search

hartleys / JunctionSeq

The JunctionSeq R package is a powerful tool for testing and visualizing differential usage of exons and splice junctions in next-generation, high-throughput RNA-Seq experiments.
28 stars 15 forks source link

Normalized gene counts of 0? #57

Open alanbergland opened 3 years ago

alanbergland commented 3 years ago

Hi -

I am trying to use JunctionSeq to test for differential splicing at novel splice junctions identified in a gene. This gene of interest is strongly differentially expressed (as identified by DESeq2). However, the exonic feature of this gene is flagged as not-testable because of "ALL_ZERO". JunctionSeq shows that there are reads in the "normCount" column, but not in the "normGeneCount" column. Why?

image

hartleys commented 3 years ago

Heh. That is a very oddball case indeed.

So the two counts for each locus are: (1) the number of reads that intersect with exon X. This is the feature count. (2) the number of reads that intersect with the gene but are NOT on exon X. This is the "gene count".

In other words, the gene count = (# reads on the gene) - (# reads on exon X)

So in your case, it looks like all of the reads on this gene also intersect with exon 1. Hence the zero.

This clearly is going to make it impossible to detect differential usage of that exon. All your data for the gene is on that exon so you have nothing to compare it to.

-steve

On Sun, Mar 7, 2021, 6:17 PM Alan Bergland notifications@github.com wrote:

Hi -

I am trying to use JunctionSeq to test for differential splicing at novel splice junctions identified in a gene. This gene of interest is strongly differentially expressed (as identified by DESeq2). However, the exonic feature of this gene is flagged as not-testable because of "ALL_ZERO". JunctionSeq shows that there are reads in the "normCount" column, but not in the "normGeneCount" column. Why?

[image: image] https://user-images.githubusercontent.com/41924004/110258569-7c10dd00-7f60-11eb-940d-e610b834e11d.png

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/hartleys/JunctionSeq/issues/57, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAWC53B6ZUWSOKAWZCVYTM3TCQCPLANCNFSM4YYLVTAA .