solvent masking

[Marc]

Hi there!

A certain material full of solvent of which I could locate nothing from the difference Fouriers kind of forced me to use a squeezing routine. My choice was the solvent masking procedure in Olex2. Is there any information on how the method works and how it differs from Platon SQUEEZE and also, is there any advice on how to document the outcome in for example the cif file?

thanks,

Marc

[rjgildea]

Hi Marc,

The method implemented in Olex2 is based on that described by van der Sluis and Spek, Acta Cryst. (1990). A46, 194-201, upon which the SQUEEZE routine in PLATON is also based.

Olex2 will automatically take care of ensuring the correct information about the procedure appears in the CIF file if you use the merge cif or report buttons in the GUI. It will add a loop detailing the position, volume and electron count for each void found in the unit cell. It will also ensure that the correct reflection merging statistics appear in your CIF (based on the original hkl file, not the modified hkl intensities required for refinement with SHELXL). You can see what Olex2 will add to the CIF after merging by clicking the "edit cif info" button in the GUI (here you can also edit, delete or add new CIF items).

Cheers,

Richard

[Marc]

thanks Richard, that was the info I was looking for.

Marc

[Marc]

I noticed that the hkl ranges in the -mask.cif file differ from the actual values in the -mask.hkl file which in turn differ from the original hkl file. The system in question is cubic so you can swap hkl around but is this what is actually happening?

C:\Xray\MJR0872\optbusf02>perlx hklranges mjr0872-mask.hkl
h_min=-24 h_max=24 k_min=-42 k_max=42 l_min=-42 l_max=42

mjr0872-mask.cif:
_diffrn_reflns_limit_h_min -29
_diffrn_reflns_limit_h_max 29
_diffrn_reflns_limit_k_min 0
_diffrn_reflns_limit_k_max 30
_diffrn_reflns_limit_l_min 2
_diffrn_reflns_limit_l_max 42

and the original file:
C:\Xray\MJR0872\optbusf02>perlx hklranges mjr0872.hkl
h_min=-42 h_max=42 k_min=-42 k_max=32 l_min=-42 l_max=35

[rjgildea]

Hi Marc,

The routine is carried out after merging in the current space group, this is essential for the procedure to work correctly. The *-mask.hkl contains the modified intensities after merging - to do anything else would result in meaningless merging statistics, so it is better to use the merging statistics based on the original hkl file.

The hkl limits in the -mask.cif are clearly wrong - they are the limits for the reflections used in the refinement (CIF item _reflns_limit_*), rather than for the original intensities (CIF item _diffrn_reflns_limit_*) - this will be fixed in the next release. The values used here should be those you obtained for the original file, if you wish you can click "edit CIF items" under the report menu and edit the CIF items here to the correct value. The fact there are only non-negative values for the k and l indices is a consequence of how the indices are mapped to the asymmetric unit in the merging procedure.

In Olex2 you can type

Code: Select all

HklStat

which will list you the full merging statistics for the current reflection file - including the hkl ranges for both the entire file, and the range that would be included in the refinement after merging.

[celineb]

I have a question concerning the merging. I have a structure in P1 and it seems that the mask procedure is merging the Friedel pairs. Is it possible to avoid it.

Thanks a lot

Celine

[Marc]

Hi Celine,

if you generate the mask through tools->maps->mask Friedel pairs are not merged. work->refine->recompute mask before refinement will, however, merge Friedel pairs.

Marc

[celineb]

Many thanks Marc, this solved my problem.

Celine

[AmyS]

Does anyone know if there is an attempt to make the CheckCIF routine from the IUCr understand the _smtbx_masks_... loop items for solvent masking?

Currently if you append the CIF items in filename.sqf to your CIF file, CheckCIF will understand this as SQUEEZE having been run and demote your "solvent voids" A-level alert to a C- or lower level alert. It would be nice if it would also recognize the _smtbx_ solvent mask in the same way.

[Marc]

AmyS,

I have no answer to that question. Squeezing should trigger plenty related alerts, so one extra alert A should not be a big deal, and you have to comment on the very fact of squeezing anyway. As far as I am concerned the detection of a SQUEEZE-like routine should give a default A** alert instead of a downgrade to code C-G. But then, and just to show how one should treat the presence (and absence) of alerts in general, you might be lucky and get away with a code G alert: "PLAT603_ALERT_4_G Unit Cell TOO large for VOID SEARCH in Structure".

Marc

[Marc]

to comment further on my previous comment, I would rather like to see that the IUCR implements additional cif entries to document reflection statistics, R values, residual densities, etc for refinement against the solvent masking corrected data AND measured data.

[AmyS]

Marc,

I agree with you on that point...about adding extra cif entries. I don't mind the A-alerts SQUEEZE generates. The students who submit CIFs to CheckCIF and then balk at the A-level alerts (even the ones with author responses) tend to like the C-level void alert. But what can you do?

Still, since IUCr does recognize the Platon-SQUEEZE generated .sqf entries, it would be nice to see recognition of the smtbx entries, too.

A.

[jacobo]

Hi experts in Olex2
having done the solvent masking is quite easy, but I am not sure how I get Olex2 to make a new hkl file for me with the solvent contribution subtracted?
Cheers,
Jacob

[jacobo]

Hi all
looks quite easy to generate the solvent mask, but how do I modify the hkl file and obtain a new one for subsequent refinements? That is a little unclear to me
Cheers,
Jacob

[oleg]

When you expand the Refinement settings, you will see the 'Use solvent mask' option, just tick it,

Cheers,

o