On 24/05/12 11:03, Ian Tickle wrote:
> Hi Graeme
> This has nothing to do with Randy's 1986 paper on minimally-biased map
> coefficients from partial structures (which I assume is the one you're
> referring to) since as you said yourself these are experimental, not
> partial structure, phases. Rather, this is Blow& Crick's (1959)
> result where they showed that coefficients m*F with the centroid phase
> give the electron density with the least mean-square error over the
> unit cell. But maybe the error in the density is not something you
> need to worry about, since if the signal/noise ratio is sufficiently
> high the peaks will stand out against the background anyway?
> Paul's solution looks like the best one (once tested!).
Hmm... either that or read in the shelx output directly and let Coot
calculate the sigmaa map... (which is the thing I'd try first - Coot may
well be picky about the right type of Shelx* output).