[gmx-users] Re:Re: how to calculate the non-bond interaction between solvent and solvent in pure solvent system (XAvier Periole)

[Mark Abraham]

Jinyao Wang wrote:
> 	
>> From: XAvier Periole <x.periole at rug.nl>
>> Subject: Re: [gmx-users] how to calculate the non-bond interaction
>> 	between	solvent and solvent in pure solvent system
>> To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>> Message-ID: <BF11406F-B3B4-42E2-BCD0-011CE303F863 at rug.nl>
>> Content-Type: text/plain; charset=GB2312; format=flowed; delsp=yes
>>
>>> I want to calculate the non-bond interaction (LJ+Coulomb) between  
>>> solvent and solvent in pure solvent system.
>>>  After the simulation of my pure solvent system, I calculated the  
>>> interaction between solvent and solvent using the rerun option of  
>>> the "mdrun program".
>>>  First, I set the  energygrps = sol sol  in rerun.mdp file.
>>>  Then, I make the grompp program like this,
>>>  grompp -f rerun.mdp -c conf.gro -n index.ndx -p pure.top -o rerun.tpr
>>>  But this step has a fatal error: Atom 1 in multiple energy Mon.  
>>> groups (1 and 2)
>> This option can only calculate energies between non-overlapping groups.
>> That what he message is telling you: one atom can not be part of
>> more than one group.
>>
>> If you actually have a simulation of pure solvent the total non-bonded  
>> energy
>> of your system seems to be what you are looking for!	
>  Hello XAvier.
>   Thank you for your reply.
>  Because I don't know how to set the option "energygrps" in rerun.mdp file, 
>   Now I still have no ideal for how to calculate the intermolecular interaction energy in pure solvent system.
> Could you give me some details.　
>  Thanks in advance for your suggestion.　　　　　　　

Divide the total non-bonded energy by the number of molecules. In the
limits of large enough sampling and system size, this should be
constant. Whether it correlates with anything experimentally measurable
will depend on the quality of your non-bonded model.

Mark