JBrowse 2 visualizations gallery
Synteny and whole-genome comparison
Whole-genome dotplot
Two yeast assemblies from a minimap2 PAF: a single diagonal running corner to corner where the two agree, stepping off it once where they do not. The dotplot pans and zooms like any other view, and drag-selecting a block opens that region as a linear synteny view.
Human vs mouse synteny
Whole-genome liftOver chains between human (hs1) and mouse (mm39), auto-diagonalized so one genome reorders to match the other, and colored by query chromosome to follow where each one lands.
hg38 vs CHM13 liftOver alignment
The UCSC chain at TNNT3, colored by strand. The one reverse-strand block is a segment the two assemblies place on opposite sides of the gene, the rearrangement from Fig 5C of the T2T human variation paper.
Multi-way synteny
Grape, peach, and cacao stacked in one synteny view from MCScan gene pairs, one track per adjacent pair of genomes, auto-diagonalized and colored by the genome shared between both tracks.
Structural variants
SV inspector
SKBR3 translocations in a sortable, filterable table beside a circular whole-genome overview that mirrors the table filters. A row dropdown or an arc click opens a breakpoint split view on that call.
Breakpoint split view
An SKBR3 translocation with each side in its own panel and the connections drawn across them: supporting reads as black curves, the variant call in green with feet showing directionality.
1000 Genomes SV genotypes
One row per sample straight from a multi-sample VCF, colored by genotype. Carriers of a large inversion appear as a solid block across the 1000 Genomes SV callset.
Pair orientation coloring
Read pairs colored by orientation and joined to their mates by arcs. Green LL, navy RR, and magenta split reads mark the inverted segment of this 1000 Genomes INVdup call, clustered at the breakpoints of an otherwise grey concordant pileup.
Tumor/normal translocation evidence
A C-GIAB cancer benchmark translocation between two chromosomes, one per panel. Black splines connect tumor PacBio HiFi reads that map partway to each side, the read-level evidence for the call.
K562 translocation in Hi-C
ENCODE Hi-C for a leukaemia line and a normal-karyotype line, each with a chr9 window and a chr22 window open at once. The wedge between the two panels is chr9 against chr22: empty in the normal cell, a solid block in K562, where the two chromosomes are fused.
Read cloud
Each mate pair and split-read chain collapses to one mark placed by its insert size, so short-insert pairs sit above the concordant baseline instead of hiding in the pileup below. Marks are colored by insert size and orientation together.
Alignments, long reads, and base modifications
Group reads by tag
GIAB nanopore reads colored and grouped by HP tag, which splits the pileup into a labeled block per tag value. The deletion falls in one block only. Grouping works off any BAM tag, not just HP.
Insertion across platforms
The same GIAB insertion in Nanopore, PacBio, and Illumina reads. Soft clipping on the Illumina reads marks the boundaries the long reads span.
RNA-seq splice junctions
RNA-seq over ACTB: coverage histogram, strand-colored junction arcs, the spliced read pileup, and the gene model below. The arcs come from the N skips in the BAM, so there is no separate junction file to load.
Fiber-seq base modifications
ONT fiber-seq over the GAPDH promoter, each read colored base by base from its MM/ML modification tags (6mA), with single-cell ATAC above. The coloring is per read, so modifications stay at single-molecule resolution instead of collapsing into an aggregate track.
Nanopore methylation coloring
Human nanopore reads over a CpG island: red CpGs where the 5mC call is methylated, blue where it is not. The calls come straight from the CRAM modification tags, with no bedMethyl track involved.
Bisulfite read coloring
Arabidopsis WGBS colored per read from its C→T conversions against the reference rather than from MM/ML tags, with aggregate CpG/CHG/CHH tracks and the gene annotation alongside.
Variants and populations
1000 Genomes trio SV coverage
Coverage for a 1000 Genomes trio (mother, child, father) beneath the ensemble structural-variant VCF.
1000 Genomes phased trio matrix
A phased trio genotype matrix: child, mother, and father each as two haplotype rows, so matching blocks reveal which parental haplotype the child inherited.
Dog10K body size scan
Fst between toy/small and giant dog breeds in 200 kb windows across all 38 autosomes. The differentiated windows sit on known body-size genes, labeled on the figure.
GWAS with LD coloring
An SLE GWAS at the STAT4 locus, with points shaded by r² to the lead SNP from a PLINK .ld file beside the summary statistics.
BXD QTL with haplotype painting
A BXD mouse coat-color QTL scan with the strain haplotype painting below it. Sorting the rows by genotype at the peak over Tyrp1 separates the mixed block into B (red) above D (blue), directly under the Manhattan peak.
Drosophila inversion Fst
Every Drosophila dm6 chromosome arm at once, with the In(2L)t inversion extent on top and Fst between inverted and standard lines below it. Fst is elevated across the whole inverted arm and low elsewhere.
More examples
- More GWAS examples (Storybook) ↗ — The same LD-colored Manhattan track inside an embedded React linear genome view, with its source alongside.
