Ab-initio protocol

Hello,
I’m trying to do protein (1008 residues) and antibody (1300 residues) docking in three different ways, in order to answer t different biological questions:

1- dock the to proteins putting as active both the epitope and the paratope
2- dock the to protein putting as active all the CDR of the antibody and passive the accessible surface of the protein
3- dock the to proteins without any active residue: and so I have to choose between random AIRs, surface restrains and central mass restrains.

in the 2 and 3, I also changed those parameters:
Random exclusion=1.1111
Number of structures for rigid body docking =10000
Number of structures for the final refinement=1000
Number of structures to analyze =1000
Full or limited analysis of results? Full

My question is in particular on the 3 protocol: so I want to sample all the possible docking pose taking in account the entire surface of both the two proteins. I know that the space to sampling is very huge, but I’m trying to do a protocol that could be the better one to do that.
I’m not understanding which ambiguous restrains is better to chose between random AIRs, surface restrains and central mass restrains, or use a combination of them.

I’m using the Guru profile on the web server. Thank you!

Hi,

first, let me say that using the whole antibody structure for docking purposes is not optimal, it’s better to focus on the variable region (combination of light and heavy chain variable regions), thus “cutting away” the rest of the antibody.
As for your question, I would not consider the 3 protocol: you always know that the antibody will bind using the CDR loops, so the minimum level of information would be protocol 2.
If you really want to perform a protocol 3, I would not know which is the best option between the three, but probably center-of-mass restraints, as surface and random restraints won’t have much effect given the huge space to be sampled.
to wrap up, cut the antibody and forget about protocol 3 :slight_smile:

Thank you for the quick answer,
I understand well your points, but actually, we want to see if the antibody can do nonspecific interactions (so non with the CDR) also with other parts of the proteins (so non in the epitope), this is why we want to do the protocol 3.

Can you explain better why you would prefer the center of mass restrains?

Thank you for the quick answer,
I understand well your points, but actually, we want to see if the antibody can do nonspecific interactions (so non with the CDR) also with other parts of the proteins (so non in the epitope), this is why we want to do the protocol 3.

You will always get models in other location is my guess. Experiment 3 won’t prove anything…

Also the full analysis will be very time-consuming with refining 1000 models… I would not do it.
What kind of additional info do you expect to get compared to clustering only?

the idea was to get enough docking poses to be able to do statistical analysis and see which part of the antibody surface interacts most with which part of the protein surface.
I, too, have a strong doubt that protocol 3 will work, but we wanted to try the best possible protocol.

My gut feeling is that COM restraints might work better than random and surface restraints, as the surface of both molecules is big thus giving rise to a huge combinatorial complexity for a random/surface sampling…but it’s just a gut feeling, I wouldn’t do any of the three in this case

Ok, thank you for the explanation. I understand your suggestions, so I will no do the protocol 3

Thank you!

1 Like