Hi,
the exact answer to your question is here:
http://cci.lbl.gov/~afonine/for_ak/validation.pdf
In particular, your question below is spelled out is discussed staring
from slide #2.
Pavel.
On 12/14/09 7:50 AM, rui wrote:
Hi, All,
I have a general question for refinement. I tried to refine a dataset
that is about 2.3A and got R/Rfree is about o.18/0.24. RMS(angles):
0.77, RMS(bonds): 0.003
however, when I check with molprobity, There are still some outliers
and some bad contact that is overlap > 0.4 A.
Ramachandran outliers: 0.2% (Goal: < 0.2%)
Ramachandran favored: 97.7% (Goal: > 98%)
Rotamer outliers: 1.4% (Goal: 1%)
C-beta outliers: 0 (Goal: 0)
Clashscore: 14.16
Generally speaking, should I remove all those bad contact or is there
any miminum requirement that ok, the structure is acceptable? I know
the refinement can be endless, but I just want to get an idea how far
away I'm . Thanks.
R
