Find and plot similarity structure between archetypes

Find similarity structure between archetypes by comparing archetype weights for cells, measuring similarity of archetypes in expression space and in marker gene. genesXarch option shows

plot_arch_struct(arc, type = c("cells", "space", "marker_genes",
  "genesXarch")[1], dist_fun = function(x) dist(x, method = "euclidean"),
  marker_genes = NULL, marker_genes_mean = F,
  marker_genes_mean_col = "mean_diff")

Arguments

arc	object containing archetypal analysis results (class "pch_fit")
type	one of c("cells", "space", "marker_genes", "genesXarch"). "cells": distances are computed over archetype weights of cells (S matrix). "space": measures distances between archetypes in space, PCs or gene expression, depending on which space was used to find archetypes (XC matrix). "marker_genes": distance is computed base on lists marker genes. "genesXarch": show both archetypes and dimensions, both clustered with hierarchical clustering. When marker_genes is provided "genesXarch" shows marker gene membership across archetypes rather than expression values of those genes (PCs) at archetypes.
dist_fun	function use to compute distance, should take one argument and compute distances between rows.
marker_genes	filtered data.table containing markers for each archetype, normally the output of get_top_decreasing stored in $enriched_genes. When type = "genesXarch" dimensions are filtered using gene names supplied in marker_genes.
marker_genes_mean	show the strength of marker gene association with archetypes? (mean_diff in the output of get_top_decreasing stored in $enriched_genes). By default is FALSE - show only marker gene membership (0/1).
marker_genes_mean_col	which column in marker_genes stores the strength of marker gene association with archetypes?

Value

Matrix of the same dimention as the original matrix but with values in each column permuted.