An Attention-Aware Multi-Task Learning Framework Identifies Candidate Targets for Drug Repurposing in Sarcopenia.

Area of Science:

• Genomics and Bioinformatics
• Aging Research
• Pharmacology

Background:

• Sarcopenia is a significant public health issue linked to aging, causing muscle mass and strength decline.
• Current treatments for sarcopenia are limited, highlighting the urgent need for new pharmacological interventions.

Purpose of the Study:

• To develop a novel computational method for identifying drug targets and therapies for sarcopenia.
• To analyze multi-omics data, protein-protein interactions, and drug-target networks to discover novel therapeutic strategies.

Main Methods:

• A Multi-Task Attention-aware method for Multi-Omics data (MTA-MO) was developed to integrate transcriptome, methylome, and genome data.
• The MTA-MO method incorporated human protein-protein interaction (PPI) and drug-target networks for enhanced target identification.
• The approach was applied to a dataset of 1055 participants, analyzing genetic, epigenetic, and clinical data.

Main Results:

• The study identified seven potential hub genes associated with sarcopenia, including ESR1, ATM, CDC42, EP300, PIK3CA, EGF, and PTK2B.
• Pathway analysis revealed key functions and signaling pathways relevant to skeletal muscle health.
• In silico screening identified canagliflozin as a promising drug candidate, showing strong binding affinity to PTK2B.

Conclusions:

• The developed MTA-MO approach effectively integrates multi-omics data for sarcopenia research.
• Canagliflozin shows potential as a therapeutic agent for treating sarcopenia, warranting further investigation.