[gmx-users] Different bonding energies for two almost the same systems

[Mark Abraham]

Reza Salari wrote:
> Dear Gromacs Users,
> 
> I've been doing some calculations on methane-methane system in "bound" 
> and "unbound" states to see how the energies compare. I set up 
> everything almost the same between two system but I still get different 
> bonding energies for two system. Here is what I did:
> 
> I first created two methanes separated for 4A to make my bound state. 
> For unbound state, I increased the distance to 10A. The simulations were 
> done in vacuum, all atoms were kept frozen and all non-bonding 
> interactions were excluded (the coulombic and LJ energies are zero). So 
> basically I have two system withe same atoms, bond lengths and angles. 
> In each system, the two methanes don't "see" each other and are fixed in 
> their positions. However I still get slightly different bonding energies 
> for bound and unbound states. My question is, for this two systems, 
> shouldn't the bonding energies (bond and angle energies) be same? Or 
> more basically, isn't it expected that the potential energies for these 
> two system be same at least for the first step?
> 
> I hope I am not missing something obvious. I attached mdp, pdb and log 
> files. I also repeated the test with steep algorithm and also md 
> dyanmics but the energies are still different.

Try turning off freeze groups.

Mark