Tomasz Cierpicki wrote:
>
> Hi
>
> I have some problems to generate correct structure files in xplor-nih.
It's not clear to me what you're trying to do here. Are you trying to
get
initial coordinates to start a structure determination? Are you trying
to
refine a known structure, starting from its PDB entry's coordinates?
I'm going to assume that you mean that you've created a PSF file from
the
sequence, and now need to get some starting coordinates with good
covalent
geometry.
> Minimization of original pdb file stops with very high bond/angle/improper
> energies and error message:
>
> --------------- cycle= 35 ------ stepsize= 0.0000 --------------
> | Etotal =2675.554 grad(E)=17.466 E(BOND)=175.944 E(ANGL)=863.045 |
> | E(DIHE)=0.000 E(IMPR)=1998.340 E(VDW )=-150.215 E(ELEC)=-211.560 |
> -----------------------------------------------------------------------
> %POWELL-ERR: Line search abandoned
There are two problems with what you seem to be doing. First, it
appears
that you forgot to issue a flags statement before starting the
minimization.
The energy terms shown in your output are the ones that are turned on by
default.
For NMR structure determination purposes, you should never use the
electric
field term, and the van der Waals term needs to be set up for the purely
repulsive potential shape.
Second, you appear to be running the minimizer with everything turned on
at once.
The conjugate gradient energy minimizer in xplor gets stuck easily in
local
minima. In order to make it work well, you need to minimize a complex
energy
potential in stages, building up from the most geometrically simple
terms to the more
complex.
In the old xplor scripting language, getting starting coordinates
requires something
like this:
struct @afile.psf end
param @aparameterfile.pro end
vector do (x = random() * 10) (all)
vector do (y = random() * 10) (all)
vector do (z = random() * 10) (all)
flags exclude * include bond end
mini powell nstep 100 nprint 10 end
flags exclude * include bond angle end
mini powell nstep 500 nprint 10 end
flags exclude * include bond angle impr end
mini powell nstep 500 nprint 10 end
param nbonds repel 0.8 rcon 0.01 wmin 0.0 end end
flags exclude * include bond angle impr vdw end
mini powell nstep 500 nprint 10 end
write coords output = "start.pdb" end
stop
This will work reasonably well. If it still gets stuck after the last
minimization, you
could have it repeat some of the simpler stages (eg., after
bond+angle+impr+vdw, go back
to bond+angle, bond+angle+impr, and then bond+angle+impr+vdw).
For those who've been experimenting with the new xplor-nih scripting
interfaces, you could
use some standard procedures I've defined in TCL to make this a bit
shorter and a lot more
reliable:
package require marvin
initParamPsfPdb \
-paramFileName $env(XPLOR_DIR)/toppar/aParameterFile \
-psfFileName afile.psf
XplorCommand "vector do (x = random()*10) (all)"
XplorCommand "vector do (y = random()*10) (all)"
XplorCommand "vector do (z = random()*10) (all)"
cleanCovalentGeom
writePDB -fileName "start.pdb"
exit
I hope this helps. If I misunderstood your question, or you're still
having trouble,
please write again and include whatever xplor script is creating the
problem.
--JK
--
Dr John Kuszewski
301-594-5831 (voice)
301-402-2867 (fax)
[email protected]
