Hi,
> I have calculated the RMS by subtracting the RMS of the ligand in water
> from that of the complexed ligand. This was then fed into the script on
Do you mean that you subtracted the deltaGs of water vs. complexed?
Subtracting the rms-deviations of the two individual results is not
usefull at all. To propagate the error you would rather add them to each
other.
> the last page of the tutorial. Is this the correct way to calculate the
> RMS? Would better error values be obtained by increasing the number of
> lambda points sampled, or by increasing the equilibration time?
Normally adding lambda-points or simulating longer will increase your
statistical certainty, i.e. decrease the error. But note that my scripts
only give you standard deviations for each step's result, which are not
identical to errors. Figuring out what the true statistical error from a
TI calculation is is not that easy at all.
> My input files for the complex in step2 are as follows:
[...]
> crgmask=':158.H14',
> scmask=':158.H14',
[...]
> crgmask=':158.CT7,H15,H16,H7',
> scmask=':158.CT7,H15,H16,H7',
[...]
> crgmask=':158.H14',
> scmask=':158.H14',
Why are your crgmask and scmask different between the sequential steps?
This is equivalent to changing the potential function after each run,
which makes your equilibration invalid.
> Total DV/DL:-0.011+--0.0715
> Total DV/DL:-0.8995+--0.242
> Total DV/DL:-0.198+--0.122
So these are the differences between the water and complexed steps? I
normally compute a total transformation energy for each step and finally
subtract them from each other, but the alternative approach (compute
deltadeltaGs for each step and sum them up) works as well.
> Total free energy diff is -1.1085
The value may be reasonable, but make sure that you really simulated the
change that you think you are simulating. Your errors are fairly small, so
check carefully how you computed them. It is quite rare to get errors
below 0.5 kcal/mol in most TI transformations.
Kind Regards,
Thomas
Dr. Thomas Steinbrecher
BioMaps Institute
Rutgers University
610 Taylor Rd.
Piscataway, NJ 08854
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Feb 09 2010 - 23:30:02 PST
