GCAC: galaxy workflow system for predictive model building for virtual screening.

Area of Science:

• Computational chemistry
• Cheminformatics
• Drug discovery

Background:

• Traditional drug discovery is lengthy, costly, and suffers from inconsistent high throughput screening (HTS) results.
• Reproducible research is crucial for validating computational findings in drug discovery.
• Current methods for identifying drug-like molecules via HTS lack confidence and reproducibility.

Purpose of the Study:

• To develop a reproducible knowledge-based predictive model building system for drug discovery.
• To address the challenge of tool availability in high throughput virtual screening.
• To enhance the reliability and efficiency of identifying potential drug candidates.

Main Methods:

• Development of a web-enabled data mining analysis pipeline named "Galaxy for Compound Activity Classification (GCAC)".
• Leveraging R statistical packages and literate programming within the Galaxy workflow system.
• Automated configuration for descriptor calculation, feature selection, model building, and screening.

Main Results:

• GCAC integrates R packages and literate programming in a reproducible workflow.
• The pipeline automates key steps in virtual screening, from descriptor calculation to candidate extraction.
• Demonstrates a standardized approach for predictive model building in virtual screening.

Conclusions:

• GCAC provides a reproducible standard for screening drug candidates using predictive modeling.
• The system facilitates easy installation and ensures reproducibility in computational drug discovery.
• A demo site for GCAC is available for public access and evaluation.