Friday, 25 November 2011

Merging with CAD files

From: Ramanuj Banerjee
Date: 6 November 2011 05:35


Hi,

  I used CAD for merging datasets during MIR. I faced the same problem. The solution is: the datasets you are trying to merge should have different labels i.e if dataset 1 has labels: F and SigF, dataset 2 should be F_d1 and SigF_d1. Mention the labels during the cad run.

Ramanuj Banerjee
Crystallography and molecular Biology Division
Saha Institute of Nuclear Physics,
Kolkata, INDIA

----------
From: Clemens Vonrhein

That might make life easier - but a simple

 cad hklin1 one.mtz hklin2 two.mtz hklout onetwo.mtz <<e
 LABI FILE 1 E1=F  E2=SIGF
 LABO FILE 1 E1=F1 E2=SIGF1
 LABI FILE 2 E1=F  E2=SIGF
 LABO FILE 2 E1=F2 E2=SIGF2
 e

can do something similar. That is for 'merging' two MTZ files
(ie. gluing together columns), not merging two datasets into one (see
pointless/scala answers).

You can also do

 cad hklin1 one.mtz hklout one_tmp.mtz <<e
 LABI FILE 1 E1=F    E2=SIGF
 LABO FILE 1 E1=Fcmb E2=SIGFcmb

 cad hklin1 two.mtz hklout two_tmp.mtz <<e
 LABI FILE 1 E1=F    E2=SIGF
 LABO FILE 1 E1=Fcmb E2=SIGFcmb
 e

 mtzutils hklin1 one_tmp.mtz hklin2 two_tmp.mtz hklout onetwo.mtz <<e
 UNIQ
 e

which will fill up missing data in one file with the data present in
another MTZ file.

There's not a lot one can't achieve with those low-level utilities in
CCP4, but be careful of the order of doing things (both in CAD and
MTZUTILS), sorting issues etc. it can be very educational to have a
look at the input files and the output file with

 % mtzdmp some.mtz -n -1 > some.mtzdmp

to check it does exactly what you want to do.

Cheers

Clemens

--

***************************************************************
* Clemens Vonrhein, Ph.D.     vonrhein AT GlobalPhasing DOT com
*
*  Global Phasing Ltd.
*  Sheraton House, Castle Park
*  Cambridge CB3 0AX, UK
*--------------------------------------------------------------
* BUSTER Development Group      (http://www.globalphasing.com)
***************************************************************


No comments:

Post a Comment