To add to Garibs answer - it is a good idea if your data set is pretty
complete - but can enhance model bias if it is not .
Eleanor
Garib Murshudov wrote:
>
> On 18 Mar 2009, at 01:47, Bernhard Rupp wrote:
>
>> I mean *absent* reflections here with fobs=0
>> ----------------------------
>> Dear All,
>>
>> during processing some mtz files I noted that some programs complete
>> the reflection list by padding missing reflections.
>>
>> a) for Refmac (which did not pad reflections automatically) I
>> read that
>> 1) for 2mFo-DFc: Fwt = DFc in case Fo=absent
>> 2) for mfo-fc : Fwt=0 for Fo=absent
>>
>
> It is true. Although there are other options (i.e. ignore them). Many
> tests showed that recovering missing reflections this way (i.e.
> replacing them with their approximately expected values) give better
> maps that not using them at all.
>
>
>
>
>> b) how does PHENIX/refine handle the fobs = absent reflections
>> in either case?
>>
>> c) Prime and switch maps also have padded reflections, but here
>> I am at a loss,
>> because the maps seem to acquire very high positive
>> density - i.e. the distribution
>> seems atypical when read into say Xtalview. Might be a
>> result of the density modification
>> procedure? Can I get the observed reflections only from P&S?
>>
>> Could the program authors kindly shed some light or references on the
>> treatment of absent reflections in their
>> respective programs? In most cases, the padded reflections have not
>> much effect, as expected from their weights.
>>
>> Thx, br
>>
>> Bernhard Rupp
>> 001 (925) 209-7429
>> +43 (676) 571-0536
>> [log in to unmask]
>> [log in to unmask]
>> http://www.ruppweb.org/
>> -----------------------------------------------------------------
>> People can be divided in three classes:
>> The few who make things happen
>> The many who watch things happen
>> And the overwhelming majority
>> who have no idea what is happening.
>> -----------------------------------------------------------------
>>
>
>
|