XDS-CORRECT and Aimless use
1. different scaling models - CORRECT includes a poorly documented correction across the detector plane not present in Aimless: this may or may not be a Good Thing
2. different outlier rejection algorithms - XDS seems to reject more observations
3. different “correction” of the sigma(I) estimates - XDS seems to do a better job at this
In practice, the differences are likely to be marginal, and it is hard to decide which is better
Phil
> On 21 Nov 2016, at 11:13, Tim Gruene <[log in to unmask]> wrote:
>
> Dear Nishant,
>
> XDS_ASCII.HKL contains corrected, scaled, but not merged reflections.
> You can specifically ask XDS to merge your data, but I would not do so unless
> really necessary - you loose a lot of information.
>
> I would like to offer a different opinion to Graeme's:
> You can read XDS_ASCII.HKL into pointless and aimless and provide aimless with
> the option 'onlymerge'. This way aimless merges the data, but it does not
> rescale them.
>
> XDS performs a couple of corrections in the CORRECT step, the output of which
> is XDS_ASCII.HKL. And while XDS is extremely well documented, I am not sure
> aimless takes into account how XDS treats the data. I would therefore trust
> the step of scaling to the same author and continue with XDS_ASCII.HKL.
>
> Best,
> Tim
>
>
> On Monday, November 21, 2016 11:37:15 AM Nishant Varshney wrote:
>> Dear All,
>>
>> Just to understand more, the XDS_ASCII.HKL file generated after running XDS
>> contains scaled and merged reflections?
>>
>> Moreover, what happens exactly, if you use XDS_ASCII.HKL file in AIMLESS
>> instead of INTEGRATE.HKL file??
>>
>> I ran AIMLESS separately, one using already scaled XDS_ASCII.HKL and
>> another using INTEGRATE.HKL and I found that in the run using XDS_ASCII.HKL
>> little lesser total number of observation but marginally better statistics.
>>
>> Thanks
>> Nishant
>>
>> On Thu, Nov 17, 2016 at 10:08 PM, Andreas Forster <[log in to unmask]>
>>
>> wrote:
>>> Dear Wei,
>>>
>>> if you process your data with XDS, the best is probably to do the scaling
>>> in XDS (CORRECT) and be done with it. If you want to use Aimless for
>>> merging, you can turn off scaling with the ONLYMERGE keyword or use SCALES
>>> CONSTANT.
>>>
>>> All best.
>>>
>>>
>>> Andreas
>>>
>>> On Thu, Nov 17, 2016 at 9:40 PM, Wei Wang <[log in to unmask]> wrote:
>>>> Hi,
>>>>
>>>> Is there a way to let xds_par use less than all processors/threads on the
>>>> machine? Sometimes I would like to process something else while XDS is
>>>> running.
>>>>
>>>> Another question is related to the scaling procedure. My understanding is
>>>> that the XDS already does the scaling during correction. So if I follow
>>>> the
>>>> XDS-Aimless route, then probably I should let Aimless do "skip scaling
>>>> and
>>>> only merge"? Please elucidate me on this issue.
>>>>
>>>> Regards,
>>>>
>>>> Wei
>>
>> --
>> Dr. Nishant Kumar Varshney,
>> IISc-ICTP Fellow
>> XRD2 Beamline, Elettra-Sincrotrone,
>> In Area Science Park,
>> Basovizza, S.S. 14, Km 163,5,
>> 34012 Trieste, Italy
>> +39-040-375 8737 (office ESP4 P1 031)
>> +39-040-375-8435 (XRD2 beamline)
>> +39 3318809798 (Mobile)
> --
> --
> Paul Scherrer Institut
> Dr. Tim Gruene
> - persoenlich -
> Principal Investigator
> Biology and Chemistry
> OFLC/102
> CH-5232 Villigen PSI
>
> Phone: +41 (0)56 310 5297
>
> GPG Key ID = A46BEE1A
>
|