Skip to contents

Run NMF (non-negative matrix factorization)

Source: R/RunNMF.R
RunNMF.Rd

Run NMF (non-negative matrix factorization)

Usage

RunNMF(object, ...)

# S3 method for class 'Seurat'
RunNMF(
  object,
  assay = NULL,
  layer = "data",
  features = NULL,
  nbes = 50,
  nmf.method = "RcppML",
  tol = 1e-05,
  maxit = 100,
  rev.nmf = FALSE,
  ndims.print = 1:5,
  nfeatures.print = 30,
  reduction.name = "nmf",
  reduction.key = "BE_",
  verbose = TRUE,
  seed.use = 11,
  cores = 0,
  ...
)

# S3 method for class 'Assay'
RunNMF(
  object,
  assay = NULL,
  layer = "data",
  features = NULL,
  nbes = 50,
  nmf.method = "RcppML",
  tol = 1e-05,
  maxit = 100,
  rev.nmf = FALSE,
  ndims.print = 1:5,
  nfeatures.print = 30,
  reduction.key = "BE_",
  verbose = TRUE,
  seed.use = 11,
  cores = 0,
  ...
)

# S3 method for class 'Assay5'
RunNMF(
  object,
  assay = NULL,
  layer = "data",
  features = NULL,
  nbes = 50,
  nmf.method = "RcppML",
  tol = 1e-05,
  maxit = 100,
  rev.nmf = FALSE,
  ndims.print = 1:5,
  nfeatures.print = 30,
  reduction.key = "BE_",
  verbose = TRUE,
  seed.use = 11,
  cores = 0,
  ...
)

# Default S3 method
RunNMF(
  object,
  assay = NULL,
  layer = "data",
  nbes = 50,
  nmf.method = "RcppML",
  tol = 1e-05,
  maxit = 100,
  rev.nmf = FALSE,
  ndims.print = 1:5,
  nfeatures.print = 30,
  reduction.key = "BE_",
  verbose = TRUE,
  cores = 0,
  seed.use = 11,
  ...
)

Arguments

object

An object. This can be a Seurat object, an Assay object, or a matrix-like object.

...

Additional arguments passed to RcppML::nmf or NMF::nmf.

assay

Which assay to use. If NULL, the default assay of the Seurat object will be used. When the object also contains ChromatinAssay, the default assay and additional ChromatinAssay will be preprocessed sequentially.

layer

Which layer to use. Default is data.

features

A character vector of features to use. Default is NULL.

nbes

The number of basis vectors (components) to be computed. Default is 50.

nmf.method

The NMF algorithm to be used. Currently supported values are "RcppML" and "NMF". Default is "RcppML".

tol

The tolerance for convergence (only applicable when nmf.method is "RcppML"). Default is 1e-5.

maxit

The maximum number of iterations for convergence (only applicable when nmf.method is "RcppML"). Default is 100.

rev.nmf

Whether to perform reverse NMF (i.e., transpose the input matrix) before running the analysis. Default is FALSE.

ndims.print

The dimensions (number of basis vectors) to print in the output. Default is 1:5.

nfeatures.print

The number of features to print in the output. Default is 30.

reduction.name

The name of the reduction to be stored in the Seurat object. Default is "nmf".

reduction.key

The prefix for the column names of the basis vectors. Default is "BE_".

verbose

Whether to print the message. Default is TRUE.

seed.use

Random seed for reproducibility. Default is 11.

cores

The number of threads to be used in RcppML functions that are parallelized with OpenMP. If 0, the number of threads will be automatically determined by RcppML::setRcppMLthreads(). Default is 0.

Examples

library(Matrix)
data(pancreas_sub)
pancreas_sub <- standard_scop(pancreas_sub)
pancreas_sub <- RunNMF(pancreas_sub)
CellDimPlot(
  pancreas_sub,
  group.by = "CellType",
  reduction = "nmf"
)

FeatureDimPlot(
  pancreas_sub,
  features = c("BE_1", "BE_2", "BE_3"),
  reduction = "UMAP",
  palette = "RdBu",
  xlab = "UMAP_1",
  ylab = "UMAP_2",
  theme_use = "theme_blank"
)