Built-in track kinds

A track’s kind says what it is in domain terms — features, variants, a confidence ramp — and ProtVista maps it to the right renderer and data adapter for you. Write kind: features, not a component name.

This page lists every built-in kind. For the exact data shape each one consumes, see the Adapter reference; for the boundary between what your config controls and what the data must provide, see Configuration vs data.

The kinds

A kind’s name tells you whether you can bring your own data to it:

• A plain domain word — features, variants, peptides — takes a file from you as readily as it takes a provider API. Point data: at your CSV, TSV, or JSON and the kind reads it.
• A name that starts with a provider — alphafold-confidence, alphamissense-heatmap — reads that provider’s feed and nothing else. Those tracks are assembled from two API responses plus a further fetch, which no single file can stand in for.

kind	What it draws	Your own data	Typical provider
features	Sequence feature regions (domains, chains, binding sites, signal peptides…) drawn as shapes along the sequence.	.csv .tsv .json .bed	UniProt features API
interpro-features	InterPro domain and family predictions.	.csv .tsv .json .bed	InterPro entries API
variants	Single-residue variants (amino-acid changes), with disease association.	.csv .tsv .json	UniProt variation API
variant-counts	A line graph of variant frequency along the sequence.	.csv .tsv .json	UniProt variation counts
rna-editing	RNA-editing sites, drawn like variants.	.csv .tsv .json	UniProt RNA-editing API
rna-editing-counts	A line graph of RNA-editing frequency.	.csv .tsv .json	UniProt RNA-editing counts
linegraph	A line graph of numeric values you supply, one per position.	.csv .tsv .json, or inline	— (yours by design)
peptides	Mass-spectrometry-detected peptides.	.csv .tsv .json .bed	UniProt proteomics API
peptides-ptm	Post-translational modifications detected by MS.	.csv .tsv .json .bed	UniProt proteomics PTM API
structure-coverage	Which stretches of the sequence are covered by known 3D structures.	.csv .tsv .json .bed	UniProt/PDBe coverage
alphafold-confidence	AlphaFold per-residue confidence (pLDDT), drawn as a colour gradient over the sequence.	—	AlphaFold prediction
alphamissense-pathogenicity	AlphaMissense average pathogenicity per residue, as a colour gradient.	—	AlphaMissense average
alphamissense-heatmap	AlphaMissense pathogenicity as a full heatmap (every substitution × position).	—	AlphaMissense full matrix

Three record shapes cover every “your own data” cell above — features (type,start,end, plus a description column in CSV/TSV), points (position,value), and residue changes (position,variant). Which one a kind reads follows from what it draws, so variants and rna-editing share one, and every feature kind shares another.

Colour scales

The score kinds (alphafold-confidence, alphamissense-pathogenicity) come with colour ramps built in — AlphaFold’s confidence ramp (orange = low confidence, through to blue = very high) and AlphaMissense’s benign-to-pathogenic scale (blue = benign, through to red = pathogenic). Their colour-blind and contrast results are on Accessibility and known limitations. You can define and apply your own ramps; see Theme the viewer and Escape hatches (registerTheme).

Using a kind with your own data

Swap the data: and keep everything else. The kind still decides the renderer, the tooltips, and the colours — only the parser changes:

# The UniProt variation API…
- id: variants
  kind: variants
  data: variation

# …and your own calls, same kind, same track.
- id: variants
  kind: variants
  data: ./my-variants.csv

Pair a kind with a format it has no parser for — kind: alphafold-confidence at a ./x.csv — and the config is rejected before anything loads, with a message naming the kinds that do read that format. See Load your own data for each shape’s columns.

Isoform rows

A features track can also draw one row per isoform of a reviewed UniProt entry, on the canonical sequence’s coordinates. Point it at the entry and name the built-in adapter uniprot-isoforms-json:

sources:
  uniprotEntry: https://rest.uniprot.org/uniprotkb/{accession}.json?fields=sequence,ft_var_seq,cc_alternative_products
rows:
  - id: isoforms
    label: Isoforms
    kind: features
    data:
      source: uniprotEntry
      adapter: uniprot-isoforms-json
    rendering:
      height: 160

Each isoform UniProt lists gets a row: the canonical in blue, the others in grey. A gap in a row is a stretch of the canonical sequence that the isoform lacks. A darker mark is a canonical residue the isoform replaces with different ones; at an insertion, it marks the residue the new ones are added next to. Hover a row to see its changes, e.g. P05067-4 (APP695): 289: E → V; 290-364: missing.

An External isoform, which UniProt describes in another entry, gets no row, and neither does one whose sequence UniProt doesn’t know (“Not described”). The canonical row’s tooltip names them.

Use the canonical accession, such as P10636. An isoform’s own entry (P10636-8) has no alternative sequences, so the track stays empty.

Ask for at least the fields above: the adapter needs the sequence, the alternative sequences (ft_var_seq) and the “Alternative products” comment (cc_alternative_products). UniProt rejects an unknown field name, such as features, with an error.

The rows share the track’s height, 40 px unless you set one, so give the track about 15 px per isoform. Leave layout unset: the default, non-overlapping, keeps each isoform on its own row.

The track runs only when a config names the adapter, so the default UniProt view makes no extra request. The isoforms-app example in the repository’s examples/ folder is this config for APP (P05067), and the playground offers it as APP isoforms under Community views.

Licensed under CC BY 4.0.