Skip to main content

JBrowse 2 feature overview

JBrowse 2 is a pluggable, GPU-accelerated, fully client-side genome browser. The same core runs as a web app on static hosting (no server required), as a cross-platform desktop app, and as embeddable components inside other websites, notebooks, and R.

This page is a curated tour, not an exhaustive checklist. For the definitive, always-current lists, see the track and display types and supported file types references, both generated directly from the source. For visuals, browse the gallery.

What sets JBrowse apart

  • GPU-accelerated rendering - each track uploads its data to the GPU once, then zooming and panning redraw from that data instead of re-fetching and re-rendering. Scroll-zoom and drag-pan are continuous and immediate on every track type, even on whole-genome views. Browsers without WebGPU or WebGL2 fall back to a Canvas2D renderer with the same behavior.
  • Fully client-side, no server - host the whole app on static storage like S3 or GitHub Pages. Indexed files (BAM, CRAM, BigWig, tabix) are read by HTTP byte-range requests, so only the bytes in view are fetched and multi-gigabyte files work without a backend. Your data can stay on your own machine.
  • Local files that survive a refresh - open files from disk in the web app, and JBrowse retains access across page reloads using the File System Access API (handles persisted in IndexedDB), so you do not have to re-pick them every time.
  • Reproducible, shareable sessions - the full state of a session encodes into a URL. Share a link or bookmark it, and it reopens exactly as you left it.
  • Runtime plugins - install community plugins from the in-app plugin store without rebuilding or redeploying anything.
  • Exports match the screen - SVG export reuses the same drawing code as the on-screen renderer, so a figure looks exactly like what you saw.

View types

Any number of these can be open at once and linked together, and plugins can add more. See the user guide for walkthroughs of each.

  • Linear genome view - browse tracks along a reference, with split view and multiple regions side by side
  • Circular view - whole-genome overview of translocations and other rearrangements, drawn as arcs, with zoom-to-cursor
  • Dotplot view - zoomable whole-genome alignment and synteny comparison, with lockable aspect ratio and cursor-anchored zoom
  • Linear synteny view - stacked genomes with their syntenic alignments drawn between them, including multi-way and all-vs-all comparisons
  • Breakpoint split view - connect split or paired-end reads across chromosomes with stacked linear views
  • SV inspector - a sortable, filterable spreadsheet of structural variants with a linked circular overview
  • Spreadsheet view - BED, VCF, CSV/TSV, or bespoke formats like STAR-fusion output in a sortable, filterable table

Track and data types

JBrowse renders a wide range of data types, each with display modes tuned to it. For the complete track-to-display mapping, see the track and display types reference.

  • Sequence - reference sequence with six-frame translation, GC content, and GC-skew
  • Genes and features - GFF3, GTF, BED, and BigBed with gene glyphs, subfeatures, collapse-introns, and automatic itemRgb / bigBed coloring
  • Alignments - BAM, CRAM, and htsget in pileup, coverage, and paired/arc modes, with sorting, grouping by sample/tag/chain, tag coloring, per-base quality, sashimi junction arcs, and a discordant-pairs SAMplot mode
  • Base modifications and methylation - per-modification-type toggles, two-color and “show only” modes, bisulfite / EM-seq mode, and CpG/CHG/CHH context
  • Quantitative signal - BigWig and bedGraph as XY, line, density, or scatter plots, with multi-wiggle overlays, hierarchical-clustering group-by, and local or global autoscaling
  • Variants - VCF variant tracks, structural variants, and population-scale multi-sample variant matrices with virtual scrolling
  • GWAS / Manhattan - genome-wide association results as a Manhattan plot
  • Hi-C - .hic contact matrices with selectable, zoom-tracking resolution
  • Multiple alignment (MAF) - per-sample rows with CDS frame overlays and a per-row percent-identity / conservation readout
  • Synteny - PAF, chain, delta, PIF, and MCScan alignments between assemblies, colored by identity or mapping quality
  • Multi-row feature / chromosome painting - many feature rows in one track with clustering and per-position sorting, for local-ancestry and ChromHMM views

