[gmx-users] tpr without top-file

[Peter Zoon]

I've succesfully made a topology with x2top, but I'm still not able to 
get a tpr file from this topfile.
While running grompp a get a lot of warnings like this:
      WARNING 385 [file "output.top", line 1205]:
      No default Angle types, using zeroes
When I ignore the warnings finally grompp crashes on
ERROR: One of the box vectors is shorter than twice the cut-off length. 
Increase the box size or decrease rcoulomb.

But increasing the box size (I started from a pdb, which has no box 
size? later converted to gro) or decreasing the rcoulomb in the mdp 
doesn't seem to have any effect.

Any suggestions apart from the ones already mentioned?

Peter

Anton Feenstra wrote:

> Peter Zoon wrote:
>
>> sounds like a good plan, but would this also work for a somewhat more 
>> complicated system?
>> my system looks more or less like this:
>>
>> thread= napthalimide-(CH2)12-succinamide-CH2-C-(phenyl)2
>>
>> the macrocycle consists of a coupleof phenyl rings bonded together 
>> via amide groups.
>>
>> I think most of the atoms are in ffgmx.n2t so I'll give it a go.
>
>
> Alternatively, it shouldn't be a problem to write your own program (or 
> perl
> or awk script), that generates a dummy topology, with the proper atom 
> names
> (which you can read from the .gro file) and only bonds for the 
> connectivity.
> For the bonds, you'll probably want to parse those out of the topology 
> file
> you used in your 'other' md program. You'll need to add in fake atom 
> types
> and fake parameters for the bonds as well, since grompp requires them.
>
> I once wrote an amber2gromacs topology converting script (in awk...), 
> which
> you may want to use as a template. It has the code to write topology 
> files.
> It should be in the contributions section, otherwise it'll be in the mail
> list archives somewhere. If that fails, ask me.
>
>