On Thursday 03 January 2008 01:25, Boaz Shaanan wrote: > > Is the wrong value pointed out by Ethan the reason why we tend > to nearly always see +ve difference peaks over Se-met after refinement ?
Could be.
I have attached a plot of the net scattering factor for the two
sets of values.
I am curious as to exactly which version[s] of atomsf had the wrong
value for A1. The various old copies I have lying around on my disks
all contain the correct value.
Ethan
>
> Just curious.
>
> Regards,
>
> Boaz
>
> ----- Original Message -----
> From: Garib Murshudov <[EMAIL PROTECTED]>
> Date: Wednesday, January 2, 2008 19:19
> Subject: Re: [ccp4bb] Refmac 5.4.0066 versus Refmac 5.4.0034
> To: CCP4BB@JISCMAIL.AC.UK
>
> > On 2 Jan 2008, at 17:15, Ethan Merritt wrote:
> >
> > > On Tuesday 01 January 2008 06:40, Mark A Saper wrote:
> > >>
> > >> Well, I guess one shouldn't change versions of refmac in the middle
> > >> of a structure refinement. What are the major
> > differences between .
> > >> 0034 and .0066 ? I noted that form factors for the Se
> > atom have
> > >> changed dramatically even though I'm using the same " anom
> > formfactor>> SE -8.13 5.05 " command. Is this because the
> > program is now using
> > >> the wavelength on the .MTZ file ?
> > >>
> > >> with .0034:
> > >>
> > >> SE 8.8706
> > 2.4098 5.8196 0.2726
> > 3.9731 15.2372
> > >> 4.3543 43.8163 2.8409
> > >
> > > The first coefficient here is wrong.
> > >
> > >> with .0066:
> > >>
> > >> SE 17.0006
> > 2.4098 5.8196 0.2726
> > 3.9731 15.2372
> > >> 4.3543 43.8163 -11.7969
> > >
> > > These are the values given for a1, b1, ... a4, b4 in Table 6.1.1.4
> > > of International Tables for Crystallography Volume C (1995).
> > >
> > > The value of c (-11.7969) corresponds to an f' value of
> > > -14.6378 (= -11.7969 - 2.8409)
> > >
> > > This could only be possible at the inflection point of the Se
> > > anomalous scattering curve, and only for a beam with a very narrow
> > > bandwidth. Current generation beamlines more
> > typically yield an
> > > effective f' of -5e to -10e in practice.
> >
> > We have corrected it. It was due to double addition of f' (one
> > from
> > anom formfactor and one from a la crossec calculation). Now it
> > should
> > not happen
> >
> > Garib
> >
> >
> > >
> > > --
> > > Ethan A
> > Merritt Courier Deliveries: 1959 NE Pacific
> > > Dept of Biochemistry
> > > Health Sciences Building
> > > University of Washington - Seattle WA 98195-7742
> > >
> >
>
> Boaz Shaanan, Ph.D.
> Dept. of Life Sciences
> Ben-Gurion University of the Negev
> Beer-Sheva 84105
> Israel
> Phone: 972-8-647-2220 ; Fax: 646-1710
> Skype: boaz.shaanan
>
--
Ethan A Merritt Courier Deliveries: 1959 NE Pacific
Dept of Biochemistry
Health Sciences Building
University of Washington - Seattle WA 98195-7742
set term pdfcairo mono dash set output 'Se_scatter.pdf' set title "Net Se scattering (includes f' term for λ=1.0Å)" set xlabel "sin(θ)/λ" set ylabel "e" norotate set encoding utf8 A0 = 8.8706 # wrong value A1 = 17.0006 # correct value B1 = 2.4098 A2 = 5.8196 B2 = 0.2726 A3 = 3.9731 B3 = 15.2372 A4 = 4.3543 B4 = 43.8163 C = 2.8409 - 3.2 # includes f' term at 1.0A wavelength f(x) = A1*exp(-B1*x*x) + A2*exp(-B2*x*x) + A3*exp(-B3*x*x) + A4*exp(-B4*x*x) + C g(x) = A0*exp(-B1*x*x) + A2*exp(-B2*x*x) + A3*exp(-B3*x*x) + A4*exp(-B4*x*x) + C set xrange [0:1] plot f(x) title "a1=17.0006" lt 1, g(x) title "a1=8.8706" lt 3
Se_scatter.pdf
Description: Adobe PDF document
