Thanks very much for all of the helpful suggestions! The most useful solution for us was a python script provided by Ed Pozharski at the ccp4wiki link below, which reads a .sca file and produces a table with both <I/sigma> and <I>/<Sigma> for a user-defined number of shells. http://strucbio.biologie.uni-konstanz.de/ccp4wiki/index.php/Calculate_av erage_I/sigma_from_.sca_file Other suggestions are listed below: >From Ed Pozharski: ctruncate outputs the table with Mn(F/sd) (which will be twice Mn(I/sd)) when it does anisotropy analysis >From Phil Jeffrey: I personally report <I/sig(I)> but wrote my own program to do it from the .sca files. >From Paul Smith: I have a perl utility that does much of what you would like. It will run scalepack for you iteratively until the number of rejections converges, or it will parse scalepack output. The output has <I/sigI> by shell gleaned from the scale log with the same resolution bins used in scaling. The usage for parsing is "autoscale.pl -e scale.log"
>From Wladek Minor: Click report on the top panel and you will get all relevant statistics that are necessary for publications. >From Graeme Winter: You can still get this analysis with Scala even after scaling with scalepack. If you output the measurements unmerged (no merge original index) you can convert them to MTZ using pointless, then remerge the data as follows: scala hkiin from_pointless.mtz hklout merged.mtz << eof run 1 all scales constant sdcorrection noadjust norefine both 1 0 0 cycles 0 eof (pointless -c scain ... - you will also need to assign the cell and symmetry) This will just remerge the measurements and give you the usual merging analysis from Scala. Very useful. Same trick also works with data from XDS/XSCALE. ________________________________ From: CCP4 bulletin board [mailto:ccp...@jiscmail.ac.uk] On Behalf Of Radisky, Evette S., Ph.D. Sent: Monday, November 01, 2010 4:18 PM To: CCP4BB@JISCMAIL.AC.UK Subject: [ccp4bb] finding <I/Sigma(I)> from HKL Scalepack Dear all, I have previously used SCALA for data reduction, and in publications and pdb depositions, reported the "Mn(I/sd)" output from SCALA for the whole data set and for the highest resolution shell. We now have some data that has instead been reduced using the HKL suite, and I am confused about how to find the value that would be equivalent to Mn(I/sd) from SCALA. For I/Sigma(I) I've been advised by a colleague more familiar with HKL to manually calculate from average I divided by average error (sigma). As pointed out in a previous ccp4bb thread, this would give me <I>/<Sigma(I)>, which is not the same as <I/Sigma(I)>. Two questions: (1) Is this <I>/<Sigma(I)> what is generally reported in the literature for data processed with the HKL suite? (2) Since I would also like to know the Mn(I/sd) by shell anyway so that I can compare to previous data sets, is there a way to extract this value from the scalepack log, or is there a simple reflection file analysis utility that could read the .sca or .mtz file to extract this information? Thanks for any clarifications or suggestions! Evette Evette S. Radisky, Ph.D. Assistant Professor Mayo Clinic Cancer Center Griffin Cancer Research Building, Rm 310 4500 San Pablo Road Jacksonville, FL 32224 (904) 953-6372