You are here

defining interface in interface analyzer

6 posts / 0 new
Last post
defining interface in interface analyzer

How does interface analyzer identify the residues at the interface and can it be restricted using resfile? I mean can I define the residues that make the interface. What would be an appropriate syntax for such a resfile?

For example, in a 6 chain (ABCDEF) PDB containing 174 aa, how do I define the interface as being between residues 8A and 8D and calculate the output results based on just these two residues.



Post Situation: 
Fri, 2013-09-27 11:08

InterfaceAnalyzer uses InterfaceNeighborDefinitionCalculator, which uses very simple "does this residue have a neighbor within X angstroms on a different chain" to determine if the residue is in the interface. X is -pose_metrics::interface_cutoff, which defaults to 10 angstroms.

You can't "tell it" what's in the interface - the point of InterfaceAnalyzer is to TELL YOU what is in the interface. If you want it to report with a more aggressive definition of interface, I guess you can push the interface_cutoff up, but since most of the scorefunction terms taper off beyond that distance, it won't have huge effects.

What does "calculate on just those residues" mean? Do you want just the interaction energies of those two residues? InterfaceAnalyzer can't do that. The fastest way to estimate this would be to copy JUST those two residues to their own PDB file and score it; it will be "wrong" for the burial-dependent scorefunction terms but good for VDW, at least. Calculating residue pair energies is common in Rosetta but it's not common for it to be "exposed" to the user interface, so I'm not sure there's a great commandline solution.

Fri, 2013-09-27 11:24

I have a series of PDBs that contain different combinations of residues ALIV at these two positions. The burial-dependent terms from other interfacial residues make it difficult to isolate the effect of mutations at just these two positions. Anyway, thank you very much for your help.

Fri, 2013-09-27 13:08

It sounds like you should just run all of them through InterfaceAnalyzer, then. You'll get binding energies, etc, for each pair. You won't get per-residue energies for your pair, bu you'll have delta-delta-Gs of mutation/binding between pairs, which is what you really need anyway, I think. They'll all have a large constant contribution from the unchanged positions, which is fine because it factors out when you rank-order or do DDGs.

Fri, 2013-09-27 14:14

That worked! Thanks a lot. When looking at the results, I was focusing entirely on packstat which obviously didn't change much from one pdb to another. While I am at it, I was wondering if there is a way to define two groups of chains using -fixedchain flag. I have a PDB with six chains A, B, C, D, E and F and I want to define the interface as being between ABC and DEF so that the interface within chains A, B and C or D, E and F are ignored.
thanks for your help!

Sat, 2013-09-28 12:37

Having not used the -fixedchains myself, it looks like the syntax should be:

-fixedchains (string) - which chains are in the two groups to define the interface? example: -fixedchains A B to keep chains A and B together, and C separate, out of a pose that contains A, B, and C. Note a space between A and B. Not tested beyond three chains.

So, in your case, -fixedchains A B C (or, equivalently, -fixedchains D E F), ought to work.

Sun, 2013-09-29 09:40