VMD-L Mailing List

From: Peter Freddolino (petefred_at_ks.uiuc.edu)
Date: Tue Nov 10 2009 - 23:24:02 CST

Hi Luis,
I believe that what you're running into now is not a bug, but a feature.
Please let me explain.

There are two ways in which the autopsf gui is intended to be used.
Either you go through the dialogue boxes in order, step by step,
splitting the chains, inspecting them, and so on, *or* at the beginning
you click "I'm feeling lucky" and autopsf tries to guess what you wanted
to do. So you're creating a bit of extra work for yourself to go through
all of the steps and then click "I'm feeling lucky" -- really, you
should just at that point click "Create chains" and then finish the psf
creation manually.

Here is the catch, though -- if you walk through the autopsf gui step by
step, it'll present you with all of the chains that it finds, and you
can decide which ones you want. However, if you click "I'm feeling
lucky", then its behavior changes -- if it sees protein and/or nucleic
acid, it generates a "clean" structure, devoid of any
non-protein/nucleic acid residues present in the input structure. This
is a desirable behavior in many cases where one is doing high throughput
structure building, and thus is how "I'm feeling lucky" and the text
mode interface behave (the text interface allows this behavior to be
overridden with explicit flags specifying the segments to be
considered). Since you want to keep non-protein cofactors, though, you
should work through the gui step by step rather than clicking "I'm
feeling lucky".

Please let me know if you have further questions.

Best,
Peter

Luis Agullo (LAB) wrote:
> Of course! I include one of the pdb used attached to this mail.
>
> The process is simple, because I use most of the automatic options found
> in autopsfgen: I open autopsfgen, load the specific stream file for ATP in
> addition to the standard topology file, select 'Guess and Split Chains'
> and afterwards I 'Feel Lucky' and obtain the final autopsf file.
>
> Using the same procedure, the previous version maintains Mg2+ in the
> output files.
>
> IMPORTANT NOTES:
> - I use VMD 1.8.7 for Windows!
> - I observed a difference between VMDs when splitting chains (previous
> version of VMD recognize the 2 atoms of Mg2+ in positions 6018 and 6019 of
> the pdb file; however, the present version found 2 atoms of Mg2+, but at
> positions 6018-12037!?; I attached 2 images to show these different
> results).
>
> Luis
>
>
>
>
>> Hi Luis,
>> could you please send me your input pdb and give me an exact set of
>> steps for reproducing the problem?
>> Thanks,
>> Peter
>>
>> Luis Agullo (LAB) wrote:
>>
>>> Hello,
>>>
>>> I have problems with the last version of VMD (VMD 1.8.7). When using
>>> autopsfgen with a file including Mg2+ (or Ca2+) cation is excluded from
>>> the output file. This does not take place with the previous of VMD
>>> version
>>> using the same files. I am using the new version for Windows, I have not
>>> been able to check this in Linux version.
>>>
>>> Cheers,
>>> Luis
>>>
>>>
>>>
>>
>>
>>
>>
>> ------------------------------------------------------------------------
>>
>>
>> ------------------------------------------------------------------------
>>