I am trying to come up with a protocol (xml script) that will allow me to perform positive design for one substrate and negative design for the inhibitor. Our goal is to create a protein that maintains functionality and is resistant to the inhibitor.
I am open to any suggestion. I have not been able to find any documentation on how to perform docking_design in this manner.
You want multistate design. There isn't really a way to get Rosetta to do negative design on its own except via the multistate code. There is a demo for this (https://www.rosettacommons.org/demos/latest/demos-by-category#protocol_captures, but you need the input files that are in the demos directory of your copy of Rosetta). The paper is https://www.ncbi.nlm.nih.gov/pubmed/24463572; the supplemental has a description of using multistate design.
Thanks for the reply. I have looked into that documentation and I'll give it a try. My one concern is how to model the best "negative state?" Is there a way to via filter or some other means to get Rosetta to prefer mutations that make the interactions with the small molecule worse? If not what would you recommend as the best way to achieve that portion of it.
From what I understand multistate design only works when you have good examples of each state. So how do you set up the worsening of interactions for my "negative state?"
I hope I am making myself clear.
There isn't a good way to directly look for bad mutations. Clearly you can make structures arbitarily bad by introducing clashes but it's not physical.
Multistate works by letting you define a "fitness function". In the inner loops the packer and minimizer work as normal, looking for good conformations. In an outer loop, the fitness function can consider the broader experimental concern. Specifically you set up a fitness function that favors the gap between the two states being large: you tell it to find a sequence that improves the energy of the "good state" and makes the energy of the "bad state" as bad as possible. The inner loop will try to keep the "bad state" honest (so that its badness is due to irresolvable problems not simple clashes) but as the documentation explains you do want to be careful ensuring your negative states are still as good as they can be.
You do need starting conformations for both states.