Extracting statistically significant differentially expressed exons from DEXseq resu

[pm2012]

Hi

I am trying to extract some data from my DEXseq results. I was able to get the final HTML report for my analysis. However, I was wondering if there's a way I could extract specific exons that are significantly differentially expressed between two conditions. I am interested in getting the gene id (ensembl gene id ok), exon or exon bin id (if applicable), their exon boundaries, fold change & adjusted p value. I am a R/bioconductor newbie. Any help is appreciated.

Thanks

[amandinette]

see the DEUresultTable function

Hi,

You can have a look at the DEUresultTable function, to create a table with gene id and number of exon bin, log2(fold change), adjusted p value...

You can see the DEXSeq vignette on page 18 for usage of the DEUresultTable function :

> res1 <- DEUresultTable(ecs)
> head( res1 )
geneID exonID dispersion pvalue padjust meanBase
FBgn0000256:E001 FBgn0000256 E001 0.046 0.94 1 58.34
FBgn0000256:E002 FBgn0000256 E002 0.040 0.47 1 103.33
FBgn0000256:E003 FBgn0000256 E003 0.035 0.92 1 326.48
FBgn0000256:E004 FBgn0000256 E004 0.035 0.81 1 253.65
FBgn0000256:E005 FBgn0000256 E005 0.046 0.87 1 60.64
FBgn0000256:E006 FBgn0000256 E006 1.067 NA NA 0.79
log2fold(untreated/treated)
FBgn0000256:E001 0.025
FBgn0000256:E002 -0.045
FBgn0000256:E003 0.025
FBgn0000256:E004 0.038
FBgn0000256:E005 -0.013
FBgn0000256:E006 0.126

I think you can extract exon boundaries from the file you made earlier when preparing the annotation (the xxx_flattened.gff file).

Hope that helps
Best regards

[amandinette]

Another piece of information : to extract a table with only significant results (FDR < 10 % for example), you can try something like this

 na.omit(res1[res1$padjust < 0.1, ])