Dear All, If in a Phenix.refine proces, the Rfree continuously lower than Rwork, what does it mean? Smith
It means that your test set is not independent of your model. If you
solved this structure by MR using a closely related model, you should use
the same test set for the new structure, extending it to higher resolution
if appropriate. The test set can be transferred using data manipulation
utilities in phenix or ccp4.
If it's not possible to maintain the same test set, for example your new
crystal is in a different space group, then simulated annealing of your
model can help remove model bias. You'll probably have to do quite a bit
of rebuilding to remove errors introduced by the simulated annealing.
Best wishes
Kevin
On Mon, Mar 30, 2015 at 8:01 AM, Smith Liu
Dear All,
If in a Phenix.refine proces, the Rfree continuously lower than Rwork, what does it mean?
Smith
_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected]
It may mean a variety of things, such as: - troubled refinement, incorrect target weights for instance; - suboptimal model parametrization or/and optimization method choice; - incorrect test/work flag use in refinement; - this sometimes happens if you refine against ultra-high resolution (higher than say 0.7A). If you send more information such as at least refinement log and eff files then it may be possible to tell more. Pavel On 3/30/15 8:01 AM, Smith Liu wrote:
Dear All, If in a Phenix.refine proces, the Rfree continuously lower than Rwork, what does it mean? Smith
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 03/31/2015 05:35 AM, Pavel Afonine wrote:
[...] - this sometimes happens if you refine against ultra-high resolution (higher than say 0.7A).
At this resolution your data to parameter ratio is often so high, though, that you don't need a test set. Of course, the ratio depends on the unit cell volume. Best, Tim
On 3/30/15 8:01 AM, Smith Liu wrote:
Dear All, If in a Phenix.refine proces, the Rfree continuously lower than Rwork, what does it mean? Smith
_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected]
- -- - -- Dr Tim Gruene Institut fuer anorganische Chemie Tammannstr. 4 D-37077 Goettingen GPG Key ID = A46BEE1A -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iD8DBQFVGkfiUxlJ7aRr7hoRAvZCAKCFN4dgGVF9HxDKHKXkGT3wsVfpAgCffKun 9EwgB5UZgqj0+CDI7UQfD4w= =ZYj7 -----END PGP SIGNATURE-----
participants (4)
-
Kevin Jude -
Pavel Afonine -
Smith Liu -
Tim Gruene