16 Mar
2018
16 Mar
'18
3:47 p.m.
Hi Radu,
Related to Mark's resolution bin point: Pavel, is there a way to specify the resolution bins Phenix uses? For example asking for an equal number of reflections to be present in each bin? Stats get bad (and probably meaningless) when outer bins are cut very thin and contain a tiny number of reflections.... a particular problem with anisotropic datasets....
not really... The use of bins is constrained by requirements for ML calculations (need to have certain free reflections per bin and bind are ought to be thin enough) and by http://journals.iucr.org/d/issues/2013/04/00/dz5273/dz5273.pdf see Table 1 and comments. Do you have a specific example where currently used binning is problematic? Thanks, Pavel