Accurate Drug Repositioning through Non-tissue-Specific Core Signatures from Cancer Transcriptomes

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Experimental drug repositioning screens are limited by in vitro conditions and lack human in vivo applicability.
• Existing methods struggle with coverage and specificity for predicting drug efficacy in complex biological systems.

Purpose of the Study:

• To develop an in silico screening method for drug repositioning under human in vivo conditions.
• To create a computational tool, core-signature drug-to-gene (csD2G), for scanning drug profiles against transcriptome signatures.
• To enhance the accuracy and applicability of drug repositioning strategies.

Main Methods:

• Generated a reference of 8,476 core transcriptome signatures (CSs) from TCGA database for single gene mutations.
• Developed the core-signature drug-to-gene (csD2G) software to analyze 3,546 drug treatment profiles against CSs.
• Validated predictions through applications in pathway inhibition, clinical electronic medical records, and model organism lifespan regulation.

Main Results:

• The csD2G software demonstrated superior performance over conventional cell line-based methods in coverage and specificity.
• Identified psychiatric drugs for TGF-β pathway inhibition.
• Predicted and validated antihypertensive calcium channel blockers impacting AMPK and AKT pathways, supported by clinical data.

Conclusions:

• The developed in silico approach offers a powerful and accurate tool for drug repositioning in human in vivo contexts.
• The csD2G software significantly advances drug discovery by providing higher specificity and broader coverage.
• Demonstrated applications highlight the potential for repurposing drugs across various therapeutic areas and biological models.