[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [ccp4bb]: A simple question of resolution/scaling weights



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

> Bernard Santarsiero writes:

>  > I also don't monkey around with the weights when processing the data.

On Thu, 25 Apr 2002, Phil Evans wrote:

> Not monkeying around assumes the initial estimate is correct & it isn't

I have found that a well-behaved xtal at a synchrotron (CHESS, Brookhaven
for most of my data) collected on a CCD and processed with HKL gives
chi**2 of ~1.0 if I use an error scale factor of 1.5 in my scalepack file.
The error scale factor seems more like 2.2 on home sources with IP
detectors.

This is over many tens of datasets at a variety of resolutions and
redundancies and mosaicities, so for xtals without problems the initial
estimate of error seems relatively consistent.

(The actual value for error scale factor will vary according to one's
"estimated error" values, I think - mine lie in the range 0.02-0.04 in 10
zones).

Using these error scale factors, I find that if chi**2 >> 1 then something
is awry with either the data or the data processing, and warrants a much
closer look.  If chi**2 << 1, I take this as a warning that my redundancy
is probably low and the Rsymms are over-optimistic.

Phil Jeffrey