haddocking / haddock-antibody-antigen Goto Github PK

Hello,
Thanks for putting up this great tool to annotate the Ab pdb.

I just try to annotate the Ab structure with IMGT format, and it seems there is a small bag there as I got the error as shown below:

python ImmunoPDB.py -i test.pdb -o test2.pdb -s IMGT --fvonly --rename --splitscfv
--> test.pdb could not be parsed: Cannot change id from (' ', 10, ' ') to (' ', 11, ' '). The id (' ', 11, ' ') is already used for a sibling of this entity

And rest of the annotation format works just fine :)
Will it be nice if IGMT format could work as well !
Thanks
Best,
Che

So far, the haddock-antibody-antigen pipeline does not generate all the files required for haddock-local run, in particular: ambig and unambig files, as well as passive residue list as well as run.param, are not generated.
I solved this issue in my antibody-antigen docker container which I published as a separate repository https://github.com/antonkulaga/haddock-antibody-docker . It needs some polishing but it kind of works.
In this issue I want to know your opinion whether you think it is good to back-port tbl files and run.param generation to haddock-antibody-antigen repository or it is better just to reference my container in readme, so people who want to use local version will know that they can save some time by using my code or maybe you will suggest something totally different.

I noticed that final pdb-s that are produced have no header that causes warnings in some of the software

The numbering script produces error with the example PDB file;

python ImmunoPDB.py -i 4G6K.pdb -o 4G6K_i.pdb --scheme i --fvonly --rename --splitscfv

4G6K.pdb could not be parsed: Cannot change id from `(' ', 10, ' ')` to `(' ', 11, ' ')`. The id `(' ', 11, ' ')` is already used for a sibling of this entity.

As it has been noted by some users (#7, #10) not all naming schemes are allowed (e748733).

We should disable the options in the re-destributed ImmunoPDB.py script to only allow for Chotia naming scheme.

Since the ANARCI installation seems to be broken (oxpig/ANARCI#41), we could fallback to installing via conda instead.