Thanks Charles,
This appears to have worked, however, my second issue is when I load the files
into the 'refine_eefx.py' script. If I simply run the script as per the
/eginput/gb1_rdc/refine_eefx.py example by loading the output files from
refine.py with the command
==> inputStructures="refine_*.sa"
the script fails with the error
==> "SystemError: xplor-nih error: error reading restraint: selection string
resid 1 and name HN selects no atoms"
Reading through my .out file it appears to trace back to the line
==> # load PDB to generate PSF
==> import glob
==> protocol.loadPDB( glob.glob(inputStructures)[0],
==> deleteUnknownAtoms=True )
This part of the code appears to be removing atoms on my N and C terminus
because my sequences are head to tail cyclic and leads to my missing atom error
above.
To get around this issue, I thought I could simply generate a new cyclic .psf
file using the eefx2.par force fields and loading this .psf file into the
refine_eefx.py script by adding the lines
==> import protocol
==> protocol.initStruct("circ_eefx.psf")
==> inputStructures="refine_*.sa" # refine each structure with implicit solvent
I was happy when this appeared to have worked; my circ_eefx.psf file loads and
the refine_eefx.py script will run. Sadly however, I end up with a lot of
missing and broken bonds along with the atoms in the generated pdb files are
scattered all over the place. I assume this issue is related to a duplication
error between my manually upload circ_eefx.psf file and the IMPORT GLOB command
trying to generate a new .psf file for each of the refine.sa structures??? Is
there a better way I could load my correct psf file to get rid of this issue
and maintain all my bonds?
Thanks,
Matt
___________________________
Matthew D. Shortridge, PhD
Research Associate
University of Washington
Department of Chemistry
http://depts.washington.edu/varani2/
From: Charles Schwieters <[email protected]>
Sent: Monday, June 5, 2017 7:09 AM
To: Matthew D Shortridge
Cc: [email protected]
Subject: Re: [Xplor-nih] water refine cyclic peptides
Hello Matt--
>
> I started this project using a set of legacy scripts which were
> handed down to me, they've worked well, but I've recently migrated
> to using the python based scripts from /eginput/gb1_rdc. Thus far
> I've been able to generate clean .psf files for my cyclic sequences
> and calculate structures using the anneal.py and refine.py scripts
> with little issues. As an added bonus, the python generated
> structures match well with legacy script structures.
>
> However, when I try to load the structures into the water refinement
> scripts (refine_eefx.py) I start having issues. From what I can
> tell, it appears I need to make new .psf files with the eefx force
> fields to load into these scripts?
That is correct.
> I first tried to do this by calling in the protein_eef.top &
> protein_eef.par force fields in the genCircpep.py script compared to
> the protein-3.1 versions. This approach appears to generate a
> cyclic peptide .psf file, however it appears to skip over LtoD
> patches with these force fields. With the LtoD patch turned on, the
> script fails at protocol.fixupCovalentGeom with the error. With the
> LtoD patch turned off I do not get this error.
>
> %CODIMP-ERR: missing improper parameters %%%%%%%%%%%%%%%%%%%%%
> improper energy constant missing.
> target improper value missing.
> periodicity missing.
> ATOM1: SEGId=" ", RESId="2 ", NAME="HA ", CHEMical="HB "
> ATOM2: SEGId=" ", RESId="2 ", NAME="C ", CHEMical="C "
> ATOM3: SEGId=" ", RESId="2 ", NAME="N ", CHEMical="NH1 "
> ATOM4: SEGId=" ", RESId="2 ", NAME="CB ", CHEMical="CT2 "
I do apologize. These parameters were missing from
toppar/eefx/protein_eefx2.par:
IMPRoper HB C NH1 CT1 500 {sd=0.031} 0 66.2535
IMPRoper HB C NH1 CT2 500 {sd=0.031} 0 66.1640
IMPRoper HB C NH1 CT3 500 {sd=0.031} 0 65.9907
I've placed a full version of the updated file here:
https://nmr.cit.nih.gov/out/protein_eefx2.par
Please copy this to the toppar/eefx directory under you Xplor-NIH
distribution.
Please let me know if you continue to have problems.
best regards--
Charles
_______________________________________________
Xplor-nih mailing list
[email protected]
https://dcb.cit.nih.gov/mailman/listinfo/xplor-nih
