[gmx-users] All-bonds vs. H-bonds using CHARMM36

[Justin Lemkul]

On 10/31/13 2:21 PM, rajat desikan wrote:
> Hi,
> In the CHARMM36 paper (Klauda et al., JPCB 2010), only the hydrogen bonds
> are constrained for the lipid simulations using SHAKE (excerpt from the
> paper below)
>
> "Consistent for all of these simulations was the use of a 1 fs time step
> and constraining of the hydrogen atoms using the SHAKE algorithm."
>
> For a membrane-protein system, is constraining all-bonds via LINCS the
> right option while using CHARMM36?
>

Normally, as the paper states, only bonds involving H are constrained with 
CHARMM.  LINCS is a suitable replacement for SHAKE, though you can use SHAKE in 
Gromacs if you want.  LINCS is generally more robust.

> There was a mention somewhere (I forgot) that constraining all-bonds
> probably prevents alkane isomerisations in membranes, which could lower the
> melting temperature. I intend to simulate a POPC bilayer. Can someone with
> experience please shed some light on this?
>

Can't comment on this, but I doubt there is any issue if you don't constrain all 
bonds.

> P.S.: Klauda et al., has posted their .mdp for POPE in gromacs on
> lipidbook. Their .mdp constrains h-bonds
> http://lipidbook.bioch.ox.ac.uk/uploads/package/CHARMM36/48-POPE-wurl/v1/charmm_npt.mdp
>

It is very uncommon that such input files exist without specifically requesting 
them; if this is exactly what the authors used, I see no reason to deviate from 
it unless you have a demonstrably superior protocol.

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 601
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441

==================================================