| Title: | RNA-seq workflow for differential transcript usage following Salmon quantification |
|---|---|
| Description: | RNA-seq workflow for differential transcript usage (DTU) following Salmon quantification. This workflow uses Bioconductor packages tximport, DRIMSeq, and DEXSeq to perform a DTU analysis on simulated data. It also shows how to use stageR to perform two-stage testing of DTU, a statistical framework to screen at the gene level and then confirm which transcripts within the significant genes show evidence of DTU. |
| Authors: | Michael Love [aut, cre], Charlotte Soneson [aut], Rob Patro [aut] |
| Maintainer: | Michael Love <[email protected]> |
| License: | Artistic-2.0 |
| Version: | 1.32.0 |
| Built: | 2026-10-04 17:13:48 UTC |
| Source: | https://github.com/bioc/rnaseqDTU |
The estimated Salmon counts (scaledTPM) generated as shown in the early steps in the workflow.
ctscts
matrix
http://dx.doi.org/10.12688/f1000research.15398.1
The names of the genes in which the expression of all transcripts was mulitplied by a common fold change.
dge.genesdge.genes
character
http://dx.doi.org/10.12688/f1000research.15398.1
The dispersion parameter used to generate the transcript-level
Negative Binomial counts. Drawn from estimated dispersion from
GEUVADIS project samples. Matches the transcripts in sim_counts.
dispsdisps
numeric
http://dx.doi.org/10.12688/f1000research.15398.1
The names of the genes in which the expression of a single transcript was multiplied by a fold change.
dte.genesdte.genes
character
http://dx.doi.org/10.12688/f1000research.15398.1
The names of the genes in which two transcripts had their abundance (TPM) swapped. Two expressed transcripts were chosen if possible, or one expressed and one non-expressed.
dtu.genesdtu.genes
character
http://dx.doi.org/10.12688/f1000research.15398.1
DEXSeq transcript-level results - see workflow for details.
dxrdxr
data.frame
http://dx.doi.org/10.12688/f1000research.15398.1
DEXSeq gene-level results - see workflow for details.
dxr.gdxr.g
data.frame
http://dx.doi.org/10.12688/f1000research.15398.1
A matrix of fold changes for group 1 and group 2, as in the
format required for polyester input. The majority of the
fold changes are 1. Matches the transcripts in sim_counts.
fold_changesfold_changes
matrix
http://dx.doi.org/10.12688/f1000research.15398.1
The GC bias matrix as returned by plotGC in the alpine package. The columns correspond to GEUVADIS samples as indicated in this workflow publication, and each row corresponds to a percentile of GC, with 101 rows in total. The first 12 columns were used to generate realistic fragment GC bias using alpine and polyester.
frag_GC_biasfrag_GC_bias
matrix
http://dx.doi.org/10.12688/f1000research.15398.1
A logical vector along all transcripts indicating which transcirpts are in the genes with differential gene expression.
iso.dgeiso.dge
logical
http://dx.doi.org/10.12688/f1000research.15398.1
A logical vector along all transcripts indicating which transcripts are chosen for differential transcript expression.
iso.dteiso.dte
logical
http://dx.doi.org/10.12688/f1000research.15398.1
A logical vector along all transcripts indicating which transcripts are chosen for differential transcript expression, and which are the only expressed transcript in the gene, thus not participating in differential transcript usage.
iso.dte.onlyiso.dte.only
logical
http://dx.doi.org/10.12688/f1000research.15398.1
A logical vector along all transcripts indicating which transcripts in the genes chosen for differential transcript usage are chosen to have their abundance (TPM) swapped.
iso.dtuiso.dtu
logical
http://dx.doi.org/10.12688/f1000research.15398.1
DRIMSeq gene-level results - see workflow for details.
resres
data.frame
http://dx.doi.org/10.12688/f1000research.15398.1
DRIMSeq transcript-level results - see workflow for details.
res.txpres.txp
data.frame
http://dx.doi.org/10.12688/f1000research.15398.1
The simulated counts as provided to polyester (representing the base line counts for group 1). The represented transcripts are a subset of all transcripts in the reference, as we do not simulate from transcripts with an expected count of 5 or more.
sim_countssim_counts
numeric
http://dx.doi.org/10.12688/f1000research.15398.1
The simulated counts for the two groups (columns).
Matches the transcripts in sim_counts.
sim.counts.matsim.counts.mat
matrix
http://dx.doi.org/10.12688/f1000research.15398.1
The simulated abundances (TPM) for the two groups (columns). The transcripts correspond to the human transcripts from GENCODE version 28.
tpmstpms
matrix
http://dx.doi.org/10.12688/f1000research.15398.1
A table matching transcript to gene, and providing the number of transcripts per gene. The GTF for human from GENCODE version 28 was used to generate this data.frame.
txdftxdf
data.frame
http://dx.doi.org/10.12688/f1000research.15398.1
The tximport object, summarized to gene level, generated as shown in the DGE section of the workflow.
txi.gtxi.g
list
http://dx.doi.org/10.12688/f1000research.15398.1