Calculating rate constants (GT and fastest path)

So, what do you do with the database of stationary points when you've finished the sampling part of the process? Usually, calculate rate constants (of various flavours) and examine mechanistically discrete paths with significant rate constants (i.e., make nice movies!).

Additionally, it's often helpful as you're building up the database to monitor the various rate constants that pathsample produces, by running separate one-processor pathsample jobs as described below. For safety, there should only be one running pathsample job in a directory at a time, though, so do this on a copy or between restarts of the sampling jobs.

It's also possible to ask pathsample to calculate the GT rate constant every n cycles of a sampling run (put n as the second integer on the GT keyword line in the pathdata input file), but this can be prohibitively expensive if n is small and the database is large.

pathdata should contain the usual keywords that set up information about the system: NATOMS, PERMDIST, etc., as used in the database-generating runs; however, remember to remove or comment out the sampling commands: SHORTCUT, FREEPAIRS, etc.

The following keywords must be present:

CYCLES 0
TEMPERATURE x

where x is the reduced temperature, i.e. for CHARMM, it's the energy in units of kcal/mol corresponding to the desired temperature: 0.592 kcal/mol for 298 K. The CYCLES 0 ensures that no further sampling is performed. The temperature need not be the same as that used for the sampling runs; MAXTSENERGY can also changed (or included) in the rate constant calculations.

Headline text

The different flavours of rate constant calculation...

Find the largest contribution to the phenomenological DPS steady-state rate constant (equation ?? in Wales02)

DIJKSTRA y

More to come...

--jmc49 21:45, 2 July 2008 (BST)