StackHCV

StackHCV identifies inhibitors of the hepatitis C virus (HCV) NS5B polymerase using a stacked machine-learning meta-predictor that integrates models built from diverse molecular fingerprints.

Key Features:

• Machine learning meta-predictor: Integrates multiple baseline models constructed from diverse molecular fingerprints to enhance prediction accuracy and stability.
• Diverse algorithms: Builds baseline models using k-nearest neighbor, multi-layer perceptron, partial least squares, random forest, and support vector machines.
• Stacking strategy: Combines baseline models via a stacking approach to form the final meta-predictor.
• Benchmarking performance: Evaluated on training datasets and independent test datasets, showing superior accuracy and stability relative to existing predictors.

Scientific Applications:

• Drug discovery: Prioritizes and predicts compounds with potential inhibitory activity against HCV NS5B polymerase for screening efforts.
• Liver cancer therapy: Supports identification of HCV-targeting inhibitors relevant to therapeutic strategies where viral replication contributes to disease.
• Clinical research: Facilitates efficient identification of candidate NS5B inhibitors for studies addressing hepatitis C and related clinical conditions.

Methodology:

Construct baseline models from diverse molecular fingerprints using k-nearest neighbor, multi-layer perceptron, partial least squares, random forest, and support vector machines; integrate baseline models via a stacking strategy to form a meta-predictor; perform benchmarking on training and independent test datasets.