{moduleScoreR}

Description

This is a simple package that, given an object with RNAseq values (which can be in a data.frame, DESeqDataObject from {DESeq2}, or a DGEList from {edgeR}) and a list of gene modules, will assign a score based on the eigenvalues of the first principle component of the genes for each module. By default, {moduleScoreR} uses the rsvd function from the {rsvd} package to calculate PC1, but svd from {base} R and nmf from the {NMF} package can also be used.

Gene modules need to be in the form of a named list of lists, i.e.

$MODULE_A
 [1] "GENE_1" "GENE_2" "GENE_3"

$MODULE_B
 [1] "GENE_4" "GENE_5" "GENE_6"

Alternatively, one can prepare the necessary list using prepGMT to format a gmt file from the Molecular Signatures Database or from a two-column data.frame, with the module names in a column named term and the genes for each module in a column named gene

Installation

devtools::install_github("milescsmith/moduleScoreR")

Usage

Basic usage takes the form of:

module_list <- prepGMT("path/to/gmt_file")
scoreEigengenes(
  object      = obj,
  module_list = module_list,
  score_func  = "rsvd"
)

Optional arguments include:

• normalize_func: choose between using vst and rlog for DESeqDataSets and cpm and rpkm for DGELists (note, though, that rpkm requires gene lengths to be present in the DGEList) for preparing expression levels
• return_self: if TRUE, a copy of the object containing expression values will be returned with the scores present in the metadata component of the respective objects