upsetly

Interactive UpSet plots with plotly, with tooltip information that can be copied easily in HTML documents (Quarto, R Markdown, Shiny).

upsetly focuses on UpSet-style visualization of set intersections (e.g. differential gene lists), and adds JavaScript hooks so that all tooltip text (Intersection, Size, Members, etc.) can be synced into a separate text box for copy‑and‑paste.

---

Installation

You can install the development version from GitHub:

# install.packages("remotes")
remotes::install_github("Anbunensi/upsetly")

Then load:

library(upsetly)

---

Basic idea

Given a data frame with:

• one ID column (e.g. gene), and
• several 0/1 (or logical / character) columns indicating membership in sets
  (e.g. comparisons A_vs_B & B_vs_C, …),

upsetly() will:

1. Compute all non‑empty intersections across the set columns.
2. Keep only intersections with size ≥ min_intersection_size.
3. Optionally keep only the largest max_n_intersections intersections.
4. Draw:
   • Left bar chart: size of each set.
   • Top bar chart: size of each intersection.
   • Bottom dot matrix: which sets are present in each intersection.
5. Build rich tooltip text for each intersection, including:
   • Intersection name (e.g. A_vs_B & B_vs_C)
   • Intersection size
   • Members (element IDs), wrapped onto multiple lines.

In HTML output, you can also display the full tooltip text in a separate box which supports copy‑and‑paste.

---

Quick example

library(upsetly)

set.seed(1)
df <- data.frame(
  gene = paste0("g", 1:100),
  A = rbinom(100, 1, 0.3),
  B = rbinom(100, 1, 0.4),
  C = rbinom(100, 1, 0.2),
  stringsAsFactors = FALSE
)

p <- upsetly(
  x = df,
  set_cols = c("A", "B", "C"),
  id_col = "gene",
  max_n_intersections = 50,
  members_per_line = 10
)

p  # interactive plotly widget

This works in a regular RStudio viewer (or browser) as an interactive UpSet plot.

---

Using in Quarto / R Markdown with a copyable text box

To make the tooltip text (Intersection, Size, Members, etc.) easy to copy, you can add a small HTML element with id members_box and let upsetly() sync to it.

Example Quarto document

---
title: "upsetly demo"
format: html
---

```{r}
library(upsetly)
library(dplyr)
```
```{r}
set.seed(1)
df <- data.frame(
  gene = paste0("g", 1:100),
  A = rbinom(100, 1, 0.3),
  B = rbinom(100, 1, 0.4),
  C = rbinom(100, 1, 0.2),
  stringsAsFactors = FALSE
)

upsetly(
  x = df,
  set_cols = c("A", "B", "C"),
  id_col = "gene",
  max_n_intersections = 50,
  box_id = "members_box",
  members_per_line = 10
)
```
```{=html}
<pre id="members_box" style="
  border: 1px solid #ccc;
  padding: 8px;
  min-height: 120px;
  white-space: pre-wrap;
  font-family: monospace;
"></pre>
```

Interaction behavior

In the rendered HTML:

• Hover over an intersection bar:
  • The tooltip appears as usual.
  • The full tooltip text (Intersection, Size, Members, …) is also written into #members_box.
• Single click on a bar:
  • The current text is written into #members_box and locked.
  • Moving the mouse to other bars will not change the box content.
• Double click on the plot background:
  • The lock is released and the box is cleared.
  • You can then hover/click another bar to update again.

The content of members_box is pure text, so you can select and copy it easily.

---

Key arguments

upsetly(
  x,
  set_cols = NULL,
  id_col = NULL,
  min_intersection_size = 1,
  max_n_intersections = NULL,
  point_size = 8,
  bar_color_sets = "black",
  bar_color_inters = "black",
  active_color = "black",
  inactive_color = "lightgray",
  line_color = "black",
  title = "UpSet (plotly)",
  height = 500,
  width = NULL,
  members_per_line = 20
)

• x: data frame containing set columns and optional ID column.
• set_cols: character vector of column names that define sets.
• id_col: column name for element IDs (e.g. genes). If NULL, a numeric .elem_id is created.
• min_intersection_size: minimum number of elements for an intersection to be kept.
• max_n_intersections:
  • Limits how many intersections are kept (largest first).
  • Also limits how many member IDs are shown for each intersection in the tooltip.
• members_per_line: how many member IDs to show per line in the tooltip / text box.
• point_size: dot size in the dot matrix.
• Color and layout arguments allow customization of bar colors, dot colors, and overall size.

---

License

MIT © Anbunensi