Integration of RNA molecules data with prior-knowledge driven Joint Deep Semi-Negative Matrix Factorization for heart...

Area of Science:

• Biochemistry
• Genomics
• Computational Biology

Background:

• Heart failure (HF) is a primary manifestation of cardiovascular disease.
• RNA molecules and their interactions are crucial in HF pathogenesis and progression.

Purpose of the Study:

• To identify key RNA molecules associated with heart failure.
• To develop a robust computational model for HF biomarker discovery.

Main Methods:

• Proposed a Prior-knowledge Driven Joint Deep Semi-Negative Matrix Factorization (PD-JDSNMF) model.
• Integrated mRNA, lncRNA, and miRNA data with PPI information and Laplacian constraints.
• Identified significant co-expression modules and performed bioinformatics analysis.

Main Results:

• The PD-JDSNMF algorithm robustly selected HF-associated biomarkers.
• A diagnostic model using the top 13 genes achieved an AUC of 0.8714 (internal) and 0.8329 (external).

Conclusions:

• The PD-JDSNMF algorithm is effective for identifying HF biomarkers.
• The developed diagnostic model demonstrates high accuracy and validates the algorithm's utility.