Entering edit mode
Just out of curiosity, are there any plans for transcript level differential expression (I.e. Supporting fractional counts), with DESeq2??
Just out of curiosity, are there any plans for transcript level differential expression (I.e. Supporting fractional counts), with DESeq2??
We don't have current plans to extend beyond gene level. The issue is the estimation variance at transcript level, not the fractional part. This can be very high (much higher than the gene-level expression estimation variance) in cases where, for example, 2 isoforms of a gene differ by only a small region. Here, we recommend the DEXSeq approach.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Isn't that what DEXSeq is doing?
DEXSeq specifically looks specifically at exon usage, rather than differential transcript expression doesn't it?
Yes -- DEXSeq only calls differential expression at the exon level and makes no attempt to "roll up" a transcript level expression estimate.