*** For details on how to be removed from this list visit the *** *** CCP4 home page http://www.ccp4.ac.uk ***
Ron I don't think you should knock what the EM people do (at least for the reason you give!). The grid they pick will be decided by all sorts of practical criteria determined by factors like tilt requirements/constraints, radiation damamge (not that again) etc. They will often oversample the data and this is good practise. The question you raise is however relevent. The Niquist-Shannon sampling theorem should provide the answer regarding the number of unique observations if you know the thickness perpedicular to the 2D crystal. If the thickness is A and you are measuring intensities the minimum sampling interval is 1/2A (1/A if measuring amplitudes with phases). This interval essentially defines the number of unique observations (with phases the real and imaginary components can each be regarded as unique observations) and we can compare what they have with the 3D crystallographic case. Note that for the crystallographic case (1D, 2D or 3D), the sampling given by a reciprocal lattice (real lattice dimension A) is 1/A - fine if you have phases but undersampled if you only have intensities. Of course if they quoted the large number of (oversampled)observations as being unique that would be misleading. Colin -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of R. Stenkamp Sent: 28 April 2006 15:59 To: Bart Hazes Cc: [EMAIL PROTECTED]; [email protected] Subject: Re: [ccp4bb]: Acceptable Rmerge values in last resolution shell *** For details on how to be removed from this list visit the *** *** CCP4 home page http://www.ccp4.ac.uk ***g Three comments about "observations" (just to point out a few other aspects of "observations", etc): 1. Reflections with zero intensities are useful in refinement, even if they don't contribute much to Fo maps. They provide important restraints on the electron density. 2. I've slowly discovered that some data reduction programs count partials as "measurements". I understand why you'd do that, but it's a tad confusing when the number of measured reflections gets so large. 3. And another slow realization on my part is that our cryo-EM colleagues have a wonderful way to increase their number of observations. They pick an arbitrary (or at least it looks that way to my un-trained eyes) grid in reciprocal space to sample their diffraction pattern perpendicular to the 2-d crystals. There must be some guidelines for acceptable choices for that sampling interval, but if they change it a little, they can change their number of observations by large numbers. Ron
