AutoDock CrankPep

AutoDock CrankPep or ADCP is an AutoDock docking engine specialized for docking peptides. 
It combines technology form the protein folding filed with an efficient representation of
a rigid receptor as affinity grids to fold the peptide in the context of the energy landscape
created by the receptor. A Monte-Carlo search is used to fold the peptide while concurrently
optimizing the interaction between the peptide and the receptor molecule, yielding docked 
peptides. The program can dock peptides provides as 3D structures in PDB files or from a 
sequence string. It has been show to successfully re-dock peptides with up to 20 amino acids in length.
 
ADCP is developed based on CRANKITE.
Podtelezhnikov,A.A. and Wild,D.L. (2008) Source Code Biol. Med., 3, 12.

ADCP is available under the GNU LGPL v2.0 OpenSource license.
Please visit adcp.scripps.edu for more details.

There are two ways of using ADCP. The first and recommended way is to use the prepared Python wrapper runADCP.py.
runADCP works directly with a zipped target file (.trg) prepared by AGFR.
Please see adcp.scripps.edu/tutorial for more details.

The other way is to use the compiled binary directly.
Note this approach required unzipped map files within the excuation folder
example usage:

adcp_Linux-x86_64 -r 10000x3000 -t 2 apgvgvapgvgv -p Bias=NULL,external=5,con8,2,1.0,external2=4,con8,2,1.0,Opt=1,0.5,0.5,-0.5 -o output.pdb

external=5,con8,2,1.0
This calls the autodock grid maps. It will look for rigidReceptor.*.map. 
con8 is the a file indicating the residues to score, usually just all residue number.
2 is a thermo factor that is an addition to the original temperature.
1.0 calls the reconstruction of side chain during the scoring.

external2=4,con8,2,1.0 
This calls the cyclic procedure to create an artificial peptide bond between first and last residue.
con8 is meaningless right now, but can be used to indicate the location of the cyclic bond in the future.
2 is the thermo factor, and it will overwrite the previous one in external.
1.0 is meaningless right now, but can be used to flag the cyclic bond type, etc.

Opt=1,0.5,0.5,-0.5
This calls the optimizing/docking procedure. Opt=0 is to use regular MC.
The following 3 numbers is the weight of the target energy in the optimizing procedure.
In the example above:
targetE = 0.5 * totalE + 0.5 * externalE + (-0.5) * firstlastE
Note that totalE includes internalE, externalE and firstlastE. So the above weights scale down the internal energy by a factor of 2 and remove the energy between the first and last residues. The default weights are set to be 1, 0, 0.