GSEA: None of the gene sets passed the size thresholds

[id0]

I am trying to use GSEA for some RNA-Seq data. I've used it previously for microarray data and it worked fine. My guess is that using well-defined probe set IDs helped.

For my data, I tried using the provided gene symbols as well as symbols exactly matching what I have. It seems no matter what I do, I end up with "None of the gene sets passed the size thresholds" error.

Using GenePattern (which I assume should be the safest option), I get the following output:

1286 [INFO ] Begun importing: Chip from: /xchip/gpprod-upload/servers/genepattern/users/uploads/tmp/run4907393233174326312.tmp/chip.platform.file/1/same.chip
1334 [WARN ] Missing chip file: >/xchip/gpprod-upload/servers/genepattern/users/uploads/tmp/run4907393233174326312.tmp/chip.platform.file/1/same.chip<	at edu.mit.broad.vdb.chip.FileInMemoryChip.initHere(?:?)
1502 [INFO ] Parsed from dotchip : 21862
1350 [WARN ] Missing chip file: >/xchip/gpprod-upload/servers/genepattern/users/uploads/tmp/run4907393233174326312.tmp/chip.platform.file/1/same.chip<	at edu.mit.broad.vdb.chip.FileInMemoryChip.initHere(?:?)
1738 [INFO ] Collapsing dataset was done. Original: 21862x2 (ann: 21862,2,same.chip) collapsed: 860x2 (ann: 860,2,GENE_SYMBOL)
to parse>c5.all.v4.0.symbols.gmt< got: [c5.all.v4.0.symbols.gmt]
1763 [INFO ] Begun importing: GeneSetMatrix from: c5.all.v4.0.symbols.gmt
2110 [INFO ] Got gsets: 1454 now preprocessing them ... min: 3 max: 500
Done removeGeneSetsSmallerThan: 3 for: 501 / 1454
Done removeGeneSetsSmallerThan: 3 for: 1001 / 1454
2259 [INFO ] Done preproc for smaller than: 3
2428 [INFO ] Renaming rpt dir on error to: error_.
2276 [WARN ] Could not rename for error to: error_.	at edu.mit.broad.genome.reports.api.ToolReport.setErroredOut(?:?)

Those warnings look questionable, but they are not exactly informative. Why would it say "Missing chip file" when the chip file is obviously present?

[snaporaz]

hi!

i just started encountering the same error. ('Renaming rpt dir on error to: error_.')
differently from you i'm still working on some microarray data, and i also tried using collapse option and also symbols that mach exactly my dataset probes.

were you able to resolve the problem? could anyone have some advice?

thanks!

p.s. i'm sure that threshold is well above number of the genes in my geneset.

[Elsie]

I haven't used the Broads GSEA in a while, but for RNAseq data you could try a Bioconductor package:

http://www.bioconductor.org/packages/release/bioc/html/SeqGSEA.html

[jamesjcai]

Set 'Collapse dataset to gene symbols' false -- might help.