In that thread Clemens reported his extensive investigation of datasets processed by XDS and found serious issues with apparent twinning. Look for "Warning about twinning (POINTLESS)" in https://www.globalphasing.com/autoproc/wiki/index.cgi?ComparisonProcessing202409 .
I. On Tue, Oct 8, 2024 at 6:21 PM Eleanor Dodson <eleanor.dod...@york.ac.uk> wrote: > Yes - I read that but didn't attach pseudo-twinning problems to it.. > E > > On Tue, 8 Oct 2024 at 18:18, Ian Tickle <ianj...@gmail.com> wrote: > >> >> Dear Eleanor >> There was a whole CCP4BB thread "[ccp4bb] Problem of data reprocessing >> with XDS" on this last week (started by Yimin Hu on Oct 2). >> >> Cheers >> >> -- Ian >> >> >> On Tue, Oct 8, 2024 at 6:08 PM Eleanor Dodson < >> 0000176a9d5ebad7-dmarc-requ...@jiscmail.ac.uk> wrote: >> >>> Maybe we need the two BBs to cross fertalise? >>> E >>> >>> On Tue, 8 Oct 2024 at 17:52, Rafael Marques <rafael_mmsi...@hotmail.com> >>> wrote: >>> >>>> Hi Eleanor, >>>> >>>> I have been seeing complaints about the new XDS software quite a lot, >>>> although I do not remember exactly where. If my memory is not failing me, >>>> it is on the phenixbb. They even released a correction for the updated XDS. >>>> >>>> Best wishes >>>> >>>> >>>> >>>> ______________________________________________________ >>>> >>>> Rafael Marques da Silva >>>> >>>> PhD Student – Structural Biology >>>> >>>> University of Leicester >>>> >>>> Mestrando em Física Biomolecular >>>> Universidade de São Paulo >>>> >>>> Bacharel em Ciências Biológicas >>>> Universidade Federal de São Carlos >>>> >>>> phone: +55 16 99766-0021 >>>> >>>> * "A sorte acompanha uma mente bem treinada"* >>>> *________________________________________________* >>>> ------------------------------ >>>> *De:* CCP4 bulletin board <CCP4BB@JISCMAIL.AC.UK> em nome de Eleanor >>>> Dodson <0000176a9d5ebad7-dmarc-requ...@jiscmail.ac.uk> >>>> *Enviado:* terça-feira, 8 de outubro de 2024 17:46 >>>> *Para:* CCP4BB@JISCMAIL.AC.UK <CCP4BB@JISCMAIL.AC.UK> >>>> *Assunto:* Re: [ccp4bb] a twinning question.. >>>> >>>> I think Gerard has probably hit the jackpot. The data was sent to me by >>>> a student and here is a bit of the header: >>>> >>>> >>>> >>>> >>>> >>>> >>>> * OPENED INPUT MTZ FILE Logical Name: HKLIN Filename: >>>> 105456487_mx32736v69_xGSTPBBRETLGHA00BA54831_scaled_unmerged.mtz * Title: >>>> From XDS file SCALED_SAD_SWEEP1.HKL, XDS run on 19-Sep-2024 from image* >>>> >>>> >>>> I will check what version of XDS was used then but I suspect it is the >>>> one recently updated! >>>> The structure solves quite well in P212121 and refines to ~ 23% >>>> Not good data - not a perfect structure but probably NOT twinned. >>>> >>>> Can someone warn the community about this? >>>> >>>> Eleanor >>>> >>>> >>>> On Tue, 8 Oct 2024 at 17:38, Ian Tickle <ianj...@gmail.com> wrote: >>>> >>>> PS you'll find a link to a description of all the twinning tests it >>>> runs on the server home page. >>>> >>>> I. >>>> >>>> >>>> On Tue, Oct 8, 2024 at 4:20 PM Ian Tickle <ianj...@gmail.com> wrote: >>>> >>>> >>>> Dear Eleanor >>>> >>>> Try submitting it to the STARANISO server? That does a whole slew of >>>> twinning tests (so you can take your pick from sometimes totally different >>>> twin fractions from the various tests!). Also it gives you the twin >>>> fraction rather than having you understand how to use the tests! But I >>>> agree that the L test is one of the best, and the server uses a variant of >>>> it which gives more accurate estimates of the twin fraction. >>>> >>>> Cheers >>>> >>>> -- Ian >>>> >>>> >>>> On Tue, Oct 8, 2024 at 4:01 PM Eleanor Dodson < >>>> 0000176a9d5ebad7-dmarc-requ...@jiscmail.ac.uk> wrote: >>>> >>>> I have always had great faith in the L test and other statistics to >>>> detect twinning.. >>>> But now I have a puzzling data set. >>>> Seems to be orthorhombic with a, b and c all different >>>> One molecule/asymm unit and no NC translation.. >>>> BUT clear indication of twinning given by second moments, etc.. >>>> >>>> I have seen this before when the spacegroup was actually monoclinic and >>>> the pointless analysis of the two fold axes along a b or c showed one was >>>> better than the other two and in fact the crystal proved not to be twinned >>>> when the data was processed as monoclinic.. >>>> >>>> But in this case no amount of reprocessing gets rid of the twin >>>> indicator .. >>>> Does anyone have any suggestions of why this might be? >>>> ( The data quality isnt great - 2.5A but the Rmerges are reasonable..) >>>> Eleanor >>>> >>>> >>>> ------------------------------ >>>> >>>> To unsubscribe from the CCP4BB list, click the following link: >>>> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1 >>>> >>>> >>>> ------------------------------ >>>> >>>> To unsubscribe from the CCP4BB list, click the following link: >>>> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1 >>>> >>> >>> ------------------------------ >>> >>> To unsubscribe from the CCP4BB list, click the following link: >>> https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1 >>> >> ######################################################################## To unsubscribe from the CCP4BB list, click the following link: https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=CCP4BB&A=1 This message was issued to members of www.jiscmail.ac.uk/CCP4BB, a mailing list hosted by www.jiscmail.ac.uk, terms & conditions are available at https://www.jiscmail.ac.uk/policyandsecurity/
