Re: [ccp4bb]: Refmac vs Ramachandran (low resolution)

[Anastassis Perrakis]

On Aug 23, 2006, at 15:29, EATON E LATTMAN wrote: Imagine the same structure initially solved at low resolution, and then re-solved with very high resolution  data.  The RMSD given by the software will be about the same, but the actual values of particular bond  lengths (say) will change,  with more accurate values coming from the atomic resolution structure.  So we  should be careful how we interpret RMSD.   Most important, and as Clemens explains, RMSD on bond lengths   and standard deviation of the Engh&Huber distributions, are different things.  And expected coordinate error (and associated errors of bonds, angles, that can be derived from that etc) is yet another totally different thing !!! The expected coordinate error which would be different in low and hi res structures is not actually correlated to RMSD at all (well only very indirectly). Its like a 'B factor': As Dale says : "the slush-bucket of protein crystallography" There are 5-6 things we call B-factors and the terms are used interchangeably. Same now for deviations ? i hope not. A.