From: Gumbart, JC (gumbart_at_physics.gatech.edu)
Date: Sun Aug 08 2021 - 07:40:49 CDT

"rigidBonds all” is indeed the recommended approach. The benefit of HMR is related to the angle fluctuations, which are slowed down by virtue of the heavier hydrogens.

TIP3P waters are always completely rigid - no angle fluctuations either, which is why they don’t need HMR. This is why you sometimes see an H-H bond as a cheap way to enforce this rigidity.

Best,
JC

On Aug 7, 2021, at 7:53 PM, René Hafner TUK <hamburge_at_physik.uni-kl.de<mailto:hamburge_at_physik.uni-kl.de>> wrote:


Dear JC,

 may you step in here? :)

   Though do I now understand correctly that:

            HMR can solve the limitations of the RATTLE algorithm (involving masses and not working good enough with low masses like hydrogen masses and timesteps >2fs),

        but the oscillations frequencies of hydrogen bonds (non solvent) would still be too high to be captured correctly and we therefore need "rigidBonds all"?

Best

René

On 8/7/2021 10:17 PM, Geist, Norman wrote:
Rigidbonds has limitations, with HMR it allows a timestep of 4fs, otherwise only 2fs. Sry for being short, but am typing on the phone right now.

Am Samstag, den 07-08-2021 um 21:14 schrieb René Hafner TUK:

Hi Norman,

    what would the phyiscal meaning of HMR be, if we keep all non-solvent bonds rigid anyway?

Best

René

On 8/7/2021 8:43 PM, Geist, Norman wrote:
No, you should keep rigid bonds to all. What timestep do you use? Have you checked the PSF contains rescaled masses? (easy to see on hydrogen atoms having a mass of 3)

Im using it for all my simulation without any problems. Don't scale the water, only solute.

Bests
Norman Geist

Am Samstag, den 07-08-2021 um 11:20 schrieb René Hafner TUK:

Hi Geo,

    When using HMR you need to apply rigidBonds to water only. The whole point of repartitioning masses is to reduce fast oscillations frequencies (of hydrogen bonds) you would not allow before due to rigidBonds and with HMR let them oscilliate less fast (which allows higher timestep).

Kind regards

René


On 8/6/2021 8:52 PM, Ropon-Palacios G. wrote:
Dear user,

I’m using HMR for a protein with glycosilation (N-glycosilation), make HMR using function implement into psfgen 2.0,
I’ve check rigth patch but when run md simulation (minimization, equlibration good), get error in rattle algorithm to an hydrogen atoms into atom N involved into glycosilation, as can fix it?

Please help me!.


Best,

Geo.



--
--
Dipl.-Phys. René Hafner
TU Kaiserslautern
Germany

--
--
Dipl.-Phys. René Hafner
TU Kaiserslautern
Germany

--
--
Dipl.-Phys. René Hafner
TU Kaiserslautern
Germany