From: Peter Freddolino (
Date: Tue Jul 28 2020 - 23:50:47 CDT

Glad to hear that things are working in VMD 1.9.3.
If a clean install of the latest VMD 1.9.4 alphas does not work, can you
please send the log/error files that I requested earlier? If there is a bug
floating around in the alpha builds of psfgen that would be very good to
If you need a bigger coarse grained water box, you should just be able to
adapt the solvate command in 03a-solvate.tcl of the M2 example, perhaps
making the values for the -z and +z padding larger. Behind the scenes, that
script is using the nonstandard solvent box capability of vmd solvate, so
all normal capabilities of solvate apply.

On Tue, Jul 28, 2020 at 8:21 PM Manas Kohli <> wrote:

> Dear Peter,
> Thank you for your email. It was actually an issue to do with VMD
> versioning. The tutorial files are unable to run on VMD 1.9.4 or psfgen
> 2.0. I reinstalled VMD 1.9.3 and can now run the tutorial files properly. I
> then tried to apply it to a new system to do with a voltage gated sodium
> channel which was my initial goal for the CG simulation. When I ran the
> simulation the minimisation worked initially but the simulation crashed due
> to atoms moving too fast. When I examined the minimization structure and
> dcd file it was apparent that the water box contracted too far in so parts
> of the protein were sticking out which I think causes the simulation to
> fail. My structure is rather large and I would require a larger coarse
> grain water box than the one provided in the tutorial files. Could you
> please help me with this? I think this would be the last issue before I
> could run the CG simulation for my channel.
> Thanks and Best Regards,
> Manas
> On Tue, Jul 28, 2020 at 9:02 PM Peter Freddolino <>
> wrote:
>> Can you please redirect the stdout and stderr from those steps to files
>> and send them? Let's focus on the VMD1.9.3 machine. Have you ever
>> installed a 1.9.4 alpha on that machine? The psfgen version bundled with
>> vmd 1.9.3 should be 1.6.4 I believe, so I'm worried about where that
>> message about psfgen 2.0 is coming from.
>> Thanks,
>> Peter
>> On Tue, Jul 28, 2020 at 10:11 AM Manas Kohli <>
>> wrote:
>>> Dear Peter,
>>> Thank you for your reply. I have a linux computer with VMD 1.9.3 and
>>> another freshly built computer with VMD 1.9.4 (new version of Ubuntu also)
>>> and am getting the same error on both computers. When I execute the scripts
>>> for either the POPC or M2 channel example, it initially runs until the fix
>>> tcl script (typically until step 2). After that during the solvate step I
>>> get:
>>> version conflict for package "psfgen": have 2.0, need 1.4
>>> and a similar error for solvate. I changed the version in the script so
>>> the tcl script can execute but after this I end up getting the error for
>>> the D3B bead I described before. For example:
>>> ATOM 20 D3B POPCL 2 0.000 0.000 0.000 -1.00 0.00
>>> L1
>>> I'm also happy to send over any files with these so you could take a
>>> look but I have changed nothing.
>>> Thanks and Best Regards,
>>> Manas
>>> On Tue, Jul 28, 2020 at 4:25 AM Peter Freddolino <>
>>> wrote:
>>>> To get started looking at this, can you please let us know the version
>>>> of VMD that you are using? I am not able to reproduce the problem; simply
>>>> downloading the tutorial files and running the listed commands from the
>>>> tutorial for the M2 example works fine for me using VMD 1.9.3 (64 bit
>>>> linux).
>>>> Thanks,
>>>> Peter
>>>> On Mon, Jul 27, 2020 at 7:54 AM Manas Kohli <>
>>>> wrote:
>>>>> Dear all,
>>>>> I have been having troubles running a CG simulation for the examples
>>>>> provided for the residue based CG tutorial. The initial lipoprotein example
>>>>> works fine but for the M2 channel and POPC membrane patch I am unable to
>>>>> run them. Specifically the issue seems to lie with coarse graining the POPC
>>>>> lipid itself to do with the D3B bead. psfgen can't seem to assign an atom
>>>>> or coordinates for this bead and as a result it sits at position 0. This
>>>>> then results in the following error message when running the simulation:
>>>>> ERROR: Atom 1853 velocity is 4.59989e+15 1.18741e+14 7.46308e+13
>>>>> (limit is 1200, atom 81 of 84 on patch 47 pe 0)
>>>>> ERROR: Atom 1879 velocity is 3.4509e+15 -3.417e+15 1.32714e+13 (limit
>>>>> is 1200, atom 82 of 84 on patch 47 pe 0)
>>>>> ERROR: Atom 1892 velocity is -8.05079e+15 3.29826e+15 -8.79022e+13
>>>>> (limit is 1200, atom 83 of 84 on patch 47 pe 0)
>>>>> ERROR: Atoms moving too fast; simulation has become unstable (3 atoms
>>>>> on patch 47 pe 0).
>>>>> Warning: Low global exclusion count! (1536 vs 1752) System unstable
>>>>> or pairlistdist or cutoff too small.
>>>>> ERROR: Exiting prematurely; see error messages above.
>>>>> Any other membrane protein embedded in POPC seems to fail because of
>>>>> the same reason. I'm wondering how exactly to solve this issue with POPC
>>>>> for CG simulations.