Dear All,
We have been using DESeq2 on our RNA-seq data to look for differential expression of genes and it works well.
One issue that keeps on cropping up is the allocation of the EnsemblIDs per row in results(dds), frequently we are getting multiple IDs per row, for example:
ENSG00000001084+ENSG00000231683 6.325517e+02 -0.2914254001 0.10554200 7.586041e+00 5.882200e-03 0.0813763865
Obviously this interfers with annotation so have split it by + and annotated both for gene names etc.
However with many of them per dataset I wondered how best to handle them quickly and easily? I know I could manually check each but with a 100 or so like this per dataset it seems to be not a great use of time.
Searching around I have seen little posted about it, although one suggested just ignore them which seems strange?
How do people handle this? Thanks in advance for any advice.
Cheers
Thanks for the information, appreciate it.
From the sorted BAM we have removed the marked duplicates using samtools view -F 0x0400 and piped this into the dexseq count.py to get the counts for each sample, then prepared the matrix from this data. I can see that output has the multiple ENSG per line, although we have paired end data its not stranded. What would be the best way to handle this? Again thanks in advance .
As far as I know, the dexseq_count.py file from DEXSeq is for exon-level analysis, so not for preparing count matrices for gene-level analysis with DESeq2. Can you take a look at our workflow which describes a number of ways to prepare count matrices for gene-level analysis with DESeq2:
http://www.bioconductor.org/help/workflows/rnaseqGene/
Or is there more to it than dexseq_count.py?