Stein, ND (Norman) schrieb:
> Dear Justin
>
> I have been unable to reproduce either of these problems. Perhaps
> something went wrong during the self compilation process? Something odd
> seems to have gone on even in the logfiles you classify as good. For
> example, in
>
> http://dev.gentooexperimental.org/~jlec/paste/good.log.html
>
> there are nan's in the L test data and the cumulative intensity
> distribution, and the acentric moments of E all seem to be exactly 1.
>
> Note that if your input mtz file contains anomalous data, you will get
> different results depending on whether or not you specify '-colano'. If
> you do specify it, F(+) and F(-) will be calculated from I(+) and I(-)
> respectively, and FMEAN will be calculated from FMEAN = 0.5*[F(+) +
> F(-)]; if you don't FMEAN will be calculated from IMEAN.
>
> You can download a linux binary for the latest ctruncate (including the
> patch) from
>
> ftp://ftp.ccp4.ac.uk/nds/bin/ctruncate
>
> Best wishes
>
> Norman
>
> Norman Stein
> CCP4
> Daresbury Laboratory
>
>
>
>
>
>
> -----Original Message-----
> From: CCP4 bulletin board [mailto:[log in to unmask]] On Behalf Of
> Justin Lecher
> Sent: 08 July 2009 07:57
> To: [log in to unmask]
> Subject: [ccp4bb] ctruncate problems
>
> Dear all,
>
> I have two problems with ctruncate. I tested two versions, the "old" one
> from the 6.1.1 binary distro and a self compiled version including the
> patch.
> As dataset I used data/mtz/X13089.mtz from the test-framework, which
> should be a twinned dataset.
>
> problem 1:
> The new version isn't able to detect the twinning anymore.
>
> problem 2:
> When running from ccp4i the used command is /usr/bin/ctruncate -mtzin
> data/mtz/X13089.mtz -mtzout out.mtz -colin "/*/*/[IMEAN,SIGIMEAN]"
> -colano "/*/*/[I(+),SIGI(+),I(-),SIGI(-)]"
> Either running it from the interface or the shell it works fine but the
> resulting mtz[2] is broken.
> From the manpage it says that those -colin or -colan are the default
> values and passing them should give no difference compared to without.
> But it actually does.
> If I run the command like it is used in the test-framework
> /usr/bin/ctruncate -mtzin data/mtz/X13089.mtz -mtzout out.mtz the
> resulting mtz is fine [1].
>
> Does anybody experienced similar problems?
>
> These are the links to the mtzdumps and build.logs of the different
> runs.
>
> [1]
> mtzdump from "unbroken" mtz
> http://dev.gentooexperimental.org/~jlec/paste/mad.html
>
> [2]
> mtzdump from "broken" mtz
> http://dev.gentooexperimental.org/~jlec/paste/good.html
I mixed the broken and unbroken mtz. [2] is of course the unbroken one
and [1] is the "good" one.
--
Justin Lecher
Institute for Neuroscience and Biophysics
ISB3 - structural biochemistry
Research Centre Juelich GmbH,
52425 Juelich,Germany
phone: +49 2461 61 5385
|