Dear Marcus,
I have a few comments:
- we do not suggest any fixed value (like 0.125) for a CC1/2 cutoff. One reason why a fixed value should not be carved in stone is that future data processing and refinement programs might extract more information from the weak data than current ones do. Another reason is that the significance of a given CC1/2 value depends on the number of pairs it is calculated from. Generally, we suggest to do paired refinements as a technique to validate that some cutoff is better than another. We really suggest you convince yourself (and colleagues/reviewers) of the value of your high-resolution data, by performing the comparison like we document in the paper (and its supporting online material which has all technical details!).
- in the same way, you can can assess the value of your incomplete high-resolution shell. It is easy enough to do. I expect that you will see model improvement when you include the data. The general rule is: throwing away significant data is bad.
- CC1/2 is a better guide than <I/sigI> because the latter depends on the sigma estimate of the data processing program, and its adjustment by the scaling program. Different programs give (sometimes very) different sigma values.
Hope that helps,
Kay
On Mon, 6 Aug 2012 13:21:19 +0200, Marcus Fislage <[log in to unmask]> wrote:
>Dear all,
>
>We have in our lab a data set collected and are discussing where to cut
>the resolution for refinement. According to the work of Kai Diederichs
>and Andy Karplus one should use CC 1/2 of 12.5% (in case it is
>significant) to determine the highest resolution independent of the
>I/sigI and R factor rules used earlier. But I would like to know if this
>also counts for low completeness data?
>The problem is that we have in the highest resolution shell an I/sigI of
>4, a good cc1/2 but only a completeness of 30%. Which I guess means we
>measured the high resolution data very accurate but not complete. Would
>you still use the low complete data in the highest resolution shell or
>should that be still a valid argument to cut your data towards lower
>resolution?
>My guess would be to use the data still even if the completeness drops,
>since the data we measured is good and according to CC1/2 significant.
>Are we right to do so or would you disagree?
>
>Thanks for any input
>Marcus
>
>--
>Marcus Fislage
>Structural Biology Brussels
>Vrije Universiteit Brussel
>Department of Structural Biology, VIB
>Oefenplein, Gebouw E
>Pleinlaan 2,
>1050 Brussel
>Belgium
>Tel: +32-2-629 18 51
>Email : [log in to unmask]
>Url: http://www.verseeslab.structuralbiology.be
|