[gmx-users] about DMSO c-c bond and c-o bond in ffg53a6 force field

[Mark Abraham]

wang kelvin wrote:
> Thanks MARK for your reply,and your suggestion is very helpful.
> 
>  
> 
> I don't write back until mdrun ends just now.
> 
> I want to simulate DMSO property as it is cooled. So I put 500 DMSO 
> moleculers in a box. Ref_t =300  ref_P=1.
> 
> During the first 60ps ,the 500 DMSO moleculers seems to be diffusing 
> outside with box getting larger and larger.
> 
>  
> 
> Then box size become stable until 290ps from when box get smaller and 
> smaller, DMSO meoleculers form several clusters. Gradually clusters get 
> bigger and bigger ,at last there are three big moleculer blocks in box 
> with two relatively smaller blocks.
> 
>  
> 
> During 1000 -1300ps,big moleculer continue gathering and form two main 
> blocks ,one is in the upside of the box ,and another is in the underside 
> of the box. This two blocks' size keep changing ,but the number of 
> blocks never changed . the whole simulating system has already been 
> stable when 1500ps comes.
> 
>  
> 
> 300 K temperature and 1 bar pressure means DMSO is in liquid phase. So I 
> think DMSO moleculers shoud gather and form relatively big blocks. But I 
> expect there is only one block in box,and simulation results give two 
> blocks .

Thanks for the detailed description - it sounds like a classic case of 
normal behaviour under periodic boundary conditions :-) Check out 
http://wiki.gromacs.org/index.php/Periodic_Boundary_Conditions

> By the way, I have searched many DMSO literaures ,but never seen 
> information about c-c bond and c-o bonds in gmx force field . maybe I am 
> quite new ,and not good at English .

Maybe it was just done for gmx for efficiency with constraint 
algorithms, but I'm fairly sure I read about it in a paper for some DMSO 
model.

Mark