Proteins on genomes.jbrowse.org
3 videos in this tutorial
TL;DR: genomes.jbrowse.org loads the protein3d and msaview plugins, so any gene in a linear genome view can be taken to a 3D structure or to a cross-species protein MSA with nothing prepared beforehand. Both views stay linked to the genome, so hovering a variant highlights the residue it lands on.
Prerequisites
- nothing to install: this is a click-path through hosted sites, and no sequence, structure, alignment or tree is prepared by hand
- to add these views to your own JBrowse instead, see Adding the plugins to your own instance
Where the data comes from
genomes.jbrowse.org's hosted hg38 config carries the gene track the examples below click through; the protein3d and msaview plugins resolve everything else live, per gene, from the services beside them.
- hg38: https://jbrowse.org/ucsc/hg38/config.json
- AlphaFold DB, where a launched structure comes from: https://alphafold.ebi.ac.uk/
- UniProt, the isoform mapping and the projected Domains, Chains, Mutations and AA Modifications tracks: https://www.uniprot.org/
- NCBI's ortholog report, the MSA's rows: https://www.ncbi.nlm.nih.gov/datasets/docs/v2/reference-docs/rest-api/
- NCBI's Conserved Domain Database, the alignment's overlay: https://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml
Two ways to open a protein view
Two hosted sites reach the same three linked views.
genomes.jbrowse.org suits starting from a genome. It hosts a JBrowse instance for every UCSC genome, each with that genome's UCSC track catalog and both protein plugins. Search a gene, right-click it, and the two launchers are in the menu. The rest of this page walks that route.
The JBrowseMSA Gene Explorer suits
starting from a gene. Pick a species, type a gene symbol, and Open in
JBrowse opens a session with all three views built and connected. It takes a
gene and a taxon in its own URL, so
?gene=TP53&taxon=9606
arrives with the gene resolved.
The Gene Explorer's genome view collapses the introns, so the whole CDS is on screen at residue zoom. Its catalog is seven species: human genes arrive with all three views, and mouse, zebrafish, fly, worm, plant and yeast arrive as a genome view and a structure.
Launching a structure
Open hg38 on genomes.jbrowse.org and type TP53
into the location box. The hosted config ships a name index, so gene symbols
work with no setup.
Right-click the gene. The menu carries Launch protein view from protein3d and Launch MSA view from msaview, which the next section takes. Choose Launch protein view.
The dialog opens on its AlphaFoldDB search tab with Auto-detect using UniProt ID mapping API selected. It maps the transcript's accession to UniProt entries and picks the reviewed one. Choose transcript isoform picks which transcript becomes the query, tagging the isoforms whose translation matches the structure's residues. Launch renders the structure with Mol*.
The structure carries its own sequence, which often differs from the transcript's translation: an AlphaFold model covers one UniProt isoform, and a PDB entry can be a construct, a fragment or another species. When the two differ the dialog aligns them in the browser before mapping any position. The gear beside that notice opens Alignment settings: Smith-Waterman (local alignment), the default and the one to keep for a structure covering part of the protein, Needleman-Wunsch (global alignment), or Import manual alignment... for a pairwise alignment in Clustal format. The same options sit under Advanced... in the view menu afterwards.
The structure arrives with the genome view still above it. Hovering a genomic position highlights the matching residue on the structure, on the pairwise alignment above it, and in the per-residue tracks; hovering the structure highlights the genomic position.
A genomic position maps to a residue through the transcript's CDS with g2p_mapper, so introns, UTRs and residues the structure is missing highlight nothing. A missing residue shows as a gap in the Pairwise alignment panel above the structure, which carries the transcript row against the structure row with a consensus line. Folding the transcript's own sequence with AlphaFold gives an exact correspondence.
The lookup needs a gene feature carrying a recognizable protein or transcript ID, which the RefSeq gene tracks on the hosted configs carry.
The protein view holds the AlphaFold structure, the genome-to-structure alignment, and per-residue tracks for pLDDT confidence, domains, helices and hydrophobicity. The gear beside the dialog's Launch button opens Launch settings, whose side-by-side option puts the protein view beside the genome view.
Other views the Launch menu builds
The arrow beside Launch lists everything the dialog can build.
Launch 1D protein annotation view opens a linear genome view whose genome is the protein. The plugin registers the UniProt accession as a temporary assembly whose reference sequence is the amino-acid sequence, then adds a track per UniProt feature type, plus Antigen, Variation, AlphaFold confidence (pLDDT) and AlphaMissense scores. The view opens empty, with the tracks in its selector under Session tracks. Coordinates are residues. It needs a session it can add tracks to, so it is absent from the single-view embedded components.
The dialog's other two tabs take a structure from elsewhere: Foldseek search finds structures resembling the protein's own, and Open file manually takes a PDB or mmCIF file of yours.
Launching an MSA
Launch MSA view on the same right-click menu builds a cross-species protein MSA. The dialog opens on its Orthologs (fast) tab, and three fields matter:
- Query species is the species the gene came from, free text resolved against NCBI's taxonomy: a scientific name, a common name or a taxon id
- Rows to align is how many species to build. NCBI orders its ortholog report from the reference organisms outward, so this takes the closest N
- Choose isoform picks which transcript becomes the query row, the one the genome view stays linked to
MSA Algorithm is what EBI is asked to run, Clustal Omega by default, and it is where the wait is. The cost scales with the row count.
NCBI publishes one ortholog gene per species for most annotated genes. A gene with no resolvable symbol goes through the dialog's NCBI BLAST query tab.
