[ccp4bb]: I's and F's

[Bernard Santarsiero <bds@uic.edu>]

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

One comment about negatives I's.  If you're only going to use F's, than
you probably have a 0-2 sigma cutoff in refinement, and it's acceptable to
reduce the negative intensities to zero.  I think the better option is to
flag them somehow.  One option is to assign the F as negative, and make
sure that the calculations only use positive F's for sigma tests, etc.  

Negative intensities are a valid measurement even if not physically
meaningful.  They simply imply that a intensity is weak, and very close to
zero, as would a measurement of exactly zero.  Negative intensities can be
included in a refinement on F^2 just as positive intensities, since you
are minimizing the difference between I(obs) and I(calc).  You also have
the added advantage that there is no sigma cutoff with a F^2 refinement,
and you use all of the data.  With proper weighting (real weights), the
goodness of fit is a substantially better indicator of how the
refinement is proceeding, vs. the various R's.

Bernie Santarsiero