celda Cell and Gene Clustering Model
celda_CG(counts, sample.label = NULL, K, L, alpha = 1, beta = 1,
delta = 1, gamma = 1, count.checksum = NULL, max.iter = 25,
seed = 12345, z.split.on.iter = 3, z.num.splits = 3,
y.split.on.iter = 3, y.num.splits = 3, z.init = NULL, y.init = NULL,
logfile = NULL, ...)A numeric count matrix.
A vector indicating the sample for each cell in the count matrix
The number of cell populations
The number of gene clusters being considered
Non-zero concentration parameter for sample Dirichlet distribution
The Dirichlet distribution parameter for Phi; adds a pseudocount to each transcriptional state within each cell. Default to 1
The Dirichlet distribution parameter for Eta; adds a gene pseudocount to the numbers of genes each state. Default to 1
The Dirichlet distribution parameter for Psi; adds a pseudocount to each gene within each transcriptional state. Default to 1
An MD5 checksum for the provided counts matrix
Maximum iterations of Gibbs sampling to perform. Defaults to 25
Parameter to set.seed() for random number generation
On z.split.on.iter-th iterations, a heuristic will be applied using hierarchical clustering to determine if a cell cluster should be merged with another cell cluster and a third cell cluster should be split into two clusters. This helps avoid local optimum during the initialization. Default to be 3
Maximum number of times to perform the heuristic described in z.split.on.iter
On every y.split.on.iter iteration, a heuristic will be applied using hierarchical clustering to determine if a gene cluster should be merged with another gene cluster and a third gene cluster should be split into two clusters. This helps avoid local optimum during the initialization. Default to be 3
Maximum number of times to perform the heuristic described in y.split.on.iter
Initial values of z. If NULL, z will be randomly sampled. Default NULL.
Initial values of y. If NULL, y will be randomly sampled. Default NULL.
The name of the logfile to redirect messages to.
Additional parameters