Comparing DEXSeq counting bin and real exons

[alittleboy]

Hi All:

I use DEXSeq for detecting differential exon usage (DEU), and the results seem to be good. One thing I am not sure is that, I know which gene has DEU, and I know which "exonic region (counting bin)" is differentially expressed. However, I wonder if I can accurately tell which "real exons" are differentially expressed. The authors of DEXSeq kindly pointed out that we could overlap the exon bins with the real exons to see to which specific "real" exons my exon bins overlap -- is there any R functions that can:

1. Pass an Ensembl gene ID
2. Output a transcript profile plot (with annotated exons) associated with that gene

For now, I just search Ensembl website and get the profile, but most time, it's hard to tell the correspondences...

Thanks a lot!!

[areyes]

HI a little boy, you could use findOverlaps from the Bioconductor package GenomicRanges to associate exon bins to real exons.

[alittleboy]

HI a little boy, you could use findOverlaps from the Bioconductor package GenomicRanges to associate exon bins to real exons.

Hi @areyes:

Would you please give me some examples on how to get the real exons from the fData(ExonCountSet_obj) output (which contains exon bins start and end information)? Thanks a lot!

[areyes]

It would be something like this:

library(DEXSeq)
library(GenomicFeatures)
library(GenomicRanges)

data("pasillaExons", package="pasilla")
exonBins <- GRanges(
   seqnames=fData(pasillaExons)$chr, 
   ranges=IRanges( 
      fData(pasillaExons)$start, fData(pasillaExons)$end ),    
   strand=fData(pasillaExons)$strand)
   transcriptDb <- makeTranscriptDbFromGFF("/home/alejandro/Work/Graveley/Reanalisis/Annotations/Drosophila_melanogaster.BDGP5.25.62.mychr_tss.gtf", format="gtf")

exonsByTranscript <- exonsBy(transcriptDb, "tx", use.names=TRUE)

findOverlaps( unlist(exonsByTranscript), exonBins )

[alittleboy]

It would be something like this:

library(DEXSeq)
library(GenomicFeatures)
library(GenomicRanges)

data("pasillaExons", package="pasilla")
exonBins <- GRanges(
   seqnames=fData(pasillaExons)$chr, 
   ranges=IRanges( 
      fData(pasillaExons)$start, fData(pasillaExons)$end ),    
   strand=fData(pasillaExons)$strand)
   transcriptDb <- makeTranscriptDbFromGFF("/home/alejandro/Work/Graveley/Reanalisis/Annotations/Drosophila_melanogaster.BDGP5.25.62.mychr_tss.gtf", format="gtf")

exonsByTranscript <- exonsBy(transcriptDb, "tx", use.names=TRUE)

findOverlaps( unlist(exonsByTranscript), exonBins )

Hi @areyes:

Thanks so much! I tried that on my own ExonCountSet object, and it works well. I personally think that exonsByTranscript is more informative as I can pass a particular transcript name to it and see all its exon information as the output. Combined with the plotDEXSeq(), I can tell which real exons are differentially expressed.

[wanfahmi]

Hi areyes, can this script applied to the current DEXSeq version? How can I get the real exon? Thanks

It would be something like this:

library(DEXSeq)
library(GenomicFeatures)
library(GenomicRanges)

data("pasillaExons", package="pasilla")
exonBins <- GRanges(
   seqnames=fData(pasillaExons)$chr, 
   ranges=IRanges( 
      fData(pasillaExons)$start, fData(pasillaExons)$end ),    
   strand=fData(pasillaExons)$strand)
   transcriptDb <- makeTranscriptDbFromGFF("/home/alejandro/Work/Graveley/Reanalisis/Annotations/Drosophila_melanogaster.BDGP5.25.62.mychr_tss.gtf", format="gtf")

exonsByTranscript <- exonsBy(transcriptDb, "tx", use.names=TRUE)

findOverlaps( unlist(exonsByTranscript), exonBins )