Skip to main content

Pangenome (cattle)

The bovine super-pangenome aligns twelve cattle assemblies against ARS-UCD1.2. The panel is unusually wide for a livestock pangenome, with taurine and indicine breeds beside yak, bison and gaur, so a locus that varies here varies across most of the genus.

Each of those assemblies walks the graph as a named path, and vg deconstruct turns those paths into a VCF. The same locus then reads twice: as a graph showing where sequence is present and absent, and as a callset naming who carries it.

We serve the graph as the same rGFA projections the HPRC pangenome page opens, so the tracks, the adapters and the coarse tier are that page's.

Experimental

The graph view is a beta plugin, and this tutorial covers experimental ideas. Where a step below says the view works a particular way today, the step describes a current limit of the view. We welcome your feedback.

Prerequisites

Where the data comes from

The projections are hosted beside the graph:

Preparing your own graph describes what each of those files holds and how a graph produces them.

Where the graph and the callset show different things

In the BoLA class II region the graph shows a lot of sequence present in some assemblies and absent from the reference. It attributes that sequence only by convention. These graphs record no construction rank, so firstSeenIn in the allele file names the first assembly in a fixed list, and that assembly need not carry the sequence.

The callset attributes each allele. One vg deconstruct call per chromosome over the same graph gives a genotype per assembly:

vg convert -g "$TMPDIR/$k.renamed.gfa" -p > "$TMPDIR/$k.vg"
vg deconstruct -p "chr$k" -a -t "$THREADS" "$TMPDIR/$k.vg" > "vcf/chr$k.vcf.tmp"
The BoLA class II region on ARS-UCD1.2: RefSeq genes, the bubbles lane, the deconstructed callset with one row per assembly, and the allele inventory. The callset resolves alleles the graph lanes can only show as bulk presence and absence. Open this view in JBrowse ↗ (large dataset — this session takes a while to load)

The steps behind the figure above, run against your own data.

  1. Open your genome: on the JBrowse Desktop start screen click Open new genome (or Show all available genomes to pick a hosted one).This figure uses bosTau9, loaded from a .2bit genome.
  2. Add your own track: File → Open track..., then paste a URL or choose a local file. This one needs a bgzip-compressed, tabix-indexed GFF3 (.gff.gz + .tbi).This figure uses “NCBI RefSeq genes (bosTau9)”.
  3. Drag the bar at the bottom edge of the track to resize it.“NCBI RefSeq genes (bosTau9)”.
  4. Add your own track: File → Open track..., then paste a URL or choose a local file.This figure uses “Bovine super-pangenome bubbles”.
  5. Drag the bar at the bottom edge of the track to resize it.“Bovine super-pangenome bubbles”.
  6. Add your own track: File → Open track..., then paste a URL or choose a local file. This one needs a bgzip-compressed, tabix-indexed VCF (.vcf.gz + .tbi).This figure uses “Bovine super-pangenome variants (12 assemblies)”.
  7. Drag the bar at the bottom edge of the track to resize it.“Bovine super-pangenome variants (12 assemblies)”.
  8. Add your own track: File → Open track..., then paste a URL or choose a local file. This one needs a bgzip-compressed, tabix-indexed BED (.bed.gz + .tbi).This figure uses “Bovine super-pangenome: allele inventory”.
  9. Drag the bar at the bottom edge of the track to resize it.“Bovine super-pangenome: allele inventory”.
  10. Type your region of interest into the location box and press Enter.This figure is at chr23:25,844,769-25,968,809.

Open this view in JBrowse Desktop ↗

Opens JBrowse Desktop (JBrowse Desktop 5.0+) at this view and saves it as a reopenable session — swap in your own files afterwards.

