When I submit ROSIE docking2 jobs at http://rosie.rosettacommons.org/docking2/submit
there are options for local_docking and docking_local_refine docking protocols.
What is the difference between these options?
Where can I find documentation about them?
Documentation for ROSIE protocols can be found under the Documentation tab on the ROSIE website, or on the far right (paper and globe icon) of the page for the protocol. Here's a direct link to the Docking2 documentation: http://rosie.rosettacommons.org/docking2/documentation
Unfortunately, it doesn't look like the difference between the local_docking and docking_local_refine options are listed there.
local_docking is the full protocol described under point 4, which active sampling of the orientation of the two binding partners followed by high resolution refinement.
docking_local_refine skips the active translation/rotation sampling, and only does high resolution optimization and minimization of the starting structures. This is useful if you have a structure where the two partners are already in more or less the docked conformation, but where you want to optimize the sidechain conformations and do *slight* refinement of the rigid body orientations of the two partners.
Thanks for the info, rmoretti.
I have tried a few ROSIE docking2 docking_local_refine jobs and have noticed the following about them:
In each job's tar archive is an output\trigger-00001.dock folder containing the files proteins_0001.pdb to
proteins_0025.pdb, with just one copy of each. Nevertheless, the spreadsheet for each job often has multiple
rows with the same decoy proteins_00*.pdb file name listed, but not all of these rows list the same scores.
It is as if there were several versions of the proteins_0010.pdb file, for example. Also, sometimes under Results,
several of the models shown have the same proteins_00*.pdb file name. If you click on Model-* under Results,
a point on one of the plots, or a file name on the spreadsheet, it opens a pdb file. If you look at the bottom of
this pdb file, there are lines like below (here for proteins_0010.pdb from job 16568):
If you scroll up from the bottom of the pdb file, you can find lines like below:
# All scores below are weighted scores, not raw scores.
label fa_atr fa_rep fa_sol fa_elec fa_pair hbond_sr_bb hbond_lr_bb hbond_bb_sc hbond_sc dslf_ss_dst dslf_cs_ang dslf_ss_dih dslf_ca_dih fa_dun total
weights 0.338 0.044 0.242 0.026 0.164 0.245 0.245 0.245 0.245 0.5 2 5 5 0.036 NA
pose -239.252 5.68926 96.4564 -2.18161 -2.38554 -6.63757 -13.0112 -3.61172 -2.54153 0 0 0 0 7.78452 -159.691
These are for a structure with total_score -159.691, rms 0.0221701, and I_sc -11.3776. Sometimes these
numbers from the pdb file don't agree with the rest of the row on the spreadsheet or the location of the
point on the plot, which is confusing. It is also inconvenient when you want to get the structure for a certain
set of scores, but the only pdb file available with the desired name has a different set of scores. For example,
for job 16568, I would like to get the proteins_0010.pdb file with total_score -159.757, rms 0.177, and
I_sc -11.509. The plots and spreadsheet all show that this structure did exist at some point, it was just not
included in the tar archive for the job.
Can you fix the docking_local_refine protocol so that each structure generated by a job has its own unique
pdb file name, and all of these pdb files are included in the tar archive for the job?