VMD-L Mailing List

From: Jim Phillips (jim_at_ks.uiuc.edu)
Date: Mon Nov 28 2005 - 19:14:35 CST

Hi folks,

The patch CT1 looks like this:

  PRES CT1 0.00 ! methylated C-terminus from methyl acetate
  GROUP ! use in generate statement
  ATOM N NH1 -0.47 !
  ATOM HN H 0.31 ! OT1
  ATOM CA CT1 0.17 ! | //
  ATOM HA HB 0.09 ! -N--CA--C HT1
  ATOM C CD 0.63 ! | | \ /
  ATOM OT1 OB -0.52 ! HN HA OT2--CT--HT2
  ATOM OT2 OS -0.34 ! \
  ATOM CT CT3 -0.14 ! HT3
  ATOM HT1 HA 0.09 !
  ATOM HT2 HA 0.09 !
  ATOM HT3 HA 0.09 !
  DELETE ATOM O
  ...

Most residues, like ALA below, this would be fine for:

  RESI ALA 0.00
  GROUP
  ATOM N NH1 -0.47 ! |
  ATOM HN H 0.31 ! HN-N
  ATOM CA CT1 0.07 ! | HB1
  ATOM HA HB 0.09 ! | /
  GROUP ! HA-CA--CB-HB2
  ATOM CB CT3 -0.27 ! | \
  ATOM HB1 HA 0.09 ! | HB3
  ATOM HB2 HA 0.09 ! O=C
  ATOM HB3 HA 0.09 ! |
  GROUP !
  ATOM C C 0.51
  ATOM O O -0.51
  ...

For ALA, the CT1 line "ATOM HA HB" changes the type/charge of the existing
atom HA (or it would if they weren't the same already).

The sequence is just HIS GLY, so the C-terminal residue is GLY:

  RESI GLY 0.00
  GROUP
  ATOM N NH1 -0.47 ! |
  ATOM HN H 0.31 ! N-H
  ATOM CA CT2 -0.02 ! |
  ATOM HA1 HB 0.09 ! |
  ATOM HA2 HB 0.09 ! HA1-CA-HA2
  GROUP ! |
  ATOM C C 0.51 ! |
  ATOM O O -0.51 ! C=O
                          ! |
  ...

In this case, the CT1 line "ATOM HA HB" adds a *new* atom, HA, which is
just wrong. You could delete that line and things would be better, but
you'd still get the angle parameter error because CT1 line "ATOM CA CT1"
changes the type of the GLY CA from CT2 to CT1 (which, I am guessing, is
wrong for an atom bonded to two hydrogens) so you get an HB-CT1-HB
parameter error (you're asking for something that shouldn't exist).

You could easily modify the CT1 patch to use the correct atom types, but
getting the charges right is beyond my expertise.

-Jim

On Tue, 22 Nov 2005, hodak_at_chips.ncsu.edu wrote:

