On Mon, Sep 8, 2014 at 9:58 AM, CPMAS Chen <cpmasmit@gmail.com> wrote:

A difference map is used to identify whether there is anything not-modeled, say some ligands, ions. Then when I generate anomalous difference map, I should not put the ligand(which contains Br) in the model, right? Or in phenix, after I put the Br-ligand in the model, I should not see the anomalous difference density at the site, right?

An "anomalous difference map" is a map of the anomalous differences DANO (Fobs(+) - Fobs(-)). It's independent of Fcalc, so it doesn't matter if you put the ligand in or not. An "anomalous residual map", or the log-likelihood gradient map used by Phaser (and also available from phenix.refine/phenix.maps), uses DANOobs - DANOcalc. For these, what really matters is whether you've refined anomalous scattering for the atom(s) in question. I've uploaded some slides showing practical examples:

http://cci.lbl.gov/~nat/slides/llgmaps.pdf

Regarding Phil's comment, it probably isn't having much of an effect on phases (and certainly very little risk of bias), but it will have a significant impact on amplitudes and resulting peak heights.

-Nat