run_cacomp: Internal function for 'cacomp'

In ClemensKohl/APL: Association Plots

Internal function for 'cacomp'

Description

'run_cacomp' performs correspondence analysis on a matrix and returns the transformed data.

Usage

run_cacomp(
  obj,
  coords = TRUE,
  princ_coords = 3,
  python = FALSE,
  dims = NULL,
  top = 5000,
  inertia = TRUE,
  rm_zeros = TRUE,
  ...
)

Arguments

obj	A numeric matrix or Seurat/SingleCellExperiment object. For sequencing a count matrix, gene expression values with genes in rows and samples/cells in columns. Should contain row and column names.
coords	Logical. Indicates whether CA standard coordinates should be calculated. Default TRUE
princ_coords	Integer. Number indicating whether principal coordinates should be calculated for the rows (=1), columns (=2), both (=3) or none (=0). Default 1.
python	A logical value indicating whether to use singular-value decomposition from the python package torch. This implementation dramatically speeds up computation compared to 'svd()' in R.
dims	Integer. Number of CA dimensions to retain. Default NULL (keeps all dimensions).
top	Integer. Number of most variable rows to retain. Default 5000.
inertia	Logical.. Whether total, row and column inertias should be calculated and returned. Default TRUE.
rm_zeros	Logical. Whether rows & cols containing only 0s should be removed. Keeping zero only rows/cols might lead to unexpected results. Default TRUE.
...	Arguments forwarded to methods.

Details

The calculation is performed according to the work of Michael Greenacre. Singular value decomposition can be performed either with the base R function 'svd' or preferably by the faster pytorch implementation (python = TRUE). When working with large matrices, CA coordinates and principal coordinates should only be computed when needed to save computational time.

Value

Returns a named list of class "cacomp" with components U, V and D: The results from the SVD. row_masses and col_masses: Row and columns masses. top_rows: How many of the most variable rows/genes were retained for the analysis. tot_inertia, row_inertia and col_inertia: Only if inertia = TRUE. Total, row and column inertia respectively.

References

Greenacre, M. Correspondence Analysis in Practice, Third Edition, 2017.

ClemensKohl/APL documentation built on May 25, 2024, 3:15 p.m.