=======================================================================================================================
HowTo: global blind docking prediction
=======================================================================================================================
Here are three basic steps for docking:

1. cat the pair of proteins into one pdb file, say 1xxx.pdb, with each protein ending with a "TER". Put the larger protein first since it is the one which will fixed during docking.

2. prepacking: rosetta.exe aa 1xxx 1 -dock -prepack_full -prepack_rtmin -s 1xxx.pdb -ex1 -ex2aro_only -find_disulf -unboundrot.

This command line will separate two docking partners and refine their sidechains individually. Be sure to make a copy of 1xxx.pdb as 1xxx.unbound.pdb and have it location specified in "pdb2" in paths.txt. As other normal Rosetta runs, location of 1xxx.pdb should be specified in "pdb1" and "starting structure" lines. The run will result in a model called "1xxx.ppk.pdb", and this is the one used as the starting structure for docking.

3. docking: rosetta.exe bb 1xxx 1 -dock -dock_mcm -randomize1 -randomize2 -s 1xxx.ppk.pdb -ex1 -ex2aro_only -find_disulf -unboundrot -dock_rtmin -nstruct 5000 -norepack_disulf -fake_native -docking_score_norepack.

This command line performs full atom global docking. replace "-randomize1 -randomize2" by "-dock_pert 3 8 8" if you only need to perform a local perturbation docking. For global docking, we usually generate 5000 structures first and take top 1% score cutoff. Then repeat with the same command line by adding "-scorefilter <cutoff>". This will only output structures which have lower energies than the cutoff. We usually generate about 500~1000 structures (i.e., -nstruct 500) and those structures will be subjected to clustering analysis.

=======================================================================================================================
Here are four basic steps for clustering: clustering scripts can be found in rosetta_release/rosetta_scripts/docking/

1. organized your directory like "1xxx/aa/*.pdb" and "1xxx/aa1xxx.fasc". "1xxx" need to be replaced by your pdb name and "aa" needs to replaced by the two letter series code you used in your docking runs. If you have multiple series codes/runs, such as aa, bb, cc..., put them all under "1xxx" directory.

2. outside "1xxx" directory, run "pp_pdb.sh 1xxx". You should have "scorefiles", "top10" and "top200" created in "1xxx" dir.

3. Copy those directories to a machine where R software is installed, still under "1xxx" dir.

4. outside "1xxx", run "pp_pdb2.sh 1xxx". The clustering should start and in top10 and top200, you should have files regarding clustering information.

Since pp_pdb.sh and pp_pdb2.sh will call other children scripts in the tar ball, you may need to edit your PATH env variable properly. You can also add clustering cutoff value as the second argument for pp_pdb2.sh. The default we use is 2.5A.
=======================================================================================================================
