Do you really want atomradius 20A?
Molecules separated by 40 A atomcenter to atomcenter will
be in contact, exclude solvent? Maybe you should tell us what you
are trying to do?
Using fft mode atommap to make a protein mask you could use a
low threshold when converting map to mask, which would expand the
atoms somewhat, but not to 20A.
The uppsala program mama lets you make a protein mask with
setting the atom radius. It has all kind of other neat tools
which may be useful for whatever you are trying to do.
http://xray.bmc.uu.se/usf/mama_man.html
########################################
#Make a mask in ref cell, grid, etc:
setenv MASKSIZE 4573536
setenv MApSIZE 4573536
mama -b <<eof
new Cell 170.900 181.400 240.200 90.000 90.000 90.000
new Grid 168 168 240
new radius 2.0
new pdb m1 ref.pdb
smooth m1 10
smooth m1 10
smooth m1 10
island m1
fill m1
write m1 ref.msk
eof
I think the Uppsala mask format is different from the ccp4 one, but that it is
easy
to convert.
Hailiang Zhang wrote:
Hi Edward:
Yes, this is really a good way to do it. Now I am trying to generate a
solvent map using CCP4 sfall (MODE ATMMAP). The thing is I want to specify
a large probe radius (~20A), but it seems that sfall can't change the
probe radius at all. Do you know any other tools to do that?
Thanks again for your time!
Best Regards, Hailiang
Hailiang Zhang wrote:
Thanks Edward! Actually Areaimol works well for my problem.
But now I have a new issue looking for some advice. I want to randomly
generate some points in the unit cell and make a quick judgment whether
it
is outside of the solvent mask or not. It seems that Areaimol doesn't
help
at this point, and wonder whether some others tools in CCP4 can help to
make it.
Convert solvent mask to a map, exress the random points as dummy atoms in
a
pdb file, and see reent thread on "program to calculate electron density
at x,y,z"
for methods to print out density at arbitrary points in a map.
Thanks again for your help!
Best Regards, Hailiang
Areaimol is good for determining the contact area from the difference
you
mentioned. If you want to distinguish real clashes from comfortable
van-der-Waals
contacts, you can use pdbdist3:
http://sb20.lbl.gov/berry/for/pdbdist3.for
The two molecules have to be in separate pdb files. You give a
threshold distance. For every atom in the first structure, every
atom in the second structure that is closer than the threshold distance
results in printing out the pair of atoms and the distance separating
them.
this gives a list of all contacts within the threshold distance.
For v-d-w contacts are around 3.4 A, H-bonds 2.7, and anything
closer than 2.0 could be considered a serious clash.
Hailiang Zhang wrote:
Hi,
I have 2 rigid and fixed proteins and want to quickly judge whether
there
are some steric clashes. One quick way I am thinking is using CCP4
AREAIMOL to calculate the surfaces of each individual protein as well
as
the heterodimer, and check whether the sum of the two individual
surfaces
is larger then the dimer. I am wondering whether I can get some
advices
about this method.
I also know there must be some other tools to quickly do it since this
is
kinda a simple docking problem, and I appreciate if suggested some
more
direct methods.
Finally, I am also wondering whether AREAIMOL considers the assymetric
unit during calculation.
Thanks!
Hailiang
