Integrating transcriptomic data with a novel drug efficacy prediction model for TCM active compound discovery

Area of Science:

• Computational biology
• Pharmacology
• Bioinformatics

Background:

• Drug discovery from natural products faces challenges in identifying active compounds.
• Predictive algorithms are needed for complex natural product analysis.
• Existing network topology analyses have limitations in drug-disease interaction prediction.

Purpose of the Study:

• To develop Meta-paths-based Drug Efficacy Prediction (Meta-DEP) for identifying active natural compounds.
• To predict drug-disease relationships using a drug-protein-disease heterogeneity network.
• To validate Meta-DEP's efficacy in discovering active compounds from traditional Chinese medicine.

Main Methods:

• Constructed a drug-protein-disease heterogeneity network.
• Utilized Meta-paths representing shortest paths between drug targets and disease proteins.
• Measured drug efficacy via predictive scores based on network proximity.
• Applied Meta-DEP to Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP) data.

Main Results:

• Meta-DEP outperformed traditional network topology analysis in predicting drug-disease interactions.
• Identified key drug targets consistent with clinical pharmacological evidence.
• Accurately predicted most drug-disease pairs within the TCMSP database.
• Biological experiments confirmed Meta-DEP's ability to mine active compounds from traditional Chinese medicine, integrating transcriptomic data.

Conclusions:

• Meta-DEP offers a novel approach for predicting active ingredients from natural products.
• The model facilitates the discovery of natural compounds as innovative lead molecules.
• Meta-DEP shows significant potential in advancing drug discovery and development from natural sources.