Yeah, I really dont need to use the softcore potential.
The carbon atom and oxygen atom are in the formaldehye.
And I found that the distance between them reached about 18 angstroms.
In my simulations, the box size is 15 angstroms and the QM cutoff is 8 angstroms.
So, in the above described situation, why did only the lamda=0.005 simulation fail?
Pengfei Li
On Fri, Mar 11, 2016, lipengfei_mail wrote:
> I performed a series of simulations with a potential that
> is a mixture of pure MM and QM/MM (i.e. of the type V(¦Ë)=
> (1-¦Ë)*V_MM + ¦Ë*V_QM/MM ) using sander, where ¦Ë is an adjustable
> parameter used in standard free-energy perturbations in AMBER
> (¦Ë=0.005,0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,0.995 respectively in my
> simulations).
> The initial state is a pure MM state of a formaldehyde molecule solvated
> by the TIP3P waters, and the end state is a QM state of a formaldehyde
> molecule solvated by the same TIP3P waters. In other words, both states
> are described by the same prmtop and inpcrd files. Only the both mdin
> files have the difference.
> The softcore potential is used to transform the pure MM state of the
> formaldehyde molecule to the QM state.
Why do you need or want a softcore potential? It sounds like you have the
same atoms in the beginning and ending states. If nothing is "appearing" or
"disappearing", there should be no need for softcores. What did you set for
your softcore mask?
> However, the lamda=0.005 simulation is not carried on. The mdout file gives that
> "SANDER BOMB in subroutine QM_CHECK_PERIODIC<qm_mm.f>
> QM region + cutoff larger than box
> cannot continue, need larger box".
> I solvate the formaldehyde molecule with a 15 angstroms TIP3P box and
> set up cutoff=8,
> so I visualize the structure of solute and find that the distance
> between the carbon atom and the oxygen atom reaches about 18 angstroms.
Not sure what carbon atom and oxygen atom you are referring to: are they both
in the formaldehye? What is your box size? What is your QM cutoff?
Hard to tell what happened, but the above ideas should give you some ideas for
debugging.
...dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Mar 11 2016 - 23:00:03 PST
