RNAseq reads are displayed differently at different zoom levels and in different versions.

[monicacecilia]

How does Apollo load RNAseq reads? And why does their detailed view change so much depending on the zoom level?

This is the instance on staging, centered at Group1.33:409181..437190 (28.01 Kb):

The exact same data, seen at the same coordinates (Group1.33:409181..437190 (28.01 Kb)) looks different in v2.0.1. Note the higher number of light blue "connectors":

Also, it looks like the zoom level on RC3v2.0.2 (and on v2.0.1 as well) affects how the reads are displayed. See the next 3 figures for the same region, and notice how the number of reads that skip the two exons in the middle (count the light blue connectors, again) changes depending on how zoomed into the view the user is:

Figure 1.

Figure 2. Closer

Figure 3. Even closer

[cmdcolin]

This problem (although maybe not apparently similar) is similar to an older bug that we discussed about how the features on the user-created annotation track are not organized on the y-axis intuitively. In general, the y-axis layout algorithm is hard to optimize. It's jbrowse code though, it's not really apollo.

Note that it is possible to override the layout algorithm, even in a plugin (for example, Apollo does this in the "collapse features" code by just making everything y=0) but making the layout better is hard to define

[cmdcolin]

On a related note, the "sashimiplot" plugin we started at the hackathon is a pretty good way to view alternate splicing too. It reduces the deep coverage into more simple colored arcs