Introgression plots

Plot introgression summaries from Dsuite localFstats/Dinvestigate windows, Dsuite BBAA/Dmin trio tables, ADMIXTOOLS qpdstat/f3/f4ratio tables, TreeMix lightweight graph summaries, and legacy genomics_general or qpGraph inputs. Windowed statistics use chromosome-wise Manhattan-like points or regional point-and-line traces, Dsuite trio summaries default to a raster-style P2 x P3 matrix with significant cells outlined, ADMIXTOOLS statistic tables use a horizontal forest/lollipop display with standard-error bars when se is present, mixed D/f3/f4-ratio summaries are split by statistic family, fixed-difference trio summaries use a boxplot-style display, and graph inputs use a compact edge diagram that emphasizes population tips and migration arrows. TreeMix edge graphs preserve drift coordinates when companion vertex output is present; covariance residual heatmaps and model-comparison diagnostics are intentionally outside this generic layer.

Usage

geom_introgression(
  mapping = NULL,
  data = NULL,
  ...,
  stat = "all",
  analysis = c("auto", "window", "trio", "graph"),
  chr = NULL,
  start = NULL,
  end = NULL,
  style = c("auto", "window", "manhattan", "region", "matrix", "raster", "trio", "graph"),
  colour_by = c("stat", "chr"),
  point_size = NULL,
  point_alpha = 0.9,
  base_size = 11,
  base_family = "",
  palette = "publication",
  na.rm = FALSE,
  show.legend = NA,
  inherit.aes = TRUE
)

plot_introgression(
  data,
  stat = "all",
  analysis = c("auto", "window", "trio", "graph"),
  chr = NULL,
  start = NULL,
  end = NULL,
  style = c("auto", "window", "manhattan", "region", "matrix", "raster", "trio", "graph"),
  title = NULL,
  subtitle = NULL,
  caption = NULL,
  base_size = 11,
  base_family = "",
  palette = "publication",
  point_size = NULL,
  point_alpha = 0.9,
  ...
)

Arguments

mapping

Optional aesthetic mapping. Defaults are chosen from the resolved plot style.

data

A ggpop_introgression object from import_introgression().

...

Additional geom parameters.

stat

Introgression statistic names to plot, e.g. c("D", "fdM") or "all".

analysis

Optional analysis class filter: "window", "trio", or "graph".

chr

Optional chromosome vector for windowed results.

start, end

Optional genomic region boundaries in base pairs for windowed results.

style

Plot layout. "auto" uses chromosome-wise Manhattan-like points for genome-wide window statistics, a point-and-line regional axis for local window calls, a raster-style P2 x P3 matrix for Dsuite BBAA/Dmin trio summaries, a horizontal forest/lollipop summary for ADMIXTOOLS statistic tables with standard-error bars when se is present, a boxplot-style trio summary for fixed-difference rows, and an edge diagram for lightweight graph data. "manhattan" is accepted as a compatibility alias for "window" and "raster" is an alias for the matrix view; use style = "trio" explicitly to draw Dsuite trio tables as ordered forest/lollipop summaries. Matrix views draw a complete background grid, observed D-statistic tiles, direct labels for small matrices, and outlines for significant cells when P values are present. Repeated P2/P3 combinations keep the row with the largest absolute D value while preserving its sign. Trio and matrix axes clean underscores in population labels for display only; mixed trio statistic families are faceted by stat; graph views label population tips and suppress internal node identifiers. TreeMix *.edges.gz imports use matching *.vertices.gz coordinates when available, so the graph view can show the drift-parameter axis rather than a generic network layout.

colour_by

Colour window plots by statistic or chromosome.

point_size, point_alpha

Layer appearance. Genome-wide window plots use this value as point size; region plots use it as point size and draw a thinner connecting line.

base_size, base_family

Base theme font controls.

palette

ggpop discrete palette.

na.rm, show.legend, inherit.aes

Standard ggplot2 layer arguments.

title, subtitle, caption

Optional plot labels. No default title is added.

Value

geom_introgression() returns a list of ggplot layers. plot_introgression() returns a ggplot object.

Examples

intro_file <- system.file(
  "extdata", "introgression", "Dsuite",
  "PopB_PopC_PopA_localFstats_run1_100_50.txt",
  package = "ggPopi"
)
intro <- import_introgression(intro_file, type = "dsuite_dinvestigate")
intro |> plot_introgression(stat = c("D", "fdM"))

intro |> plot_introgression(stat = "D", chr = "1")

intro |> ggpop() + geom_introgression(stat = "D")