Over-Representation Analysis Using enrichGO
ora_enrichgo.RdPerforms Gene Ontology (GO) over-representation analysis (ORA) using
enrichGO on differential expression results.
Genes can be analyzed separately based on their direction of differential
expression (upregulated or downregulated).
Usage
ora_enrichgo(
filtered_results,
direction = c("up", "down"),
log2fc_thresh = 1,
pval_thresh = 0.05,
top_terms = 15,
OrgDb = org.Hs.eg.db,
keyType = "ENSEMBL",
ont = "BP"
)Arguments
- filtered_results
list of tables or a table of differential expression results
- direction
Which direction +log2fc -log2fc should be looked at ? e.g.
"up","down"- log2fc_thresh
what is log2FC cutoff you would want for the analysis
- pval_thresh
after running fisher exact test for ORA, what should the distribution of pvalues be?
- top_terms
How many resulting terms should be display in the visuals.
- OrgDb
rganism database. Default is human:
org.Hs.eg.db. Other organism annotation databases are available from https://bioconductor.org/packages/3.23/data/annotation/.- keyType
which type of geneid are being entered into the cluster
- ont
which ontological DB will you be using ? options
c("BP", "CC", "MF")
Value
A list containing:
enrichgo_results: Results from the GO over-representation analysis returned byenrichGO.visuals: A collection of visualizations summarizing the enriched GO terms, including dotplots and barplots.
Details
Genes are first separated according to their direction of differential
expression and filtered using the supplied log2 fold-change and p-value
thresholds. Over-representation analysis is then performed using
enrichGO.
Examples
if (FALSE) { # \dontrun{
enrich_ora_results <- ora_enrichgo(filtered_results,
direction = c("up", "down"),
log2fc_thresh = 1,
pval_thresh = 0.05,
top_terms = 15,
OrgDb = org.Hs.eg.db,
keyType = "ENSEMBL",
ont = "BP")
} # }