Hello Tang,
1) For MR, you might want to try a range of homologs, or even a stack of
overlapping homologs. A normal modes server like elNemo might also help
if it can predict the "bend" your molecule undergoes upon binding. A
long shot perhaps, but stranger things have happened. You also might be
able to find the DNA by molecular replacement.
2) radiation damage increases with photons/area, not time. So no matter
what your degrees/image you want the total shuttter-open time at the end
of the data set to be below the damage limit of interest. A little web
app I made once might help: http://bl831.als.lbl.gov/xtallife.html .
These days, there is no reason not to know how long your crystal will
last before you push "collect", and it is definitely worth knowing.
-James Holton
MAD Scientist
On 7/28/2017 12:21 AM, Tang Chenjun wrote:
Hi,
Thanks to all who gave me suggestions concerning the weird diffraction pattern
and I really appreciate it that Kay Diederichs help me processing my data set
and answer my questions. Although the data set can be processed using HKL3000,
XDS without problems, the Rwork/Rfree values are still above 0.5 after
molecular replacement. There can be several reasons.
1) The structure change a lot after binding DNA, so it is not possible to find
a solution using molecular replacement.
2) Strong radiation damage and 1.0 degree image widths prevent good integration
results. It may be better to use 0.1 degree image widths.
3) Streaky spots appearing in certain directions because of anisotropy or
lattice translocation disorder, or one very large unit cell dimension lying
along the X-ray beam may also have an affect on data processing.
Now I am optimizing the crystals to address these problems.
Best wishes and thanks again for your help,
Chenjun Tang
