QKDTI A quantum kernel based machine learning model for drug target interaction prediction.

Area of Science:

• Computational chemistry
• Quantum machine learning
• Drug discovery

Background:

• Drug-target interaction (DTI) prediction is vital for drug discovery, but traditional methods face computational and generalization challenges.
• Quantum Machine Learning (QML) offers enhanced accuracy, scalability, and efficiency through quantum computing principles like superposition and entanglement.

Purpose of the Study:

• To introduce QKDTI, a novel quantum-enhanced framework for DTI prediction.
• To leverage Quantum Support Vector Regression (QSVR) with quantum feature mapping for improved binding affinity predictions.

Main Methods:

• Developed QKDTI using QSVR with quantum feature mapping for molecular and protein features.
• Integrated the Nystrom approximation for efficient kernel approximation and reduced computational overhead.
• Evaluated QKDTI on benchmark datasets (Davis, KIBA) and validated on BindingDB.

Main Results:

• QKDTI achieved high accuracy: 94.21% on Davis, 99.99% on KIBA, and 89.26% on BindingDB.
• The model significantly outperformed classical and other quantum DTI prediction models.
• Statistical tests confirmed the reliability and superiority of the QKDTI results.

Conclusions:

• QKDTI demonstrates the potential of quantum computing to revolutionize computational drug discovery.
• The framework offers improved predictive accuracy and generalization capabilities for DTI prediction.
• This approach can accelerate drug repurposing, precision medicine, and virtual screening.