 |
 |
On Tue, May 25, 2010 at 6:22 AM, Frank von Delft <[log in to unmask]> wrote: > Hi Ian, I read with great interest. But got stumped here: > >>> - how you compute sigma(rho)? >> >> See my reply to George Sheldrick's post. >> > > I think your reply did not make it out to the BB, certainly neither to my > inbox nor to the archive. Do you think you could post it again? Hi Frank, my apologies, I failed to spot that my reply wasn't being copied to the BB, here it is: I just use the EXTENDS program (I modified Phil Evans' original EXTEND program), which gets an initial estimate as the RMS difference density of the map from FFT (i.e. either an a.u. or a complete cell) as you say, then makes a correction for outliers (> 3 sigma), replacing the RMSD value in the map header with the estimated standard uncertainty (i.e. sigma). This is not perfect in terms of outlier rejection: to do it rigorously takes significantly more CPU time (i.e. more than a few seconds), and I wanted to keep map calculation fast (since we do a lot of it!), but it's probably good enough. > By the way, I'm afraid you're wrong with this statement: > >> Note that I'm not proposing anything new, this is all explained in >> standard statistics textbooks (Kendall's Advanced Theory of Statistics >> by Stuart& Ord is probably the best). In fact this is exactly my >> point: why re-invent the wheel (and likely end up with a square one!) >> when the appropriate statistics is all there in the textbooks and has >> been for ~ 80 years? > > Any time someone spots that an old technique is relevant in a new context, > that is "something new". And since you bemoan that this is not being used, > that by definition makes it a new context :) I'm happy to go along with that! > So I'll be bugging Paul to put this in coot, e.g. for sidechain fitting. I think he already got the message! Cheers -- Ian