Re: Error: psfgen) unknown atom type H

From: Gumbart, JC (gumbart_at_physics.gatech.edu)
Date: Fri Aug 27 2021 - 10:27:52 CDT

Hi Adam,

Rebuilding a PSF from CHARMM-GUI output is certainly possible, but is often quite challenging as you have found. Unfortunately, there’s no quick solution. Ideally, you would find a way to avoid it altogether, just using the output of CHARMM-GUI directly. We find this is sufficient for almost all of our use cases.

If all you need is to combine parts of your system, you can do this without the trouble of rebuilding by using the “readpsf ___.psf pdb ___.pdb” command in psfgen. This will just take the parameters and coordinates directly from the CHARMM-GUI output without modifying anything.

Best,
JC

On Aug 26, 2021, at 9:14 PM, Mohammad Mehdi Pirnia <mpdezfouli_at_gmail.com<mailto:mpdezfouli_at_gmail.com>> wrote:

Dear users,

Please see the output below and all my codes and files uploaded in the link at the bottom.

psfgen) $Id: top_allxx_sugar.inp,v 1.106 2014/08/19 19:07:43 alex Exp $
psfgen) >>>>>>>>>>>> All-hydrogen topologies used in the <<<<<<<<<<<<<<<<
psfgen) >>>>> development of the CHARMM carbohydrate force field<<<<<<<<
psfgen) >>>>>>>>>>>>>>>>>>>>>>>>> June 2009 <<<<<<<<<<<<<<<<<<<<<<<<<<<<<
psfgen) >>>>>>>> Direct comments to Alexander D. MacKerell Jr. <<<<<<<<<<
psfgen) >>>>>>>>>> via the CHARMM web site: https://urldefense.com/v3/__http://www.charmm.org__;!!DZ3fjg!vO0uc_anTOzmqkjSf2C59dllukgfC0GMxBt9EO-qGcmwDXT6qKcK82D60kuBPnFWHA$ <https://urldefense.com/v3/__http://www.charmm.org__;!!DZ3fjg!sN1ZQPHyGqtzMAVzV4NGwbp0DYwKNoHjDVsVHnWD_BqbggqVptqiKvozjI3iSuOirQ$> <<<<<<<<<<<<<<
psfgen) >>>>>>>>>>>>>>> parameter set discussion forum <<<<<<<<<<<<<<<<<<
psfgen)
psfgen) Created by CHARMM version 32 1
psfgen) reading topology file ../temporary2/toppar_all36_lipid_lps.str

psfgen) reading topology from stream file
psfgen) Topology for lipidA from LPS
psfgen)
psfgen) skipping parameters in stream file
psfgen) new parametes
psfgen)
psfgen) reading topology file ../temporary2/toppar_all36_carb_imlab.str

psfgen) reading topology from stream file
psfgen) glycan chemical modification patches
psfgen)
psfgen) Created by CHARMM version 41 1
psfgen) cross-term entries present in topology definitions
psfgen) skipping parameters in stream file
psfgen) glycan chemical modification patches
psfgen)
2.0
psfgen) building segment L
psfgen) reading residues from pdb file upper_leaflet.pdb
psfgen) extracted 540 residues from pdb file
Info: generating structure...
psfgen) unknown atom type H
psfgen) add atom failed in residue ECLI:1
Segmentation fault (core dumped)

----------------------------

I've been getting this error for days. I didn't see anything wrong with my stream files and atom type H is definitely defined!

I'm simulating peptide interactions with a bacterial membrane. I used https://urldefense.com/v3/__http://www.charmm-gui.org__;!!DZ3fjg!vO0uc_anTOzmqkjSf2C59dllukgfC0GMxBt9EO-qGcmwDXT6qKcK82D60kvsyb-qHw$ <https://urldefense.com/v3/__http://www.charmm-gui.org__;!!DZ3fjg!sN1ZQPHyGqtzMAVzV4NGwbp0DYwKNoHjDVsVHnWD_BqbggqVptqiKvozjI39vxcbvQ$> in order to build my membrane and it includes LPSA, PMCL, PMPE, PMPG, PVPE, and PVPG.

- Is there a way to create such membranes using VMD? I just saw POPC and POPE in the membrane builder.

- After downloading, I modified all residues' res-numbers to be consecutive, they were repetitive. Now, I'm trying to create my membrane's PSF file. The output is shown below.

my VMD version is 1.9.4a.51

Here is the link on which I've uploaded all my files:
https://urldefense.com/v3/__https://files.fm/u/bcvap3g8g__;!!DZ3fjg!vO0uc_anTOzmqkjSf2C59dllukgfC0GMxBt9EO-qGcmwDXT6qKcK82D60ktIPKOTWA$ <https://urldefense.com/v3/__https://files.fm/u/bcvap3g8g__;!!DZ3fjg!sN1ZQPHyGqtzMAVzV4NGwbp0DYwKNoHjDVsVHnWD_BqbggqVptqiKvozjI2YfcU4HQ$>

Thank you all
Sincerely
Adam

This archive was generated by hypermail 2.1.6 : Fri Dec 31 2021 - 23:17:11 CST