countSubsetNorm: RNAseq counts subsetting and normalizing.

Description Usage Arguments Details Value

View source: R/calc_norm_counts.R

This function performs a number of operations on a counts object. It filters the counts to include only samples matching a sample annotation object. It optionally filters out low-count genes. It optionally normalizes, log2-transforms, and/or transposes counts. Finally, it can return either a DGEList object or a data frame.

calc_norm_counts(
  counts, design, libID_col="lib.id",
  min_count=NULL, min_cpm=NULL, min_libs_perc=0.15,
  normalize=TRUE, norm_method="TMM",
  log2_transform=FALSE, transpose=FALSE, return_DGEcounts=FALSE,
  ...)

`counts`	a matrix or data frame of gene expression counts. Should have sample in columns and genes in rows.
`design`	a data frame of sample information. At minimum, must contain a column corresponding to sample identifiers matching column names of `counts`. Passed to `design_filter_counts`.
`libID_col`	numeric index or character name of column in `design` containing sample identifers matching column names of `counts`. Passed to `design_filter_counts`.
`min_count`	numeric, the minimum count for a library to be considered passing threshold for a given gene.
`min_cpm`	numeric, the minimum counts per million for a library to be considered passing threshold for a given gene.
`min_libs_perc`	numeric, the minimum percentage of libraries that must meet `min_count` or `min_cpm` threshold for a gene to be retained.
`normalize`	logical, whether to normalize counts using `edgeR::calcNormFactors`.
`norm_method`	character, the method by which to normalize counts; passed to `edgeR::calcNormFactors`. Used only if `normalize` is TRUE. Defaults to "TMM".
`log2_transform`	logical, whether to log2 transform the counts. Defaults to `FALSE`.
`transpose`	logical, whether to transpose the matrix or data frame of counts.
`return_DGEcounts`	logical, whether to return return counts as a `DGEList` object. If FALSE, counts are returned as a data frame. Defaults to FALSE.
`...`	(optional) parameters passed to normalization functions.

This function utilizes design_filter_counts and (optionally) min_filter_counts to filter the counts object. Genes are filtered by expression if a non-Null value is provided for either min_count or min_cpm. It then (optionally) normalizes the counts, using edgeR::calcNormFactors. It then (optionally) log2-transforms the counts and/or transposes the counts object. Finally, it returns the counts either as a data frame, or as a DGEList object. For WGCNA analyses, should use log2_transform=TRUE, transpose=TRUE, return_DGECounts=FALSE. For limma analyses, should use log2_transform=FALSE, transpose=FALSE, return_DGECounts=TRUE.

a data frame or DGEList object containing the processed counts.

mjdufort/countSubsetNorm documentation built on Nov. 7, 2019, 6:37 p.m.