Skip to main content

AlphaGenome predictions

TL;DR: AlphaGenome reads a megabase of DNA and says what it does: where it is transcribed, where the chromatin is open, how it splices, and how it folds. This page asks it about TAL1 in two cell lines, stacks the answers that share units on one axis, and then scores an oncogenic insertion against the reference to see the enhancer it creates.

Experimental

The AlphaGenome plugin is beta, and every track on this page is a model's answer rather than a measurement. We welcome your feedback.

Prerequisites

  • the AlphaGenome plugin, which contributes the adapters these tracks use, the query panel, and the variant right-click item
  • nothing installed locally, and no AlphaGenome API key: the two predictions this page reads were recorded once and stored, so following it spends no quota

Where the data comes from

Two predictions over the TAL1 locus, made once against AlphaGenome and frozen under stable tokens, plus the hg38 annotation they are read against. The locus and the variant follow the AlphaGenome team's worked example.

A model instead of an experiment

AlphaGenome is given a window of reference sequence and returns what an RNA-seq, DNase, ATAC, CAGE, PRO-cap or ChIP-seq experiment in a named cell type would have produced over it, plus splice junctions and a contact map. One request carries all of them, per biosample, so asking about K562 and GM12878 together gives two cell types to read against each other.

The model takes sequence, so it answers just as readily for sequence that does not exist: change one base, ask again, and the difference between the two answers is the variant's predicted effect.

The window is a megabase centered on TAL1 on chr1, a transcription factor whose misexpression drives T-cell acute lymphoblastic leukemia. It is on in K562, an erythroleukemia line, and off in GM12878, a lymphoblastoid line, the control in every figure below.

The AlphaGenome plugin

The plugin is beta and not in the plugin store yet, so it loads by URL from the plugins array in config.json (see configuring plugins):

{
  "plugins": [
    {
      "name": "AlphaGenome",
      "umdUrl": "https://jbrowse.org/demos/alphagenome-plugin/jbrowse-plugin-alphagenome.umd.js"
    }
  ]
}

On JBrowse Desktop, install it once from the start screen at Global plugins... → Add custom plugin, with that URL as the plugin URL and the name AlphaGenome.

The plugin holds no API key. It talks to a small service that runs the prediction, stores the arrays, and hands the browser a manifest of byte ranges. The public instance is the default; setApiRoot points it at your own.

The plugin is Apache-2.0, but AlphaGenome's API is offered as a free service for non-commercial use, so the public instance, and any instance run with your own key, carries that restriction.

Ask for a prediction

Open the session below and the locus is already in view, with RefSeq genes and the oncogenic TAL1 variants above it and no predictions yet.

{
  "defaultSession": {
    "name": "TAL1",
    "views": [
      {
        "id": "alphagenome_lgv",
        "type": "LinearGenomeView",
        "assembly": "hg38",
        "loc": "chr1:47,189,833..47,259,832",
        "tracks": ["genes", "tal1_variants"]
      }
    ]
  }
}
jbrowse set-default-session --session - << 'EOF'
{
  "name": "TAL1",
  "views": [
    {
      "id": "alphagenome_lgv",
      "type": "LinearGenomeView",
      "assembly": "hg38",
      "loc": "chr1:47,189,833..47,259,832",
      "tracks": ["genes", "tal1_variants"]
    }
  ]
}
EOF

Open this session in JBrowse

AlphaGenome predictions… in the view menu opens the query panel. Four things go into a request:

  • what to predict. Eleven output types; asking for all of them is one call, and the presets pick the usual groups
  • which cell types and tissues. The box searches the catalog of predictable tracks by biosample, and marks a biosample that has none of the output types asked for. K562 and GM12878 are the two this page uses
  • how wide a window. 16 kb, 100 kb, 500 kb or 1 Mb, centered on the view. The panel prints the interval it will ask about
  • a variant, or none. Left empty, the locus is predicted as it is

A wide request takes minutes, so it is registered and the browser polls for it. Requests are keyed by content with the window rounded to 4 kb, so asking again for what this page already asked returns at once, which is why the prediction behind the figures below is instant. The rounding matters because a view's region comes from an integer pixel offset, and two browser windows of different widths ask about the same locus in coordinates tens of bases apart. A hit can therefore return a window up to about 2 kb off the one asked for, and the track list names the interval that came back.

Two cell lines on one axis

A finished prediction is a list of tracks, often thousands, and none cost anything to add: the arrays are stored, and a track is an HTTP range request into one. Filter the list to polyA plus, tick K562 and GM12878, and Add selected puts both in a single multi-wiggle track.

Predicted RNA-seq in two cell types is a comparison only if the rows share a y-axis. Everything measured in the same units goes onto one axis: several biosamples of one assay, and DNase beside ATAC, splice sites beside splice site usage, CAGE beside PRO-cap, histone ChIP beside TF ChIP. Untick Stack on a shared scale for one track per pick instead.

On that shared axis, TAL1 carries predicted transcription in K562 and essentially none in GM12878, on the annotated exons. The lane to the right is STIL, predicted in both lines, which says the flat GM12878 row is a prediction rather than a track that failed to load.

