reduce_genesets {goat}

Reduce the set of significant genesets to a minimum

Description

Analyses are performed independently per 'source' of genesets. The result of this function is the geneset table with a newly appended column 'signif_and_reduced'

Usage

reduce_genesets(
  clusters,
  simscore_threshold = 0.9,
  universe_fraction = 0.25,
  signifgenes_fraction = 0.9
)

Arguments

clusters

results from cluster_genesets()

simscore_threshold

similarity score (0~1) that is required to consider one geneset to be a "parent term" of another. Setting a lower value will yield fewer genesets / stronger summarization. Typical settings for this parameter are 0.8~0.99 (0.9 is default)

universe_fraction

discard genesets that cover more than X fraction of all genes in the universe (unique set of genes covered by all significant genesets). Setting this to 0.25 will deprioritize genesets that cover 25% of all genes (in significant genesets). This prevents very generic GO terms like "protein-containing complex" to be included in results. Typical settings for this parameter are 0.1~0.5 (0.25 is default)

signifgenes_fraction

the minimum fraction of "foreground genes" ('genes_signif' column) found across all significant genesets that should be covered by the reduced geneset collection. This parameter doesn't do anything if there are fewer than 5 "foreground genes" alltogether. Typical settings for this parameter are 0.75~0.95 (0.9 is default)

Value

the genesets table from the clusters parameter, with results in column "signif_and_reduced"