VMD-L Mailing List

From: Low Soo Mei (lowsm_at_bii.a-star.edu.sg)
Date: Mon Apr 07 2008 - 21:02:37 CDT

Hi, will do, over private email.

Cheers,
Soo Mei

John Stone wrote:
> Hi,
> Can you send me the exact structure files and commands needed to
> replicate your problem? We had asked Cesar to provide this information
> previously, but never got any reply, so we were never able to reproduce
> the problem. (and thus it didn't get fixed)
> The only bugs that get fixed are the ones that we're able to reproduce.
>
> From the sound of it, this is actually an issue with psfgen and not VMD,
> but we'll still need the structure files and command sequence you used
> that let to the trouble.
>
> Cheers,
> John Stone
> vmd_at_ks.uiuc.edu
>
> On Mon, Apr 07, 2008 at 06:31:52PM +0800, Low Soo Mei wrote:
>
>> Hi VMD users,
>>
>> I have had the same problem as Cesar described here
>> http://www.ks.uiuc.edu/Research/vmd/mailing_list/vmd-l/8395.html in Dec
>> 2006, but my VMD version is 1.8.6 for Linux OpenGL, so it appears that
>> the bug reported there by Cesar has not been fixed yet...
>>
>> In any case I have resolved the problem of the extra X altering the
>> resid positions the same way as Cesar did, by hand-deleting the resname
>> X column. Much thanks to him and to the list for being available online
>> to Google!
>>
>> Briefly, my system had 6831 waters and I could not get the psf generated
>> with the correct number of waters by either AutoPSF or by psfgen. I
>> then thought the waters might be too numerous to be taken as one segment
>> and remembered seeing something about using only 999 waters per segment,
>> so I hand-separated them into segments of 999 waters each, giving each
>> segment a different segname. Then I used VMD atomselect to write out
>> each of these segments to their own pdb file each. And thereafter I
>> tried psfgen again. The results are: Psfgen was only able to read in
>> the first segment of 999 waters and then thereafter truncated all my
>> remaining six 999-water segments to read only 100 or 101 waters each
>> when building the segments. Segment-building, coordpdb, and
>> writepsf/writepdb went smoothly for all segments with no errors, aside
>> from the 100 or 101 waters seen in the tkconsole. When I opened the
>> original 999-water segment pdb files I had written out by VMD
>> atomselect, all had 999 waters and an X column tacked on to the resname
>> by VMD. When I opened the final erroneous pdb file, as described the
>> first 999-water segment had all waters present but the following six
>> segments had only 100 or 101 waters present, and all seven segments
>> still had the X column as well. I noticed the 2nd digit of all the
>> 4-digit resids had been truncated such that e.g. my 2014th residue of
>> water was only "214". And I had lots of waters missing because of this
>> resid issue.
>>
>> I then went back to my segment pdb files that VMD atomselect had written
>> out, and hand-deleted the X column tacked on to the resname by VMD. I
>> tried psfgen again and this time all my waters were present in the final
>> psf and pdb. I left the X column untouched throughout in the other
>> components of my system (the lipids, the ions) and no complications
>> ensued from that.
>>
>> Perhaps this email is too long, but it was initially very frustrating
>> wondering why my waters had disappeared somewhere between segment
>> building and the final psf/pdb, in the innards of psfgen. I had spent a
>> long time reading through various VMD and NAMD mails on the mailing list
>> before finding Cesar's that helped me. The mails regarding the solvate
>> and autoionize plugins were not applicable to my system because it was
>> already nicely solvated and ionized, previously created by someone else
>> :) I simply needed the psf from it. I hope this email has succeeded in
>> having all the relevant keywords implanted in it :) such that I wish it
>> may help others with the same problem.
>>
>> Thank you for your time and kind attention.
>>
>> Cheers,
>> Soo Mei
>>
>
>