Known Issues

From GeneSetEnrichmentAnalysisWiki

Revision as of 16:22, 14 February 2007 by Hkuehn (Talk | contribs)
Jump to: navigation, search

<a href="http://www.broad.mit.edu/gsea/">GSEA Home</a> | <a href="../../software/software_index.html">Software</a> | <a href="../../msigdb/msigdb_index.html">MSigDB</a> | Documentation | <a href="../../resources/resources_index.html">Resources</a>


Error in memory.size when running GSEA-R

Problem: When running the example programs provided for R, the following error occurs:

[1] " *** Running GSEA Analysis..."
Error in memory.size(size) : don't be silly!: your machine has a 4Gb address limit

Solution: This is produced by the following line early in the GSEA.1.R file:

 memory.limit(6000000000)

This line set the memory limit to a large size as a work around to a platform problem with an earlier R version.
The easiest fix is just to comment out that line:

 #      memory.limit(6000000000)

This will allocate the default amount of memory. If after this change the program runs out of memory, change the line to:

 memory.limit(max. size in Mbytes available)<br /><br />

Error running leading edge analysis (GSEA v2)

Problem: When you select a report for leading edge analysis, the following error sometimes occurs:

                   java.lang.NullPointerException
at org.genepattern.gsea.LeadingEdgeWidget.setData(EIKM)
at xapps.gsea.LeadingEdgeReportWidget.setData(EIKM)
at xapps.gsea.LeadingEdgeReportWidget$1.run(EIKM)
at java.lang.Thread.run(Unknown Source)

Solution
: Corrected in GSEA v2.0.1.


java.lang.OutOfMemoryError (GSEA v1)

Problem: On the Mac, you can run GSEA from the command line, but when you attempt to use the GSEA application from the desktop you receive errors similar to the following:


Full Error Message ----

na


Stack Trace ----
  1. of exceptions: 1

null------

java.lang.OutOfMemoryError

Solution: Corrected in GSEA v2. In GSEA v1, this is a memory issue with the gsea installer on the Mac. As a workaround, use the following command to launch the GSEA application rather than double clicking the icon:

java -Xmx1800m xapps.gsea.Main


java.lang.NullPointerException (GSEA v1)

Problem: By default, a gene set enrichment analysis uses phenotype permutations. If you have too few samples for phenotype permutation, the following error occurs:


Stack Trace ----
  1. of exceptions: 1

null------

java.lang.NullPointerException
    at edu.mit.broad.genome.alg.DatasetStatsCore.calc2ClassCategoricalMetricMarkerScores(DatasetStatsCore.java:236)
    at edu.mit.broad.genome.alg.markers.PermutationTestBuilder.<init>(PermutationTestBuilder.java:94)
    at edu.mit.broad.genome.alg.gsea.KSTests.shuffleTemplate_canned_templates(KSTests.java:360)
    at edu.mit.broad.genome.alg.gsea.KSTests.shuffleTemplate(KSTests.java:291)
    at edu.mit.broad.genome.alg.gsea.KSTests.executeGsea(KSTests.java:156)
    at edu.mit.broad.genome.alg.gsea.KSTests.executeGsea(KSTests.java:130)
    at xtools.gsea.AbstractGsea2Tool.execute_one(AbstractGsea2Tool.java:103)
    at xtools.gsea.AbstractGsea2Tool.execute_one_with_reporting(AbstractGsea2Tool.java:137)
    at xtools.gsea.Gsea.execute(Gsea.java:111)
    at edu.mit.broad.xbench.tui.TaskManager$ToolRunnable.run(TaskManager.java:468)
    at java.lang.Thread.run(Unknown Source)

Solution: Corrected in GSEA v2. In GSEA v1, use gene_set permutation rather than phenotype permutation. For more information, see the description of the Permutation type parameter on the Run GSEA Page in the GSEA User Guide.

Personal tools