Using cummeRbund with Galaxy output files

[TinaS87]

Hello,

I'm a new user. I would like to compare gene expression data between "WT" and "sample". I processed my data using the following pipeline in Galaxy: Tophat->Cufflinks->Cuffmerge->Cuffdiff.

*NOTE* My colleague has run my RNA-seq data through the same pipeline using a UNIX interface. With his UNIX Cuffdiff data, I do not run into the problem below. This is a specific problem with Cuffdiff data from Galaxy.

So, with the Galaxy Cuffdiff data, I am able to do the following:


WT_v_sample <- readCufflinks(dir = Cuffdiff_WT_v_sample,
gtfFile = Cuffmerge_Exp1,
dbFile="cuffData.db",
geneFPKM = "gene_fpkm_tracking.tabular",
geneDiff = "genes_differential_expression_testing.tabular",
geneRep = "genes_read_group_tracking.tabular",
isoformFPKM = "transcript_FPKM_tracking.tabular",
isoformDiff = "transcript_differential_expression_testing.tabular",
isoformRep = "isoforms_read_group_tracking.tabular",
TSSFPKM = "TSS_groups_FPKM_tracking.tabular",
TSSDiff = "TSS_groups_differential_expression_testing.tabular",
TSSRep = "TSS_groups_read_group_tracking.tabular",
CDSFPKM = "CDS_FPKM_tracking.tabular",
CDSExpDiff = "CDS_FPKM_differential_expression_testing.tabular",
CDSRep = "CDS_read_group_tracking.tabular",
CDSDiff = "CDS_overloading_differential_expression_testing.tabular",
PromoterFile = "promoters_differential_expression_testing.tabular",
SplicingFile = "splicing_differential_expression_testing.tabular",
genome = genomePath,
rebuild = T)


I am able to make a density plot of FPKM (using csDensity) and a box plot of FPKM (using brep) of my data, which suggests that the above code worked fine.

The issue comes when I try to make a heatmap. This is the code I want to use:


SigGeneIds<-getSig(WT_v_sample,alpha=0.05,level='genes')
SigGenes<-getGenes(WT_v_sample,SigGeneIds)
sigHeat<-csHeatmap(SigGenes,cluster='both')


However, when I run "SigGeneIds", it says "character(0)", which means no significant genes were found. This is incorrect as I know I should have 29 significant genes and am able to pull them using the UNIX Cuffdiff data. I have no idea why this doesn't work with the Galaxy Cuffdiff files since they have identical column headings and roughly the same values as the Unix Cuffdiff files.

If you have a suggestion on how I can fix this problem with Galaxy output data, please let me know. Thanks in advance for your help.