It is worth adding a note that if you are remerging data it is also a great idea to have "sdcorrection noadjust 1.0 0.0" somewhere in your script so that the errors are not inflated a second time by the default values (sdadd of 0.02) - I am pretty sure that this is necessary, certainly seemed to be.
Yes, people appear to do this. I hadn't been bitten by this one quite yet as I have been using "ONLYMERGE RESTORE" with the SCALES file and rescaling / remerging the data but without recalculating the scales. Seems like this wasn't as bad an idea as I thought it was ;o) Perhaps naming the SCALE columns differently when they have been applied so that they are not reapplied would be a good idea, something like SCALEUSED. This takes us back to the beginning of this thread, always fun. Having output of both merged and unmerged reflection files by default could be a cause for some confusion if someone tries to take one of the unmerged files and put it into truncate... Bang! Cheers, Graeme -----Original Message----- From: CCP4 bulletin board [mailto:[EMAIL PROTECTED] On Behalf Of Phil Evans Sent: 20 November 2007 15:51 To: CCP4BB@JISCMAIL.AC.UK Subject: Re: [ccp4bb] Unmerged output from Scala If you don't put INSCALE OFF, the scales will be applied a second time & you will get (much) worse data I didn't know anyone was doing this You should get the same statistics (Rmerge etc) if you reinput an OUTPUT UNMERGED file with ONLYMERGE; INSCALE OFF I'll look into how complicated it would be to write both files (but if you have eg multi-wavelength you might get a lot of files out :-( ) Phil On 20 Nov 2007, at 15:44, Frank von Delft wrote: > Does it make a big difference if you *don't* have "INSCALE OFF"? I > just notice I've been doing it for years. > > I use OUTPUT UNMERGED as well in my standard scaling script, in which > I run two scalas in sequence: the first writes out the unmerged data > and scales, the second only merges. The purpose is to have both > unmerged and merged data after I've scaled, the former > for xprep, pointless etc, the latter for sharp, refinement, etc. > If scala could dump out both versions I wouldn't bother either. > > phx. > > > > > > Clemens Vonrhein wrote: >> Hi Graeme, >> >> even with these options (ONLYMERGE and SCALES CONSTANT) you will have >> the SCALE column applied again to the intensities (not good) - at >> least that's how I understood Phil. You need to use >> >> ONLYMERGE >> INSCALE OFF >> >> to use the already scaled intensities and avoid applying the SCALE >> column again - Phil was the one telling me that. I was scratching my >> head for a long time trying to understand those various >> INTIAL/ONLYMERGE/INSCALE/NOSCALE options and how they relate to each >> other ... >> >> Cheers >> >> Clemens >> >> On Mon, Nov 19, 2007 at 08:07:02PM -0000, Winter, G (Graeme) wrote: >> >>> Hi Phil, >>> I use this option but not these columns. The only time I feed the >>> file back I use "ONLYMERGE" and "SCALES CONSTANT", to remerge the >>> reflections. >>> Cheers, >>> Graeme >>> >>> ________________________________ >>> >>> From: CCP4 bulletin board on behalf of Phil Evans >>> Sent: Mon 19/11/2007 5:07 PM >>> To: CCP4BB@JISCMAIL.AC.UK >>> Subject: [ccp4bb] Unmerged output from Scala >>> >>> >>> >>> Is anyone using the OUTPUT UNMERGED option in Scala? >>> >>> This file contains columns called SCALE & SIGSCALE which are the >>> applied scale and its SD >>> >>> I propose to change the names of these columns so that if you put >>> the file back into Scala the scales do not get re-applied by default >>> (which is wrong since they have been applied already) >>> >>> Will this cause anyone problems? I suspect that very few people or >>> programs are using this file >>> >>> Phil Evans >>> >>> >> >>
