You are here

Modelling at defined pH - for pmutscan - pH:mode not working

2 posts / 0 new
Last post
Modelling at defined pH - for pmutscan - pH:mode not working
#1

Hi all, 

I am new to Rosetta and using various Rosetta protocols, primarily remodel and pmutscan, to assess the impact of various mutations on protein stability. I would like to model the protein and associated mutations at acidic pH as this may affect salt bridge formation and overall REU score.

I have tried using the -pH_mode true and -value_pH 2.5 flags as suggested in other posts (https://www.rosettacommons.org/node/3711). However, these flags do not have any effect on the protonation states of side chains in output pdb structures or on the calculated energy values.

Does anybody know how I could make these flags work? Do they still work on the most recent versions of Rosetta? Are there any other ways to take pH into account in Rosetta?

Thank you for any help!

My command looks like this:

pmut_scan_parallel.static.macosclangrelease -s 1tie_0008_0002_0004.pdb -database database -output_mutant_structures -DDG_cutoff 999 -ignore_unrecognized_res -no_his_his_pairE -multi_cool_annealer 10 -pH_mode -value_pH 2.5 -mute basic core -ex1 -ex2 > .sc

 

Post Situation: 
Tue, 2022-03-15 04:05
cagfa1

The flags for the pH function are not working because not all score functions have the energy term. It can be made to work by adding an extra term "e_pH" to the score function. A command line option can be as follows:
 /path_to_rosetta/Rosetta/main/source/bin/relax.linuxgccrelease -overwrite -in:file:s name_of_pdb.pdb  -score:weights talaris2013 -set_weights e_pH 1.0  -corrections::restore_talaris_behavior true -pH_mode true -value_pH 4
This command line will just fix the score, one has to use relevant movers/protocols to enable the sampling of protonated states. Please note I am using talaris2013/2014 here. The current default Rosetta score function ref2015 uses optimized charge values for all atoms. However, the charge values for the protonated/deprotonated  AA in the database are non-optimized (specifically from CHARMM27). Therefore to keep everything in the same data frame  (in this case charges from CHARMM), it is better to use any score function which uses real charges, not the scaled ones. The protocols which can work with ref2015 and further developments are under progress and in personal branches. I am happy to discuss more and help further if needed.

--

Rituparna Samanta (Gray Lab)

Wed, 2022-03-23 11:21
Rituparna_Saman...