Adding ligands into the SCOPe pdbs

The SCOPe database classifies the structures from the PDB database according to their structure and topological connections hierarchically. This process occurs via a combination of automation and manual curation, and in doing so the heteroatoms from the PDB structures are discarded.
We wanted to analyze our fragments also in the context of their functions. To that end, we needed to include information on how the fragments interact with biologically relevant molecules such as cofactors, ions, metals, or drugs. We mapped each of the domains in SCOPe to its corresponding PDB and retrieved those non-protein molecules that are within 4 Å of any domain atom. This way, we preserve the domain definition from SCOPe and include the ligand information from the original PDB.
Domains in SCOPe (such as the protein in white) were mapped to their corresponding PDBs. All HETATM records within 4 Å of any protein heavy atoms where included

Searching ligands in Fuzzle

PDB search

Ligands can be searched using their PDB identifier. Each ligand that appears in any PDB X-ray structure presents a unique three-letter code identifier (or two and one-letter identifier for ions and atoms), such as HEM (Heme group). The output will be a table with the representative domains that bind that ligand. It is also possible to retrieve all domains that bind this ligand, not only those that are representatives of their clusters in SCOPe95.

SMILE search

Ligands can also be searched via their SMILE. SMILES are representations of ligand structures in the form of a line notation. They uniquely describe a ligand and can be reverted to their 2D structures, or can be used to quickly search similar compounds.

Examples of superstructure and substructures for ligand ATP
The search will always provide a list of domains binding these compounds. Therefore PDBs that do not appear in SCOPe will not be listed, and neither ligands that do not bind SCOPe ligands are searchable.