Drug-target interaction prediction: databases, web servers and computational models.

Area of Science:

• Computational biology
• Drug discovery
• Bioinformatics

Background:

• Experimental drug-target interaction identification is costly and time-consuming.
• Computational models offer scalable solutions for predicting drug-target associations.
• Databases and web servers are crucial resources in this field.

Purpose of the Study:

• To review existing databases and web servers for drug-target identification.
• To introduce state-of-the-art computational models for predicting drug-target interactions.
• To discuss future directions and advanced evaluation frameworks.

Main Methods:

• Summary of databases and web servers.
• Introduction to network-based computational models.
• Detailed explanation of machine learning-based models (supervised and semi-supervised).

Main Results:

• Significant improvements in drug-target interaction prediction using computational models.
• Identification of limitations in current network-based and machine learning methods.
• Discussion of novel evaluation frameworks and regression formulations.

Conclusions:

• Computational approaches are vital for efficient drug discovery.
• Future research should focus on personalized drug discovery and improved validation.
• Advanced methods offer promise for more accurate drug-target interaction prediction.