Dear Joe,
On Mon, Oct 27, 2008 at 06:04:38AM +0000, joe wrote:
> Hi, all:
>
> I have a poor density after phase extension problem.
> Usually I use HKL2000 to scale the data and use the macro at the integrate
> and scale steps. I got a good native data( P6422 130 130 150 , 90 90 120) at
> 3.0A (completeness >90%, I/sigma >5, redandance>10, Linear R-fac<0.5, but
> square R-fac >0.6; one problem is Linear R-fac smaller than squear R-fac in
> my scaled data. I used macro: sigma cutoff 0 during scaling stage. if i
> don't
> use macro, the data will be getting bad less than 4.0A) and my reflection
> data has strong anisotropy.
>
> I used this native data to do phase extension from 3.6A to 3.0A by CCP4i DM
> program(the phase got from sharp program), but I cannot find any density is
> improved.(density of sidechain cannot be seen)
Two points that I would try:
- if your data is very anisotropic, try and remove that anisotropy
from your data before going into the (real-space) density
modification part. I found over the last 1-2 years that this can
have an impact.
autoSHARP will do that automatically for you - but only if the
reference dataset has at least 2.5A resolution. This might be a
very conservative limit and you could run SHARP itself instead:
specify to estimate (but not refine) the anisotropic scale of the
first (reference) dataset. This will then give you 'isotropic'
amplitudes in the output eden.mtz file.
- when coming from SHARP I would recommend using the SOLOMON
procedure we have inside our system: it is tightly connected to
the SHARP output, makes sure to use the right columns and
information from the SHARP file and generally gives as good
density (and very often better) than running DM outside our
system.
However, if you have NCS and want to do NCS averaging you should
still use something different (our SOLOMON procedure doesn't deal
with NCS yet): but you should start these density modification and
averaging jobs from a map coming out of SHARP+SOLOMON.
There are obviously also other programs out there you could try:
parrot and pirate from Kevin Cowtan or resove from Tom
Terwilliger/Phenix. Make sure to use the correct column labels
from the eden.mtz file though (if in doubt: drop me a line)!
Cheers
Clemens
--
***************************************************************
* Clemens Vonrhein, Ph.D. vonrhein AT GlobalPhasing DOT com
*
* Global Phasing Ltd.
* Sheraton House, Castle Park
* Cambridge CB3 0AX, UK
*--------------------------------------------------------------
* BUSTER Development Group (http://www.globalphasing.com)
***************************************************************
