The Computational Models of Drug-target Interaction Prediction

Area of Science:

• Bioinformatics
• Computational Biology
• Drug Discovery

Background:

• Identifying Drug-Target Interactions (DTIs) is crucial for drug discovery but traditional experimental methods are costly and time-consuming.
• Computational methods offer a faster, more economical alternative for predicting potential DTIs.

Purpose of the Study:

• To summarize and introduce state-of-the-art computational methods for DTI identification.
• To compare network-based and machine learning-based approaches for DTI prediction.

Main Methods:

• Review of existing literature on computational DTI identification methods.
• Introduction of network-based and machine learning-based (supervised and semi-supervised) models.
• Evaluation of methods using Area Under the Precision Recall curve (AUPR) on benchmark datasets.

Main Results:

• Computational methods, particularly network-based and machine learning, show significant improvements in DTI identification.
• Supervised and semi-supervised machine learning models exhibit distinct strategies for handling negative samples.
• Both network-based and machine learning methods possess inherent limitations.

Conclusions:

• Computational approaches provide valuable tools for accelerating DTI identification in drug discovery.
• Understanding the nuances of different computational models, especially in negative sample handling, is key for effective DTI prediction.
• Further research is needed to overcome the limitations of current computational methods for DTI identification.