ANALYZE_MAD: Analyzing a MAD dataset
ANALYZE_MAD is a routine to run MADMRG and MADBST. You usually do not have to worry about this routine at all because it is ordinarily called right after running SCALE_MAD for automatic structure determination, and all the parameters are already set for you. It is ordinarily followed by running SOLVE with the solve_mad.script file written out by this routine or by using the keyword "SOLVE" after running this routine.
ANALYZE_MAD assumes that you have a datafile ("mad_fbar.scl") that contains (Fbar,sigma,DelAno,sigma) for each wavelength of MAD data. That is, there are exactly 4 data columns for each wavelength. This is a dorgbn file. It is ordinarily created by SCALE_MAD, but you can create your own if you like.
To run ANALYZE_MAD, you need to input your standard setup file ("solve.setup"), and the wavelength to define as the standard (JSTD, see below). You need to input the scattering factors for the anomalously scattering atom at the various wavelengths too. You also need to input an estimate of the number of anomalously scattering atoms in the asymmetric unit (NANOMALOUS) and the number of protein residues in the asymmetric unit (NRES). It is only the ratio of NANOMALOUS to NRES that is important.
ANALYZE_MAD will run MADMRG (output ="madmrg.out") and MADBST ("madbst.out"). See the writeups below for these two routines. It will calculate origin-removed difference Pattersons for all dispersive and anomalous differences and Pattersons based on the output of MADMRG and MADBST. These maps are all compared to each other and the correlation coefficients are displayed in a table. In a typical MAD experiment the anomalous and dispersive difference Pattersons have correlations with each other on the order of 0.1 to 0.3 or so (pretty low, so don't be worried).
A typical input script for ANALYZE_MAD follows.
Script for ANALYZE_MAD
!--------------------------Run ANALYZE_MAD----------------------- -----------
@solve.setup ! standard information for this crystal
madfbarfile mad_fbar.scl ! input dorgbn file with (Fbar,sig,Delano,Sig)
! for each wavelength
madfpfmfile mad_fpfm.scl ! input dorgbn file with (F+,sig,F-,Sig)
! for each wavelength
logfile analyze_mad.logfile ! write out most information to this file
mad_atom se ! atom type is selenium
fixscattfactors ! do not refine scattering factors (you can if
! you want and the data is good)
jstd 1 ! Define wavelength #1 as the reference wavelength
lambda 1 ! wavelength #1 information is to follow
label Wavelength #1 ! label for lambda 1
wavelength 0.9000 ! wavelength value
fprimv_mad -1.6 ! f' value at this wavelength
fprprv_mad 3.4 ! f" value at this wavelength
lambda 2
wavelength 0.9794
fprimv_mad -8.5
fprprv_mad 4.8
lambda 3
wavelength 0.9797
fprimv_mad -9.85
fprprv_mad 2.86
nres 100 [approx # of residues in asymmetric unit]
nanomalous 2 [approx # of anomalously scattering atoms in a.u.]
madmrgfile madmrg.out ! write the SIRAS-like MAD dataset to madmrg.out
madbstfile madbst.out ! write the Bayesian Patterson to madbst.out
ANALYZE_MAD ! run MADMRG and MADBST and analyze all the Pattersons
!-----------------------------------------------------------------------------
LOGFILE xx.logfile log file for output will be xx.logfile
madfbarfile xx.scl input dorgbn file with (Fbar,sig,Delano,Sig)
for each wavelength will be xx.scl
[Default="mad_fbar.scl"]
madfpfmfile xx.scl input dorgbn file with (F+,sig,F-,Sig)
for each wavelength will be xx.scl
[Default="mad_fpfm.scl"]
madmrgfile xxx.out write the SIRAS-like MAD dataset to xxx.out
[ default="madmrg.out"]
madbstfile yyy.out write coefficients for a Bayesian Patterson to yyy.out
[default="madbst.out"]
SOLVEDATAFILE xxx Output datafile with MADMRG output and MADBST
output combined together, suitable for use
with routine SOLVE, will be xxx. (DEFAULT file name =
"solve.data") The datafile has the following
columns of data:
1 Fnative (from MADMRG)
2 Sigma of Fnative (from MADMRG)
3 Fderiv (from MADMRG)
4 Sigma of Fderiv (from MADMRG)
5 DelAno (from MADMRG)
6 Sigma of DelAno (from MADMRG)
7 <Fh cos(theta)> (from MADBST)
8 <Fh sin(theta)> (from MADBST)
SCRIPTFILE xxx Output script file containing instructions
for running SOLVE written to xxx
[default="solve_mad.script"]. Starting SOLVE with
this script is equivalent to following
ANALYZE_MAD with SOLVE.
