Cant answer your first Q.
For the second - SFCHECK is only a "checking" procedure. It does not
calculate the scale factor in the same way as REFMAC and will very often
give differences (of up to 5%) in Rfactors..
Eleanor
Julien Lafrance-Vanasse wrote:
*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
Hi
Sorry for these real-ccp4 questions:
1. Remote jobs
I tried to run a remote job on a distant computer. It didn't work. I
guess it has to run on a local network with the /home and /tmp shared
by NFS?!? Does anyone has a script integrated in ccp4 that transfers
file to another machine, runs the job and re-transfers the ouput?
2. R and r_free under REFMAC
I decided to run refmac with TLS at the end of my refinement process.
I'm quite new to ccp4, because I usually use cns. The conversion
between the two went well.
I ran SFcheck on the initial model, to mesure r and free_r which gives
0.186/0.208. There's a small difference with cns (0.1854/0.2060), but
that doesn't bother. What really bothers is those values at the end of
the refinement with REFMAC. The values go down during the refinement
process (if I look at the log graph) and the pdb output gives:
REMARK 3 CROSS-VALIDATION METHOD : THROUGHOUT
REMARK 3 FREE R VALUE TEST SET SELECTION : RANDOM
REMARK 3 R VALUE (WORKING + TEST SET) : 0.16981
REMARK 3 R VALUE (WORKING SET) : 0.16822
REMARK 3 FREE R VALUE : 0.20006
REMARK 3 FREE R VALUE TEST SET SIZE (%) : 5.0
REMARK 3 FREE R VALUE TEST SET COUNT : 8921
But if I run SFcheck on the output pdb, it gives me 0.187/0.217, which
is approx 0.02 difference with REFMAC! I really don't understand what
happens. I'm asking myself if the refinement went well and from two
programs, I have different values!
Thank you,
Julien
begin:vcard
fn:Eleanor Dodson
n:Dodson;Eleanor
email;internet:[EMAIL PROTECTED]
tel;work:+44 (0) 1904 328259
tel;fax:+44 (0) 1904 328266
tel;home:+44 (0) 1904 424449
version:2.1
end:vcard
