On Mon, Feb 11, 2013 at 9:03 AM, 李翔 wrote:

I got a native and derivative dataset (by soaking) of my crystal and want to do MIR in phenix. I am wondering whether there is a way for me to export the difference Patterson map so I could have a better idea that whether the heavy atom is really in the crystal?

Okay, with a lot of help from Tom Terwilliger, there are two new commands (starting with build 1296 or newer, hopefully available tomorrow): cctbx.patterson_map data.mtz which will generate either a simple or an anomalous Patterson map in CCP4 format, and: cctbx.isomorphous_difference_patterson deriv.mtz native.mtz which does exactly what it sounds like. Run without arguments or with "--help" to see a full list of options. I am not 100% confident that these are entirely correct implementations - in particular, the anomalous Patterson does not perfectly agree with the equivalent map from CCP4, I suspect in part due to different filtering options - but the output is approximately what we expect. We will continue testing these when we have a chance (after we release a new version of Phenix). One other comment: depending on the heavy atom used and the wavelength, you may also have anomalous signal, so it's worth checking for that too - the first tool to try is Xtriage, but you can also look at the anomalous (single-dataset) Patterson. (Having never done one of these experiments, I don't know if there are any good reasons to use pure MIR phasing, as opposed to MIRAS.) -Nat