Adding palmitoyl chains to a protein

[mdoktorova]

Hi all,

I want to add palmitoyl chains to a protein and simulate it with Elnedyn22. I have an all-atom structure of the protein and can easily convert it to CG with the martinize scripts. However, I can't find a version that would let me add the PALM chains and produce an .itp file with Elnedyn22. I could add the chains manually but I don't know how they (and the CYS residues they are attached to) are treated in an elastic network. Any advice is welcome!

Thanks,
Milka

[pckroon]

Hi!
Martinize2 can definitely do this. You'll need to add the PALM residue though. You'll need to add it in 3 places: 1) the input forcefield (default charmm); 2) the output forcefield (elnedyn22 in your case); and 3) the mapping. See #535 for some hints on how to do that.
In the output forcefield you'll also need to add a link describing the bonded parameters between a PALM residue and other residues (i.e. CYS).

As for your actual question, by default elastic network bonds are only applied between beads with the atomname 'BB' (https://github.com/marrink-lab/vermouth-martinize/blob/master/vermouth/selectors.py#L63), but you can tune this with the -eb CLI flag. Note that the elastic network is only made if you provide the -elastic flag.

As to what the elastic network should do with PALM residues... I have no idea :)

[pckroon]

Bit of both, it happens because we simply took the existing backmapping files as a starting point, and never really bothered to clean them up.
Is your CYP mapping mentioned in that list? Also, please run martinize2 -list-ff -ff-dir newparam -map-dir newparam/mapping -vv instead, otherwise it'll never look at your files.

[mdoktorova]

I tried many things but cannot fix this problem. What happens is that in the output of something like:

martinize2 -ff-dir newparam -from charmm -ff elnedyn22 -map-dir newparam/mapping -list-blocks

I can see that the new CYP 'block' is known to both charmm and elnedyn22 (as expected). Yet, when I run:

martinize2 -list-ff -ff-dir newparam -map-dir newparam/mapping -vv

I get a bunch of 'inconsistent data' messages, including:

Level 5 - inconsistent-data - vermouth.map_input - Can't find block CYP for FF universal
Level 5 - inconsistent-data - vermouth.map_input - Can't find block CYP for FF charmm36

Also, I noticed that in my cyp.charmm36.map file (in the new mapping folder), if I change the [ from ] field from 'universal' to 'charmm', I get the following error when I try to run the program on my protein:

Traceback (most recent call last):
File "/usr/local/bin/martinize2", line 824, in
entry()
File "/usr/local/bin/martinize2", line 539, in entry
known_force_fields)
File "/usr/local/lib/python3.6/site-packages/vermouth/map_input.py", line 426, in read_mapping_directory
new_mappings = read_backmapping_file(infile, force_fields)
File "/usr/local/lib/python3.6/site-packages/vermouth/map_input.py", line 110, in read_backmapping_file
weights, extra, name_to_index)
File "/usr/local/lib/python3.6/site-packages/vermouth/map_input.py", line 165, in make_mapping_object
idx_to = to_name_to_idx[atname_to]
KeyError: 'BB'

So the new mapping file is being read but something else is off (maybe some link in the default vermouth/data folder?). Any ideas where this may be coming from?

[pckroon]

That looks like progress.

I get a bunch of 'inconsistent data' messages, including:

Level 5 - inconsistent-data - vermouth.map_input - Can't find block CYP for FF universal
Level 5 - inconsistent-data - vermouth.map_input - Can't find block CYP for FF charmm36

These sound correct if you defined CYP for charmm (not charmm36 and not universal).

The traceback means that the elnedyn22 CYP block doesn't have a bead named BB. Could you post your residue (you can leave out any bonded parameters. If you prefer you can swap all beadtypes for e.g. C1 or another placeholder).
If that doesn't shed any light on it I will suggest you send me the files you have.

[mdoktorova]

This worked!! It turned out that the atom names in my CYP block in the output force field were different (that's why the 'no BB' error). When I fixed them the program worked without any errors and warnings. Thanks a lot!

I noticed that in many of the elnedyn22 blocks the non-CA atoms mapped to 'BB' had an exclamation mark in front: '!BB'. Do you know what this is used for? I can build my palmitoylated protein now either way (with or without these exclamation marks) but I can't visually identify any differences.

[pckroon]

Great news :)

The exclamation mark in the mapping indicates a mapping weight of 0. So the position of the BB bead is not influenced by the non-CA atoms, as prescribed by elnedyn.