PyRMD

PyRMD performs ligand-based virtual screening using the Random Matrix Discriminant (RMD) algorithm and machine learning to identify biologically active compounds for drug discovery.

Key Features:

• Implementation: Implemented in Python and leveraging machine learning techniques.
• Algorithm: Implements the Random Matrix Discriminant (RMD) algorithm tailored for ligand identification.
• Automated workflow: Provides an automated computational workflow for virtual screening tasks.
• Data integration: Integrates target bioactivity data directly from the ChEMBL repository.
• Active/inactive partitioning: Processes bioactivity data by splitting compounds into active and inactive sets for model learning.
• High-throughput screening: Screens large compound libraries, including millions of compounds, within short time frames.
• Parameter tuning: Supports customization and tuning of calculation parameters.
• Performance metrics: Computes benchmark metrics for evaluating model performance and predictive potential.

Scientific Applications:

• Ligand identification: Identification of new ligands with high biological activity in drug discovery projects.
• High-throughput virtual screening: Rapid evaluation of large chemical libraries to prioritize candidate compounds.
• Model benchmarking: Assessment and benchmarking of predictive performance for screening models using computed metrics.

Methodology:

Implemented in Python; leverages machine learning and the Random Matrix Discriminant (RMD) algorithm; integrates bioactivity data from ChEMBL and splits compounds into active and inactive sets; performs high-throughput screening of compound libraries; allows parameter tuning and computes benchmark performance metrics.