Hi Mark,

I tried the "Isomorphous difference map” action and specified a file for input 1 containing F1+, F1– at one wavelength, and the file input 2 containing F2+,F2– at another wavelength. I selected the appropriate Labels for the columns containing the anomalous data.  I then specified in "Advanced Options” to use anomalous differences instead of amplitudes.  Will this procedure use the following coefficients for the difference Fourier?

( F1(+) – F1(–) ) – ( F2(+) – F2(–) ), alpha(calc)

Can someone confirm if this is indeed what is calculated?

no, not quite.. Given two data sets, Fobs1 and Fobs2, and a phase source (eg., atomic model), it computes

(Fobs1-scale*Fobs2, Phase)

Fourier map. Both data sets are assumed to be collected from isomorphous crystals (same space groups, unit cells can deviate up to a small tolerance).

Phenix treats Fobs(+) and Fobs(-) as individual data points, so in the map above they will participate as, for example

(Fobs1(+)-scale*Fobs2(+), Phase)

(Fobs1(-)-scale*Fobs2(-), Phase)

Pavel