MTPrior: A Multi-Task Hierarchical Graph Embedding Framework for Prioritizing Hepatocellular Carcinoma-Associated...

Area of Science:

• Oncology
• Bioinformatics
• Genomics

Background:

• Hepatocellular carcinoma (HCC) is a prevalent liver cancer with significant global health implications.
• Identifying key coding genes and noncoding RNAs (ncRNAs), like long noncoding RNAs (lncRNAs), is crucial for HCC understanding and treatment.
• Current computational models often focus on single RNA types, neglecting crucial interactions between coding and noncoding RNAs.

Purpose of the Study:

• To introduce MTPrior, a multi-task graph embedding model for prioritizing cancer-associated genes and lncRNAs in HCC.
• To develop a computational approach that accounts for interactions between coding and noncoding RNAs.
• To provide a more efficient method for identifying key RNA candidates for HCC research.

Main Methods:

• Developed MTPrior, a multi-task graph embedding model.
• Constructed an adaptable framework to accommodate diverse prioritization tasks.
• Incorporated interactions between coding and noncoding RNAs within a biological pathway context.
• Utilized extensive HCC patient datasets and comprehensive gene/lncRNA inventories.

Main Results:

• MTPrior successfully prioritized and identified relevant genes and lncRNAs associated with HCC.
• An ablation study confirmed the effectiveness of individual model components.
• MTPrior demonstrated superior performance compared to existing state-of-the-art methods in predicting disease-related RNAs.

Conclusions:

• MTPrior offers an efficient and effective approach for prioritizing multiple RNA types in HCC.
• The model's ability to consider coding-noncoding RNA interactions advances RNA-based cancer research.
• MTPrior streamlines the identification of critical RNA candidates for further investigation in HCC.