Pharmacophore searching: A potential solution for correcting unknown ligands (UNK) labelling errors in Protein Data...

Area of Science:

• Structural Biology
• Biochemistry
• Computational Chemistry

Background:

• The Protein Data Bank (PDB) is a crucial resource for biological structural data.
• Ensuring the accuracy of PDB data, particularly ligand identification, is vital for research.
• Mislabeled ligands, such as 'Unknown' (UNK), can hinder downstream analyses.

Purpose of the Study:

• To identify and correct mislabeled ligands in the Protein Data Bank.
• To evaluate the efficacy of pharmacophore searching and electron density analysis for ligand identification.
• To improve the reliability of structural data in the PDB.

Main Methods:

• Generated pharmacophores for 292 protein structures using Discovery Studio Visualizer and Accelrys Catalyst.
• Performed database searches with generated pharmacophores against reported ligands.
• Validated potential ligand fits against observed electron density maps using Coot.

Main Results:

• Pharmacophore searching identified suitable known ligands that fit electron density maps better than PDB-reported ligands.
• At least 20 out of 292 examined PDB structures (as of May 2016) contained ligands with clear electron density but were incorrectly labeled as UNK.
• The study confirmed the utility of pharmacophore searching and Coot for accurate ligand identification.

Conclusions:

• Pharmacophore searching and Coot analysis are effective tools for identifying and correcting mislabeled ligands in the PDB.
• These methods enhance the accuracy and reliability of structural data by ensuring correct ligand annotation.
• Accurate ligand identification is essential for advancing research in structural biology and drug discovery.