Hi, another possibility to avoid the problems Dirk is mentioning: mask the beamstop correctly during data processing. Usually, each data processing package has some tools for doing this (and sometimes even an automatic procedure might actually work ... but I wouldn't rely on it without checking).
Obviously, even then there could be missing overloaded reflections that would result in all strong reflections missing (even if it is only a few). Collecting a second low-dose dataset can be very beneficial. I'm learning not to call it 'low-resolution' dataset any longer, because it sounds as if it won't be necessary when collecting 4A data (when it still might be important). Cheers Clemens On Wed, Dec 03, 2008 at 01:18:41PM +0100, Dirk Kostrewa wrote: > This reminds me of a potential problem with low resolution data that you > might have: if any strong reflection (or a couple of them) is only > partially measured, because it is in the half-shadow of the beamstop, then > the measured intensity is too small, which in turn results in a too low > overall B-factor. For bulk-solvent correction, the Bsol is then usually at > its lower limit and the overall B might still be too low. An easy way to > check this is to re-run TLS, bulk solvent correction and refinement > excluding the lowest resolution shell (say, use 20, 30, 40 A instead of 50 > A or 100 A, or whatever) and check whether the overall Bs are still too > small. If not, you could try to find out which reflections are the culprits > and exclude them. > > Good luck, > > Dirk. > > Am 03.12.2008 um 12:17 schrieb Eleanor Dodson: > >> Then I think it is the problem of fitting the TF parameters into low >> resolution data. >> REFMAC tries to use the lowest ( < 9A) data to get the solvent B factor >> and then after applying that correction to get an overall B. >> Many Wilson plot programs only use the data from 4A out, with the belief >> that the lower resolution stuff is so entangled with cell content >> distribution and solvent scattering that it is un-scramble-able. >> For higher resolution data sets the two methods usually agree pretty well. >> At low resolution I am not sure which is the better.. Arp/wARP tries to be >> smarter but onerarely uses it at that resolution.. >> >> So it is a bit like comparing apples and pears - I doubt if any sensible >> conclusions can be drawn. >> Eleanor >> >> Wim Burmeister wrote: >>> Eleanor Dodson a écrit : >>>> Wim Burmeister wrote: >>>>> Dear all, >>>>> >>>>> I have a 3 A structure refined with REFMAC which gives consistently >>>>> average atomic B-factors of 40 A2, whereas the B factor from a Wilson >>>>> plot is about 60 A2. Is there any explanation for such a discrepancy? >>>>> There are no obvious problems: >>>>> No twinning, spacegroup P21 with two molecules in the asu, no proper >>>>> ncs symmetry. No pathologic Wilson plot, complete and redundant dataset >>>>> (although collected on several crystals with serious problems due to >>>>> radiation damage). >>>>> Interestingly, the Wilson plot of the Fcalc values is about 60 A2 as >>>>> for Fobs in the output dataset. >>>>> >>>>> Yours >>>>> >>>>> Wim >>>>> >>>> Are you using TLS? B factors in the pdb are always relativre to the TLS >>>> paramters unless you have run TLSANL >>>> Eleanor >>>> >>>> >>>> >>> Dear Eleanor, there is no TLS used. >>> >>> Yours >>> >>> Wim >>> > > > ******************************************************* > Dirk Kostrewa > Gene Center, A 5.07 > Ludwig-Maximilians-University > Feodor-Lynen-Str. 25 > 81377 Munich > Germany > Phone: +49-89-2180-76845 > Fax: +49-89-2180-76999 > E-mail: [EMAIL PROTECTED] > ******************************************************* -- *************************************************************** * Clemens Vonrhein, Ph.D. vonrhein AT GlobalPhasing DOT com * * Global Phasing Ltd. * Sheraton House, Castle Park * Cambridge CB3 0AX, UK *-------------------------------------------------------------- * BUSTER Development Group (http://www.globalphasing.com) ***************************************************************
