Proteome-Scale Drug-Target Interaction Predictions: Approaches and Applications.

Area of Science:

• Computational drug discovery
• Pharmacology
• Bioinformatics

Background:

• Drug-target interaction (DTI) prediction is crucial for computer-aided drug discovery.
• Proteome-scale DTI models are a recent advancement in the field.
• Existing literature often lacks sufficient detail to assess the effectiveness of proposed DTI prediction techniques.

Purpose of the Study:

• To provide context for understanding advances in proteome-scale drug-target interaction prediction.
• To evaluate emerging technologies for their suitability in DTI prediction tasks.
• To offer guidelines for designing, implementing, and validating new DTI prediction models.

Main Methods:

• Overview of current state-of-the-art in drug-target interaction prediction.
• Evaluation of emerging technologies for computational drug discovery.
• Discussion of validation approaches and potential biases in cross-validation methods.

Main Results:

• Identified deficiencies in the existing literature regarding the practical effectiveness of DTI prediction methods.
• Highlighted sources of bias in commonly used cross-validation techniques.
• Proposed key criteria for evaluating the validity of DTI prediction models.

Conclusions:

• There is a need for standardized validation and reporting in drug-target interaction prediction to improve model reliability.
• Awareness of biases in cross-validation is essential for accurate assessment of DTI prediction models.
• This review aims to improve the understanding and development of robust drug-target interaction prediction models.