Cohort composition heatmap (

WHAT: for each population, the share of its cells contributed by each cohort AFTER equalising every cohort to the same notional cell count.

Usage

fig_cohort_confusion(
  cells,
  outfile,
  group_col = "cohort",
  pop_col = "population_label",
  norm_to = 1000,
  panel_label = "",
  dpi = 300,
  colors = fcs_colors()
)

Arguments

cells: Embedded cell table: one row per cell, carrying umap_1/umap_2, sample_id, population_label and one column per marker.
outfile: Path to write the figure to.
group_col: Name of the column holding the biological grouping (cohort). Default "cohort".
pop_col: Name of the column holding the population label. Default "population_label".
norm_to: Notional cell count each group is normalised to before shares are taken. Default 1000.
panel_label: Marker-panel name added to the figure title; empty for none. Default "".
dpi: Resolution in dots per inch. May be reduced automatically to respect the raster ceiling; see safe_ggsave(). Default 300.
colors: Named list of colours; defaults to the package palette. See fcs_colors(). Default fcs_colors().

Details

WHY THE EQUALISATION IS THE WHOLE POINT: cohorts differ in sample count and in acquisition depth, so raw contributions are dominated by whichever cohort has the most cells and every population looks like it belongs to that cohort. The normalisation – the same device used by the cohort-confusion heatmaps elsewhere in the field, which equalise to a fixed cell count per group – removes that, so a row that is not near-uniform means a population is genuinely cohort-skewed.

WHY IT IS NOT A SUBSTITUTE FOR group_comparison.png: this pools cells across donors, so it shows a pattern, not a tested effect. A population that looks skewed here is a lead to check against the sample-level abundance test, which is the figure that carries statistics.

Cohort composition heatmap ("confusion matrix")

Usage

Arguments

Details