Coverage, copy number, and epigenomics
Tumor/normal copy number
COLO829 tumor and normal mosdepth BigWigs as one multi-quantitative track in scatter rendering, sharing an autoscaled y-axis, with every main chromosome open at once.
1000 Genomes copy number
Copy number over one megabase, one row per 1000 Genomes individual. "Cluster rows by score" in the track menu reorders the rows on the window in view, so matching rows stack together.
TCGA-BRCA cohort copy number
Copy number across every primary tumor in the cohort, one row per tumor, clustered by profile. Blue is loss and red is gain on the log2 ratio, so recurrent events appear as vertical stripes through the stack.
C-GIAB subclonal copy number
One row per clone grown from a single tumor cell, partitioned out of one CNVkit BED by its clone column, under the bulk HiFi depth and the benchmark's absolute call. The benchmark calls one state across the arm and the rows mostly hold it; the one that departs at the p-terminus is copy number the bulk lanes average away.
ChromHMM chromatin states
Roadmap Epigenomics chromatin states from a single BigBed. The multi-row feature display splits features by cell type into one row each, colors them with the itemRgb values the file carries, and derives the state legend from the data. Clustered over the HOXA cluster, the 127 epigenomes separate into those that open its anterior half and those holding all of it repressed.
Single cells under their pseudobulk
Nine pseudobulk coverage rows from a 10x PBMC experiment, above the 4390 cells they sum over, read one row per cell from a cells-by-bins Zarr matrix. At LYZ the monocyte block is solid and the lymphocyte rows are ambient speckle, one UMI per cell.
Hi-C contact matrix
A Hi-C contact matrix with the RefSeq gene track above. JBrowse reads the .hic file in place over HTTP range requests and picks the binning resolution from the zoom level. The track menu steps that resolution and sets the color ramp.
Genes and proteins
Horizontally flipped view
The whole view mirrored so coordinates run right to left and the reverse-strand ACTB runs in its own 5′→3′ direction. Reads, sequence, and annotations flip with it, and the location box records the state as a [rev] locstring.
Collapsed introns
PTEN's introns collapsed so its exons sit side by side, with the NA12878 direct-RNA sashimi arcs spanning adjacent exons. Right-click the gene in the live view to run it yourself.
SARS-CoV-2 polyprotein subfeatures
The SARS-CoV-2 ORF1ab polyprotein with CDS-frame coloring and subfeature labels on, so each mature_protein_region in the GFF is drawn and named on its own instead of collapsing into one CDS block.
Selenocysteine translation
Per-codon amino-acid lettering on GPX1, where the in-frame UGA shows as selenocysteine on orange rather than a stop. Translation follows the annotation's exception instead of a fixed codon table.
Differential transcript usage
ATP5F1C with ENCODE muscle and liver RNA-seq coverage over GENCODE transcripts colored by the change in isoform fraction satuRn measured between the two tissues. The boxed exon has no muscle reads and a liver peak, and only the liver-preferred transcript draws an exon there.
Genome linked to a 3D structure
TP53 with RefSeq models and ClinVar variants beside its AlphaFold structure, connected through the genome-to-structure alignment. Hovering a variant in the genome highlights the residue it hits on the structure.
Cross-species protein alignment
NLRP1 aligned across the species NCBI has an ortholog gene for, built from the gene symbol rather than from an alignment file, with NCBI's conserved-domain calls drawn in alignment columns. The pyrin domain is called on the human, chimpanzee, gorilla and marmoset rows.
Pangenomes
HPRC pangenome locus as a graph
The C4 locus of the HPRC release 2 pangenome as a force-directed graph, above lanes of the same window on GRCh38. Both panels are colored by reference position, so the thread winding through the graph runs red to magenta with the segment blocks above it.
HPRC graph beside its callset
One deletion site in the HPRC release 2 MHC, marked in a linear view over all 464 haplotypes clustered by genotype, with an arrow from the band to the reference node it removes in the force-directed graph below.
E. coli pangenome (pggb)
A pggb pangenome graph projected onto the K12 reference as per-strain presence, one row per non-reference strain across the whole chromosome, dropping to zero over the stretches that strain does not carry.
E. coli pangenome (Minigraph-Cactus)
Per-strain presence in the Minigraph-Cactus graph, one row per strain in the colors odgi viz gives them, drawn on K12's coordinates instead of the graph's node order. White is sequence that strain does not carry.
E. coli all-vs-all alignment (minimap2)
One all-vs-all PAF read twice on the same axis: as one-vs-all lanes in an ordinary linear view (K-12 against every other strain, grouped per strain) and as the ribbon bands of the stack below. Inversions are blue in both.
Hosted genomes and track hubs
phyloP conservation
The UCSC phyloP conservation score over TP53 as a quantitative track under the NCBI RefSeq gene model, its per-base peaks lining up with the coding exons. Both tracks come from the hosted hg38 config on genomes.jbrowse.org, opened from the track selector with no display settings changed.
More examples
- UCSC GenArk hub import — A UCSC track hub opened from a hubURL parameter alone, no JBrowse config: the assembly and its tracks all come from hub.txt.