Scholarlink

Protein–protein interaction (PPI) prediction is essential for understanding disease mechanisms, drug discovery, and therapeutic interventions, particularly in neurodegenerative diseases like Alzheimer’s disease and Behavioral and Psychological Symptoms of Dementia (BPSD). In this study, we propose a novel graph-based approach that integrates GraphSAGE, Graph Convolutional Networks (GCN), and Graph Attention Networks (GAT) within a Variational Graph Autoencoder (VGAE) framework to predict PPIs. By utilizing datasets from BPSD and the Integrated Interactions Database (IID), our model identifies potential interactions between shared proteins. This unique combination of GNN layers captures both local and global protein interaction features, while the VGAE introduces probabilistic inference for improved generalization. Comprehensive evaluation shows that our approach outperforms traditional models, offering a more accurate and robust method for link prediction in PPI networks. This work contributes a powerful tool for understanding the molecular underpinnings of Alzheimer’s disease and BPSD, potentially uncovering novel targets for further biological research.

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

Link Prediction in Protein Interaction Networks for Alzheimer’s Disease and Dementia Using Graph-Based Approaches

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • K. G. Gagana,
  • Aishwarya Rajkumar,
  • Alekhya Ponnekanti,
  • Arti Arya

摘要

Protein–protein interaction (PPI) prediction is essential for understanding disease mechanisms, drug discovery, and therapeutic interventions, particularly in neurodegenerative diseases like Alzheimer’s disease and Behavioral and Psychological Symptoms of Dementia (BPSD). In this study, we propose a novel graph-based approach that integrates GraphSAGE, Graph Convolutional Networks (GCN), and Graph Attention Networks (GAT) within a Variational Graph Autoencoder (VGAE) framework to predict PPIs. By utilizing datasets from BPSD and the Integrated Interactions Database (IID), our model identifies potential interactions between shared proteins. This unique combination of GNN layers captures both local and global protein interaction features, while the VGAE introduces probabilistic inference for improved generalization. Comprehensive evaluation shows that our approach outperforms traditional models, offering a more accurate and robust method for link prediction in PPI networks. This work contributes a powerful tool for understanding the molecular underpinnings of Alzheimer’s disease and BPSD, potentially uncovering novel targets for further biological research.