Use either the primary manifold embedding of cell similarity to find the root cell by cluster-cell rank-correlation or use the flattened representation of this embedding to find the root cell by maximum separation heuristic. Possible values for cluster_order_by are predecessor and distance. Possible values for cell_order_by are index and distance.
FindRootCell( use_flat_dist = TRUE, cluster_order_by = "distance", cell_order_by = "distance", graph_cluster_mst = NULL, dist_graph = NULL, dist_flat, cluster_labels = NULL, root_cluster = NULL )
| use_flat_dist | use max path length heuristic on flat embedding |
|---|---|
| cluster_order_by | cluster rank parameter for measuring cluster/cell rank correlation |
| cell_order_by | cell-wise pseudotime parameter for measuring cluster/cell rank correlation |
| graph_cluster_mst | igraph object representing an mst on the cluster-cluster graph |
| dist_graph | a distance matrix of cells embedded in a graph |
| dist_flat | the manifold embedding of the cells |
| cluster_labels | the cluster label for each cell |
| root_cluster | the id of the root cluster on the cluster-cluster graph |
integer index of the root cell