Re: [ccp4bb]: Question about I to F conversion

[Bart Hazes <bhazes@ualberta.ca>]

***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***

On Tue, 18 Sep 2001, Doug Ohlendorf wrote:

> Fro Bart's commnets it appears that the I -> F conversion does not properly
> handle reflections or observations with I<0. All intensity obervations should
> be combined when merging data (I prefer I>-5sigma cutoff). Once merged,
> reflections with I<0 must be set to zero. Truncating only the observations with
> I<0 before merging skews the distribution and produces incorrect results.

Truncate truncates ALL reflections, not just the ones with I<0. However, for
large intensities (large I/Sigma to be precise) the effect is extremely small.
I would expect the effect to be noticable for reflections with I/SigI < 1-2
since for those reflections there will be a significant fraction of the
probability curve that extends into the negative. (I think this is actually
reported in the paper).

Setting all negative intensities to zero is, at least theoretically, less
accurate since you will be treating I/SigI=-3 reflections the same as
I/SigI=-0.1 reflections. Truncate may not be perfect, but the only thing
better today seems to be refining against intensities. This has several
theoretical advantages, but I haven't heard any of the experts proclaim it to
be the recommended procedure in practise yet. Maybe Randy, Garib, Raj,
Eleanor...  can comment on this???

Bart

===============================================================================

Dept. of Medical Microbiology & Immunology
University of Alberta
1-15 Medical Sciences Building
Edmonton, Alberta, T6G 2H7, Canada
phone:	1-780-492-0042
fax:	1-780-492-7521

===============================================================================