[gmx-developers] Strange default group behavior

[Justin Lemkul]

On 8/21/12 5:11 AM, Jochen Hub wrote:
>
>
> Am 8/20/12 7:35 PM, schrieb Justin Lemkul:
>>
>>
>> On 8/20/12 1:23 PM, Jochen Hub wrote:
>>> Hi,
>>>
>>> I noticed that, since gmx 4.5, some groups are twice in the default group
>>> selection. For example, I have a pdb file with residues SOL, HEX, SOQ
>>> and NA:
>>>
>>> ATOM      1  C1  HEX     1      22.920  23.660  65.120  1.00  0.00      C
>>> ATOM   5688  MW  SOL  1182       8.140   7.030   0.760  1.00  0.00
>>> ATOM   5692  MW  SOQ  1183      21.970  25.660   4.180  1.00  0.00
>>> ATOM   5693  NA  NA   1184      10.010   5.080   4.740  1.00  0.00
>>>
>>>
>>> The default groups are:
>>>
>>>    0 System              :     4 atoms
>>>    1 Other               :     2 atoms
>>>    2 HEX                 :     1 atoms
>>>    3 SOQ                 :     1 atoms
>>>    4 NA                  :     1 atoms
>>>    5 Water               :     1 atoms
>>>    6 SOL                 :     1 atoms
>>>    7 non-Water           :     3 atoms
>>>    8 Ion                 :     1 atoms
>>>    9 HEX                 :     1 atoms
>>>   10 SOQ                 :     1 atoms
>>>   11 NA                  :     1 atoms
>>>   12 Water_and_ions      :     2 atoms
>>>
>>> The problem is now, that make_ndx and other tools throw an error if I
>>> try to
>>> select the group "HEX":
>>>
>>> Error: Multiple groups 'HEX' selected
>>>
>>> This is a very annoying behavior in scripts where different groups
>>> don't always
>>> have the same group number, so one wants so select a group by its name.
>>>
>>> Btw, the problem disappears if one removes the "NA" line from the PDB
>>> file,
>>> which is even stranger.
>>>
>>> Is this an intended behavior or should I file a redmine?
>>>
>>
>> It's probably related to this issue:
>>
>> http://redmine.gromacs.org/issues/490
>>
>> Perhaps we need to re-open that one rather than start a new issue
>> report.  I have always thought that the multiple groups was somewhat
>> inconvenient, but never had an error arise from it.  I'm assuming you're
>> only selecting one group when prompted?  The groups are redundant, but
>> in theory, you can fill any index file with junk numbers with whatever
>> names and you can select any viable group.
>
> Hi,
>
> I don't really understand "only selecting one group when prompted?" here. It
> seems that Gromacs does not allow the selection of a group using the group name
> *if* several groups have the same name.
>
> If the groups would "only" be redundant the double groups would not be a
> problem. The problem is that the group selection does not work properly this way.
>

I see the problem now.  Can you select the desired group by its number rather 
than its name?  That should be a viable workaround while waiting for a fix.

-Justin

-- 
========================================

Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================