Open-source chemogenomic data-driven algorithms for predicting drug-target interactions.

Area of Science:

• Computational biology
• Drug discovery
• Cheminformatics

Background:

• Despite advances in high-throughput screening, drug development success rates remain low.
• Drug repositioning presents a viable strategy to address this challenge.
• Experimental drug target identification is costly and time-consuming.

Purpose of the Study:

• To review and compare publicly available, data-driven computational algorithms for predicting drug-target interactions (DTIs).
• To provide guidance for researchers in selecting appropriate algorithms for drug repositioning and DTI prediction.

Main Methods:

• Systematic review of chemogenomic data-driven computational algorithms for DTI prediction.
• Organization of algorithms based on model properties and evolutionary relationships.
• Re-implementation of five representative algorithms in R and performance comparison using mean percentile ranking.

Main Results:

• Comparison of five representative DTI prediction algorithms based on mean percentile ranking.
• Evaluation of algorithm performance using a novel recall-based metric.

Conclusions:

• The review provides an objective comparison of computational DTI prediction algorithms.
• This work aims to assist researchers in improving existing algorithms or selecting suitable ones for drug repositioning projects.