Skip to main content
An UpSet plot is the scalable successor to Venn diagrams for showing set intersections when there are more than three or four sets. It has three visual components: intersection size bars (top), dot matrix (middle), and optional set size bars (left).

​
When to Use

  • Differential expression overlap: Show genes called significant by multiple methods (DESeq2, edgeR, limma)
  • Variant calling concordance: Compare SNPs/indels detected by GATK, FreeBayes, Strelka
  • Drug response: Identify compounds effective across multiple cell lines
  • Pathway enrichment: Show overlap of enriched pathways across conditions
  • Any set intersection: When Venn diagrams become cluttered (>3 sets)
  • Set cardinality: When exact counts matter more than spatial overlap

​
Basic Example

use kuva::plot::UpSetPlot;
use kuva::backend::svg::SvgBackend;
use kuva::render::render::render_multiple;
use kuva::render::layout::Layout;
use kuva::render::plots::Plot;

let up = UpSetPlot::new()
    .with_sets(vec![
        ("Set A", vec!["apple", "banana", "cherry", "date"]),
        ("Set B", vec!["banana", "cherry", "elderberry", "fig"]),
        ("Set C", vec!["cherry", "fig", "grape"]),
    ]);

let plots = vec![Plot::UpSet(up)];
let layout = Layout::auto_from_plots(&plots)
    .with_title("Set Intersections");

let svg = SvgBackend.render_scene(&render_multiple(plots, layout));
std::fs::write("upset.svg", svg).unwrap();

​
Input Modes

​
Raw Sets: with_sets

Provide (name, elements) pairs. Intersection counts computed automatically. 
let up = UpSetPlot::new().with_sets(vec![
    ("DESeq2",   vec!["BRCA1", "TP53", "EGFR", "MYC",  "CDK4"]),
    ("edgeR",    vec!["TP53",  "EGFR", "RB1",  "PTEN", "CDK4"]),
    ("limma",    vec!["EGFR",  "MYC",  "RB1",  "CDK2"]),
]);
Elements can be any type implementing Eq + Hash (strings, integers, gene IDs, etc.). Intersection counts follow bitmask semantics: an element in sets A and C but not B is counted under A∩C — not A, not C, not A∩B∩C.

​
Precomputed: with_data

Provide set names, total sizes, and (mask, count) pairs directly. 
// Three variant callers — bit 0 = GATK, bit 1 = FreeBayes, bit 2 = Strelka
let up = UpSetPlot::new().with_data(
    ["GATK", "FreeBayes", "Strelka"],
    [280usize, 263, 249],
    vec![
        (0b001, 45),   // GATK only
        (0b010, 35),   // FreeBayes only
        (0b100, 28),   // Strelka only
        (0b011, 62),   // GATK ∩ FreeBayes
        (0b101, 55),   // GATK ∩ Strelka
        (0b110, 48),   // FreeBayes ∩ Strelka
        (0b111, 118),  // all three
    ],
);
Bit i of mask set means set_names[i] participates. Use when intersection counts come from external source (database, enrichment tool).

​
Key Methods

​
Sorting

​
with_sort(sort: UpSetSort)

Set intersection ordering (default ByFrequency).
  • UpSetSort::ByFrequency (default) — Largest count leftmost
  • UpSetSort::ByDegree — Most complex (most sets) leftmost, ties broken by count
  • UpSetSort::Natural — Preserve supplied order
use kuva::plot::UpSetSort;

let up = UpSetPlot::new()
    .with_sets(sets)
    .with_sort(UpSetSort::ByDegree);

​
with_max_visible(max: usize)

Show only top max intersections after sorting. 
let up = UpSetPlot::new()
    .with_sets(sets)
    .with_max_visible(10);  // show only 10 largest

​
Styling

​
without_set_sizes()

Hide horizontal set-size bars on left panel. 
let up = UpSetPlot::new()
    .with_sets(sets)
    .without_set_sizes();
Produces more compact layout focused on intersections.