Nothing happens? Paste this link into Desktop's Open JBrowse Web link... (start screen, or File → Session):

https://jbrowse.org/code/jb2/main/?config=test_data%2Fgraphgenomeview%2Fpangenome_nonhuman.json&session=spec-%7B%22views%22%3A%5B%7B%22type%22%3A%22LinearGenomeView%22%2C%22id%22%3A%22bovine-bola-lgv%22%2C%22assembly%22%3A%22bosTau9%22%2C%22loc%22%3A%22chr23%3A25%2C844%2C769-25%2C968%2C809%22%2C%22tracks%22%3A%5B%7B%22trackId%22%3A%22bosTau9_ncbiRefSeq_ucsc%22%2C%22type%22%3A%22LinearBasicDisplay%22%2C%22height%22%3A90%7D%2C%7B%22trackId%22%3A%22bovine_minigraph_bubbles%22%2C%22type%22%3A%22LinearBasicDisplay%22%2C%22height%22%3A90%7D%2C%7B%22trackId%22%3A%22bovine_pangenome_vcf%22%2C%22type%22%3A%22LinearMultiSampleVariantDisplay%22%2C%22height%22%3A190%7D%2C%7B%22trackId%22%3A%22bovine_minigraph_alleles%22%2C%22type%22%3A%22LinearPileupDisplay%22%2C%22height%22%3A90%7D%5D%7D%5D%7D&sessionName=JBrowse+docs%3A+pangenome%2Fbovine_bola

This session spec pastes after &session=spec- on any JBrowse Web instance.

{
  "views": [
    {
      "type": "LinearGenomeView",
      "id": "bovine-bola-lgv",
      "assembly": "bosTau9",
      "loc": "chr23:25,844,769-25,968,809",
      "tracks": [
        {
          "trackId": "bosTau9_ncbiRefSeq_ucsc",
          "type": "LinearBasicDisplay",
          "height": 90
        },
        {
          "trackId": "bovine_minigraph_bubbles",
          "type": "LinearBasicDisplay",
          "height": 90
        },
        {
          "trackId": "bovine_pangenome_vcf",
          "type": "LinearMultiSampleVariantDisplay",
          "height": 190
        },
        {
          "trackId": "bovine_minigraph_alleles",
          "type": "LinearPileupDisplay",
          "height": 90
        }
      ]
    }
  ]
}

Loaded against config: test_data/graphgenomeview/pangenome_nonhuman.json

The same view with jbrowse-anywidget. Point the adapter at your own file.

from jbrowse_anywidget import LinearGenomeView, make_assembly

view = LinearGenomeView(
    assembly=make_assembly("bosTau9", "https://your-server/your-genome.fa.gz"),
    location="chr23:25,844,769-25,968,809",
)

view.add_track({
    "type": "FeatureTrack",
    "trackId": "my_track_1",
    "name": "My track 1",
    "assemblyNames": ["bosTau9"],
    "adapter": {"type": "Gff3TabixAdapter", "uri": "https://your-server/your-file"},
})
view.add_track({
    "type": "FeatureTrack",
    "trackId": "my_track_2",
    "name": "My track 2",
    "assemblyNames": ["bosTau9"],
    "adapter": {"type": "MinigraphBubbleAdapter", "uri": "https://your-server/your-file"},
})
view.add_track({
    "type": "VariantTrack",
    "trackId": "my_track_3",
    "name": "My track 3",
    "assemblyNames": ["bosTau9"],
    "adapter": {"type": "VcfTabixAdapter", "uri": "https://your-server/your-file"},
})
view.add_track({
    "type": "AlignmentsTrack",
    "trackId": "my_track_4",
    "name": "My track 4",
    "assemblyNames": ["bosTau9"],
    "adapter": {"type": "BedTabixAdapter", "uri": "https://your-server/your-file"},
})

view  # display the widget

Hand this to a coding agent. It rebuilds the figure above headlessly, and the session file is then the thing to edit — swap an adapter uri for your own file and rerun.

