Re: vmd-l: Re: namd minimization looking for wrong angle

From: Francesco Pietra (chiendarret_at_gmail.com)
Date: Fri Nov 19 2021 - 15:15:00 CST

I thank you for your help. Your mode or running autopsf from the Tk console
worked also for me with vmd 1.9.4a55, and also for the box.

As usual with me, I run autopsf from the gui after having loaded the
solvent molecule (or solvent box). Then, in a sequence, Add the .rtf file,
Load input files, Guess and split chains, Create chains.

Thanks indeed

On Fri, Nov 19, 2021 at 6:03 PM Peter Freddolino <petefred_at_umich.edu> wrote:

> How were you running autopsf, and with what version of vmd?
>
> autopsf -mol 0 -top 1009279.rtf
> (after having loaded chcl3.pdb) works fine for me with the files that you
> provided, using vmd1.9.4a51
>
> Best,
> Peter
>
> On Fri, Nov 19, 2021 at 11:56 AM Francesco Pietra <chiendarret_at_gmail.com>
> wrote:
>
>> Unfortunately I have no access to that journal. At any event, I never use
>> SHAKE, so that I parameterized CHCL3 on the MATCH server, getting the
>> attached rtf. On that basis, I first tried to generate psf/pdb for a single
>> CHCL3 molecule. With autopsf the error.log was
>>
>>
>> ERROR: failed on end of segment
>>> MOLECULE DESTROYED BY FATAL ERROR! Use resetpsf to start over.
>>> ERROR: failed on end of segment
>>> MOLECULE DESTROYED BY FATAL ERROR! Use resetpsf to start over.
>>> while executing
>>> "segment $segid {
>>> pdb $segfile
>>>
>>
>> # We alias the C-terminal OXT atoms to OT2 so that psfgen has to
>>> guess one atom less.
>>> # Otherwise psf..."
>>> (procedure "psfsegments" line 37)
>>> invoked from within
>>> "psfsegments $logfileout"
>>> (procedure "::autopsf::afterchains_gui" line 66)
>>> invoked from within
>>> "::autopsf::afterchains_gui"
>>> invoked from within
>>> ".autopsf.chains.finish invoke"
>>> ("uplevel" body line 1)
>>> invoked from within
>>> "uplevel #0 [list $w invoke]"
>>> (procedure "tk::ButtonUp" line 22)
>>> invoked from within
>>> "tk::ButtonUp .autopsf.chains.finish"
>>> (command bound to event)
>>
>>
>> while the terminal last lines told
>>
>> segfiles CHCl3_autopsf-temp.pdb_XO1.pdb
>>> CHCl3_autopsf-temp.pdb_XO1.pdb
>>> 1 4
>>> psfgen) building segment XO1
>>> psfgen) reading residues from pdb file CHCl3_autopsf-temp.pdb_XO1.pdb
>>> psfgen) unknown residue type CCL3
>>> psfgen) extracted 1 residues from pdb file
>>> psfgen) setting patch for first residue to none
>>> psfgen) setting patch for last residue to none
>>> Info: generating structure...
>>> psfgen) unknown residue type CCL3
>>>
>>
>> Then, I made recourse to vmd text mode, getting attached chcl3.psf/pdb
>> without any issue.
>>
>> However, I needed psf/pdb for the solvent box, again from the box
>> generated with packmol that I reported previously. In this case I was not
>> aware of any trick to avoid writing 1000-fold inputs on the .gen file, so
>> that I tried with autopsf. The last lines on the terminal were:
>>
>> psfgen) extracted 1000 residues from pdb file
>>> psfgen) setting patch for first residue to none
>>> psfgen) setting patch for last residue to none
>>> Info: generating structure...
>>> psfgen) unknown residue type CCL3
>>> Segmentation fault
>>>
>>
>> Obviously CCL3 was and is the residue. Probably, without additional
>> information no one will be able to suggest a remedy for autopsf (perhaps,
>> however, suggesting how to deal with vmd text mode for the solvent box,
>> i.e. a shortcut for the onethousand blocks)
>>
>> Thanks for your attention
>> francesco
>>
>>
>> On Fri, Nov 19, 2021 at 2:57 PM Peter Freddolino <petefred_at_umich.edu>
>> wrote:
>>
>>> Based on the comments in that topology file (which I had to google -- it
>>> would really help to include such things in the mailing list discussion),
>>> it looks like those CL-CL and other problematic bonds are present only for
>>> SHAKE. That suggests to me that this is supposed to be a rigid chloroform
