Add directionality to pathway enrichment

Usage

add_pathway_direction(
  x,
  gene_hits,
  delim = "/",
  dir_colname = "z-score",
  verbose = FALSE,
  ...
)

Arguments

x: enrichResult object as returned by clusterProfiler::enricher()
gene_hits: numeric vector, with values c(-1, 1) to indicate up or down, respectively, and whose names(gene_hits) should match the gene symbols (or values) in x@result$geneID. If no names(gene_hits) match values in the geneID column, the direction score will ultimately be populated with NA values.
delim: character string with delimiter used in the enrichResult object, default: "/"
dir_colname: character string indicating the column name to create, default "z-score".
verbose: logical indicating whether to print verbose output.
...: additional arguments are ignored.

Value

enrichResult object, with directional values populated into the result column name defined by dir_colname. When no genes in a gene set are present in gene_hits the expected directional value is NA.

Details

This function uses pathway enrichment data provided as an enrichResult object. Data is added as a new column (default "z-score") with numeric values indicating the direction of change.

It requires gene_hits with the direction of change for each gene, which is used to determine the up- and down-regulated genes in each gene set.

The default equation uses the IPA z-score of activation, since the z-score has beneficial properties in downstream visualization functions. It is symmetric around zero, and implies reasonable thresholds.

Note that when no genes involved in enrichment are present in the gene_hits data, the resulting directional value is expected to be NA, which will be treated as "non-directional" in downstream visualization functions.

IPA z-score activation calculation

The IPA Activation "z-score" uses the following formula:

z <- (N_genes_up - N_genes_down) / sqrt(N_genes_up + N_genes_down)

Citation for IPA formula: Kramer 2014 https://doi.org/10.1093/bioinformatics/btt703, as referenced: IPA FAQ - Statistical Calculations

Direction calculation (deprecated)

The deprecated calculation for "direction" is shown below:

direction_1 <- (N_genes_up - N_genes_down) / (N_genes);
# then log2(1.2 + x) transformed
direction <- log2(1.2 + abs(direction_1)) * sign(direction_1)