Hi Robert,
The problem is that you've given each MarvinNOEPotential object the
same instance name--"3dC". Those instance names are used inside of
the network filter to keep track of which peaks and shift assignments
came from which spectrum.
I obviously need to add a test to ensure that they're all unique. (I
didn't bother because I thought that the create_MarvinNOEPotential
method was catching that error. Doh.)
While you're at it, please make sure that the name prefixes that
you're giving to each peak and shift assignment are different for
each spectrum. Those are set in the initMatch scripts.
Thanks for catching the bug!
--JK
On Apr 24, 2008, at 2:04 PM, Robert Schneider wrote:
> Hi everyone,
>
> I'm trying to run PASD in xplor version 2.18 on my data. I have 4
> spectra (recorded with different mixing times) that I'm putting in;
> initial matching scripts (one for each spectrum) ran fine, logfiles
> look
> promising in terms of assignments found. I then rewrote
> jointFilter.tcl
> to read in shiftAssignments and peaks from all spectra (since in
> eginput/marvin/cvn there's also only one jointFilter script that reads
> in both 3dC and 3dN), but once running, it stays at the netFilter step
> forever. The last entry in the logfile shows that netFilter was
> started,
> but nothing else is logged; my calculation was then aborted by the
> queueing system after 48h (!). Is it just too much for the
> netFilter to
> handle data from 4 spectra? (Each spectrum has ~300-500 peaks,
> there are
> ~350 shiftAssignments.) Would it help to split jointFilter in two or
> four to only handle two spectra or one spectrum each? Or have I
> overlooked something?
>
> I'm attaching the jointFilter script I'm using up to the netFilter
> step.
> Any help would be greatly appreciated. Thanks very much in advance!
>
> Best greetings,
> --robert.
>
>
>
>
>
> package require aeneas
>
> #
> # this entire example takes about 12 hours to run on a 60-processor
> cluster.
> # If you just want to verify that things run, run all the scripts with
> # the flag -quick
> #
>
> set quickMode [flagExists $argv -quick]
>
> #
> # initialization
> #
>
> initParamPsfPdb \
> -psfFileName ./prot.psf \
> -randomCoords
>
> #
> # quick mode will only calculate the structure of the first 36
> residues
> of cyanovirin
> #
>
> if {$quickMode} {
> XplorCommand "delete sele (not (resid 1:36)) end"
> }
>
> #
> # make sure the initial coords have good covalent geometry
> #
>
> cleanCovalentGeom
>
>
> #
> # set up the 3dC NOEs
> #
>
> XplorCommand "set message off echo off end"
>
> set noe_3dc1 [create_MarvinNOEPotential 3dC]
>
> readShiftAssignments \
> -fileName "prot_3dc1_pass1.shiftAssignments" \
> -pot $noe_3dc1
>
> readMarvinPeaks \
> -fileName "prot_3dc1_pass1.peaks" \
> -pot $noe_3dc1
>
>
> XplorCommand "set message off echo off end"
>
> set noe_3dc2 [create_MarvinNOEPotential 3dC]
>
> readShiftAssignments \
> -fileName "prot_3dc2_pass1.shiftAssignments" \
> -pot $noe_3dc2
>
> readMarvinPeaks \
> -fileName "prot_3dc2_pass1.peaks" \
> -pot $noe_3dc2
>
>
> XplorCommand "set message off echo off end"
>
> set noe_3dc3 [create_MarvinNOEPotential 3dC]
>
> readShiftAssignments \
> -fileName "prot_3dc3_pass1.shiftAssignments" \
> -pot $noe_3dc3
>
> readMarvinPeaks \
> -fileName "prot_3dc3_pass1.peaks" \
> -pot $noe_3dc3
>
> XplorCommand "set message off echo off end"
>
> set noe_3dc4 [create_MarvinNOEPotential 3dC]
>
> readShiftAssignments \
> -fileName "prot_3dc4_pass1.shiftAssignments" \
> -pot $noe_3dc4
>
> readMarvinPeaks \
> -fileName "prot_3dc4_pass1.peaks" \
> -pot $noe_3dc4
>
>
> set peakRemarks3dC1 [list]
> set saRemarks3dC1 [list]
> set peakRemarks3dC2 [list]
> set saRemarks3dC2 [list]
> set peakRemarks3dC3 [list]
> set saRemarks3dC3 [list]
> set peakRemarks3dC4 [list]
> set saRemarks3dC4 [list]
> set peakRemarksJoint [list]
>
>
> initializePeakAssignmentNumFiltersFailed -pot $noe_3dc1
> initializePeakAssignmentNumFiltersFailed -pot $noe_3dc2
> initializePeakAssignmentNumFiltersFailed -pot $noe_3dc3
> initializePeakAssignmentNumFiltersFailed -pot $noe_3dc4
>
> currentFilterSummary $noe_3dc1
> currentFilterSummary $noe_3dc2
> currentFilterSummary $noe_3dc3
> currentFilterSummary $noe_3dc4
>
>
>
> netFilter \
> -pot [list $noe_3dc1 $noe_3dc2 $noe_3dc3 $noe_3dc4] \
> -intraresidueNeighborhoods \
> -extendNeighborhoodsWithHighScoringPeakAssignments \
> -failedFiltersCutoff 0 \
> -numIterations 5 \
> -expectedContactsPerShiftAssignment [list 10 10 10 10] \
> -createInverseExceptions \
> -preferIntra \
> -resetNumFailedFilters \
> -remarksVariableName peakRemarksJoint
> _______________________________________________
> Xplor-nih mailing list
> Xplor-nih at nmr.cit.nih.gov
> http://dcb.cit.nih.gov/mailman/listinfo/xplor-nih
