MUBD-DecoyMaker 2.0

MUBD-DecoyMaker 2.0 generates maximal unbiased benchmarking data sets for virtual screening to support reliable evaluation of ligand enrichment in drug discovery.

Key Features:

• Maximal Unbiased Benchmarking: Produces benchmarking data sets specifically constructed to minimize bias for reliable assessment of virtual screening methods.
• De-biasing Algorithm: Implements a de-biasing algorithm previously validated in pharmacological studies.
• Detect 2D Bias: Identifies and mitigates two-dimensional biases in molecular data sets.
• Quality Control: Applies quality-control procedures to ensure benchmarking data sets meet established standards from prior applications.
• Python Implementation: Delivered as a Python-based implementation for computational generation of decoy sets.
• MUBD-HDACs Contribution: Produces data sets that contribute to the MUBD-HDACs benchmark set.

Scientific Applications:

• Virtual Screening Benchmarking: Provides unbiased data sets for evaluating and optimizing virtual screening algorithms and workflows.
• Ligand Enrichment Assessment: Supports ligand enrichment analyses used to identify drug-like molecules in screening campaigns.

Methodology:

Implements a previously published de-biasing algorithm (validated in pharmacological studies) with explicit detection and mitigation of two-dimensional dataset biases and produces data sets used in the MUBD-HDACs benchmark.