-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Here is how I view this stuff. In general the diffraction pattern does not obey Fridel's law, which means that the actual electron density map is a set of complex numbers. The map that consists of just the real part of these density values is what we call the normal electron density map. The purely imaginary part is the anomalous map. We don't (yet) have FFT programs that calculate complex electron density (although that is quite easy) nor do we have graphics programs to display them, so we make due with two separate maps. The normal map is easy, just manipulate the structure factors to ensure compliance with Fridel's law and calculate the map. The anomalous map is harder. If you create coefficients that purely disobey Fridel's law (F(h) = -F*(-h)) they will correspond to the imaginary part of the total map, but our FFT programs cannot calculate imaginary density. If we note that the imaginary density is just a real value multiplied by "i", we can absorb the troublesome "i" into the phase by recognizing that i = Exp[Pi/2 i] and rotate the phase by 90 deg. This is nothing fundamental to diffraction. It is just a trick to get a program that is designed to calculate purely real values to calculate a map of imaginary density. Whether you rotate by +90 or - -90 only changes the sign of the result. It would be really handy if Coot could take a general set of structure factors, calculate the complex density and display, separately, contours for the real and imaginary components of the density. Then, even if you didn't notice the presence of "anomalous" scattering in your diffraction, you would still see the anomalous peaks on your display. Dale Tronrud On 8/29/2014 11:43 AM, Alexander Aleshin wrote: > Could anyone remind me how to calculate anomalous difference > Fourier maps using model-calculated phases? I was doing it by (1) > calculating PHcalc from a pdb file using Sfall, then (2) merging > PHcalc with Dano of experimental SFs, then (3) calculating a map > with Dano and PHcalc using FFT program of CCP4. > > Now, I've read Z. Dauter's et all paper > http://mcl1.ncifcrf.gov/dauter_pubs/175.pdf, and it said that their > anomalous maps were calculated using (delF, PHcalc-90degrees). Why > did they use -90 degrees? How does it relay to a (delF, phcalc) > map? > > Thank you for an advice. > > Alex Aleshin > -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (MingW32) iEYEARECAAYFAlQA07QACgkQU5C0gGfAG100OgCfWJmiPBmTsm//P7VcqjKBcamm ZZoAoIi6vAe0fQMLwIXaQGxAsGjIMe5v =pky4 -----END PGP SIGNATURE-----
