Re: [ccp4bb] differences between Rsym and Rmerge

Protein crystallography Programs for crystallography X-ray detectors
Basic tutorials:
Protein Peptide

CCP4bb navigation

CCP4bb <-- 2008 <-- January 2008 <-- 18 January 2008

Previous message:
Subject: Re: differences between Rsym and Rmerge
From: "R {- dot -} M {- dot -} Garavito" garavito {- at -} MSU {- dot -} EDU
Date: 2008-01-18

Next message:
Subject: Re: differences between Rsym and Rmerge
From: "Santarsiero, Bernard D {- dot -} " bds {- at -} UIC {- dot -} EDU
Date: 2008-01-18

Subject: Re: differences between Rsym and Rmerge
From: "Manfred S {- dot -} Weiss" msweiss {- at -} EMBL-HAMBURG {- dot -} DE
Date: 2008-01-18

Dear Mohd and all others,

Well, I guess it is time again to define and talk about R-factors.

The term R_sym goes back to the times, when X-ray data were
recorded by precession photography on film. Except for the central
cone, each reflection was observed only once and R_sym described
the agreement strictly between symetry-related reflections, hence
R_sym.

In diffractometer times, when reflections were measured one by one,
only for a subset of reflections (typically a plane in reciprocal
space) were symmetry-related reflections measured two times. R_sym
became R_int, where int stands for internal agreement. These
additional reflections were just measured to calculate R_int,
later on they were discarded.

Nowadays, where a diffraction data set typically consists of dozens
or hundreds of images recorded from some sort of an area detector,
multiple measurements of the same reflection AND symmetry-related
reflections are merged together to calculate the mean intensity
for a given reflection. Hence, the agreement factor becomes R_merge
or merging R-factor.

As you can see, R_merge is more general than R_sym, and is (as Kay
pointed out) the preferred term.

However, when talking about R-factor I can never refrain from
mentioning that R_merge should actually NEVER EVER be used,
because it is inherently flawed. As the redundancy or the
multiplicity of the data increases, R_merge will also increase,
although the mean intensity will be more precisely determined.
As was postulated by Kay and myself about 10 years ago, R_merge
should be replaced by a redundancy-independent merging R-factor
(termed R_rim or R_meas). Unfortunately, only SCALA and XDS
produce this R-factor, SCALEPACK does not (not yet, I hope -->
pun to Dallas). If you want to calculate R_rim or R_meas based
on scaled but unmerged data, I have my own program, which you
can download from my web site, as does Kay.

I hope this clarifies things.

Cheers, Manfred.

********************************************************************
* *
* Dr. Manfred S. Weiss *
* *
* Team Leader *
* *
* EMBL Hamburg Outstation Fon: +49-40-89902-170 *
* c/o DESY, Notkestr. 85 Fax: +49-40-89902-149 *
* D-22603 Hamburg Email: msweiss@embl-hamburg.de *
* GERMANY Web: www.embl-hamburg.de/~msweiss/ *
* *
********************************************************************

On Fri, 18 Jan 2008, Salameh, Mohd A., Ph.D. wrote:

> Hi everybody!
> I will appreciate it if anybody can clarify to me the differences
> between Rmerge and Rsym. Many thanks, M
>
> ****************************************************
> Mohammed A. Salameh, Ph.D.
> Mayo Clinic Cancer Center
> Griffin Cancer Research Building
> 4500 San Pablo Road
> Jacksonville, FL 32224
> Tel:(904) 953-0046
> Fax:(904) 953-0277
> salameh.mohd@mayo.edu
> ****************************************************
>
>
>

CCP4bb navigation

CCP4bb <-- 2008 <-- January 2008 <-- 18 January 2008

Previous message:
Subject: Re: differences between Rsym and Rmerge
From: "R {- dot -} M {- dot -} Garavito" garavito {- at -} MSU {- dot -} EDU
Date: 2008-01-18

Next message:
Subject: Re: differences between Rsym and Rmerge
From: "Santarsiero, Bernard D {- dot -} " bds {- at -} UIC {- dot -} EDU
Date: 2008-01-18