vermouth-martinize
vermouth-martinize copied to clipboard
marrink-lab
Raise warning if mutation not found
At the moment, if a -mutate flag is given, but the target residue is not found, then martinize will process the input structure as given without raising a warning.
Eg, taking the martinize tutorial:
martinize2 -f 181L_clean.pdb -o t4l_only.top -x t4l_cg.pdb -dssp /usr/bin/dssp -p backbone -ff martini3001 -mutate A-MET1:LYS
will correctly mutate the first residue from MET to LYS, but
martinize2 -f 181L_clean.pdb -o t4l_only.top -x t4l_cg.pdb -dssp /usr/bin/dssp -p backbone -ff martini3001 -mutate A-MET2:LYS
will do nothing (ie. keep residue 1, 2 as MET, ASN as in the input structure). If a user is expecting the mutation (eg the above could easily be a typo) and uses the topology naively (even though they should really check...) then that's a problem.
It's probably about time I put in a PR here, so I can try and do something about it this week, just thought I'd open the issue first, or in case I don't get to it as soon as I'd like.
vermouth 0.9.6
I think this would be a great first PR :)
The best place is probably in vermouth/processors/annotate_mut_mod.py. Beyond that, good luck, have fun, and feel free to ask for help/advice/whatever!
@pckroon Noticed one minor issue with how this was fixed. In the case where we have >1 chain in the protein, the mutation is looked for across all chains and raises a warning for each of the chains where the mutation isn't found. Then nothing is written, even though the mutation has been specified correctly. I'll come up with a fix asap...
Good find. I don't think the issue is chains per se, but rather molecules, since the AnnotateMutMod processor has a run_molecule method rather than a run_system method.
Thanks, not sure I have time to address this week but I'll try to start next. It would be good to try and address #512 at the same time, but I'm less sure about where to start on that one.
Don't. That will be a separate PR for sure, because #512, #17, #307 is a huge feature. The open choice is whether you want to rebuild coordinates/atoms at the atomistic level or the CG level. Once you can do that for mutated residues (#512) you can do it for any atom without coordinates in the input (#17)
ha, understood! I'll address this soon then, and maybe we could work something out for the rebuilding issue in the new year?
IMNSHO the rebuilding is one of the biggest missing features we have, so yes please!