You are here

Dock antibody at the epitope of antigen

4 posts / 0 new
Last post
Dock antibody at the epitope of antigen
#1

Deal all:

I am a beginner of computational protein docking. For my current project, I am going to dock Ab at Ag using PyRosseta.
Is there any option or parameters that I can constrain the region (epitope) of Ag to be docked by Ab? I would like to extract the value of scoring function based on this constrain (epitope).

Please having your comments or suggestions regarding how to do that.
Thanks so much,

Hsih-Te

Post Situation: 
Fri, 2012-06-22 05:44
styang

I can't offer advice for pyrosetta specifically.

There are two major methods for restricting docking. The first is to do non-aggressive docking. Make sure that your starting structure is relatively close to the structure you think is real, and perform docking in such a way that it samples relatively gently. In particular, do not use the spin or randomize options for docking and it will be gentler. Using the local_refine_only mode is gentler still, and minimization only will make the least change.

The second is explicit score-based constraints, where you tell the scorefunction what interactions you think exist and it evaluates models based on the existence of those constraints.

http://www.rosettacommons.org/manuals/archive/rosetta3.4_user_guide/de/d...

Fri, 2012-06-22 08:40
smlewis

Thanks for your kind reply. This is really helpfule to me.

To explain clearly, the epitopes I have are predicted from computer software. The epitope information only include linear type with secessive A.A. in length of 8-20 or discontinuos type by serveral single A.A.s with coressponseing position.
For example,
(1)linear epitope:
DRFKHLKTEAEMKASEDLKKH starting from position 44 to position 64
(2)Discontinuos eiptope:
_:R45, _:H48, _:L49, _:K50, _:T51, _:E52, _:A53, _:E54, _:M55, _:K56, _:A57, _:S58, _:E59, _:D60, _:L61, _:K63, _:H64

In this case, is it possible to incoporate such score-based constraints for scoring function?
Thank you again,
Hsih-Te

Fri, 2012-06-22 13:49
styang

It sounds like what you have is a vague idea of which resdues you want in the interface, but no precise idea of what residues ought to interact. This is a common problem and we really should have an "InterfaceConstraint" to handle it, but we don't.

The problem is that Rosetta's constraint machinery is focused on residue pairs (because it is pairwise-decomposable for rotamer packing, like much of the scorefunction). To do non-pairwise work like this, it's best to break it up into the many possible residue pairs, arranged in an "Ambiguous" constraint so that only one of each group of possible pairs of interactions across the interface has to work out.

Read about ambiguous constraints here: http://www.rosettacommons.org/manuals/archive/rosetta3.4_user_guide/de/d...

Let's say that the residue range for interaction with your epitope is 100-150. You'll make all possible residue pairs between your epitope 44-64 and the interacting region 100-150, like so:

AmbiguousConstraint
AtomPair CA 44 CA 100 [[constraint function]]
AtomPair CA 45 CA 100 [[constraint function]]
AtomPair CA 46 CA 100 [[constraint function]]
...
AtomPair CA 62 CA 100 [[constraint function]]
AtomPair CA 63 CA 100 [[constraint function]]
AtomPair CA 64 CA 100 [[constraint function]]
AtomPair CA 1 CA 2 CONSTANTFUNC 10 # capping constraint
END_AMBIGUOUS

AmbiguousConstraint
AtomPair CA 44 CA 101 [[constraint function]]
AtomPair CA 45 CA 102 [[constraint function]]
AtomPair CA 46 CA 103 [[constraint function]]
...
AtomPair CA 62 CA 148 [[constraint function]]
AtomPair CA 63 CA 149 [[constraint function]]
AtomPair CA 64 CA 150 [[constraint function]]
AtomPair CA 1 CA 2 CONSTANTFUNC 10 # capping constraint
END_AMBIGUOUS

I don't know how to pass these in to pyrosetta off the top of my head, but I can dig up the C++ if you need it.

Sun, 2012-06-24 12:32
smlewis