It may be time for our annual I/sigmaI discussion.
Please note that <I/sigmaI> is what the RCSB expects from you and it is
generally lower than <I>/<sigmaI>. Some packages do not output <I/sigmaI>
in an obvious place for you to put in your Table 1. :)
On Wed, 6 Oct 2010, Ed Pozharski wrote:
> You don't need twinning to invalidate the Rmerge as a criterion for the
> resolution cutoff, there are other reasons why you should use I/sigma
> instead. If you process data all the way to 3A, what's the I/sigma in
> the highest resolution shell?
>
> On Wed, 2010-10-06 at 11:28 +0200, fulvio saccoccia wrote:
>> Dear all,
>> I have a data set collected at 3A resolution. I processed the data but I
>> had to cut the resolution at 3.6A for the high value of Rmerge (at 3.6A
>> it is 0.18). After scaling and MR I realized that my data were twinned.
>> This is my question: can I reprocess all the data set using all the
>> reflections up to 3A resolution even if the Rmerge is very high, knowing
>> that data are twinned? and also, is the Rmerge invalidate by the twinning?
|