Cluster gene expression patterns.

Cluster genes in clusters that have similar expression patterns along all lineages in the trajectory. By default, this function uses the clusterExperiment package to do the clustering. If another clustering method is of interest, one can extract fitted values to use for clustering, see details in the vignette.

# S4 method for SingleCellExperiment
clusterExpressionPatterns(
  models,
  nPoints,
  genes,
  reduceMethod = "PCA",
  nReducedDims = 10,
  minSizes = 6,
  ncores = 1,
  random.seed = 176201,
  verbose = TRUE,
  ...
)

# S4 method for list
clusterExpressionPatterns(
  models,
  nPoints,
  genes,
  reduceMethod = "PCA",
  nReducedDims = 10,
  minSizes = 6,
  ncores = 1,
  random.seed = 176201,
  verbose = TRUE,
  ...
)

Arguments

models	The fitted GAMs, typically the output from fitGAM.
nPoints	The number of points to use for clustering the expression patterns.
genes	A numerical or character vector specifying the genes from models that should be clustered.
reduceMethod	Dimensionality reduction method used before running the clustering methods. Passed to RSEC. Defaults to PCA.
nReducedDims	Number of dimensions kept after reduceMethod. Passed to RSEC.
minSizes	Minimum size of clusters. Passed to RSEC.
ncores	Number of cores to use. Passed to RSEC
random.seed	Passed to RSEC
verbose	Passed to RSEC
...	Additional arguments to be passed to RSEC.

Value

A list containing the scaled fitted values yhatScaled(for plotting) and a clusterExperiment object, containing the clustering results.

Details

This method adopts the RSEC function from the clusterExperiment package to perform consensus clustering.

Examples

if (FALSE) {
data(gamList, package = "tradeSeq")
clusterExpressionPatterns(models = gamList, nPoints = 200, genes = seq_len(11),
                          verbose = FALSE)
}