[gmx-developers] Lambda dynamics

[Shirts, Michael R. (mrs5pt)]

> That sounds great!

Excellent!

> I think some time ago Thomas already left some comments on
>http://redmine.gromacs.org/issues/1332

Would this be the proper thread for further discussions?

I think so, especially for the multiple end states approach.


> Right. At some point we realized that PME is not very suited when it
>comes to calculating multiple alternative states (since one needs to do
>the communication-intense transforms for each alternative charge grid) so
>we started to work on providing a fast multipole method as an alternative
>to PME for such scenarios.

> This is in progress and the topology/free energy treatment can be worked
>on more or less independently.


Great.  It is likely that we will be working more on the multiple end
state approach, as well as Niels Hansen, who approached me about putting
it into gromacs more.  It will depend a bit on which students join my lab,
which will be decided in the next week or two.  We are particularly
interested in constant pH / multiple tautomer states.

Note that we have been working on a few ways of improving how to handle
free energies for multiple end states. See:
http://pubs.acs.org/doi/abs/10.1021/ct4009188.  This is similar to what
was proposed in the Grubmuller group, but should be more flexible and
computationally efficient to implement.   We have a paper on including
electrostatics that is about ready to share.

Best,
~~~~~~~~~~~~
Michael Shirts
Associate Professor
Department of Chemical Engineering
University of Virginia
michael.shirts at virginia.edu
(434)-243-1821