You are here

relax with resfile

4 posts / 0 new
Last post
relax with resfile
#1

the relax protocol in 3.2.1 doesn't seem to be recognizing my resfile. I want to relax my structure as close to the native as possible (rotamers, backbone) and slowly allow more freedom.

relax.mpi.linuxgccrelease -database /home/gw/rosetta3.2.1/rosetta_database/ -s input.pdb -ex1 -ex2 -use_input_sc -extra_res_fa /home/gw/rosetta3.2.1/rosetta_database/chemical/residue_type_sets/fa_standard/residue_types/water/TP5.params -relax:fast -resfile natro.txt -nstruct 250 -overwrite

Here is the resfile, natro.txt:

NATAA
start
45 A NATRO #i45
46 A NATRO
47 A NATRO #n47
48 A NATRO
49 A NATRO #d49
51 A NATRO
53 A NATRO #m53
56 A NATRO
75 A NATRO
79 A NATRO #h79
112 A NATRO

1 B NATRO
2 B NATRO

I tested a point mutation resfile on fixbb and the resfile was recognized, but no point mutation appeared in the relax output structure.

Post Situation: 
Fri, 2011-08-05 10:57
gw

Relax doesn't read resfiles. I think there's a way to do relax while keeping some residues fixed, would that work? I think you use the RosettaScripts relax script, which lets you set certain residues to fixed in the MoveMap.

Fri, 2011-08-05 17:47
smlewis

Where can we find out how to do this? The FastRelax example in the RosettaScripts doc doesn't show how to fix residues at all.


<FastRelax name="&string" scorefxn=(score12 &string) repeats=(8 &int) task_operations=(&string, &string, &string >
<MoveMap>
<Chain number=(&integer) chi=(&bool) bb=(&bool)/>
<Jump number=(&integer) setting=(&bool)/>
<Span begin=(&integer) end=(&integer) chi=(&bool) bb=(&bool)/>
</MoveMap>

For instance, I have 3 identical chains in my protein (homotrimer), and I would like to relax a loop on all three monomer units. Let's say each chain has 400 residues, and I would like to keep all except residues 200-215 fixed. How could I use the above script to do this for each chain?

Many thanks,
Brett

Tue, 2011-08-09 13:01
brspurri

Here is how you would use the MoveMap in your example. I don't know how the MoveMap will react to each chain having the same numbering, e.g. if the MoveMap follows pdb_numbering or rosetta_numbering. You might have to renumber so each chain runs sequentially.

< FastRelax name="&string" scorefxn=(score12 &string) repeats=(8 &int) task_operations=(&string, &string, &string >
< MoveMap>
< Chain number=1 chi=0 bb=0 /> Turns of all degrees of freedom for chain A
< Chain number=2 chi=0 bb=0 /> Turns of all degrees of freedom for chain B
< Chain number=3 chi=0 bb=0 /> Turns of all degrees of freedom for chain C
< Jump number=1 setting=0/> Turns off minimization between chain A and B
< Jump number=2 setting=0/> Turns off minimization between chain B and C
< Span begin=200 end=215 chi=1 bb=1/> This line turns the side-chain (chi) and backbone (bb) freedoms back on
< Span begin=600 end=615 chi=1 bb=1/> If PDB had to be renumbered include this separately
< Span begin=1000 end=1015 chi=1 bb=1/> If PDB had to be renumbered include this separately
< /MoveMap>
< /FastRelax>

Also, try adding the -relax:constrain_relax_to_native_coords or -relax:constrain_relax_to_start_coords flag.

Including a constraints file for specific residues might also help. However, I'm not sure if that will be over-written by the relax protocol or if it even works at all.

HTH

Wed, 2011-08-10 11:11
nannemdp