On Fri, Sep 5, 2014 at 11:20 AM, Wei Shi <wei.shi118@gmail.com> wrote:

Thank you, Nat! So, to process the data in space group P212121 in XDS generates the difference in the map compared to that processed in P222?

I have no idea - I wouldn't necessarily expect this to happen, but that's what your description sounded like. In theory P222 and P212121 are processed and scaled almost identically, but it's possible that you ended up with a slightly different indexing solution.

Then, you would suggest to use the cns.hkl file from XDS using space group P222(?), and then do molecular replacement (try all possible in same pointgroup), and then do molecular replacement using the original mtz (from XDS, space group P222).... Thank you!

I think you're making this much more complicated than necessary. As a general rule, unless you're absolutely certain about the specific space group, your original steps are exactly what I'd recommend: process in P222 and let Phaser decide. You processed in P222, obtained an MTZ file, ran Phaser on the output, it gave you a solution in P212121. Assuming this solution is correct, all you need to do is start refining/rebuilding in that space group. You do not need to reprocess your data or do anything else other than maybe convert the same MTZ file to P212121.

(Of course you need to be certain that the solution is real - just because Phaser finishes a run and writes out files does not mean it actually worked. A high TFZ score, high positive LLG, and reasonable R-free after refinement are all important.)

I am just not quite sure why there is difference in density using two different mtz file....Thank you!

This is impossible to answer using the information provided. If you send me all input files off-list along with XDS log files I can take a look.

-Nat