Predicting Drug-Target Interactions Over Heterogeneous Information Network.

Area of Science:

• Bioinformatics
• Computational Biology
• Drug Discovery

Background:

• Drug-target interactions (DTIs) are crucial for understanding disease and developing new therapies.
• Experimental DTI identification is costly and inefficient, necessitating computational approaches.
• Existing computational methods often overlook the network topology of DTIs.

Purpose of the Study:

• To propose LG-DTI, a novel network-based computational method for accurate DTI prediction.
• To leverage both local and global network information for enhanced DTI identification.
• To provide a valuable tool for discovering novel drug-target interactions.

Main Methods:

• Developed LG-DTI, a semi-supervised heterogeneous network embedding method.
• Learned local representations from drug molecular structures and protein sequences.
• Integrated local and global representations for Random Forest classification of DTIs.

Main Results:

• LG-DTI demonstrated superior performance in DTI prediction on two independent datasets.
• The method effectively utilizes heterogeneous network information for accurate predictions.
• Experimental results confirmed the method's effectiveness compared to state-of-the-art approaches.

Conclusions:

• LG-DTI offers a powerful and accurate approach for predicting drug-target interactions.
• The method's ability to incorporate network topology enhances DTI discovery.
• LG-DTI serves as a valuable tool for accelerating drug development and pathogenesis research.