When running fixbb design runs I tend to find I get a very large number of long charged sidechains such as Arg, Lys etc on the surface.
I would like to manually adjust the reference energies for these residues in order to increase the penalty for these problematic residues. I assume this is possible but haven’t been able to figure out which of parameters or weight files I need to change.
The reference energies are stored in the file rosetta_database/scoring/weights/standard.wts. If you’re not using the standard scorefunction they’ll be somewhere else in the same folder. The first line is “METHOD_WEIGHTS ref” and then 20 numbers – these are the 20 reference weights for the residues. They’re in order by single letter code (ACDEFG…).
Has anyone in the Baker group or in the wider Rosetta community parameterized a set of weights specifically for surface residues redesign?
It would potentially be very helpful as I don’t really like the surface residues you get when redesigning with the standard weights and suspect you’d get a lot of solubility problems – I often get a load of Phe, Tyr, and Trp stacking together on the surface.
Yes, someone is working on exactly this. He’s in the middle of writing up a paper on it and hopes to release the weight sets soon – but the code and weights haven’t been released yet.
You can use the POLAR command in resfiles at surface positions to restrict the residue set to only polar residues as a temporary solution?
that is excellent news – looking forward to seeing it!
I tried using POLAR and PIKAA etc but it still doesn’t really solve the problem as you still get very odd surface residue compositions so that’s why I wanted to adjust the ref energies manually.
.
.