>>> model, which might not be possible in NAMD. Did you check the original
>>> paper (W. Dietz, K. Heinzinger, Ber. Bunsen-Ges. Phys. Chem 1985, 89, 968)
>>> for how this is supposed to be implemented?
>>> Thanks,
>>> Peter
>>>
>>> On Fri, Nov 19, 2021 at 4:33 AM Francesco Pietra <chiendarret_at_gmail.com>
>>> wrote:
>>>
>>>> Hi Peter
>>>> Bonds between all couple of atoms, even CL-CL
>>>>
>>>>> 5000 !NBOND: bonds
>>>>> 1 2 1 4 1 5 1
>>>>> 6
>>>>> 2 3 7 8 7 10 7
>>>>> 11
>>>>> 7 12 8 9 13 14 13
>>>>> 16
>>>>> 13 17 13 18 14 15 19
>>>>> 20
>>>>> 19 22 19 23 19 24 20
>>>>> 21
>>>>>
>>>> where atom numbering is
>>>>
>>>>> 5000 !NATOM
>>>>> 1 AO1 1 CCL3 C CCM 0.179000 12.0110 0
>>>>> 2 AO1 1 CCL3 CL1 CLCM -0.087000 35.4500 0
>>>>> 3 AO1 1 CCL3 CL2 CLCM -0.087000 35.4500 0
>>>>> 4 AO1 1 CCL3 CL3 CLCM -0.087000 35.4500 0
>>>>> 5 AO1 1 CCL3 HX HCM 0.082000 1.0080 0
>>>>>
>>>>
>>>>> All such bonding is also evident from loading psf/pdb to vmd.
>>>>>
>>>>> I started from a box prepared with packmol, and I tried also adding
>>>>> TER between residues
>>>>> HEADER
>>>>> TITLE Built with Packmol
>>>>>
>>>>> REMARK Packmol generated pdb file
>>>>> REMARK Home-Page: https://urldefense.com/v3/__http://m3g.iqm.unicamp.br/packmol__;!!DZ3fjg!oU5LcoEmxeOd9WjCKnc2FF4OCPoJQwOBy-phxMnBFjnGqH92u1CiKsFx1Sv3YtoLmw$
>>>>> REMARK
>>>>> HETATM 1 CL1 CCL3A 1 30.522 5.429 3.173 1.00 0.00
>>>>> CL
>>>>> HETATM 2 CL2 CCL3A 1 30.670 3.053 4.852 1.00 0.00
>>>>> CL
>>>>> HETATM 3 CL3 CCL3A 1 31.018 2.823 1.969 1.00 0.00
>>>>> CL
>>>>> HETATM 4 C CCL3A 1 31.285 3.845 3.391 1.00 0.00
>>>>> C
>>>>> HETATM 5 HX CCL3A 1 32.361 3.996 3.509 1.00 0.00
>>>>> H
>>>>> TER
>>>>> HETATM 6 CL1 CCL3A 2 3.817 17.228 1.210 1.00 0.00
>>>>> CL
>>>>> HETATM 7 CL2 CCL3A 2 1.496 17.919 2.829 1.00 0.00
>>>>> CL
>>>>> HETATM 8 CL3 CCL3A 2 3.573 16.195 3.922 1.00 0.00
>>>>> CL
>>>>> HETATM 9 C CCL3A 2 3.231 17.561 2.849 1.00 0.00
>>>>> C
>>>>> HETATM 10 HX CCL3A 2 3.759 18.439 3.230 1.00 0.00
>>>>> H
>>>>> TER
>>>>>
>>>> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
>>>> No such problems encountered in making a MEOH box, starting from "RESI
>>>> MEOH" in CGenFF. Clear, not interacting MEOH molecules where obtained,
>>>> with psf indicating only the correct bonds.
>>>>
>>>> With CHARMM36/CGenFF the only source of topology for CHCL3 is
>>>> toppar_chloroform_dh.str
>>>> I can't understand where I am wrong with CHCL3
>>>> Thanks for advice
>>>> francesco
>>>>
>>>> On Thu, Nov 18, 2021 at 3:38 PM Peter Freddolino <petefred_at_umich.edu>
>>>> wrote:
>>>>
>>>>> Have you looked at the bonds in your psf to make sure they are
>>>>> correct? This looks like you've hydrogen atoms that have two bonds...
>>>>> Best,
>>>>> Peter
>>>>>
>>>>> On Thu, Nov 18, 2021 at 5:42 AM Francesco Pietra <
>>>>> chiendarret_at_gmail.com> wrote:
>>>>>
>>>>>> It seems thatb the str used is inadequate. Alsdo adjusting angles in
>>>>>> str, namd assks for dihedrals
>>>>>> fp
>>>>>>
>>>>>> On Thu, Nov 18, 2021 at 9:29 AM Francesco Pietra <
>>>>>> chiendarret_at_gmail.com> wrote:
>>>>>>
>>>>>>> While attempting to minimize a CHCL3 box built from
>>>>>>> CHARMM36-provided toppar_chloroform_dh.str, namd crashes
>>>>>>>
>>>>>>> FATAL ERROR: UNABLE TO FIND ANGLE PARAMETERS FOR CCM HCM CLCM (ATOMS
>>>>>>>> 1 5 4)
>>>>>>>> FATAL ERROR: UNABLE TO FIND ANGLE PARAMETERS FOR CCM HCM CLCM
>>>>>>>> (ATOMS 1 5 4)
>>>>>>>>
>>>>>>>
>>>>>>> Actually, the prm part of the str (which was read by namd) l
>>>>>>> correctly furnishes
>>>>>>>
>>>>>>>> ANGLES
>>>>>>>> CLCM CCM CLCM 0.0 111.30
>>>>>>>> HCM CCM CLCM 0.0 107.57
>>>>>>>>
>>>>>>>
>>>>>>> Why is namd looking for those wrong angles? The central atom is
>>>>>>> carbon not hydrogen.
>>>>>>>
>>>>>>> Thanks for advice
>>>>>>> francesco pietra
>>>>>>>
>>>>>>

This archive was generated by hypermail 2.1.6 : Fri Dec 31 2021 - 23:17:12 CST