[gmx-users] special bond

Wed Apr 30 14:05:12 CEST 2014

On 4/30/14, 4:53 AM, Nicola Staffolani wrote:
> Hi Justin,
>
> thank you very much for caring!
>
> 
> On Tue, Apr 29, 2014 at 6:19 PM, Justin Lemkul <jalemkul at vt.edu> wrote:
>
>>
>> It would be helpful to see:
>>
> 1. The actual error messages.
>> 2. The lines of the topology to which they correspond.
>> 
>>
>
> 
> *ERROR 1 [file speptide.top, line 3224]:  No default G96Bond types*
> 
>
> 
> line 3224 of speptide.top:  *1239  1264     2 *
>
> ...
> **
>
>
>
> *ERROR 911 [file speptide.top, line 12119]:  No default G96Angle types*
> line 12119 of speptide.top:*  1412  1437  1462     2 *
>
> ...
>
>
>
> *ERROR 3590 [file speptide.top, line 15794]:  No default Proper Dih. types*
>
> line 15794 of speptide.top:  *1762  1787  1788  1763     1 *
>

To which atoms do these atoms correspond?  Presumably they're related to 
whatever special bonds you're trying to create, but it is really important that 
you tell us everything as concisely as possible to efficiently troubleshoot this.

> As you can see, it looks like the errors are generated because no gb, no
> ga and no gd are there... The gb's are looked for because I have told
> gromacs in *specbond.dat* to put* special bonds *between *those* couple of
> atoms - I mean, the couple of atoms it tries to (special) bond are exactly
> those I have requested to...
>
> The thing you should know is that the Gromos force fields work in a manner
>> that is different from basically all the others in Gromacs.  The fact that
>> "gb_99" doesn't show up in the .top file is actually not really a problem -
>> the parameters will be looked up from [bondtypes].  The explicit use of
>> gb_*, ga_*, etc. is merely a convenience.  One can always override whatever
>> is listed in [bondtypes] by writing in their own parameters in the .top
>> file.  Is grompp complaining about missing parameters (i.e. "no default
>> bondtypes") for your special bond?
>>
>
> I guess I have answered this question with my previous lines, so the
> answer is yes, grompp is complaining ... 
>

Perhaps there is something screwy going on with specbond.dat; if it's writing 
the bonds it should (in theory) be writing the corresponding parameters.  If 
it's just those three lines that are complaining, then it's simple to write in 
the g[abd]_* entries, though that is of course generally not preferable.

To determine if there is actually a bug here, please send me (off-list, since 
attachments are prohibited here) the following:

1. Your input coordinate file
2. Your modified force field files
3. Your version of specbond.dat

-Justin

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

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 601
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

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