FIXSCATTFACTORS Fix scattering factors at their input values.
This is a good idea if you have a reasonable
idea of the f' and f" values. [this is the
default]
REFSCATTFACTORS refine scattering factors f' and f". If you
refine them, be sure to look at their new
values at the end of the routine MADMRG and
verify that they are reasonable.
mad_atom xx anomalously scattering atom is "xx"
The atom types recognized by SOLVE are:
H, H-1, He, Li, Li+1, Be, Be+2, B, C, Cv, N, O, O-1,
F, F-1, Ne, Na, Na+1, Mg, Mg+2, Al, Al+3, Si, Siv, Si+4,
P, S, Cl, Cl-1, Ar, K, K+1, Ca, Ca+2, Sc, Sc+3, Ti, Ti+2,
Ti+3, Ti+4, V, V+2, V+3, V+5, Cr, Cr+2, Cr+3, Mn, Mn+2, Mn+3,
Mn+4, Fe, Fe+2, Fe+3, Co, Co+2, Co+3, Ni, Ni+2, Ni+3, Cu,
Cu+1, Cu+2, Zn, Zn+2, Ga, Ga+3, Ge, Ge+4, As, Se, Br,
Br-1, Kr, Rb, Rb+1, Sr, Sr+2, Y, Y+3, Zr, Zr+4, Nb, Nb+3,
Nb+5, Mo, Mo+3, Mo+5, Mo+6, Tc, Ru, Ru+3, Ru+4, Rh, Rh+3,
Rh+4, Pd, Pd+2, Pd+4, Ag, Ag+1, Ag+2, Cd, Cd+2, In, In+3,
Sn, Sn+2, Sn+4, Sb, Sb+3, Sb+5, Te, I, I-1, Xe, Cs, Cs+1,
Ba, Ba+2, La, La+3, Ce, Ce+3, Ce+4, Pr, Pr+3, Pr+4, Nd,
Nd+3, Pm, Pm+3, Sm, Sm+3, Eu, Eu+2, Eu+3, Gd, Gd+3, Tb,
Tb+3, Dy, Dy+3, Ho, Ho+3, Er, Er+3, Tm, Tm+3, Yb, Yb+2,
Yb+3, Lu, Lu+3, Hf, Hf+4, Ta, Ta+5, W, W+6, Re, Os, Os+4,
Ir, Ir+3, Ir+4, Pt, Pt+2, Pt+4, Au, Au+1, Au+3, Hg, Hg+1,
Hg+2, Tl, Tl+1, Tl+3, Pb, Pb+2, Pb+4, Bi, Bi+3, Bi+5, Po,
At, Rn, Fr, Ra, Ra+2, Ac, Ac+3, Th, Th+4, Pa, U, U+3, U+4,
U+6, Np, Np+3, Np+4, Np+6, Pu, Pu+3, Pu+4, Pu+6, Am, Cm, Bk, Cf
If xx is not recognized by SOLVE you need to
specify a newatomtype and then refer to it
lambda n wavelength number (n=1,2,3...) for values
to be entered next of f' and f"
wavelength xx wavelength for lambda xx (i.e., 0.9798)
fprimv_mad xx 1 real number for f' value for anomalously
scattering atom at the current wavelength.
Wavelength is defined by the most recent
value of the keyword "LAMBDA". You always
need to enter f' and f" explicitly for MAD
data.
fprprv_mad xx 1 real number for f" value for anomalously
scattering atom at the current wavelength.
Wavelength is defined y the most recent value
of the keyword "LAMBDA"
JSTD n Wavelength "n" will be used as the standard
to which all data is referred (see MADMRG)
[default= the lowest wavelength]
NRES n # of residues in asymmetric unit
[default=100]
NANOMALOUS n # of anomalously scattering atoms in asymmetric unit.
Used to estimate how big the Fa values might be
NSHELLS n Number of shells for analysis is n
[default=10]
bayes Use Bayesian MAD phasing at the
very end of SOLVE. Requires that madfpfmfile
exists. It also requires that INPHASE be
specified for all wavelengths. (This is the default)
nobayes Use the compressed MADMRG datafile for all
phasing when program gets to SOLVE.