Skip to contents

Plots fitted coefficients against the authoritative stored -log10(p) values using the same group legend, colors, effect-direction symbols, theme, and annotation style as plot_phewas_manhattan(). Coefficients remain on their fitted model or link scale. Results containing different effect measures are separated into free-scale facets, each centered at zero, because their magnitudes are not directly comparable. Mean differences also require common outcome units and transformations before their magnitudes can be compared within a panel.

Usage

plot_phewas_volcano(
  x,
  significance = "bonferroni",
  label = NULL,
  highlight = NULL,
  category_order = NULL,
  x_limit = NULL,
  y_limit = NULL,
  missing = c("error", "drop"),
  title = NULL,
  label_mode = c("description", "phenotype", "both"),
  top_label_n = 0L,
  label_wrap_width = 26L,
  group_legend_columns = 5L,
  group_wrap_width = 24L,
  point_size = 2.2,
  base_size = 12,
  label_size = base_size
)

Arguments

x

A phewas_result.

significance

Testing threshold to draw: "bonferroni", "bh", or "none".

label, highlight

Explicit phenotype values to label or ring.

category_order

Optional character vector giving every plotted group in display order.

x_limit, y_limit

Optional reviewed axis limits. Limits that clip data or the testing threshold error. x_limit must be symmetric around zero.

missing

What to do when selected phenotypes are unavailable.

title

Optional in-plot title.

label_mode

Label text: phenotype description, phenotype, or "both".

top_label_n

Number of strongest associations to label automatically when label is NULL. Use zero to disable automatic labels.

label_wrap_width

Approximate characters per label line.

group_legend_columns

Number of columns in the bottom group legend.

group_wrap_width

Approximate characters per group-label line.

point_size

Point size in millimetres.

base_size, label_size

Base and association-label font sizes in points.

Value

A ggplot object with a phewas_audit attribute.

Details

The default Bonferroni threshold uses the recorded original successful testing family, including successful rows omitted from the plot because their estimate is not finite. Threshold text is placed inside the plotting panel rather than added as an axis tick.