[gmx-developers] Coming on to help in GROMACS development

[Michael Shirts]

Hi, all-

My name is Michael Shirts, and I'll be starting to do some development
work on GROMACS in consulation with Erik Lindahal.  I'm currently and
I'm a new postdoc in Rich Friesner's group at Columbia.  I recently
graduated from Stanford, where I was in Vijay Pande's group, working
on Folding at Home (http://folding.stanford.edu), and worked closely with
Erik Lindahl, then a postdoc in Michael Levitt's group at Stanford, to
help implement features in GROMACS that we required in our work.  My
CV is at http://www.columbia.edu/~mrs2144/MRSCV.html -- my work with
GROMACS is almost entirely in the three papers in press or just
submitted, so you'll have to wait a little bit for that :)

I've done some work on the GROMACS code itself already -- I rewrote
the routine for the mean field dispersion correction to the
Lennard-Jones cutoff, adding the effect of the switching or shifting
function, as well as a few other bug fixes.

Based on my previous research, I've got a few other things that I'll
be working on that I'd like to put into the next couple releases of
GROMACS.  Please comment on any of these features that you find
interesting (or think that really don't need to be implemented, etc).

1) Improved free energy features, including  
     a) decoupling free energies capabilities (there are  
         problems with decoupling energies with Coluomb in 
         GROMACS currently) 
     b) output suitible for after the fact FEP and Bennett
     acceptance ratio computations, over multiple states
     c) an improved, more comprehensive tutorial for free
     energy calculations.
     d) including forces from restraints in free energy
     computations

2) Velocity Verlet integrator as an option:
     This is important for pressure control calculations, as
     the calculation of the box size at time t+dt requires the
     pressure at time t, which in turn requires both the virial    
     and the velocities at t, which is not possible with leapfrog 
     Verlet -- approximations are necessary.

3) Adding bookkeeping so that if there is a conserved    
    quantity in the simulation (energy, or enthalpy, or 
    whatever), it will be kept track of separately, in order to 
    help judge the goodness of fit.

4) Improved pressure control, adding Andersen pressure 
    control as well as Parinello-Rahman, and fixing some bugs 
    in the current P-R implementation.
 
5) Andersen temperature control, which is a useful 
    alternative to Nose-Hoover when there are highly 
    decoupled degrees of freedom, and equipartion is
    breaking down.

Thanks for your input, and I look forward to working with all of you!

Michael