A Knowledge-Graph-Based Multimodal Deep Learning Framework for Identifying Drug-Drug Interactions.

Area of Science:

• Pharmacology and Cheminformatics
• Artificial Intelligence in Drug Discovery

Background:

• Accurate identification of drug-drug interactions (DDIs) is critical for safe and effective drug development.
• Existing methods for DDI prediction often struggle to integrate complex, heterogeneous information sources.

Purpose of the Study:

• To develop and validate a deep learning framework (KGCN_NFM) for recognizing DDIs by integrating knowledge graphs and neural factorization machines.
• To assess the synergistic anticancer effects of identified drug combinations in lung carcinoma models.

Main Methods:

• Utilized knowledge graph convolutional networks (KGCNs) to learn drug embeddings capturing structural and semantic information from knowledge graphs.
• Integrated KGCN embeddings with Morgan molecular fingerprints as input for neural factorization machines (NFMs) to predict DDIs.
• Validated predicted interactions (topotecan and dantron) using MTT assays, apoptosis experiments, cell cycle analysis, and molecular docking.

Main Results:

• KGCN_NFM demonstrated superior performance compared to state-of-the-art algorithms on two real-world datasets.
• Experimental validation confirmed the predicted interaction between topotecan and dantron.
• The combination therapy of topotecan and dantron exhibited synergistic anticancer effects against lung carcinoma.

Conclusions:

• KGCN_NFM is an effective deep learning tool for DDI identification by integrating heterogeneous information.
• The identified synergistic interaction between topotecan and dantron offers a promising therapeutic strategy for lung carcinoma.
• This approach advances computational methods in drug discovery and personalized medicine.