WARNING: At present the density fitting routines only work for orthogonal
maps with equal stepsize in all three directions.
The X-ray menu holds the options mainly meant for crystallography. At present
there are commands to convolute desity with atoms, to use symmetry matrices
to objects or parts of the soup, and to do some real
space refinement. The real space refinement options assume that you
already used the map menu to read an electron density map. The symmetry
options assume that you already obtained symmetry matrices.
In the electron density convolution options atoms are represented by a
sphere with a bell shaped radial density distribution. See:
R. Voorintholt,
M.T. Kosters, G. Vegter, G. Vriend, W.G.J. Hol, A very fast program for
visualizing protein surfaces, channels and cavities. (1989) J. Mol. Graph.
pp 243-245.
The convolution is the sum of all products rho(i)*atm(i) in
which i runs over the volume of the atom, atm(i) is the value of the
radial bell-shaped atom density at position i,
and rho(i) is the electron density map
value at that same position in space.
The results of the density convolution options are NOT on any absolute
scale. You can
only compare values within one run of this option. If you
want to compare values between different maps, then these
maps should have the same statistics, e.g. the same average
density, and the same sigma (=spread in density).
The option EVADEN will cause WHAT IF to prompt you for a residue range.
It will then convolute all atoms in this range with the density found in the
present default map. The density under each atom will be listed as well as the
summed density per residue. You will be asked if you want to store the
results in a table. If you answer with YES, you will as usual (see the
chapter on tables) be prompted for the table number and the table name.
The option EVACAS will cause WHAT IF to prompt you for a residue range.
It will then convolute all alpha carbons in this range with
the density found in the
present default map. The density under each alpha carbon will be listed.
You will be asked if you want to store the
results in a table. If you answer with YES, you will as usual (see the
chapter on tables) be prompted for the table number and the table name.
The RSRSID option will prompt you for a residue range. It
will try to optimize the convolution between the density and
bell shaped side chain atoms for the residues in this range.
You will also be prompted for the allowed Van der Waals
radius overlap. The suggested value is 0.25 Angstrom. This
number indicates how far two atoms are allowed to penetrate
each other. You will be asked to give the number of steps
per chi-angle. The suggested value is 2. This number tells
WHAT IF how many steps to try in either direction for every
free rotatable side chain torsion angle. Choosing 6 means
that for lysine 13*13*13*13 density convolutions (13 values
for each of 4 free rotatable bonds) have to be done. The
last value you are prompted for is the angle per step. In
early stages of building you might give 20 degrees or so. In
later stages a value less than 5 degrees is suggested.
The command RSRRES will cause WHAT IF to prompt you for a residue range.
For all residues in this range it will move them around in small translational
steps (maximal step per call to this option is 0.4 Angstrom) to search
for the optimal convolution between the atom positions and the electron
density. This is a rigid body motion.
The option RSRBB does the same as the option RSRRES. The difference is that
RSRBB only tries to optimize the backbone density fit. It will however
apply the translation used for the backbone to the sidechain too.
The graphical expansion of MOL-items or the FBRT or MOVITM object through
the usage of symmetry matrices is probably only useful for crystallographers.
Also, it relies heavily on some Evans and Sutherland hardware features
Todate the option only works on E and S machines, and it is not
likely that it will run equally well on other machines once
the conversion is done.
EXPSYM is used to apply symmetry matrices to a MOL-object.
These matrices can be crystallographic or
non-crystallographic. You can give `normal` symmetry matrices or real
space symmetry matrices.
You will be prompted for the number of a MOL-item.
All presently active symmetry matrices (see the chapter on symmetry matrices)
will be applied to this MOL-object.
The symmetry expanded object is stored
in MOL-object 8. That means that MOL-object 8 can not be
used as input for this option.
The flow of events to use this option is:
1) Get the symmetry matrices (see the SYM menu).
2) Create a MOL-item in a MOL-object, but not in MOL-8.
3) Use the EXPSYM option.
Step 2 and 3 can be interchanged. It is not possible to undo
the EXPSYM option. It will stay active.
FBRSYM can be used to move part of the soup as a FBRT object
(see FBRT in the chapter on graphics options) while at the
same time moving the symmetry related parts of the soup in
accordance with the symmetry matrices. These matrices can be
crystallographic or non-crystallographic. You can give `normal` symmetry
matrices or real space symmetry matrices.
Remember that the
MOVITM object is also stored in FBRT.
The flow of events to use thsi option is:
1) Get the symmetry matrices (see the SYM menu).
2) Uset the FBRSYM option.
3) Pass control to the graphics screen/window.
4) Use the FBRT option.
The symmetry related parts of the soup are automatically put
in MOL-object 0 (zero) by WHAT IF. FBRSYM can not be
switched off. It works on every next FBRT or MOVITM option. If you click
YES at the graphics screen menu the present matrix that you applied
to the FBRT/MOVITM object will be displayed. This is always a real space
matrix.
Other Xray related options are found in the menus: MAP, MAPEDT,
CHKMDF, WATER, REFINE, etc..
If the either the density or a WHAT IF checking option suggests that
a peptide plane needs flipping, you can use this option to do so.
Be aware however, that backbone bond angles are no longer OK after
such an operation, and subsequent normalisation with REFI in the
REFINE menu seems required if you still want to call your
molecule a protein afterwards.
Sometimes, upon density fitting in poor density, one builds a
stretch of residues in the wrong direction. The RNGFLP option can
invert the chain direction of a range of residues, without the
need to refit everything afterwards. This option is likely to
perform better for beta-like structures than for highly curved
secondary structure elements. The range will be returned as
poly-alanine.