What the x-axis of this box plot is measuring?

Question

[arronar]

Hi. i'm trying to understand a box plot from a paper where they used ChIP-seq to identify relations between histones modifications and exons. Here is the plot

As you can see there are some specific clusters of exons (two from more included and one form less included) that interacts with histones .

What I'm trying to understand is what the x-axis measures in order to understand why in the case of more included exons the box plots lie in positive values while the less included exons box plots lie in negative values.

Here is the DOI of the paper DOI 10.1186/s13059-015-0797-8

[karl.stamm]

Unlabeled axes mean the journal is not worth your time to read. The figure caption says it measures "ChipSeq differential signal". In the Supplemental materials, another similar figure (S3) is present and they have the axes labeled "differential signal", and its caption says "differential ChipSeq signal". The supplemental materials section talks of "signal" behind several promoters from Chipseq or Chia-pet. Might have to find the supplemental methods for those physical experiments, because it sort of looks like most of the authors and reviewers did not really understand what was going on. Take a lot of numbers and throw them at the wall until one of them appears to be p<0.05.

Answer 1

[Noushin N]

From the figure legend:

K-means clustering based on epigenetic signatures of regulated exons. a, b boxplots represent differential ChIPSeq signal for more included (a) and less included (b) exons.

Seems like the x-axis represents a measure of differential ChIPSeq signal. Each subplot in the panels left and right represent data points coming from one of the K-means clusters identified. They define what they mean by the differential signal in their results section R2.

[karl.stamm]

Thanks Noushin, I never would have looked for methods in the Results section. There they define differential signal as "difference of the averaged normalized signal" of an exon between two conditions. Still looking for what is "signal".

Later they talk about "Exonic Average Signal (EAS)" which just might be the same signal. It's described in Methods as the "sequencing-depth-normalized read density for each position on the genome" from ENCODE. So you can go to ENCODE to find out how they like to do sequencing depth normalization, or be satisfied that all these "signals" are some sort of read depth from NGS.

Why the authors thought to define an abbreviation (EAS) and never use it in the text is beyond comprehension.