> Generated protein looks much better now, but only if last is set to CT3.
> CT1 (which I like since it is simple) still generates broken pdb file (and
> namd complains about the HB CT1 HB angle). Do you know what the problem
> with CT1 is?
>
> Also, even with CT3 namd still complain about non-unique coordinates,
> but all coordinates seem to be OK. Perhaps the problem is with namd
> config file, but I first want to psfgen and CT1 (if possible) running.
>
> Thanks a lot for your help.
> Miro
>
> On Tue, 22 Nov 2005, Michel Espinoza-Fonseca wrote:
>
>> okay, I think that now it'll work... I modified a little bit the input script for psfgen. The script is:
>>
>> topology top_all27_prot_na.inp
>> alias residue HIS HSE
>> segment HG3 {pdb HG3.pdb
>> first ACE
>> last CT3
>> auto angles dihedrals}
>> coordpdb HG3.pdb HG3
>> guesscoord
>> writepdb HG3-good.pdb
>> writepsf HG3-good.psf
>>
>> you have to be sure that the protonation state of HIS is correct (HSD, HSE or HSP). I just changed "auto none" by "auto angles dihedrals" to actually reassing angles and dihedrals. hopefully you won't have problems now.
>>
>> I also attach the pdb and psf files I got.
>>
>> Cheers,
>> Michel
>>
>> ________________________________
>>
>> From: hodak_at_chips.ncsu.edu [mailto:hodak_at_chips.ncsu.edu]
>> Sent: Tue 11/22/2005 5:13 AM
>> To: Michel Espinoza-Fonseca
>> Cc: vmd-l_at_ks.uiuc.edu
>> Subject: RE: vmd-l: Uncharged terminal ends?
>>
>>
>>
>> I stripped all hydrogens from my starting dipeptide (HG) but the problem
>> is still there (namd still complains). I cannot spot the atoms that are
>> close or identitical, but 27th atom has 0,0,0 coordinates. I am attachaing
>> my starting pdb file and also the script I am using.
>>
>> Thanks,
>> Miro
>>
>> On Tue, 22 Nov 2005, Michel Espinoza-Fonseca wrote:
>>
>>> okay... I have one more question... Is your starting protein WITHOUT hydrogen atoms? if yes, remove all of them prior building your psf and pdb files. You can also check that the structure of your protein is correct (e.g., that you don't have duplicated atoms).
>>>
>>> Try to do that and tell me if you see something weird. You can also send me your pdb file and I can check it.
>>>
>>> Cheers,
>>> Michel
>>>
>>> ________________________________
>>>
>>> From: hodak_at_chips.ncsu.edu [mailto:hodak_at_chips.ncsu.edu]
>>> Sent: Tue 11/22/2005 4:40 AM
>>> To: Michel Espinoza-Fonseca
>>> Cc: vmd-l_at_ks.uiuc.edu
>>> Subject: RE: vmd-l: Uncharged terminal ends?
>>>
>>>
>>>
>>> It seems that I have one more issue with generated pdb file. When I run
>>> namd I get:
>>> Warning: Not all atoms have unique coordinates.
>>>
>>> Inspecting generated pdb file, my CA on G has two hydrogen that are very
>>> close (and CA actually has 5 bonds):
>>>
>>> ATOM 27 HA GLY 3 4.826 -0.964 -1.247 1.00 0.00 PROT H
>>> ATOM 36 HA2 GLY 3 4.806 -0.987 -1.255 0.00 0.00 PROT H
>>>
>>> Please note that 36 is my last atom.
>>> It seems like problem with psfgen's guessing of coordinates (guesscoord).
>>>
>>> What can I do to avoid this problem?
>>>
>>> Thanks,
>>> Miro
>>>
>>> On Tue, 22 Nov 2005, Michel Espinoza-Fonseca wrote:
>>>
>>>> "auto none" automatically generates angles and dihedrals based on bonds, since the topology file don't list angles and dihedrals. This is specially needed when you apply patches on your protein (such as ACE or CT1) or with water.
>>>>
>>>> Hope it is clear,
>>>> Michel
>>>>
>>>> ________________________________
>>>>
>>>> From: hodak_at_chips.ncsu.edu [mailto:hodak_at_chips.ncsu.edu]
>>>> Sent: Tue 11/22/2005 4:16 AM
>>>> To: Michel Espinoza-Fonseca
>>>> Cc: vmd-l_at_ks.uiuc.edu
>>>> Subject: RE: vmd-l: Uncharged terminal ends?
>>>>
>>>>
>>>>
>>>> With this I am namd simulation will start. Now what is does "auto none"
>>>> actually do?
>>>>
>>>> Thanks,
>>>> Miro
>>>>
>>>> On Tue, 22 Nov 2005, Michel Espinoza-Fonseca wrote:
>>>>
>>>>> mmmm... try to rebuild your system using
>>>>>
>>>>> first ACE
>>>>> last CT1
>>>>> auto none
>>>>>
>>>>> and then see if it works...
>>>>>
>>>>> ________________________________
>>>>>
>>>>> From: hodak_at_chips.ncsu.edu [mailto:hodak_at_chips.ncsu.edu]
>>>>> Sent: Tue 11/22/2005 4:00 AM
>>>>> To: Michel Espinoza-Fonseca
>>>>> Cc: vmd-l_at_ks.uiuc.edu
>>>>> Subject: RE: vmd-l: Uncharged terminal ends?
>>>>>
>>>>>
>>>>>
>>>>> I cannot minimize the system (at least not with namd), since I get that error
>>>>> about HB CT1 HB angle parameters when I try to run namd with my system.
>>>>>
>>>>> On Tue, 22 Nov 2005, Michel Espinoza-Fonseca wrote:
>>>>>
>>>>>> That's weird... Are you minimizing your system first ? Usually I get the same thing you described (multiple bonds on the ACE or NME terminals), but you can easily solve this problem by subjecting your system to a short (~1000 steps) minimization.
>>>>>>
>>>>>> After that you should't have any problem.
>>>>>>
>>>>>> Hope it helps.
>>>>>>
>>>>>> Michel
>>>>>>
>>>>>> ________________________________
>>>>>>
>>>>>> From: owner-vmd-l_at_ks.uiuc.edu on behalf of hodak_at_chips.ncsu.edu
>>>>>> Sent: Tue 11/22/2005 3:18 AM
>>>>>> To: Justin Gullingsrud
>>>>>> Cc: vmd-l_at_ks.uiuc.edu
>>>>>> Subject: Re: vmd-l: Uncharged terminal ends?
>>>>>>
>>>>>>
>>>>>>
>>>>>> This helped a lot. I decided that I want acetylated N-terminus and
>>>>>> methylated C-terminus. Using
>>>>>>
>>>>>> first ACE
>>>>>> last CT1
>>>>>>
>>>>>> I can get the ends I want, although vmd shows extra bonds that those added
>>>>>> groups have carbons connected to N or C. I am not sure if this is a
>>>>>> problem or not.
>>>>>> I went ahead and used generated psf and pdb files as input for namd
>>>>>> simulation, but I get the following error:
>>>>>>
>>>>>> FATAL ERROR: UNABLE TO FIND ANGLE PARAMETERS FOR HB CT1 HB
>>>>>> ------------- Processor 0 Exiting: Called CmiAbort ------------
>>>>>> Reason: FATAL ERROR: UNABLE TO FIND ANGLE PARAMETERS FOR HB CT1 HB
>>>>>>
>>>>>> Charm++ fatal error:
>>>>>> FATAL ERROR: UNABLE TO FIND ANGLE PARAMETERS FOR HB CT1 HB
>>>>>>
>>>>>> The problem seems to be the C-terminal end, omitting "last CT1" namd
>>>>>> simulation runs without problems.
>>>>>>
>>>>>> Any idea what the problem is and how it can be solved?
>>>>>>
>>>>>> Thanks,
>>>>>> Miro
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> On Mon, 21 Nov 2005, Justin Gullingsrud wrote:
>>>>>>
>>>>>>> Hi,
>>>>>>>
>>>>>>> You probably want an acetylated N terminus, then, which you can get by
>>>>>>> specifying "first ACE". I'm not sure what kind of C terminus you're
>>>>>>> looking for, but check out the PRES patches in the topology file to
>>>>>>> see what's available.
>>>>>>>
>>>>>>> Hope this helps,
>>>>>>> Justin
>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>>
>>>
>>
>>
>