“yes it runs!” was a really early response for the comparative modeling run I started for a multi chain protein. I use rosetta 3.4.
My problem is that the end result looks completely off compared to the template file (also to any other folded structure, the end result is basically a stretch of amino acids). Points that I messed up something at the alignment.
My query protein, has 5 chains (3 of them are unique) , my template has 5 chains and it is a homomer. The fasta file and the alignment file contains the whole sequence. I didn’t use any chain break neither for query nor for the template sequences. The template pdb file contains different chain ids and residues where numbered starting from 1 for each chain. I have another run going on with template pdb containing only one chain id.
I was thinking if it would be better to have each chain separately defined instead of merging the chains in pdb file? I am not quite sure if any of these changes would change the results?
I don’t think the threading code is meant to be used for anything other than single domains. fold_and_dock works for symmetric assemblies. Maybe you can try threading the domains individually and then docking the results using constraints from the template. (or maybe there’s a way to get this alignment to work; I don’t know how the alignment files work either…)