preprocess: PreProcess Expression Matrix
In jlaffy/statistrics: Basic Statistics For Single-Cell (or bulk) RNA Analyses

If the data is passed in TPM form, the data is first transformed to logTPM. Low quality cells are then removed. They are a filtered according to a cutoff for the number of genes detected. Next, low quality genes are removed. These are filtered according to a cutoff for their avg. expression values across all cells. Finally, the data is centered such that the average expression of each gene across all cells is 0. Note: The standard deviation is left as is – rather than being normalised to equal 1 – since s.d. values are very skewed due to how sparse the scRNA data is (many 0s).

1
2
3

preprocess(mat, pipeName = "preprocess", cachePath = ".",
  logTransform = TRUE, complexity.cutoff = 3000, genes.cutoff = 4,
  centering = TRUE)

`mat`	matrix of vars. by. obs. (in TPM or logTPM).
`pipeName`	a job ID, a name for the project/pipeline. Defaults to function name.
`cachePath`	passed to `cacheCall::cacheCall`. A character string providing path to the Cache directory.
`logTransform`	if TRUE, apply `logTPM` to matrix.
`complexity.cutoff`	if a numeric value, apply `complexityCut` to matrix with cutoff `complexity.cutoff`. Else if FALSE, do not `complexityCut`.
`genes.cutoff`	if a numeric value, apply `genesCut` to matrix with cutoff `genes.cutoff`. Else if FALSE, do not `genesCut`.
`centering`	if TRUE, apply `center` to matrix.