Computational framework for analyzing miRNA-mRNA interactions in sarcopenia: Insights into age-related muscular degeneration

Abstract

Background

Sarcopenia, an age-related loss of skeletal muscle mass and function, impairs mobility, fragility, and quality of life. Despite progress in pathophysiology, molecular processes remain unknown. Recent research has investigated miRNAs as biomarkers for sarcopenia diagnosis and therapy. This work analyses differentially expressed genes (DEGs) and predicts miRNA-mRNA interactions using ML methods like XG-Boost and SHAP to find biomarkers.

Objective

This work evaluated the function of miRNA-mRNA interactions in sarcopenia pathogenesis and identified possible biomarkers by transcriptome analysis utilizing machine learning.

Methods

High-throughput mRNA sequencing datasets (GSE111006, GSE111010, and GSE111016) from GEO database were combined, pre-processed, and normalized using TPM and DESeq2 methods. XG-Boost regression analysis used 80/20 training and testing sets. SHAP analysis was used to evaluate model data and find significant DEGs. PPI networks were created using the STRING database, while miRNA-mRNA interactions were predicted using Encori and displayed with Cytoscape. The degree scores of miRNA-mRNA interactions were utilized to find biomarkers.

Results

XG-Boost and SHAP analysis revealed 20 influential DEGs linked to sarcopenia. With 97% accuracy, the model predicted accurately. PPI network research identified six hub genes: NTRK2, PCK1, DSP, SCD, MMRN1, and EDIL3. MiRNA-mRNA interaction analysis found miR-186–5p as the highest-degree biomarker candidate (36). MiR-186–5p was linked to muscle metabolism, hypertrophy, and exercise response.

Conclusion

The study found miR-186–5p to be a promising biomarker for sarcopenia using an integrated machine learning technique. The findings show that miR-186–5p may be a diagnostic and therapeutic target for sarcopenia, revealing its pathogenesis and enabling tailored treatments. Experimental research is needed to prove its therapeutic value.