Hi Jianghai,

I do not know what Refmac uses to get this estimations (I have some ideas but I prefer to refrain from such delicate guessing) and we never did systematic comparisons. phenix.refine estimates a coordinate error as described in:

1) V.Yu., Lunin & T.P., Skovoroda. Acta Cryst. (1995). A51, 880-887. "R-free likelihood-based estimates of errors for phases calculated from atomic models"

2) V.Y., Lunin, P.V. Afonine & A.G., Urzhumtsev. Acta Cryst. (2002). A58, 270-282. "Likelihood-based refinement. I. Irremovable model errors"

The formula is actually in the second paper, but to get a general overview look at both papers.

This error estimation is more or less ok for "relatively high" resolution structures towards the "end" of refinement. Bulk solvent correction must be turned on as well.

Pavel.

Jianghai Zhu wrote:

Hi,

I am refining a 2.6 A structure using both phenix.refine and refmac5. Both programs give me similar R and Rfree values. But the estimated coordinate errors are quite different.

From phenix.refine:

maximum likelihood estimate for coordinate error: 0.42 A

From refmac5:

estimated overall coordinate error:

ESU based on R value: 0.427 A

ESU based on Free R value: 0.293 A

ESU based on maximum likelihood: 0.239 A

I am wondering why they are different.

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

+++++++++++++++++++++++++++++++
_______________________________________________
phenixbb mailing list
phenixbb@phenix-online.org
http://www.phenix-online.org/mailman/listinfo/phenixbb
  