Predicted polyA plus RNA-seq over TAL1 in K562 and GM12878, both rows on one y-axis. The K562 row carries a block of signal across the annotated exons that the GM12878 row does not. Open this view in JBrowse ↗ (large dataset — this session takes a while to load)

Where the chromatin is open

Add the DNase and ATAC tracks for both biosamples the same way. All four land in one track, because accessibility is one scale, and a DNase peak in K562 should also be an ATAC peak in K562.

The shared axis also makes the rows' floors comparable: predicted ATAC in K562 sits well above everything else across the window, while the same cell line's DNase resolves into peaks. On a per-row scale the two would look alike. The GM12878 rows are not empty either; chromatin is open at plenty of places that are not transcribed.

Predicted DNase and ATAC for K562 and GM12878, four rows on one shared y-axis because accessibility is one set of units. The K562 ATAC row runs high across the whole window; the K562 DNase row below it resolves into peaks. Open this view in JBrowse ↗ (large dataset — this session takes a while to load)

Splicing and folding

The other two output types are not quantitative rows and never join a stacked track.

Splice junctions come back as arcs, a sashimi plot. AlphaGenome returns tens of thousands for a megabase, so the adapter ships them whole and thresholds them in the browser. Add the K562 polyA junctions and the arcs land on the exon boundaries the RefSeq track draws, colored by strand, one color here because every junction over TAL1 is on the gene's strand.

Predicted splice junctions for K562 polyA plus RNA-seq, as sashimi arcs. The arcs span the introns between the TAL1 exons the RefSeq track draws above them, with a second set over STIL to the right. Open this view in JBrowse ↗ (large dataset — this session takes a while to load)

Contact maps come back as a triangle at 2 kb bins, and only for about a dozen cell lines, GM12878 among them. Predicted maps are much less skewed than sequenced ones, so the display saturates at the 95th percentile rather than at a fraction of the maximum.

At the 70 kb this page has been sitting at, a 2 kb map is thirty-five bins across; zoom out to the whole predicted megabase and the domain structure appears. Navigating after the prediction is free, since the adapter reads whatever range the view asks for from the stored arrays.

The predicted GM12878 contact map across the whole 1 Mb window, at 2 kb bins. Blocks of self-interaction meet along the diagonal, with TAL1 near the middle of the view. Open this view in JBrowse ↗ (large dataset — this session takes a while to load)

Scoring a variant

TAL1 is off in GM12878 and on in K562, and in T-ALL patients it is switched on in a lineage where it should be silent. One way that happens is a small insertion upstream of the gene that creates a binding site, and one of those is in the variant track on screen.

Right-click it and the last row of the menu reads Predict variant effect with AlphaGenome. It opens the query panel with the variant loaded: the chip names the position and the two alleles, and the Predict button renames itself to match. Run it, and the same window comes back twice, once for the reference sequence and once with the insertion in place.

Right-clicking a variant in the oncogenic TAL1 variants track. The last row of the menu is the plugin's, and it opens the query panel with the variant under the cursor already loaded. Open this view in JBrowse ↗ (large dataset — this session takes a while to load)

The recorded prediction this page reads is for the Jurkat insertion, chr1:47239297 C>CCGTTTCCTAACC, easiest to reach by typing it into the panel's variant box. Right-clicking any other row is a real API call rather than a stored answer.

Before running one yourself:

  • the variant has to be inside the window. The panel says when it is not, and offers to navigate there
  • splice junctions go sparse in variant mode. AlphaGenome reports only junctions the variant could plausibly affect, which for a variant nowhere near a splice site is none. Clear the variant to get the whole locus back

Reading the difference

Adding a track from a variant prediction gives two tracks: the reference and alternate curves together, and their difference on its own row. At this scale the two curves sit almost on top of each other, which is why the difference gets its own row and a much smaller axis.

On the difference track, positive is where the insertion raises predicted expression and negative where it lowers it. The row is flat across most of the window and moves only over TAL1 and STIL. Each carries positive and negative within the same gene body, a redistribution of predicted coverage.

The flatness everywhere else is the check: a difference track lit up across the whole megabase would be responding to the request rather than the variant.

A variant prediction adds two tracks: the reference and alternate curves together, and their difference below. The difference is flat except over TAL1 and STIL, and its axis spans a fraction of the one above it. Open this view in JBrowse ↗ (large dataset — this session takes a while to load)

What a prediction is, as configuration

The tracks the panel adds are session tracks, and their adapters address a stored array through a presigned URL that expires within the hour, so a track config copied out of one session does not load in the next. Re-open the panel instead: the request is keyed by content, so rebuilding the same query lands back on the same arrays for free.

See also

References

Avsec Ž, et al. AlphaGenome: advancing regulatory variant effect prediction with a unified DNA sequence model. bioRxiv (2025). https://doi.org/10.1101/2025.06.25.661532

Mansour MR, et al. An oncogenic super-enhancer formed through somatic mutation of a noncoding intergenic element. Science 346:1373-1377 (2014). https://doi.org/10.1126/science.1259037