ccCompare: Comparison of enriched annotations

In categoryCompare: Meta-analysis of high-throughput experiments using feature annotations

Description

Takes the results from ccEnrich and compares the enriched annotations based on the settings previously set in ccOptions. Returns a ccCompareResult or ccCompareCollection object, see Details.

Usage

ccCompare(ccEnrichResult, ccOptions)

Arguments

ccEnrichResult	The enriched annotations collection returned from ccEnrich. This can be the ccEnrichCollection, GOccEnrichResult, or KEGGccEnrichResult
ccOptions	A ccOptions object that will determine which lists to actually compare against each other. See details below.

Details

Based on the enrichments found for each gene list, we now want to compare the annotations between lists. ccCompare accesses the annotations for each enrichment performed for each list, and makes the comparisons defined in ccOptions.

Value

ccCompare generates both a graph of the comparisons (to show how the categories are linked to each list and each other) and tabular output. The tabular output is a data frame, with ID for each term that was considered as a candidate annotation for each list, as well as a long description (Desc) of what the term is, and then membership and statistics from each gene list.

For each type of comparison (GO, KEGG, etc) a ccCompareResult is generated, with the following slots:

mainGraph	Annotations arranged as a graph
mainTable	The tabular results from all enrichment calculations combined into one
allAnnotation	A list of lists, where each entry is the annotation identifier, then a list for each comparison, with the genes that are annotated to that term that also belong to each list

The default is to generate an overlap graph for GO and KEGG, where the overlap is a measure of the similarity of the features (genes) annotated to each annotation term (based on a formula from EnrichmentMap). Optionally for GO, one can generate a hierarchical layout where the parent GO terms of the significant terms will also be included in the graph, with term origin saved in the node annotation (see example below to do this).

Only those terms with more than 10 and less than 500 annotated genes (according to the GO annotation file) are included.

When using weighted overlap graphs and RCy3 for viewing, it is recommended to use breakEdges and minNodes to remove edges with low weights and nodes with only a few genes from the dataset annotated to them.

Author(s)

Robert M Flight

See Also

ccCompareResult ccCompareCollection ccOutCyt breakEdges outType ccEnrich

Examples

## Not run: 
require(GO.db)
require(KEGG.db)
require(org.Hs.eg.db)

## End(Not run)
data(ccData)

# note that enrichLists is generated from ccEnrich
# ccResults <- ccCompare(enrichLists,ccOpts)
ccResults

# use the GO hierarchy tree
graphType(enrichLists$BP) <- "hierarchical"
# ccResultsBPHier <- ccCompare(enrichLists$BP,ccOpts)
ccResultsBPHier

categoryCompare documentation built on Nov. 8, 2020, 5:59 p.m.