do.goseq: Performs goseq analysis significance of gene set membership.

In lianos/multiGSEA: A unified interface to a plethora of gene set enrichment analysis methods

Description

Genes are selected for testing against each geneset by virture of them passing a maximum FDR and minimum log fold change as perscribed by the min.logFC and max.padj parameters, respectfully.

Usage

do.goseq(
  gsd,
  x,
  design,
  contrast = ncol(design),
  feature.bias,
  goseq.method = "Wallenius",
  repcnt = 2000,
  use_genes_without_cat = TRUE,
  split.updown = TRUE,
  direction = c("over", "under"),
  plot.fit = FALSE,
  use.treat = FALSE,
  feature.min.logFC = if (use.treat) log2(1.25) else 1,
  feature.max.padj = 0.1,
  logFC = NULL,
  ...
)

Arguments

gsd	The GeneSetDb for analysis
x	The expression object
design	Experimental design
contrast	The contrast to test
feature.bias	a named vector as long as nrow(x) that has the "bias" information for the features/genes tested (ie. vector of gene lengths). names(feature.bias) should equal rownames(x). The caller MUST provide this. The goseq package provides a getlength function which facilitates getting default values for these if you do not have the correct values used in your analysis. If there is no way for you to get this information, then use ora with no feature.bias vector.
repcnt	Number of random samples to be calculated when random sampling is used. Ignored unless method="Sampling".
use_genes_without_cat	A boolean to indicate whether genes without a categorie should still be used. For example, a large number of gene may have no GO term annotated. If this option is set to FALSE, those genes will be ignored in the calculation of p-values (default behaviour). If this option is set to TRUE, then these genes will count towards the total number of genes outside the category being tested.
direction	Same as direction in GOstats
plot.fit	To plot (or not) the bias in selected genes vs. feature.bias.
logFC	The logFC data.table from calculateIndividualLogFC
...	arguments to pass down into calculateIndividualLogFC
method	The method to use to calculate the unbiased category enrichment scores

Details

Note that we are intentionally adding a hyperG.selected column by reference so that this information is kicked back to the caller multiGSEA function and included in downstream reporting.

This function is not meant to be called directly. It should only be called internally within multiGSEA().

Value

A data.table of goseq results. The "pval" column here refers to pval.over, for simplicity in other places. If split.updown=TRUE, a list of data.table's are returned named 'goseq', 'goseq.up', and 'goseq.down' which are the results of running goseq three independent times.

lianos/multiGSEA documentation built on Nov. 17, 2020, 1:26 p.m.