Why not simply scale the two data sets, and subtract corresponding Fs
from one another and then
calculate a map from those Fs.
If you want to an error-weighted map, you should also perform
error propagation on your sigF. Assuming that the two errors are
independent of one another
the formula for doing so would be sigmaFa-Fb = sqrt(sigFa^2 + sigFb^2).
My advice would be to use omit phases for this map to avoid biasing
your difference map
by model phases. In other words calculate your phases
from a model, in which you have removed atoms that show significant
peaks
in a preliminary difference map.
Provided you use the same phases for both maps (which you should to
avoid bias)
subtracting the two maps structure factor by structure factor and
subtracting them
pixel by pixel is mathematically equivalent.
Cheers,
Ulrich
On Mar 22, 2007, at 5:06 PM, Qing Xie wrote:
Hi,
I'm trying to get the difference map by subtracting the native
electron density map from the complex electron density map. MAPMASK
has a function of ADD/MULT, but I don't know how to use it?
Any other ways to attack this problem in real space?
Thanks in advance,
Qing