Analysis and interaction

  • Search - jump to a gene or feature by name using a text index, or to any region by locstring, with a “recent locations” menu
  • Sort, group, color, and filter - reshape alignment, variant, and feature tracks interactively from the track menu
  • Color variants by consequence, impact, or SV type - read SnpEff/VEP annotations, filter by no-call fraction, and view phased haplotypes
  • Linkage disequilibrium coloring - LocusZoom-style color-by-LD-to-a-SNP, backed by precomputed PLINK .ld matrices
  • BLAT and in-silico PCR - align a sequence against a genome, or find where a primer pair amplifies, against hosted UCSC databases
  • CRISPR guide and motif search - discover guide-RNA / PAM candidates (with per-guide GC% and poly-T flags) or arbitrary sequence motifs directly against the reference
  • Alternative genetic codes - NCBI transl_table resolved per reference sequence, plus transl_except for selenocysteine and other exceptions
  • Highlight or isolate features - highlight a feature, or “show only” it to mute everything else, baked into SVG export
  • Feature details - click any feature for a panel with its attributes and subfeatures

Working with your data

  • Edit any track’s settings in-app - open the configuration editor for any track, with a filter box and an “advanced settings” toggle. Non-admin changes become a personal override that travels with the session, not the shared config
  • Pin a setting as your default - pin almost any track setting (color scheme, feature height, group-by, …) as the default; it badges affected tracks and rides along in a shared link
  • Track hubs and connections - load UCSC track hubs and JBrowse connections by URL
  • Bulk and faceted track management - add many tracks at once, and filter, sort, and multi-select in the faceted track selector
  • Determinate download progress - index, BigWig, and tabix fetches report real progress instead of an indeterminate spinner
  • Pre-loaded genomes - launch any of a large database of species without configuring an assembly yourself

Supported data formats

JBrowse reads common genomics formats directly in the browser: sequence (FASTA, 2bit), alignments (BAM, CRAM, htsget), features (GFF3, GTF, BED, BigBed), quantitative signal (BigWig, bedGraph), variants (VCF), Hi-C (.hic), synteny (PAF, chain, delta, PIF, MCScan), multiple alignment (MAF), linkage disequilibrium (PLINK .ld), and tabular data (CSV, TSV, BEDPE). It also connects to UCSC track hubs, and plugins add more.

For the complete list, every format and the adapter it maps to, see supported file types.

Sharing and export

  • Shareable sessions - capture a session as a compressed share link or as plaintext JSON, or bookmark it to return later. Bookmarks can travel with the session too
  • Publication-ready export - export any view as an SVG that matches the screen, or high-resolution PNG, with a font selector. Render static images from the command line with @jbrowse/img
  • Desktop to web - export a desktop session to a shareable web link

Integration and embedding

The JBrowse 2 platform and its plugins are modular and reusable. Individual views can be packaged as prop-driven components for other applications.

  • Embedded components - drop a genome view into your own web app with @jbrowse/react-linear-genome-view and related packages, configured declaratively through props and an init field
  • R - genome views as an htmlwidget in Shiny, R Markdown, or the console with JBrowseR
  • Python - genome views as an anywidget in Jupyter, Colab, VS Code, or marimo with jbrowse-anywidget
  • Automation - drive URL, embedded, and session launches from a shared init spec

Extensibility

  • Plugins - add view types, tracks, adapters, renderers, and widgets. See the developer guide to build one
  • Plugin store - install community plugins directly from within the app, no rebuild required
  • Configuration and callbacks - drive appearance and behavior through configuration, including JEXL callbacks for dynamic coloring and filtering

See the gallery for examples of what JBrowse 2 can do, or dive into the user guide to get started.