Low completeness with twinned data

Please use this forum to post questions on general small molecule X-ray crystallography from theory to practical.
trrk
Sealed Tube
Posts: 31
Joined: 31 Aug 2011, 09:17

Low completeness with twinned data

Post by trrk » 30 May 2017, 15:58

I have a data set which I have processed normally as a single crystal and separately as a twin. The twinned refinment against the hklf5 (with BASF of ~0.2) is slightly better than the single one except that the data is only 83% complete rather than 99% when refined as a single crystal. Of course this gives an A alert in Cifcheck. Does anyone know the reason for this and whether there is any way to fix it or to adjust data collection in the future to avoid this? I am tempted just to use the single refinement as the R factor is only ~1% higher than the twin refinement - do you think this is acceptable? The data was collected and processed with Apex3.

User avatar
wlewis
Rotating Anode With Optics
Posts: 221
Joined: 23 Apr 2009, 19:48
Location: University of Nottingham

Re: Low completeness with twinned data

Post by wlewis » 31 May 2017, 11:09

In the twin handling there will be an overlap rejection somewhere I think. There should be a threshold to set at which point partially overlapped reflections will be rejected.

ecescudero
Trained Monkey
Posts: 8
Joined: 23 Oct 2012, 13:04

Re: Low completeness with twinned data

Post by ecescudero » 11 Apr 2018, 06:25

I have the same problem. hkl5 has a very low completeness. Is there any possibility of using all data of all crystal domains as in TWINABS?
Data were collected with crysalispro 39.12b

Best.
Last edited by ecescudero on 11 Apr 2018, 14:48, edited 1 time in total.

User avatar
wlewis
Rotating Anode With Optics
Posts: 221
Joined: 23 Apr 2009, 19:48
Location: University of Nottingham

Re: Low completeness with twinned data

Post by wlewis » 11 Apr 2018, 15:37

In Crysalis Pro there is a number of options regarding how you merge your data. You need to use the "Twin data finalization" tool, which is the fourth icon down the left-hand side.

You can adjust the overlap threshold in this tool, as well as control what sort of files are output etc. Of particular interest to you may be the option to inlcude beam path information in the HKL - you can then include the direction cosines, and use an external absorption correction program, such as TWINABS.

ecescudero
Trained Monkey
Posts: 8
Joined: 23 Oct 2012, 13:04

Re: Low completeness with twinned data

Post by ecescudero » 11 Apr 2018, 16:24

Thank you very much for your help wlewis. Problem solved.

Return to “General SMX”