[gmx-users] Gromos54a7 electrostatics interactions ?

[Thomas Piggot]

Hi,

Regarding the COM movement of the two leaflets, there is an option for 
gmx msd to do this (-rmcomm).

As for the combination of force fields with Berger lipids and other 
molecules, this is harder to answer and is an area that (AFAIK) hasn't 
been studied in all that much detail. Obviously you'd like everything in 
the system to have been developed in a consistent way but this isn't the 
case when you start using the Berger force field as it's such a mix of 
parameters. There is some work out there looking at different protein 
and lipid interactions with Berger (e.g. 
http://iopscience.iop.org/article/10.1088/0953-8984/18/28/S07/meta) and 
so this can probably give you some information on potential issues to 
look out for (e.g. overly strong interactions between the GROMOS/Berger 
bits). This is irrespective of whether you use a GROMOS force field for 
all of your ligand or just the non alkyl chain bit, although by having 
the chain bit as Berger you would at least have more of the system fully 
consistent.

Cheers

Tom

On 26/09/17 13:04, Sim gmx wrote:
> Hi,
>
> I tried to reply yesterday, but apparently my message was too big to be
> sent (probably because I replied within your own reply ?). I try a second
> time:
>
> Thanks for checking! Exactly, this temperature is for sure above the phase
> transition temperature of the lipid I'm studying. Except for the potential
> use of sterols discussed further, I will only use PC lipids. However, some
> of my molecules do have an alkyl chain. One point of using gromos54a7 is
> that those alkyl chains will interact with PC's alkyl chains as PC's alkyl
> chains interact between each other. What I mean is that atomtypes used for
> both lipids and molecules' alkyl chains will be identical (CH2, CH3...).
> Berger lipids use different atomtypes for the alkyl chains (LP2, LP3...).
>  From what I think I know, Berger lipids is pretty compatible with gromos'
> forcefields. So my first intention was to use berger lipids' PC together
> with gromos53a6's molecules. However, it might be better to change the
> atomtypes from the alkyl chains of my molecules so that they match the
> berger lipids alkyl chains atomtypes. It seems that they did so in this
> paper http://pubs.acs.org/doi/abs/10.1021/jp501622d
> I would be very interested in getting your opinion about this.
>
> About the removal of leaflets COM movements, I'm not sure to know how to do
> that ? "Simply" with trjconv ?
>
> I also think that Berger setup will give the fastest simulations!
>
> Thanks for the info about the condensing effect of Holtje sterol on berger
> lipids. Still not sure if I will run simulations with sterols but this will
> for sure be taken into account whenever needed.
>
> I used a shortcut, but it is true that I should try to understand why both
> forcefields don't give similar results (if it is the case).
>
> Thanks for your clear and detailed reply !

-- 
Dr Thomas Piggot
Visiting Fellow
University of Southampton, UK.