Dear users,
What is the allowed range of difference
between R and Rfree for structures with twinning?
I have a data in R3, after MR (dimer in asu) and
Refinement using Refmac-5.5.0102, I ended up getting
R ~ 0.231 and Rfree ~ 0.293, difference seemed to be
little high.
For twinning corr
Hi U Sam
When you say "R increases and Rfree decreases or vice versa" it's
important to consider which variables they are increasing/decreasing
with respect to. During a given run of Refmac (or whatever) with N
iterations, where you optimise a *fixed* set of atomic/group/overall
parameters with
It is difficult at the end of refinement - the small corrections hardly
affect the R values, but should make the maps look cleaner.
And R values are very correlated with the tightness of your geometric
restraints - maybe you are changing these slightly too?
I presume you are describing small sh
Hi Sam
Dogma would state that you are trying to maximise the correlation
between your model and the structure, *without over fitting* and Rfree
should thus be minimised.
However, you are likley using a maximum likelihood target, so you
should maximise the likelihood so that the model accurately
Hi Everybody,
I have small query.
I am feeding ions and correcting few residues into my final structure.
If R increases and Rfree decreases or vice versa in the subsequent refinement,
which one I should accept and go forward.
Thanks.
Sam
___
