[gmx-users] CNT force field and simulation

[Alex]

Also, I would really suggest that you read the Gromacs manual on the
subject of topology. This is extremely useful, because there is a
philosophical clash between biomolecular simulations and solid-state
simulations in terms of the concept of connectivity. You are effectively
using a tool that is broadly inappropriate for solid-state systems. So,
understanding the concept of topology in GMX-like packages is extremely
useful. Gromacs has a manual of outstanding quality, and it should become
your Holy Bible. :)

Alex


On Mon, Apr 13, 2015 at 6:31 PM, Alex <nedomacho at gmail.com> wrote:

> Hi Pierre,
>
> Two things:
>
> 1. rtp entries are pretty useless for crystal structures. Biochemical
> simulations use molecules with a precise atomic composition, so an rtp
> template makes sense. Not for crystals. Your rtp entry contains useless
> information and the only thing in your coordinates that can fit that
> template is one atom, which is already generous. So, you get one atom in
> your GRO. ;)
>
> 2. Andrea isn't using pdb2gmx, which relies on templates. I honestly have
> no idea why he created that rtp entry, so maybe other users will clarify
> that. Andrea uses x2top, which will look up data in ffbonded, ffnonbonded,
> and in atomname to type translator. This is the right way to treat crystals
> in GMX. Moreover, you will have to do that every time you have a new set of
> coordinates. Use what he uses and you won't get any new GRO files. All you
> will get is a topology to go with your original PDB -- and that is all you
> need to run a simulation.
>
> Alex
>
>
>
>
> On Mon, Apr 13, 2015 at 5:49 PM, Pierre Bertin <bertinp71 at gmail.com>
> wrote:
>
>> Hi Alex,
>>
>> All right, I red all this weekend and I know how to modify a force field
>> to add residues (I hope... ahah). I modified OPLSAA force field because I
>> want to add a protein to my CNT.
>>
>> So, I have edited these files with Andrea Minoia datas (and yours) :
>>
>> *aminoacids.rtp* :
>>
>> [ CNT ]
>>
>>   39  [ atoms ]
>>
>>   40      SWT    SWT    0     2
>>
>>   41      SWT    SWT    0     3
>>
>>   42  [ bonds ]
>>
>>   43    SWT  SWT  SWT
>>
>>
>>   44    SWT  SWT
>>
>> I added a new residue named CNT. 2 or 3 bonds, I do not include hydrogen
>> to start. So a carbon can have 2 (extremities) or 3 bonds (internal).
>>
>>
>>
>> *atomname2type.n2t :*
>>
>>  24 C    SWT         0      12.011  2    SWT 0.142   SWT 0.142
>>
>>  25 C    SWT         0      12.011  3    SWT 0.142   SWT 0.142   SWT
>> 0.142
>>
>> Here, conversion of C atoms from CNT.pdb file to SWT new atom type.
>>
>>
>> *atomtypes.atp :*
>>
>> 823  SWT        12.01100  ;CNT
>>
>> Here, define the new atom type : SWT with the carbon mass.
>>
>>
>> *ffbonded.itp : *
>>
>> [ bonds ]
>>
>> SWT   SWT     1    0.1420    420420.0   ; CNT
>> ​SWT SWT bond is a bond between 2 carbons in CNT. I have used your data
>> Alex. ​
>>
>> [ angletype ]
>>
>> SWT    SWT    SWT     1   120.000    659.346   ; CNT
>> Angle type between the 3 atoms of carbon with datas from you email and
>> the same in dihedral with :
>>
>> [ dihedraltypes ]
>>
>> SWT    SWT    SWT    SWT     3     17.30770   0.00000 -17.30770   0.00000
>>   0.00000   0.00000      ; CNT
>>
>> ​I have tried with SWT, with C and in n2t with a reference to oplsaa
>> carbon from benzene.
>>
>> This works, I can build the .gro file with :
>> gmx pdb2gmx -f cnt_gro.pdb -o cnt_gro.gro -p cnt_gro.top
>> And then I select OPLSAA force field and TIP3P for water.
>>
>> Problem : My cnt_gro.gro has only ONE atom of carbon. And I do not have
>> error message. What did I do wrong?
>>
>> Thanks you a lot for your consideration and time. I attach the pdb file. ​
>>
>>
>>
>>
>> *Pierre BERTIN*
>> Bioinformatics and Biostatistics student (MSc),
>> University Paris-Sud XI, ORSAY.
>>
>>
>> 2015-04-11 14:31 GMT-06:00 Alex <nedomacho at gmail.com>:
>>
>>>  Great, glad it's working.
>>>
>>>
>>> I think there was a wrong bond length somewhere in Andrea's tutorial,
>>> not sure where exactly (I could be wrong). Look at everything carefully and
>>> make sure the bond lengths are at 0.142 nm, because this is what the
>>> generator is doing.
>>>
>>>
>>> Have fun!
>>>
>>>
>>> Alex
>>>
>>>
>>>
>>>   >
>>>
>>> ​Thank you a lot, I did this and exactly as you are saying I do not have
>>> the message.
>>>
>>>
>>> I use the link posted in my first message : ​(
>>> http://www.nanotube.msu.edu/tubeASP/)
>>>
>>> And then, I converted with Babel.
>>>
>>>
>>> But I removed CONECT statements. And now I am modifying files as you
>>> said and trying to run the tutorial.
>>>
>>>
>>> Thank you a lot Alex!
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>
>