Clarification on Exon Inclusion Representation in Visualizations

[COOL005]

Hello,

I am using your visualization tool for RNA-seq data analysis and have a question about how exon inclusion is depicted. The image shows two graphs in different colors (purple on top and red on the bottom), which I assume represent different splicing events. Below each graph, there are black blocks representing exons.

My question is whether the black block diagrams directly correspond to the graphs above them, i.e., does the purple graph correspond to the set of three exons, and the red graph to the set of two exons? Intuitively, a higher PSI value indicating more inclusion should correlate with diagrams having more exons. However, I noticed that the diagrams always place included exons on top and skipped exons on the bottom. Is this a fixed representation, or should the positions be manually adjusted based on PSI values?

Thank you for your assistance.

[EricKutschera]

The image is from https://github.com/Xinglab/rmats2sashimiplot/tree/v3.0.0?tab=readme-ov-file#using-a-group-file

It's plotting the skipped exon event from -e ./rmats2sashimiplot_test_data/SE.MATS.JC.txt. The coordinates of the exons (black boxes) are taken from that file and are shown at the top of the image. All of the graphs are for that same splicing event. The set of 3 exons is the inclusion isoform and the set of 2 exons is the skipping isoform. In general there is a set of exons for each isoform. There are two graphs (purple and red) because the command used --group-info grouping.gf to define two groups. Without --group-info each sample gets its own graph. The IncLevel is based on the IncLevel columns from SE.MATS.JC.txt and represents the relative support for the inclusion isoform compared to the skipping isoform in that sample or sample group

[COOL005]

Thank you for your previous response. However, I still have some questions that weren't fully addressed. My main question is why the purple group has an IncLevel of 0.21 and the red group has an IncLevel of 0.58. Do these IncLevel values correspond respectively to the plots of three exons and two exons? In other words, does the purple IncLevel of 0.21 correspond to the inclusion of three exons, and the red IncLevel of 0.58 correspond to the skipping of one exon? Or is there a possibility that the plot is reversed, and manual adjustment is needed? I am asking because, in my own plots, it seems that the higher IncLevel is associated with fewer exons being included, which appears counterintuitive. Or is it the case that in the bottom panel of the plot, the isoform with more exons is always displayed on top? I would greatly appreciate any clarification on whether this discrepancy is an issue with the plotting or a misunderstanding on my part. Thank you for your time and assistance.

…

---- Replied Message ---- | From | Eric ***@***.***> | | Date | 10/29/2024 21:27 | | To | ***@***.***> | | Cc | ***@***.***>, ***@***.***> | | Subject | Re: [Xinglab/rmats2sashimiplot] Clarification on Exon Inclusion Representation in Visualizations (Issue #137) | The image is from https://github.com/Xinglab/rmats2sashimiplot/tree/v3.0.0?tab=readme-ov-file#using-a-group-file It's plotting the skipped exon event from -e ./rmats2sashimiplot_test_data/SE.MATS.JC.txt. The coordinates of the exons (black boxes) are taken from that file and are shown at the top of the image. All of the graphs are for that same splicing event. The set of 3 exons is the inclusion isoform and the set of 2 exons is the skipping isoform. In general there is a set of exons for each isoform. There are two graphs (purple and red) because the command used --group-info grouping.gf to define two groups. Without --group-info each sample gets its own graph. The IncLevel is based on the IncLevel columns from SE.MATS.JC.txt and represents the relative support for the inclusion isoform compared to the skipping isoform in that sample or sample group — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>

[EricKutschera]

Each plot (purple and red) is for both sets of exons and each plot shows all the exons and junctions

This post describes how IncLevel is calculated: Xinglab/rmats-turbo#349 (comment)

For the purple graph, IncLevel: 0.21 = ((63+63)/148) / (((63+63)/148) + (235/74))

148 and 74 are the effective lengths taken from SE.MATS.JC.txt. The other numbers are the junction counts shown in the plot