cat > session.json <<'JSON'
{
  "views": [
    {
      "type": "LinearGenomeView",
      "id": "bovine-bola-lgv",
      "assembly": "bosTau9",
      "loc": "chr23:25,844,769-25,968,809",
      "tracks": [
        {
          "trackId": "bosTau9_ncbiRefSeq_ucsc",
          "type": "LinearBasicDisplay",
          "height": 90
        },
        {
          "trackId": "bovine_minigraph_bubbles",
          "type": "LinearBasicDisplay",
          "height": 90
        },
        {
          "trackId": "bovine_pangenome_vcf",
          "type": "LinearMultiSampleVariantDisplay",
          "height": 190
        },
        {
          "trackId": "bovine_minigraph_alleles",
          "type": "LinearPileupDisplay",
          "height": 90
        }
      ]
    }
  ]
}
JSON

npx @jbrowse/capture --instance https://jbrowse.org/code/jb2/main/ \
  --config https://jbrowse.org/code/jb2/main/test_data/graphgenomeview/pangenome_nonhuman.json \
  --session session.json -o figure.png

Using JBrowse with AI agents covers the rest: hosted genomes to point it at, and why waiting for the render is the part that goes wrong.

Read the genotype rows across. Most assemblies carry a different allele here, much as human MHC does. The graph lanes around the callset show none of these genotypes.

The mouse graph is the other end of this. minigraph writes no path lines at all, so there is no callset to deconstruct and nothing recovers carriage. Pangenome (mouse) works with what is left.

A whole chromosome, off the coarse tier

The level-of-detail tier has one node per bubble, which makes a whole chromosome drawable. Type chr23 on ARS-UCD1.2 with the tier and the bubble curve showing. Over a full cattle chromosome the fine segments track refuses with "Too many features", and the tier draws.

A whole ARS-UCD1.2 chromosome with the RefSeq genes, the segments-per-bubble curve and the bubble tier on one axis. BoLA is the densest stretch of the curve. Open this view in JBrowse ↗ (large dataset — this session takes a while to load)

The steps behind the figure above, run against your own data.

  1. Open your genome: on the JBrowse Desktop start screen click Open new genome (or Show all available genomes to pick a hosted one).This figure uses bosTau9, loaded from a .2bit genome.
  2. Add your own track: File → Open track..., then paste a URL or choose a local file. This one needs a bgzip-compressed, tabix-indexed GFF3 (.gff.gz + .tbi).This figure uses “NCBI RefSeq genes (bosTau9)”.
  3. Add your own track: File → Open track..., then paste a URL or choose a local file.This figure uses “Bovine super-pangenome: variability (segments per bubble)”.
  4. Add your own track: File → Open track..., then paste a URL or choose a local file.This figure uses “Bovine super-pangenome: bubble tier (one node per bubble)”.
  5. Track menu → Show... → Labels → None“Bovine super-pangenome: bubble tier (one node per bubble)”.
  6. Drag the bar at the bottom edge of the track to resize it.“Bovine super-pangenome: bubble tier (one node per bubble)”.
  7. Type your region of interest into the location box and press Enter.This figure is at chr23.

Open this view in JBrowse Desktop ↗

Opens JBrowse Desktop (JBrowse Desktop 5.0+) at this view and saves it as a reopenable session — swap in your own files afterwards.

Nothing happens? Paste this link into Desktop's Open JBrowse Web link... (start screen, or File → Session):

https://jbrowse.org/code/jb2/main/?config=test_data%2Fgraphgenomeview%2Fpangenome_nonhuman.json&session=spec-%7B%22views%22%3A%5B%7B%22type%22%3A%22LinearGenomeView%22%2C%22id%22%3A%22bovine-chr23-lgv%22%2C%22assembly%22%3A%22bosTau9%22%2C%22loc%22%3A%22chr23%22%2C%22tracks%22%3A%5B%22bosTau9_ncbiRefSeq_ucsc%22%2C%22bovine_bubble_score%22%2C%7B%22trackId%22%3A%22bovine_minigraph_tier%22%2C%22type%22%3A%22LinearBasicDisplay%22%2C%22showLabels%22%3A%22none%22%2C%22height%22%3A60%7D%5D%7D%5D%7D&sessionName=JBrowse+docs%3A+pangenome%2Fbovine_whole_chromosome

This session spec pastes after &session=spec- on any JBrowse Web instance.

