Identification of Drug Compounds for Capsular Contracture Based on Text Mining and Deep Learning

Area of Science:

• Biomedical research
• Computational drug discovery
• Genomics and bioinformatics

Background:

• Capsular contracture is a frequent complication of breast implants with unclear causes.
• Current non-surgical treatments for capsular contracture lack proven effectiveness.

Purpose of the Study:

• To explore novel drug therapies for capsular contracture using computational approaches.
• To identify potential non-surgical treatments for this common breast implant complication.

Main Methods:

• Utilized text mining and bioinformatics tools (GeneCodis, STRING, Cytoscape) to identify genes associated with capsular contracture.
• Screened drug databases (Pharmaprojects) and employed machine learning (DeepPurpose) to predict drug-target interactions and binding affinity.

Main Results:

• Identified 55 relevant genes, narrowing down to eight candidate genes through pathway and interaction analyses.
• Selected 100 drugs targeting these genes, with seven candidates showing the highest predicted binding affinity, including specific antagonists and agonists.

Conclusions:

• Computational methods, including text mining and DeepPurpose, show promise for discovering non-surgical treatments for capsular contracture.
• This approach facilitates the identification of novel therapeutic strategies for managing breast implant complications.