Can I combine selenoMet data and MR model to solve the phase problem?

From: Feng Guo
Date: 15 November 2011 21:56


Hi, there,

Maybe someone asked this question before, but I couldn't find it in the archive.

 We use the native data to do molecular replacement before, but only part of the model fit the density. After collect  a new set of selenoMet data, we try to use it to solve the phase, it solve some of the phase problem other than the MR, but still not complete. Is there anyway that I can somehow combine the two phases together? Thank you.

Best,

Feng

----------
From: Bosch, Juergen

Sharp will do that. Or you can simply handpick them when looking at your anomalous map using the MR phases. With those sites you can use a variety of programs to refine the SeMet sites and hopefully get better phases.

Are you sure your MR model is correctly placed ?

Have you tried conventional methods such as Shelx C/D/E to solve your SeMet SAD (?) data ?

Jürgen

......................

Jürgen Bosch

http://web.mac.com/bosch_lab/

----------
From: Jacob Keller


How about Phaser with partial MR structure?

JPK

--


----------
From: Jon Schuermann


Phenix does this (http://www.phenix-online.org/documentation/autosol.htm#anch125)

Phaser-EP (http://www.phaser.cimr.cam.ac.uk/index.php/Experimental_Phasing#Combined_MR_and_SAD_Phasing)

and SHELX using an experimental version you have to get from George Sheldrick.

Jon

--
Jonathan P. Schuermann, Ph. D.

----------
From: Jon Schuermann


I forgot to mention... there are two ways of calculating the phases in Phenix, try both ways because the maps can look very different.

----------
From: Frederic VELLIEUX

Hi,

This hasn't been mentioned by the people who have answered so far so here we go: your molecular replacement solution and your SeMet solution to the phase problem
are not necessarily using the same origin. There is a ccp4 program (is it phistats ? and there may be other programs around) that can deal with this.

Fred.

----------
From: Navraj S. Pannu

Hi Feng,

If your native and selenomet data are isomorphous, I would use your native data, your selenomet data and your (partially correct) MR solution and your heavy atoms altogether in refmac.  At the moment, the refmac gui does not support SIRAS refinement, but you can do it via script - Pavol Skubak, Garib or I can show you.

Otherwise, you can try just the SAD selenomet data.

For SIRAS, check out this paper for more info:

Skubak et al (2009) Acta Cryst D, 1051-1061.

For SAD, check out

Skubak et al (2004) Acta Cryst D, 2196-2201

Skubak et al (2005) Acta Cryst D, 1626-1635.

A nice test case using the SAD function in refmac (at 3.8 Angstroms) on RNA Polymerase II is described here:

Meyer et al (2009) JBC 284(19), 12933-12939.

Sorry for the obvious self-promotion, but I woke up in a Pavel Afonine-ish type mood.

Best wishes,
Raj

----------
From: Eleanor Dodson


Yes - one solution is to use the MR phases to do a anom diff map to position the Se sites.

You need to
a) run CAD to combine the DANO/ SIGDANO columns from your data with the refmac output

b) use the fft utility to do a DANO map with PHIC and FOM and run a peak search.

Ideally you should find whacking peaks related to your previous Se sites
(may be different origin, and symmetry equivalent.


If that is so, then use phistats (Reflection utility ) to move the Se phases to the MR ones and you can do several things then.
I often just check the MR solution against the exptl phases first - correct obvious errors, delete wrong residues etc)

Then do some refinement using phases and you will get out combined phases from REFMAC..

Or use the SIRAS option #or ....


Eleanor

----------
From: Felix Frolow

Could DANO/SIGDANO be included to REFMAC output mtz automatically? I find it very useful on the various stages of refinement to observe anomalous map.

Last time I was trying, I failed, probably due to  GUI'sh addiction and abandonning line editor.

Dr Felix Frolow   
Professor of Structural Biology and Biotechnology
Department of Molecular Microbiology
and Biotechnology
Tel Aviv University 69978, Israel

Acta Crystallographica F, co-editor

----------
From: Bosch, Juergen

Yes you can with the

LABOUT command even in the GUI :-)

Jürgen 

......................

Jürgen Bosch