OVERVIEW
The dataset is composed of 2376 structures of ligand/protein complex. The ligand is either a drug-like ligand or a fragment.
The dataset describes 64 different proteins. A complex is represented by 2 MOL2 files:
	* ligand coordinates (suffix '_ligand')
	* protein coordinates (suffix '_protein')

A complex is identified as follow:
	<PDB id>_<Biological unit>_<Ligand_index>

Ligand index distinguishes the different compounds bound to the protein site in the same biounit of a PDB file.

DIRECTORY ARCHITECTURE

fragdock (root directory)
    |
    +--> "representative_protein_structures.tsv"
    |    description:    List of the proteins representative structures.
    |                   
    +--> <Protein 1>
    |    folded name:     Uniprot accession code 
    |    
    |
    +--> <Protein 2>
    |
    :
    :
    |
    +--> <Protein N>
             |
             +--> "druglike"
             |    content:    coordinates files 
             |
             +--> "fragment"
             |    content:    coordinates files
             |
             +--> "info.tsv"
                  content:    list of annotated structures
