[gmx-developers] Position Restraints in Thermodynamic Integration

[Berk Hess]

Hi,

The option decoupling a molecule from the rest of the system. But 
restraints are not really the rest of the system. Also, in most cases 
you want the restraints to work exactly in the opposite direction. When 
the interactions are on, the restraints should be off and vice versa. I 
think that should already work against confusion.

Cheers,

Berk

On 09/28/2017 03:24 PM, Carruthers, James wrote:
> Hi,
>
> Forgive my inexperience, I thought the restraints etc. were handled in the same way as the LJ and electrostatic interactions (i.e. couple-lambda0 and 1). I've gone through my topology and added explicit B state parameters and it's working as intended. Perhaps it would be worth emphasising in the online documentation that B parameters are only automatically generated for coulombic and vdW interactions.
>
> Thanks,
> James
> ________________________________________
> From: gromacs.org_gmx-developers-bounces at maillist.sys.kth.se [gromacs.org_gmx-developers-bounces at maillist.sys.kth.se] on behalf of Berk Hess [hess at kth.se]
> Sent: 27 September 2017 21:04
> To: gmx-developers at gromacs.org
> Subject: Re: [gmx-developers] Position Restraints in Thermodynamic      Integration
>
> Hi,
>
> But the position-restraint code really seems to use the restraint lambda.
> Are you using normal position-restraints and not flat-bottomed ones?
> Have you set the position restraint force constants to zero in either
> the A or B state?
>
> Cheers,
>
> Berk
>
> On 09/27/2017 06:11 PM, Carruthers, James wrote:
>> Hi,
>>
>> I just tried it at 0.0, 0.5 and 1.0 restraint lambdas. Position restraints are fully applied in all states and DH/Dlambda is effectively zero when the foreign lambdas only have different restraint-lambdas.
>>
>> Best,
>> James
>> ________________________________________
>> From: gromacs.org_gmx-developers-bounces at maillist.sys.kth.se [gromacs.org_gmx-developers-bounces at maillist.sys.kth.se] on behalf of Berk Hess [hess at kth.se]
>> Sent: 27 September 2017 13:59
>> To: gmx-developers at gromacs.org
>> Subject: Re: [gmx-developers] Position Restraints in Thermodynamic      Integration
>>
>> Hi,
>>
>> Looking at the code position restraints are controlled through the
>> restraint lambda. Have you actually tried it?
>> This does mean the manual needs to be updated.
>>
>> Cheers,
>>
>> Berk