A large decoy set of protein-protein complexes produced by flexible docking

reference: Guillaume Launay and Thomas Simonson

Journal of Computational Chemistry (2011) 32:106-120.
Contact: Thomas Simonson

A flexible docking procedure was used to produce a large ensemble of decoy complexes. For 202 homodimeric complexes and 41 heterodimeric complexes with known Xray structures, we produced an average of 1217 decoys each, for a total of 295,744 decoys. The links below lead to the full dataset, the heterodimer files, or the homodimer files. The decoys may be freely downloaded. Please acknowledge the above reference.

Each structure, say NNNN, is characterized by a series of physical parameters, reported in the file NNNN.inf. The columns in this file have the following meaning:

1    : decoy file number
2    : number of residue-residue contacts within the A monomer.
3    : decoy/native A monomer contact map overlap.
4    : number of residue-residue contacts within the B monomer.
5    : decoy/native B monomer contact map overlap.
6    : number of residue-residue contacts across the A:B interface
7    : decoy/native interface contact map overlap
8    : Fnat (total_interface_contact_decoy / total_interface_contact_native)
9    : Decoy Delta_ASA (Free_A_asa + Free_B_asa - AB_complex_asa)
10   : interface contact empirical score (6-class alphabet)
11   : interface contact empirical score (20-class alphabet)
12   :  Hydrophobic interface residue contact fraction (%)
13   :  Hydrophobic-Polar interface residue contact fraction (%)
14   : Polar interface residue contact fraction (%)
15   : Irmsd: rmsd of the mobile partner (B) after superimposition of the fixed one (A)

Complete dataset: one file per decoy series (243 in all), or 10 larger files which include multiple series.

Heterodimer dataset: one file per decoy series or one larger file including all the series.

Homodimer dataset: one file per decoy series or 9 larger files including multiple series.