Twilight

Twilight analyzes protein-ligand complexes from X-ray crystallography deposited in the Protein Data Bank (PDB) to assess ligand model quality using the real-space correlation coefficient (RSCC).

Key Features:

• Pre-filtered dataset: Operates on a curated set of 2,815 protein-ligand complexes from the PDB selected for ligand RSCC values below 0.6.
• RSCC-based quality assessment: Uses the real-space correlation coefficient (RSCC) to quantify the fit between atom coordinates and electron-density maps for ligand model evaluation.
• Visualization of annotated complexes: Provides visualization of protein-ligand complexes annotated by the Uppsala Electron Density Server to examine electron-density fit and molecular interactions.
• Automation: Automates identification and processing of complexes with RSCC < 0.6 for downstream inspection.
• Annotations integration: Incorporates annotations from the Uppsala Electron Density Server to enrich interpretation of electron-density interpretation challenges.

Scientific Applications:

• Structural validation of ligand models: Enables detection and critical assessment of potentially biased or poorly modeled ligands based on RSCC.
• Drug design and lead refinement: Supports identification of ligand models requiring re-evaluation during drug-design and lead-optimization efforts.
• Analysis of electron-density interpretation biases: Facilitates study of how subjective interpretation of electron-density maps can affect ligand modeling.

Methodology:

Analyzes pre-filtered PDB entries using RSCC as the primary metric, automates identification and visualization of complexes with low RSCC, and integrates annotations from the Uppsala Electron Density Server.