Classification of ligand molecules in PDB with fast heuristic graph match algorithm COMPLIG

Area of Science:

• Biochemistry and structural biology
• Computational chemistry
• Bioinformatics

Background:

• Small-molecule ligands in the Protein Data Bank (PDB) lack a standardized structural classification.
• Accurate classification is crucial for understanding protein-ligand interactions and drug discovery.

Purpose of the Study:

• To develop and apply a novel algorithm, COMPLIG, for classifying small-molecule ligands based on their structure.
• To establish optimal parameters for ligand categorization within the PDB.

Main Methods:

• A fast heuristic graph-matching algorithm, COMPLIG, was developed for concurrent classification of proteins and ligands.
• Ligands were clustered based on a criterion of >60% identity of atoms and bonds.
• The system was applied to assess the natural ligand status of human proteins.

Main Results:

• 11,585 ligand types were classified into 1946 distinct clusters.
• Large clusters predominantly contained nucleotides and amino acids, with a notable presence of drug compounds.
• At most, 37% of human protein structures were found in complex with natural ligands in the database.

Conclusions:

• The COMPLIG algorithm provides a robust structural classification for small-molecule ligands.
• Protein homology and ligand similarity modeling suggest potential natural interactions for an additional 28% of human proteins.
• This work enhances understanding of intrinsic protein-metabolite interactions and aids drug discovery efforts.