HGNNLDA: Predicting lncRNA-Drug Sensitivity Associations via a Dual Channel Hypergraph Neural Network.

Area of Science:

• Genomics
• Computational Biology
• Pharmacology

Background:

• Drug sensitivity is key for personalized cancer treatment.
• Long non-coding RNAs (lncRNAs) influence drug efficacy by regulating target genes.
• Current methods for identifying lncRNA-drug associations are limited in scale and efficiency.

Purpose of the Study:

• To develop an efficient computational framework for predicting lncRNA-drug sensitivity associations.
• To establish a novel method for exploring the relationship between lncRNAs and drug sensitivity.

Main Methods:

• Developed HGNNLDA, a dual-channel hypergraph neural network model.
• Utilized hypergraph neural networks to capture high-order interactions in lncRNA-drug networks.
• Employed a joint update mechanism for generating lncRNA and drug embeddings.

Main Results:

• HGNNLDA significantly outperformed six state-of-the-art models in predicting lncRNA-drug sensitivity associations.
• Case studies demonstrated HGNNLDA's effectiveness in identifying relevant lncRNA-drug links.
• The hypergraph approach effectively models complex, higher-order relationships.

Conclusions:

• HGNNLDA is the first computational framework for predicting lncRNA-drug sensitivity associations.
• The proposed method offers a scalable and effective approach for drug development and personalized treatment strategies.
• HGNNLDA advances the understanding of lncRNA roles in drug response.