On Sun, Mar 16, 2014 at 12:35 PM, de Waal, Parker <Parker.DeWaal.vai.org>wrote:
> Hello Everyone,
>
> I recently tried to set up a membrane bound protein simulation (AmberTools
> version 13.24 Amber version 12.21) following this tutorial ->
> http://ambermd.org/tutorials/advanced/tutorial16/An_Amber_Lipid_Force_Field_Tutorial.html
>
> Everything seems to work until I try to load the pdb into tleap at which
> point a stream of warnings about split residues appears:
>
> Warning: name change in pdb file residue 1 ;
> this residue is split into OL and PC.
> Warning: name change in pdb file residue 1 ;
> this residue is split into PC and OL.
>
I suspect this happens because the PDB file assigns the same residue number
(1) for each fragment of each lipid. I believe the Amber lipid force field
is designed to be modular, so it splits what some programs call a single
residue into multiple residues and parametrizes them separately.
The preprocessing scripts that prepare the Lipid PDB files appear to
(appropriately) assign the different fragments different residue names, but
(unfortunately in this case) do not assign them different numbers. As a
result, tleap sees multiple residue names assigned the same residue number
in contiguous atoms. It (correctly) splits them into different residues
and assigns the second chunk of atoms a new residue number. I suspect this
is desired behavior. However, since it indicates a somewhat arbitrary
decision made by tleap, tleap emits a warning letting the user know what
has been done. [1]
>
> ...
>
>
>
> To check if there was an error in my pdb I also tried to use the
> charmmlipid2amber.x processed DOPC_128_Lipid11.pdb from the tutorial and
> the same warning appears.
>
>
>
> Although I can still run a check and saveamberparm without errors is there
> a way to hide these warnings?
You would need to modify the PDB file to assign each fragment that has a
different name a new residue number. Doing this by hand for anything but a
toy system is infeasible (and all you gain is tleap output with fewer
warnings). You could write a script to do this auto-renumbering for you,
but again it's a question of what kind of benefit you get from the time it
takes to write, test, and debug said script.
HTH,
Jason
[1] This is not really an 'arbitrary' choice, as there are far more reasons
one would want to assign them to different residues as opposed to the same
residue. However, this warning can also indicate in many instances a typo
in the input structure whereby someone mistyped the name of the residue
which tleap now (mistakenly) splits into a different residue. So this is
not a warning you want to simply discard in general. If it's expected (as
in this case), then it is benign. But it often helps spot errors in the
input PDB.
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Mar 16 2014 - 12:00:03 PDT