{
  "views": [
    {
      "type": "LinearGenomeView",
      "id": "bovine-chr23-lgv",
      "assembly": "bosTau9",
      "loc": "chr23",
      "tracks": [
        "bosTau9_ncbiRefSeq_ucsc",
        "bovine_bubble_score",
        {
          "trackId": "bovine_minigraph_tier",
          "type": "LinearBasicDisplay",
          "showLabels": "none",
          "height": 60
        }
      ]
    }
  ]
}

Loaded against config: test_data/graphgenomeview/pangenome_nonhuman.json

The same view with jbrowse-anywidget. Point the adapter at your own file.

from jbrowse_anywidget import LinearGenomeView, make_assembly

view = LinearGenomeView(
    assembly=make_assembly("bosTau9", "https://your-server/your-genome.fa.gz"),
    location="chr23",
)

view.add_track({
    "type": "FeatureTrack",
    "trackId": "my_track_1",
    "name": "My track 1",
    "assemblyNames": ["bosTau9"],
    "adapter": {"type": "Gff3TabixAdapter", "uri": "https://your-server/your-file"},
})
view.add_track({
    "type": "QuantitativeTrack",
    "trackId": "my_track_2",
    "name": "My track 2",
    "assemblyNames": ["bosTau9"],
    "adapter": {"type": "MinigraphBubbleAdapter", "uri": "https://your-server/your-file"},
})
view.add_track({
    "type": "FeatureTrack",
    "trackId": "my_track_3",
    "name": "My track 3",
    "assemblyNames": ["bosTau9"],
    "adapter": {"type": "RgfaTabixAdapter", "uri": "https://your-server/your-file"},
})

view  # display the widget

Hand this to a coding agent. It rebuilds the figure above headlessly, and the session file is then the thing to edit — swap an adapter uri for your own file and rerun.

cat > session.json <<'JSON'
{
  "views": [
    {
      "type": "LinearGenomeView",
      "id": "bovine-chr23-lgv",
      "assembly": "bosTau9",
      "loc": "chr23",
      "tracks": [
        "bosTau9_ncbiRefSeq_ucsc",
        "bovine_bubble_score",
        {
          "trackId": "bovine_minigraph_tier",
          "type": "LinearBasicDisplay",
          "showLabels": "none",
          "height": 60
        }
      ]
    }
  ]
}
JSON

npx @jbrowse/capture --instance https://jbrowse.org/code/jb2/main/ \
  --config https://jbrowse.org/code/jb2/main/test_data/graphgenomeview/pangenome_nonhuman.json \
  --session session.json -o figure.png

Using JBrowse with AI agents covers the rest: hosted genomes to point it at, and why waiting for the render is the part that goes wrong.

A graph view pointed at a tier raises maxRegionBp explicitly. The view refuses a cut wider than 5 Mb. That width limit stands in for node count, and it tracks node count well only at segment granularity.

Note

This panel holds a dozen assemblies, against ninety haplotypes for HPRC. Its cuts are chains with a few loops, and the anchored layout often reads better. Check the node and edge counts in the graph pane's header before switching to the force layout. Use the force layout where the bubbles lane reports a tangled window.

BoLA came out of the same ranking the mouse page describes: rank the coarse tier by segments per bubble, then name each entry off the reference annotation. It is the densest stretch of chr23 in the figure above, and the ranking found it without a curated list.

Build it yourself

build_bovine_pangenome.sh downloads the published archive, recovers rGFA tags from its path lines with gfa_paths_to_rgfa.py, projects the files above, and deconstructs the callset. It takes about half an hour after the download.

The script writes a README.txt beside the data recording the source, the modifications, the tool versions and the audits that ran. Copy the audits into your own build. The build stops unless the reference path reproduces the reference chromosome lengths, and stops on a duplicate segment id after renumbering. Without these audits, either failure produces a graph with wrong coordinates, and every downstream check passes on it.

If your own graph has path lines, use build_pggb_tabix.sh in place of build_rgfa_tabix.sh. The PGGB script walks the paths and writes a carriage tag, which rGFA has no field for. The graph view then shows which samples cross each node.

See also

References