I am trying to fix the clashes in a homology model structure I have created of a membrane protein. I tried fixbb to fix only the sidechains, yet I still get clashes ):. Therefore, I want to try to use the relax application in rosetta3.1 to remove clashes. My questions are:
1. Is it O.K to use relax for a membrane protein? If so, what flags should I use to indicate in relax that I want the calculations done on a membrane protein (for instance, would adding -score:patch no_solvation be enough?)
2. How can I make sure that my final result will be clashes free yet not so far structurally from the original input structure?
I can't say that relax HAS been used in this manner before, but I expect it ought to work. Tweaking the scorefunction as you describe is probably sufficient for your needs for a membrane protein.
There is a flag:
-relax:constrain_relax_to_start_coords Add coordinate constraints to backbone heavy atoms, based on the input structure.
which probably will satisfy your second question.
Thank you very much,