Press Submit. A multiple sequence alignment view opens below the genome view, with a tree on the left, the alignment beside it, and the conserved-domain overlay drawn over the residues once NCBI returns it.
The view opens at residue zoom, a window on the N terminus of a long protein. Fit horizontally, under the toolbar's fit and zoom button, puts the whole alignment on screen, which is the zoom the domain blocks read at.
Both launchers place their new view beside the ones already open. Each panel's menu carries Global: tile horizontally, which lines every open view up side by side so a hover reaches all three genome-linked views at once.
The figures below take NLRP1, an inflammasome sensor whose domain architecture varies between mammals.
Reading the overlay
Each colored block is an NCBI conserved domain, drawn in alignment columns: the same domain lands in the same column in every row that has it, whatever the proteins' lengths.
Human NLRP1 carries a pyrin (PYD) death-fold domain at its N terminus. Some rows have it and some do not, mouse Nlrp1a among those that do not. The aligner's tree orders the rows, so the rows that have it sit together. The shared core is the control: NACHT, the winged helix, HD2, FIIND and CARD run across every row.
The calls ride along on NCBI's own protein records. For a protein NCBI has no calls for, File → Annotations → Open InterProScan results... reads them from a file and File → Annotations → Query InterProScan... computes them.
Checking the overlay against the residues
A whole-protein view cannot tell "no domain annotated" from "no sequence", so read the residues. Scroll back to the zoom the view opened at, or use the alignment's Zoom in button until the letters return.
An alignment is as long as its longest row, so its leftmost columns belong to whichever protein reaches furthest past the others. Hide columns w/ >N% gaps, the slider in the alignment's toolbar, brings the columns the panel shares (the pyrin among them) to the left edge.
The rows without a pyrin block come in three kinds: ordinary residues with nothing called over them; a generic death-domain call in place of the pyrin-specific one; and gap right across the frame, where the row begins further right. The third kind is the absent sequence.
The same domains in genome coordinates
UniProt annotates the same proteins independently of NCBI's conserved-domain database, and UCSC projects those annotations onto the genome, so the hosted config carries them as ordinary tracks. UniProt - Domains, under Genes and Gene Predictions, is the domain architecture in genomic coordinates.
Turn it on in the linear view you launched from. NLRP1 is transcribed right to
left, so the pyrin block sits at the right-hand end of the gene, where its N
terminus is, and NACHT, FIIND and CARD follow leftward. A domain drawn once in
the alignment is drawn once per isoform here, and the ones sharing exons
overlap. Filter by... → Edit filters... on the track menu narrows that to
one record: the figure below keeps the features whose uniProtId is the gene's
reviewed entry. UniProt - Chains, Mutations and AA Modifications are
the same projection of the rest of the record.
Trying other genes
The same click-path works on any gene whose symbol NCBI recognises. What changes is how far down the tree the panel reaches, read off the tree on the left. Every NLRP1 ortholog NCBI has is a mammal, so the panel stops at mammals however high Rows to align is set, while CFTR reaches birds, amphibians and fish. Genes annotated with an Ensembl identifier and no symbol fall through to the BLAST tab.
Where each MSA comes from
Two routes on this page open the same view type over different MSAs.
| Route | MSA | Rows |
|---|---|---|
| Launch MSA view on the gene menu | built per gene from NCBI's ortholog records, aligned at EBI | one per species, named |
| The Gene Explorer | UCSC's precomputed multiz alignment across a hundred vertebrates | one per species, named; human genes |
Sharing a connected view as a URL
A connected view can also be built as a session-spec URL, for demo links and embedded apps. This session opens the AlphaFold structure of UniProt P04637 beside a genome view of the TP53 locus with NCBI RefSeq and ClinVar loaded.
{
"views": [
{
"type": "ProteinView",
"uniprotId": "P04637",
"transcriptId": "NM_000546.6",
"sideBySide": true,
"connectedView": {
"assembly": "hg38",
"loc": "chr17:7,671,000-7,684,500",
"tracks": ["hg38-ncbiRefSeq", "clinvar_ncbi_hg38"]
}
}
]
}This is the short form: a UniProt accession plus a transcript ID, from which the
plugin derives the AlphaFold structure, finds the transcript in the
connectedView tracks at loc, and translates its CDS. The explicit form takes
a structure url, feature and protein sequence, for a transcript no loaded
track serves. See the
protein3d developer docs.
A ProteinView with only a structure url and no connectedView opens as a
standalone structure, with no genome to exchange highlights with.
Adding the plugins to your own instance
- Open the plugin store (Tools menu) and install Protein3d and MSAView, or
- As an admin, add them to your
config.jsonso they load for all users (see configuring plugins)
Both plugins add view types launched from a gene's right-click menu in JBrowse
Web and Desktop. The single-view embedded components host only a linear genome
view, so neither view type appears there. JBrowseR's JBrowseRApp
takes both runtime plugins and a views list, while
anywidget's JBrowseApp has no plugin loading yet.
The approach is described in Proteins in the Genome Browser (Journal of Molecular Biology, 2026).
See also
- Basic usage of genomes.jbrowse.org
- Synteny on genomes.jbrowse.org
- JBrowseMSA user guide
- jbrowse-plugin-protein3d
- jbrowse-plugin-msaview
- g2p_mapper
References
- AlphaFold DB
- UniProt
- NCBI Datasets gene orthologs
- NCBI Conserved Domain Database
- Proteins in the Genome Browser
- Broz P, Dixit VM. Inflammasomes: mechanism of assembly, regulation and signalling. Nat Rev Immunol 2016.
Feedback on this tutorial is welcome: contact us.