Skip to content

Variation and association

Plots for calls and the patterns they make across samples: point and structural variants, the genotypes of a cohort, copy number, variable sites, genotype matrices and heatmaps, association scans, linkage and other pairs, and site-wise selection.

How to choose

Choose by what carries the result: one position, two breakpoints, a fitted segment, a variable column, a cell for each sample and site, or one test per position.

Your question Plot Build it with
Where are the calls, and how strong is each one? Point variants VariantTrack, --variants
Which two positions does a rearrangement join? Structural variants StructuralTrack, --structural
How many copies are there, and has one allele been lost? Copy number CopyNumberTrack, --copy-number with --ploidy
Which sites tell closely related samples apart? Variable sites SnpTrack, --snps
Which call does each sample carry, from a VCF? Genotypes by sample GenotypeTrack, --genotypes
Which samples carry what, site by site? Genotype matrix MatrixTrack, --matrix
Which samples lost or gained a stretch, window by window? Heatmap of samples MatrixTrack::windows, --heatmap
Where does a scan cross its significance line? Association scan ManhattanTrack, --manhattan
Which markers of a peak travel with its strongest? Scan coloured by linkage ManhattanTrack::linkage, --manhattan with --ld
Which variants are inherited together, or which places are in contact? Pairs of positions PairTrack, --pairs
Which codons are under selection, and in which direction? Site-wise selection SelectionTrack, --selection

Plots

  • Variant lollipops coloured by consequence, below a depth profile with a dropout, a reference band and the rpoB gene model

    Point variants Point events as lollipops whose height is a value, or as plain ticks when there are too many for heads, with categories coloured in the order they first appear.

  • Forty samples ordered by a phylogeny beside them, each a row of calls across rpoB, the cells of the other allele forming blocks down the clades of the treeThe same figure on the dark page

    Genotypes by sample The call of each sample at each site of a VCF, a row per sample and each call at its position, with sites closer than a cell pooled a pixel at a time.

  • Five structural calls as arcs between their breakpoints, a deletion, a duplication, an inversion, an insertion and a translocation leaving the view, above a depth profile that drops under the deletion and steps up under the duplication

    Structural variants Each call as an arc between its two breakpoints, heavier with more supporting reads; a coverage track beneath shows whether the depth agrees.

  • Copy number gains and losses across a cohort along one arm of chromosome 8, and beneath them one tumour's segments on a ladder of whole copies, with lost heterozygosity marked along the foot

    Copy number Fitted segments drawn at their level on a ladder of whole copies, loss of heterozygosity in a lane of its own, and balanced wherever you say it is.

  • A phylogeny of twelve isolates beside lineage, resistance and year strips and a panel of thirty-four variable sites, each column labelled with its position and each row ending in its count of differences

    Variable sites Only the columns that vary, evenly spaced and each labelled with its position; a tree beside the rows lines a clade's shared changes up into a block.

  • An association scan over rpoB whose peak crosses the threshold line, the gene beneath it, and a genotype matrix showing which isolates carry the associated alleles

    Genotype matrix One row per sample and one cell per site at its real coordinate, where a sample without the allele, a sample never typed and a stretch with no site all look different.

  • Presence and absence of twenty-six genes across nine Klebsiella isolates, the rows sorted by the phylogeny beside them so two accessory islands come out as solid blocks

    Presence and absence by descent The same matrix sorted by a tree drawn beside it, which turns a speckle into blocks a clade carries.

  • An association scan and a depth profile across every contig of a draft assembly, with the contigs beneath as alternating named blocks

    Association scan One point per test, a threshold you set and the points above it ringed; laid over a Genome, the scan runs across a whole assembly.

  • A peak of an association scan, each marker coloured from grey to blue by its linkage with the strongest, which is a diamond labelled with its name, with a recombination rate with two hotspots laid behind itA peak of an association scan coloured by linkage

    Scan coloured by linkage A peak as LocusZoom draws one: every marker coloured by its r² with the lead, so a second signal beside the first stands out grey.

  • Forty samples in the order of a tree, each a row of cells along a chromosome, with one clade missing a stretch and three samples carrying another twiceA heatmap of forty samples in windows

    Heatmap of samples A value per sample per window, each sample read against its own usual value, in the order of a tree.

  • A triangle under a gene in which each pair of variants is a cell coloured by its linkage, with three dark blocks of variants inherited togetherA triangle of linkage under a gene

    Pairs of positions Linkage or contacts as a triangle hung under the axis, and a few pairs far apart as arcs.

  • A molecular selection atlas: rate classes and recurrent changes on a rectangular tree, mean branch omega on a circular tree, and two site-wise scans over protein domains with evidence above signed omega effects

    Site-wise selection Evidence, as a p-value or a posterior, in one tier and the signed log2(ω) effect in another, so a significant purifying site still reads as purifying.