Member Site › Forums › Rosetta 3 › Rosetta 3 – General › No RMS and IRMS values in protein/nucleic acid docking
- This topic has 2 replies, 2 voices, and was last updated 8 years, 9 months ago by Anonymous.
-
AuthorPosts
-
-
January 13, 2016 at 8:41 pm #2366Anonymous
Hello,
I am docking a small protein / nucleic acid structure to generate dockings. The scoring contains only -nan for rms and 0.000 for Irms. Why doesn’t the docking result in valid rms/irms values? Fnat and I_sc as well as total_score seem ok. What do I have to change in order to receive rms and irms values?
XML Rosetta Script (prot2.xml):
<ROSETTASCRIPTS>
<SCOREFXNS />
<TASKOPERATIONS />
<FILTERS>
<IRmsd name=”irmds” confidence=0/>
</FILTERS>
<LIGAND_AREAS>
<inhibitor_dock_sc chain=”C” cutoff=”6.0″ add_nbr_radius=”true” all_atom_mode=”true”/>
<inhibitor_final_sc chain=”C” cutoff=”6.0″ add_nbr_radius=”true” all_atom_mode=”true”/>
</LIGAND_AREAS>
<INTERFACE_BUILDERS>
<side_chain_for_docking ligand_areas=”inhibitor_dock_sc”/>
<side_chain_for_final ligand_areas=”inhibitor_final_sc”/>
</INTERFACE_BUILDERS>
<MOVEMAP_BUILDERS>
<docking sc_interface=”side_chain_for_docking” minimize_water=”false”/>
<final sc_interface=”side_chain_for_final” minimize_water=”false”/>
</MOVEMAP_BUILDERS>
<MOVERS>
<RigidBodyTransMover name=”fix” jump=”1″ distance=”15″ x=”0″ y=”0″ z=”0″ />
<Translate name=”trans” chain=”C” distribution=”uniform” angstroms=”5.0″ cycles=”1000″/>
<Rotate name=”rotate” chain=”C” distribution=”uniform” degrees=”360″ cycles=”1000″/>
<SlideTogether name=”together” chains=”C”/>
<HighResDocker name=”highdock” scorefxn=”score_docking” cycles=”9″ repack_every_Nth=”3″ movemap_builder=”docking” />
<FinalMinimizer name=”mimimize” scorefxn=”score_docking” movemap_builder=”final”/>
<ParsedProtocol name=”low_res_dock”>
<!–<Add mover_name=”fix”/>–>
<Add mover_name=”trans”/>
<Add mover_name=”rotate”/>
<Add mover_name=”together”/>
</ParsedProtocol>
<ParsedProtocol name=”high_res_dock”>
<Add mover_name=”highdock”/>
<Add mover_name=”mimimize”/>
</ParsedProtocol>
<InterfaceAnalyzerMover name=”iface_analyzer” packstat=0 pack_input=0 pack_separated=1 fixedchains=A tracer=0/>
<Docking name=”dock_high” score_high=”score_docking” fullatom=1 local_refine=1 optimize_fold_tree=1 conserve_foldtree=0 design=0 jumps=1/>
<DockingProtocol name=”dockp_high” docking_local_refine=”true” dock_min=”false” partners=”C_A” />
</MOVERS>
<APPLY_TO_POSE />
<PROTOCOLS>
<Add mover_name=”low_res_dock”/>
<Add mover_name=”high_res_dock”/>
<Add mover_name=”dockp_high”/>
<!– <Add filter_name=”irmds”/>–>
<!–<Add mover_name=”iface_analyzer”/>–>
<!–<Add mover_name=”add_scores”/>–>
</PROTOCOLS>
</ROSETTASCRIPTS>
Command line:
DIR/rosetta_scripts.RELEASE -native FILENAME -s FILENAME -use_input_sc -nstruct 10 -ex1 -ex2aro -parser:protocol prot2.xml -overwrite
Scores.sc output:
SEQUENCE:
SCORE: total_score rms Fnat I_sc Irms …
SCORE: -91.285 -nan 0.833 -1.473 0.000 …
SCORE: -99.662 -nan 0.778 -0.859 0.000 …
SCORE: -105.361 -nan 1.000 -1.271 0.000 …
SCORE: -79.505 -nan 1.000 -0.333 0.000 …
SCORE: -106.792 -nan 0.667 -0.158 0.000 …
…
-
January 14, 2016 at 4:05 pm #11401Anonymous
I would hazard a guess that the Irms and rms calculations are based on protein backbone c-alphas. Those atoms aren’t present in the moving side of your structure so it can’t be calculated. You’d have to go into the C++ and rewrite the calls to the RMS functions to use versions that measure some other atom set. I’ll take a look and see if it’s straightforward.
-
January 14, 2016 at 4:19 pm #11402Anonymous
OK, here’s a thing you can try. I’m assuming you’re comfortable editing code.
Open source/src/protocols/docking/metrics.cc. Search for the function “calc_Lrmsd” (which controls rmsd) and function “calc_Irmsd”. At the end of both of these functions is a line like:
Lrmsd += core::scoring::rmsd_no_super_subset( native_pose, pose, superpos_partner, is_protein_backbone );
Irmsd += core::scoring::rmsd_with_super_subset( native_docking_pose, pose, is_interface, is_protein_backbone );
In both cases, change “is_protein_backbone” to “is_heavyatom” (You will notice for Irmsd there is a nearby commented-out line just like this). Try recompiling (scons.py bin mode=release, like when you first compiled, but it won’t take nearly as long) then re-running your experiment.
-
-
AuthorPosts
- You must be logged in to reply to this topic.