Tools for straightforward visualization of alternative splicing events?

Question

[MLN_Bio]

Hi everyone!

I'd like to ask whether anyone has found a tool that allows for straightforward, publication-quality visualization of alternative splicing events. To date I have worked with IGV, ggsashimi and rmats2sashimiplot, and the sashimi plots generated by these are useful, but all 3 tools get out of hand when visualizing genes with high exon counts and more than 3-4 sample groups.

I think the easiest visualization to solve these issues would be a neutral (e.g. grey) global sashimi plot with novel alternatively spliced events colored per sample group of occurrence, but I haven't found anything similar. Any implementation of this or other approaches would be extremely helpful!

Thank you in advance!

[Jeremy Leipzig]

can you sketch up what you are depicting real quick? I bet it would be simple to modify ggsashimi or rmats2sashimiplot for this

[MLN_Bio]

Sure! This is the kind of visualization I'm trying to find, where the exon skipping event would be the part to highlight:

Answer 1

[Pierre Lindenbaum]

I wrote http://lindenb.github.io/jvarkit/PlotSashimi.html which was OK for my needs. It produces a SVG graphics.

https://twitter.com/yokofakun/status/1202587424725127168

https://twitter.com/yokofakun/status/1207337424935936001

[MLN_Bio]

Thank you for your reply! Unfortunately, this seems somewhat similar to the IGV implementation and does not incorporate joint plotting, so the same problem as above would occur for complex genes.

[Pierre Lindenbaum]

what is "joint plotting" ?

[MLN_Bio]

I mean as per the example above, where information from multiple samples is condensed in one single plot and where relevant information is immediately obvious. The usual visualizations I have seen published involve showing 2 or more sashimi plots with the splicing event pointed out somehow (e.g. below, two quick examples in IGV).

This works at a small scale, but it quickly becomes too complex for a figure if, say, we're talking about gene-wide information, and this just with two samples:

Hence why I'm looking for alternative methods to the implementations I've found! Based on the same idea as the mockup from the main thread, this would be the same gene visualized by highlighting AS events in the red sample:

I'm not certain that this would be the most efficient way to show these, but it already condenses the plots quite a bit!