Re: [AMBER-Developers] [AMBER] Segmentation fault when starting parmchk2 for a modified aminoacid residue

From: David Case <>
Date: Fri, 24 Feb 2017 07:54:24 -0500

On Thu, Feb 23, 2017, Nikolay N. Kuzmich wrote:
> The ac file, prepared from the mol2 by antechamber, shows C21 having 4
> bonds.

Sorry: my fault here: I was looking at it wrong.
> I took the pdb file with a modified serine from PDB database, then added
> hydrogens using pdb4amber and edited the pdb with Schrodinger Maestro to
> build my own modified residue. I also noticed that pdb4amber tried to
> assign the wrong protonation state for the sp2-hybrid NH group.

I guess I'd still kind of like more specifics, although it's probably a bit
irrelevant. pdb4amber generally does not add hydrogens; what flags did you
give (e.g. --reduce?).

> excessive proton was deleted. Then I copied the modified residue, added the
> absent H atom for N tail and OH group for C-tail and saved it as mol2 file
> (Schrodinger Maestro).
> After that I obtained the corresponding ac file employing antechamber,
> manually built the mc file after the B5 tutorial. After running prepgen
> application I obtained the prepin file.

I can again confirm that running parmchk2 on your prepin file:
> > parmchk2 -i Ser630_bound.prepin -f prepi -o frcmod.adf -a Y -p \
> > $AMBERHOME/dat/leap/parm/parm10.dat

doesn't give me a segfault with AmberTools16. However, I *am* getting
complaints about the dummy atoms (and entries for the dummy atoms in the
frcmod file.) I believe this means that changes to parmchk2 since the
tutorial was developed are changing the workflow here.

Specifically, I see this:

slater% parmchk2 -i Ser630_bound.prepin -f prepi -o foo -p
Warning: Atom Type 'DU' does not show up in PARMCHK.DAT, uses default values
for improper_flag [0], group_id [0] and equivalend_flag [0]
Warning: no mass information for Atom Type 'DU', set to 0.0
         It is recommended to add the new Atom Type 'DU' to PARMCHK.DAT

cc-ing to Amber developers: can someone try to go through this tutorial
with the current codes, and see what needs to be fixed up?

cc-ing also to Junmei/Scott: do you recognize the warning about dummy atoms?


AMBER-Developers mailing list
Received on Fri Feb 24 2017 - 05:00:02 PST
Custom Search