generateGDM: Computes ground distance matrix based on cell embedding

In wschen/phemd: Phenotypic EMD for comparison of single-cell samples

Computes ground distance matrix based on cell embedding

Description

Takes as input a Phemd object containing cell-state embedding object. Returns updated object with ground distance matrix representing pairwise distances between distinct cell subtypes based on cell state embedding.

Usage

generateGDM(
  obj,
  cell_model = c("monocle2", "seurat", "phate"),
  expn_type = "reduced",
  ndim = 8
)

Arguments

obj	'Phemd' object containing cell-state embedding object
cell_model	Method by which cell state was modeled (either "monocle2", "seurat", or "phate")
expn_type	Data type to use to determine cell-type dissimilarities
ndim	Number of embedding dimensions to be used for computing cell-type dissimilarity (optional)

Details

embedCells and orderCellsMonocle need to be called before calling this function. Requires 'igraph' package

Value

Phemd object with ground distance matrix (to be used in EMD computation) in @data_cluster_weights slot

Examples

my_phemdObj <- createDataObj(all_expn_data, all_genes, as.character(snames_data))
my_phemdObj_lg <- removeTinySamples(my_phemdObj, 10)
my_phemdObj_lg <- aggregateSamples(my_phemdObj_lg, max_cells=1000)
my_phemdObj_monocle <- embedCells(my_phemdObj_lg, data_model = 'gaussianff', sigma=0.02, maxIter=2)
my_phemdObj_monocle <- orderCellsMonocle(my_phemdObj_monocle)
my_phemdObj_final <- clusterIndividualSamples(my_phemdObj_monocle)
my_phemdObj_final <- generateGDM(my_phemdObj_final)

wschen/phemd documentation built on April 8, 2023, 6:27 a.m.