is this plugin available for Jbrowse2 ?

[mthang]

@cmdcolin I have been using this plugin for displaying tracks on Jbrowse1 . I need this plugin for Jbrowse2 but I can't find this plugin under the plugin store. Do you know if this plugin will be made available for the Jbrowse2 soon or it's already available that I am not aware of. Thanks

[cmdcolin]

Hi there!

JBrowse 2 has this plugin essentially built in, we call them "Multi-quantitative tracks" in jbrowse 2, so you don't need a plugin from the plugin store to use it

here is a "user-oriented guide" for multi-quantitative tracks
https://jbrowse.org/jb2/docs/user_guides/multiquantitative_track/

here is a "config oriented guide" for multi-quantitative tracks
https://jbrowse.org/jb2/docs/config_guides/multiquantitative_track/

There are also some new features for multi-quantitative tracks in JBrowse 2 including a built-in clustering mechanism, so it can cluster similar signals. The release announcement shows a screenshot of clustering a bunch of CNV bigwigs https://jbrowse.org/jb2/blog/2025/03/14/v3.1.0-release/#cluster-multi-quantitative-tracks

[mthang]

@cmdcolin Thank you for the information ! It's very helpful and great to know that this plugin is available for Jbrowse2. I will test it out. Thanks.

[mthang]

Hi @cmdcolin I just tested my jbrowse2 to use multi-quantitative tracks, but I am unable to configure it the way I want (See the screenshot) like Jbrowse1.

This is how the tracks.conf look like for the tracks above. Any suggestions how to confgiure it for Jbrowse2 using multi-quantative-tracks?

type = MultiBigWig/View/Track/MultiWiggle/MultiXYPlot
style.height = 100
storeClass = MultiBigWig/Store/SeqFeature/MultiBigWig
autoscale = local
showLabels = true
showTooltips = true
clickTooltips = true
min_score = 0
randomizeColors = true
urlTemplates += json:{"url":"../{refseq}/grain_1.bw", "name": "Grain 1"}
urlTemplates += json:{"url":"../{refseq}/grain_2.bw", "name": "Grain 2"}
urlTemplates += json:{"url":"../{refseq}/grain_3.bw", "name": "Grain 3"}

[cmdcolin]

here is an example

  {
      "type": "MultiQuantitativeTrack",
      "trackId": "mytrack",
      "name": "MultiWig",
      "assemblyNames": ["volvox"],
      "adapter": {
        "type": "MultiWiggleAdapter",
        "subadapters": [
          { "type": "BigWigAdapter", "name": "Grain1", "uri": "v1.cram.bw" },
          { "type": "BigWigAdapter", "name": "Grain2", "uri": "v2.cram.bw" },
          { "type": "BigWigAdapter", "name": "Grain3", "uri": "v3.cram.bw" }
        ]
      },
      "displays": [
        {
          "type": "MultiLinearWiggleDisplay",
          "defaultRendering": "multiline"
        }
      ]
    }

it assigns a default palette in the above config, but you can also manually add color like this...e.g. red orange yellow

  {
      "type": "MultiQuantitativeTrack",
      "trackId": "mytrack",
      "name": "MultiWig",
      "assemblyNames": ["volvox"],
      "adapter": {
        "type": "MultiWiggleAdapter",
        "subadapters": [
          { "type": "BigWigAdapter", "name": "Grain1", "uri": "v1.cram.bw", "color": "#f00" },
          { "type": "BigWigAdapter", "name": "Grain2", "uri": "v2.cram.bw", "color": "#f60" },
          { "type": "BigWigAdapter", "name": "Grain3", "uri": "v3.cram.bw", "color": "#fa0" }
        ]
      },
      "displays": [
        {
          "type": "MultiLinearWiggleDisplay",
          "defaultRendering": "multiline"
        }
      ]
    }

the other settings should mostly be satisfied by default

hope that helps :)

screenshot with the red orange yellow

[cmdcolin]

also note: I think in your setup you use {refseq} in the filename. we do not have the ability to resolve filenames with {refseq} in jbrowse 2 unfortunately, so if you have a single bigwig per chromosome, you'll probably want to merge them all into a single bigwig file and refer to that in the uri

[mthang]

@cmdcolin Thank you for your help ! I can make the tracks displayed the way I want using your instructions. There is a minor thing that is differred from the Jbrowse1 configure is the scaling and autoscale. I tried to reproduce the identical tracks display on Jbrowse2 , but I can't really do so even I set the scaletype=log. The default of maxscore of Jbrowse2 seems to be a lot bigger than Jbrowse1. is that correct?

[cmdcolin]

hmmmm...I am not sure...I could speculate but it might be helpful to see example bigwigs. If you have any available you can email me at colin.diesh@gmail.com or post here

[mthang]

Here's the example. On Jbrowse1 , the grain coverage for Chromsome 1B the maximum height is around 300 (see here ). On Jbrowse2, the grain coverage for Chromosome 1B the maximum height is around 800 (see here )

Region - Chr1B:385945001..393115000

There is a minor thing that I am not sure if it is possible to fix. I highlighted it in the attached screenshot. is it possible to move the vertical scale bar to the right or center , so the "track label" (aka Grain Coverage in my case) wont impose on the top of the scale bar?

[cmdcolin]

In this region there is actually data that extends up to 800 peak max value

You can see this in this screenshot

Background

To rewind how bigwig files work, they have, at the most zoomed in level, just a list of score values

But at zoomed out levels, they create "summary bins" where it stores the min,mean,and max values over a region

How JBrowse 1 handles bigwig files

JBrowse 1 only uses the "average" value of the summary bins by default, but this can easily miss important phenomena in "peaks" and "valleys" in bigwig files

How JBrowse 2 handles bigwig files

In JBrowse 2, we made a couple features to plot "all the values" of the bins with a "whiskers" plotting mode which plots min,mean,and max in a single plot.

This can be valuable in some cases but maybe more weird in others. In the multiline rendering mode, it actually just uses "average" value for the line, but is still choosing the scale to be the max value. This might want to get adjusted so that if it is only plotting the average value, the scale is just the average value also.

Conclusion

Ultimately, it is really confusing how to plot bigwig data correctly. The 'bin' concept is just really weird, and the 'average' value often washes out a lot of potentially important signal. JBrowse 2 made some work to try to remedy this but more improvements could be made.

[cmdcolin]

example of "whiskers" plotting mode showing min (dark blue) mean (blue) and max (light blue) plotting mode

I used "Score->coarser resolution" and "Score->summary score mode->whiskers" to make this screenshot