRNA-MiRNA interactions

Abstract

Journal

Organization