Question

exonicParts vs. disjoinExons inconsistency in usage for DEXSeq - bug?

0

Entering edit mode

story.benjamin ▴ 20

@storybenjamin-11722

Last seen 8 months ago

Germany

The current DEXSeq manual suggests using:

exonicParts = exonicParts( txdb, linked.to.single.gene.only = TRUE )

To generate an "exonic parts" feature set (GRanges object).

However, the current exonicParts call in GenomicFeatures omits a step, that was previously present for disjointExons, that determined and added the exonic_part information to the metadata of the final object.

Thus following the manual in DEXSeq using the R-centric commands does not work and results in an error:

Error in DEXSeqDataSetFromSE(SE, design = ~sample + exon + condition:exon) : make sure your SummarizedExperiment object contain the columns geneid, txname and exonic_part

DEXSeq • 1.1k views

ADD COMMENT • link updated 5.0 years ago by Hervé Pagès 16k • written 5.0 years ago by story.benjamin ▴ 20

0

Entering edit mode

Here would be reproducible code:

library(GenomicRanges)
library(GenomicFeatures)
library(GenomicAlignments)

hse = makeTxDbFromBiomart( biomart="ensembl", dataset="hsapiens_gene_ensembl", host="grch37.ensembl.org" )
exonicParts = exonicParts( hse, linked.to.single.gene.only = TRUE )
bamDir = system.file( "extdata", package="parathyroidSE", mustWork=TRUE )
fls = list.files( bamDir, pattern="bam$", full=TRUE )
bamlst = BamFileList( fls, index=character(), yieldSize=100000, obeyQname=TRUE )
SE = summarizeOverlaps( exonicParts, bamlst, mode="Union", singleEnd=FALSE, ignore.strand=TRUE, inter.feature=FALSE, fragments=TRUE )
colData(SE)$condition = c("A", "A", "B")
DEXSeqDataSetFromSE( SE, design= ~ sample + exon + condition:exon )

ADD REPLY • link 5.0 years ago story.benjamin ▴ 20

0

Entering edit mode

Here is a hacky fix that I came up with (stealing the code from disjointExons)... but maybe something more robust would be better.

ADD REPLY • link 5.0 years ago story.benjamin ▴ 20

score 2 · Answer 1 · 2019-05-23

2

Entering edit mode

Hervé Pagès 16k

@herve-pages-1542

Last seen 13 hours ago

Seattle, WA, United States

Hi,

Thanks for the report. The problem should be addressed in GenomicFeatures 1.36.1 (release) and 1.37.1 (devel). See this commit on GitHub for the details of the changes.

GenomicFeatures 1.36.1 and 1.37.1 should become available via BiocManager::install() in the next 24h or so.

Cheers, H.

ADD COMMENT • link 5.0 years ago Hervé Pagès 16k