DASPfind: new efficient method to predict drug-target interactions.

Area of Science:

• Computational drug discovery
• Bioinformatics
• Cheminformatics

Background:

• Accurate identification of novel drug-target interactions (DTIs) is crucial for drug discovery.
• Experimental DTI determination is costly and time-consuming, necessitating efficient computational prediction methods.
• Existing computational methods often yield a high rate of false positive predictions.

Purpose of the Study:

• To develop a novel computational method for predicting drug-target interactions (DTIs).
• To improve the accuracy and reliability of DTI predictions compared to existing state-of-the-art methods.
• To provide a practical tool for identifying novel DTIs, particularly for drugs with limited known targets.

Main Methods:

• Developed DASPfind, a novel computational DTI prediction method.
• DASPfind utilizes simple paths from a graph representing DTIs, drug similarities, and protein target similarities.
• Evaluated performance on four gold standard DTI datasets.

Main Results:

• DASPfind significantly outperforms existing methods in predicting DTIs, achieving 46.17% accuracy for top-ranked predictions.
• Achieved 49.22% accuracy for top-ranked predictions when considering all DTIs for a single drug.
• Demonstrated superior performance for drugs with no or few known targets.

Conclusions:

• DASPfind is an effective computational method for identifying reliable novel drug-target interactions.
• Outperforms state-of-the-art methods across six DTI datasets, especially for top-ranked and novel predictions.
• Practical application shown through novel predictions for the Ion Channel dataset, reducing experimental verification costs in drug discovery.