[gmx-users] LINCS with amber99SB ?

[Justin A. Lemkul]

Yun Shi wrote:
> Hi all,
> 
> I want to do MD simulation with amber99SB force field, which I found 
> originally developed using SHAKE algorithm to constrain all bonds 
> involving hydrogen atoms. But it seems to me that SHAKE is still not 
> supported with domain decomposition in GROMACS4.5.4, and we can only use 
> LINCS for bond constraints if we want to run MD simulations in parallel.
> 
> So what should I use in simulating with amber99SB in parallel? LINCS 
> constraining allbonds? LINCS constraining hbonds only? Or just run with 
> SHAKE in one node?
> 

You should be able to use LINCS with either all-bonds or hbonds.  Running your 
simulation in serial, even for a relatively small system, will likely be 
painfully slow.

> I would be very grateful if someone could tell me the relative accuracy 
> (or if they are all good enough) and computational costs of these 
> options above.
> 

Either should be fine.  LINCS is actually somewhat more stable than SHAKE.

> And maybe I should ask this question in the developers mailing list, but 
> would GROMACS support SHAKE in parallel in the near future, like in a 
> 4.5.5 version?
> 

There will not be any new features in 4.5.5, and the 4.6 list of new features is 
already basically agreed upon.  Gromacs is undergoing major code changes that 
limit the amount of new features that will come out soon.  You can file a 
feature request on redmine.gromacs.org, but do not expect to see any activity on 
the issue until probably version 5.0.

-Justin

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

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

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