I would like to perform sequence design of the peptide bound to the protein. I am using PyRosetta and PackRotamersMover function. I would like to ask if it is possible to use BINDING ENERGY of peptide instead of TOTAL ENERGY of complex as a criterion for accepting/rejecting given mutation during Monte Carlo simulation?
It’s very hard to do efficiently; there is some multistate design code in 3.2 (and a HUGE AWESOME new mode coming in 3.3 that solves the problem very efficiently and generically). I don’t know if any of this was wrapped for Pyrosetta, I would guess not.
The usual alternative is to re-rank results by binding energy after searching based on total energy. Results poor in total energy but great in binding energy are suspicious, so this is not a poor solution.
If you are redesigning a small peptide, your search space may be small enough to brute-force search it? How big is the sequence space you want to examine?
Score12 is the standard that most of us use for design (the Kuhlman lab, anyway). Check out the Top7 paper to see what it’s made of.
Science. 2003 Nov 21;302(5649):1364-8.
Design of a novel globular protein fold with atomic-level accuracy.
Kuhlman B, Dantas G, Ireton GC, Varani G, Stoddard BL, Baker D.
The community has had an internal focus on improving the scorefunction recently, but nobody’s published their results yet. I know the -correct flag is supposed to make things better, but I don’t know how or if it works with pyrosetta (or if it’s been tested with design as opposed to abinitio).