You are here

Camelid antibody (VHH) modelling

14 posts / 0 new
Last post
Camelid antibody (VHH) modelling

Im modelling a camelid antibody (vhh) amino acid sequence using Rosetta (rosetta_src_2016.32.58837_bundle). I cannot generate these input files : "FR02_H.pdb", "FR02.cst",  "FR02.cst_fa". Please tell me which 'bin' sould I use to get this inputs for "camelid antibody graffeting".  Thanks.

Post Situation: 
Tue, 2017-04-04 22:52
Ayoub Ksouri

I think there have been some issues with camelid antibodies; you might need to try a more recent release.  I've pinged the antibody group for more details.

Wed, 2017-04-05 07:01

Yes, there has been a few issues within the past year.  First, please try a recent weekly release.  If you still have trouble, please post your full command line and log output so that we may debug the issue.  Thanks.



Wed, 2017-04-05 10:14

Thank you for this instructions, I tired the  last weekly release "rosetta_src_2017.08.59291_bundle", same thing, I still not able to generate the inputs data (which is necessary  to run the command)   for the bin: "antibody_graft.default.linuxgccrelease"  

- Inputs data : .../Rosetta/rosetta_src_2017.08.59291_bundle/main/tests/integration/tests/antibody_H3_camelid/inputs
"aaFR02_03_05.200_v1_3"  "aaFR02_09_05.200_v1_3"  "FR02.cst"  "FR02.cst_fa"  "FR02_H.pdb"

Fri, 2017-04-07 06:59
Ayoub Ksouri

So by "not able to generate the inputs data" - you mean you don't have a problem running Rosetta, but instead you don't know what the inputs are supposed to be?

Have you read the documentation?

The "aa....." files are 3mer and 9mer fragments.  The .cst and .cst_fa files are constraints.  The last input is a PDB - I don't know the protocol specifically, but if you are grafting, I assume it is a skeleton to graft into.

Fri, 2017-04-07 08:29

I think it's not a matter of reading doumentation. 

In the documentation a "" file is required to generate all this files ( "aaFR02_03_05.200_v1_3"  "aaFR02_09_05.200_v1_3"  "FR02.cst"  "FR02.cst_fa"  "FR02_H.pdb"), the problem is that in the last version of Rosetta (3.7 & 3.8)  this script file  does not even exist.
Thank you.

Fri, 2017-04-07 09:13
Ayoub Ksouri

I have never heard of that script.  Can you give a link to where it is mentioned?  For the most recent documentation (which does NOT include the antibody_graft application), please see this paper: 

Fri, 2017-04-07 10:48
jadolfbr - great, that's a specific problem, I've farmed it out to the antibody group.  (Your original question was too vague for us to know how to help.)

Fri, 2017-04-07 09:50

Ok Thank you smlewis and jadolfbr

Sat, 2017-04-08 05:56
Ayoub Ksouri

Please I'm still waiting for the intervention of the "Rosetta antibody group";


Please how can I generate this inputs (see path) ??

Fri, 2017-04-14 04:00
Ayoub Ksouri

Hi Ayoub.  I am a part of the antibody modeling team and was on that paper I pointed to you eariler.  The code you are referring to is deprecated and unsupported.  Please use the application that is in the paper.  If you have issues with this, we will be happy to help and fix whatever needs to be fixed. 


Fri, 2017-04-14 11:52

Dear Jared

I tried to run the protocol described in this paper:, the problem is that the application is dedicated specially for classical antibodies (antibody with heavy and light chain), In my case Im working only with heavy chain antibodies (lacking light chain "VHH" or "Nanobodies"), so when I run the application, I have this error ;

ERROR: Error can't find section 'light' in fasta file /home/ksouri/antibody_example/antibody.fasta!

Thank you



Mon, 2017-04-17 07:50
Ayoub Ksouri

Sounds good.  I'll forward this to Sergey.  In the meantime, try passing an arbitrary fasta for the light chain, and then removing it from the PDB after  The core modeling is done afterward, so this will work fine.  Both antibody_H3 and snugdock should run with the camelid antibodies, as I had made them work with camelids at some point over the past few years...

Mon, 2017-04-17 11:17

Dear jadolfbr, thank you for your help, I tried an arbitrary fasta for the light chain with my 'VHH' heavy chain, I've got this outputs :

- debug-model-A-*.pdb (from 0 to 9)

- debug-model-B-*.pdb (from 0 to 9)

- frh-*.pdb (from 0 to 9)

- frl-*.pdb (from 0 to 9)

- frh_after_seqeunce_adjustment-*.pdb (from 0 to 9)

- frh_frl_oriented-*.pdb (from 0 to 9)

- model-*.relaxed.pdb (from 0 to 9)

- orientation-*.pdb (from 0 to 9)

I think that the best way is to use "frh.pdb" or "frh_after_seqeunce_adjustment.pdb" as inputs for antibody_H3 application. But what about  removing the light chain from the "model-*.relaxed.pdb " as you mentioned above.


Thank you 



Wed, 2017-04-19 15:16
Ayoub Ksouri