Per (cluster, subcluster, marker, study): how far is this study from reference

The figure shows the shift; this makes it sortable. Reports BOTH deviation modes named above – the occupancy of the subcluster and the intensity of the marker within it – so "which markers and which cells differ" is a sort on a column rather than an eyeball over 288 panels.

Usage

stats_subcluster_shifts(
  cells,
  markers,
  subcluster,
  group_col = "cohort",
  reference = NULL,
  min_cells = 20L
)

Arguments

cells

Embedded cell table: one row per cell, carrying umap_1/umap_2, sample_id, population_label and one column per marker.

markers

Character vector of marker names to use.

subcluster

Integer vector of subcluster assignments, one per cell.

group_col

Name of the column holding the biological grouping (cohort). Default "cohort".

reference

The group every other group is compared against.

min_cells

Minimum cells a sample must contribute before it is used. Default 20L.

Details

Effect size is the median difference in asinh units plus Cliff's delta, which is bounded -1 to 1 and rank-based: at these per-subcluster n a difference in means is dominated by outliers, and these distributions are not normal.