You are here

Rosetta repacking output

6 posts / 0 new
Last post
Rosetta repacking output
#1

Hi,
I've used rosetta3.1 for model repacking using energy minimization, my command line was:
score.linuxgccrelease -database /share/apps/rosetta/rosetta_database/ -s test.pdb -score_app:linmin -score:weights standard -score:patch no_solvation -in:file:repair_sidechains -out:output

After this procedure, I discovered that one Asp in my model (just one of the Aspartic acids in the model, the rest are fine as far as I've noticed) has a covalent bond between its backbone carbonyl and its side chain carbonyl, creating a ring. That is weird, can anyone explain to me why this happens?

Thank you,
Doran

Post Situation: 
Mon, 2012-12-03 03:44
doranhen

Rosetta does not think this is a covalent bond, Rosetta thinks this is a bad clash. Probably your molecular viewer (pymol?) detects that two atoms are too close, but misinterprets their distance as a bond rather than a clash.

You can verify that it's a clash by looking at the per-residue energies that may be present in your output PDB; that ASP should have a huge fa_rep score (possibly fa_intra_rep).

The score app is not suggested for doing model preparation; use fixbb for repacking (fixbb -s test.pdb -packing:repack_only -ex1 -ex2 -extrachi_cutoff 0 -use_input_sc -ndruns 10). Newer versions of Rosetta have further repacking options available in the fixbb executable (either pre- or post-minimized sidechains).

You can then rescore this properly prepared PDB with a separate scorefunction via the score app the way you tried. (You can also pass those -score:weight and -score:patch flags to fixbb, but I think using no_solvation is not a good idea for repacking).

Mon, 2012-12-03 07:24
smlewis

Hi,
Thank you very much for your elaborated reply.
I have two more questions:
1.Is fixbb suitable for membrane proteins?
2. If the answer to my first question is yes, then since using the no_solvation patch is not a good idea for repacking, what flag should I add to the fixbb command to make it suitable for transmembrane proteins?

Thank you,
Doran

Mon, 2012-12-03 08:35
doranhen

A good question! Use those scorefunction flags as-is with fixbb. Fixbb works fine for membrane proteins (I guarantee the membrane code is using the packer, which is all fixbb does), it's just a question of what scorefunction to use. Using no_solvation will be okay in this case.

Mon, 2012-12-03 08:39
smlewis

Hi,
Thank you again for your reply.
Just making sure I understand your answer correctly:
Did you mean that I should use the Fixbb with the flags you suggested (fixbb -s test.pdb -packing:repack_only -ex1 -ex2 -extrachi_cutoff 0 -use_input_sc -ndruns 10) and that it will work fine as is for membrane proteins and that I should use the no_solvation afterwards for scoring purposes?

Thank you,
Doran

Mon, 2012-12-03 08:53
doranhen

Hi,
Sorry for my previous reply, I misread your answer, now I get it, you were perfectly clear.

Thank you,
Doran

Mon, 2012-12-03 08:56
doranhen