 |
 |
Dear Sudipta—
Herman is correct, you just need to reindex your scala.mtz file to the correct space group.
Regarding translational NCS, it has been many years since I’ve dealt with this, but it is my impression that modern refinement methods treat this much better,
especially with maximum likelihood targets, as the low(er) intensity reflections will have systematically lower sig/noise than the other parity groups.
Felix F. Vajdos
Associate Research Fellow
Structural Biology & Biophysics
Pfizer Inc
Eastern Point Road
MS 8220-3273
Groton, CT 06334
860-715-6504
From: CCP4 bulletin board [mailto:[log in to unmask]]
On Behalf Of [log in to unmask]
Sent: Wednesday, July 16, 2014 3:08 AM
To: [log in to unmask]
Subject: [ccp4bb] AW: [ccp4bb] Translational NCS and molecular replacement.
Dear Sudipta,
you are correct, your original scala.mtz has the wrong space group in it, resulting in very high Rfactors (and presumably bad electron density).
In these cases, I usually reprocess (remerge) the data in the correct space group to get the statistics right (and gain probably a few extra h00 reflections
that got rejected in P212121). If you use the same a, b and c axes as before, you do not need to rerun Phaser, otherwise you have to.
If you have translational NCS, you have to live with it. The only way to get rid of it is to find another crystal with a different crystal packing.
Best,
Herman
Von: CCP4 bulletin board [mailto:[log in to unmask]]
Im Auftrag von Sudipta Bhattacharyya
Gesendet: Dienstag, 15. Juli 2014 20:32
An: [log in to unmask]
Betreff: [ccp4bb] Translational NCS and molecular replacement.
Dear Community,
I have some doubts to clarify. In a way to solve a structure by Phaser-MR, I found phaser ended up with a potentially good MR solution (with good statistics, packing and electron density, and as
we know the homolog structure so in a reasonable biological assembly also). However, the space group where phaser found the potential solution (P22121) is different what we got in pointless (P212121), and phaser also indicated the presence of translational
NCS (NCS translation vector = 0.500, 0.494 0.391). Now when we try to refine the structure with its original scala.mtz file (which is indexed and sclaled in P212121) and phaser generated pdb file, the R/Rfree is very high (around 0.5) but when I tried refining
with mtz file generated by phaser, it was reasonable, at least for first cycle of refinement (R/Rfree, 0.41/0.46). Now my question is, can I use the phaser generated mtz file instead of the original scala.mtz for further refinement? Or I have to reindex my
original data into phaser suggested spacegroup and run the MR again? Translational NCS are generally associated with high R values, is there any way to get rid of that problem?
Best regards,
Sudipta.