A colleague of mine asked me what the height in the sashimi plots created by Gviz represents. It was not intuitive and I dug around quite a bit before coming to the following conclusions:
* the height of an arc is not informative, it's just for graphical purposes.
* the Y-axis on the sashimi arcs shows the relationship between the line width the the number of reads spanning a given junction.
Below is a screenshot of the Gviz 1.8.0 (Bioc 3.4 release) vignette. It's not initally clear why the arcs have different heights.
I did look at the vignette, the documentation, the sashimi plot paper https://academic.oup.com/bioinformatics/article-abstract/31/14/2400/254271/Quantitative-visualization-of-alternative-exon and the code. Eventually I understood that the code at https://github.com/Bioconductor-mirror/Gviz/blob/eea629439789367f39908c338844b1a5d7c267d9/R/Gviz-methods.R#L2015-L2028 is for determining the height of the sashimi track (how much space is needed for the plot) and that the score (or scaled score) https://github.com/Bioconductor-mirror/Gviz/blob/eea629439789367f39908c338844b1a5d7c267d9/R/Gviz.R#L2646-L2651 was not really used for that part since `.sashimi.junctions()` is called later on https://github.com/Bioconductor-mirror/Gviz/blob/08eebe8c2c27f4ef95e53078306002d5072e011e/R/Gviz-methods.R#L2035. That's when I realized the meaning of the Y-axis label (which is not part of the original sashimi plot paper).
So, I just want to make sure that we got the interpretation correctly of the arc heights. If so, it'd be great if this information was described in the Gviz vignette (and maybe also the R docs).