FindRootCluster: Find the root cluster

In mkarikom/RSoptSC: Low-Rank Factorization for Subspace-clustering

Find the root cluster

Description

Given a weighted adjacency matrix, generate a cluser-to-cluster graph based on a similarity-matrix of connected cells. Generate the minimum spanning tree on the cluster-to-cluster graph. Use one of the max dist clusters from this graph, unless: The cluster is given, then proceed to find the root cell The root cell is given, then use its cluster

Usage

FindRootCluster(
  cluster_labels,
  flat_embedding,
  dist_graph,
  dist_flat,
  reverse = FALSE
)

Arguments

cluster_labels	the cluster label for each cell
flat_embedding	rows are cells and columns are coordinates in n-col space
dist_graph	distances on a connected graph
dist_flat	flat distance matrix for all cells
reverse	a boolean variable whether to take the root cluster based on minimum dispersion on the flat embedding

Value

a list containing:

cluster_adj_matrix	a weighted upper-triangular adjacency matrix for the clusters based on avg pseudotime between their cells
graph_cluster	an igraph object on the clusters, with a minimum spanning tree
root_cluster	index of the root cluster
cluster_mst	an igraph mst object

mkarikom/RSoptSC documentation built on May 10, 2023, 1:10 a.m.