On Dec 5, 2007, at 14:05, Jianghai Zhu wrote:
Hi all,
I updated refmac5.3 to refmac5.4 and found out that the reported
RMSDs are quite different from these two versions even the refining
protocols are the same.
refmac5.3
rmsBOND rmsANGLE rmsCHIRAL
0.006 0.694 0.046
refmac5.4
rmsBOND rmsANGLE rmsCHIRAL
0.015 1.150 0.004
The R and Rfree are very similar from these two versions of
refmac5. Does refmac5.4 have different geometry weights from
refmac5.3? Which one is more reasonable?
besides the obvious answer that Garib gave (I agree...5.4 is better),
it might be good to remind (without going to the long previous
discussions about the best
rmsBOND), that its the users responsibility to choose correct weights.
having said this, its good to say that half the people will think
0.015 is better, and half that 0.006 is better.
what I find strange is the discrepancy in CHIRAL volumes rms, which I
can attribute either to a bug fix, a print error,
or drastically changed internal weighting between bonds and chiral
volumes.
A.
Thanks.
Jianghai
+++++++++++++++++++++++++++++++
Jianghai Zhu, Ph.D
CBR Institute for Biomedical Research
Department of Pathology
Harvard Medical School
200 Longwood Ave., Boston, MA 02115
Ph: 617-278-3211
Fx: 618-278-3232
+++++++++++++++++++++++++++++++