​
with_bar_color(color: impl Into<String>)

Set color for intersection bars and set-size bars (default "#333333"). 
let up = UpSetPlot::new()
    .with_bar_color("steelblue");

​
with_dot_color(color: impl Into<String>)

Set fill color for dots in participating set (default "#333333"). 
let up = UpSetPlot::new()
    .with_dot_color("#1f77b4");
Non-participating dots remain light gray.

​
Examples

​
Differential Expression Overlap

let up = UpSetPlot::new()
    .with_sets(vec![
        ("DESeq2", vec!["BRCA1", "TP53", "EGFR", "MYC",  "CDK4", "AKT1"]),
        ("edgeR",  vec!["TP53",  "EGFR", "RB1",  "PTEN", "CDK4", "VHL"]),
        ("limma",  vec!["EGFR",  "MYC",  "RB1",  "CDK2", "AKT1"]),
    ])
    .with_sort(UpSetSort::ByFrequency)
    .with_max_visible(10);

let plots = vec![Plot::UpSet(up)];
let layout = Layout::auto_from_plots(&plots)
    .with_title("DEG Method Overlap");
Shows which genes are called by one, two, or all three methods.

​
Variant Calling Concordance (Precomputed)

let up = UpSetPlot::new().with_data(
    ["GATK", "FreeBayes", "Strelka"],
    [280usize, 263, 249],  // total variants per caller
    vec![
        (0b001, 45),   // GATK only
        (0b010, 35),   // FreeBayes only
        (0b100, 28),   // Strelka only
        (0b011, 62),   // GATK ∩ FreeBayes
        (0b101, 55),   // GATK ∩ Strelka
        (0b110, 48),   // FreeBayes ∩ Strelka
        (0b111, 118),  // all three (high-confidence)
    ],
)
.with_bar_color("#2ca02c");

let plots = vec![Plot::UpSet(up)];
let layout = Layout::auto_from_plots(&plots)
    .with_title("Variant Caller Concordance");
Precomputed counts from database query; 118 variants in all three callers.

​
By Degree Sorting

use kuva::plot::UpSetSort;

let up = UpSetPlot::new()
    .with_sets(vec![
        ("A", vec![1, 2, 3, 4, 5]),
        ("B", vec![3, 4, 5, 6, 7]),
        ("C", vec![5, 6, 7, 8, 9]),
        ("D", vec![7, 8, 9, 10]),
    ])
    .with_sort(UpSetSort::ByDegree)
    .with_max_visible(8);

let plots = vec![Plot::UpSet(up)];
let layout = Layout::auto_from_plots(&plots)
    .with_title("Sorted by Degree");
Most complex intersections (4-way, 3-way) shown first.

​
Compact Layout (No Set Sizes)

let up = UpSetPlot::new()
    .with_sets(sets)
    .without_set_sizes()
    .with_bar_color("#1f77b4")
    .with_dot_color("#1f77b4");

let plots = vec![Plot::UpSet(up)];
let layout = Layout::auto_from_plots(&plots)
    .with_title("Compact UpSet Plot");
No left panel; focuses solely on intersection sizes.

​
Visual Components

​
Intersection Size Bars (Top)

Vertical bars show element count for each intersection. Height encodes count.

​
Dot Matrix (Middle)

Grid of circles where:
  • Filled dots indicate set participates in intersection
  • Empty dots (light gray) indicate set does not participate
  • Vertical line connects filled dots in same intersection

​
Set Size Bars (Left, Optional)

Horizontal bars show total element count per set. Enabled by default; hide with without_set_sizes().

​
Pixel-Space Rendering

UpSet plots render entirely in pixel space — no standard x/y axis system. Layout::auto_from_plots skips axis computation for UpSet plots. Title set on Layout is still rendered.

​
See Also

  • [Venn diagrams] — For 2–3 sets with spatial overlap
  • Heatmap — For pairwise set similarity matrices

Build docs developers (and LLMs) love

Get started for freeTalk to us