Francis E Reyes wrote:
> Hi all
>
> I have pretty low res data (4-4.5A) and was wondering what's the
> suggestion for allowing the bfactors to vary or keep them fixed during
> scaling. If I fix them, xtriage reports wonderful anomalous signal
> whereas if they are varied, the anomalous signal is gone. I have natives
> and suspected derivatives so MIR is possible. Rmeas's are pretty similar
> independent of whether I keep them fixed or allow them to vary. Is there
> a more sensitive statistic (to whether or not to scale or vary bfactors)
> that I'm missing?
>
Chi^2?
If you are using scalepack, there is an intermediate option-
"B restrain 0.1" or so.
Look at the B factors in the first list after baddies printout-
are they jumping around randomly or increasing monotonically?
The former is probably "fitting the noise" and should be repressed,
while the latter is reporting the loss of resolution with radiation
damage. "B restrain" allows the B-factors to follow the general
trend upward but prevents big fluctuations between adjacent frames.
Ed
|