Run Gene Ontology enrichment analysis on differentially expressed genes.

This function forms genesets basing on the differential expression result, and calls gene ontology (GO) analysis method provided by gprofiler2.

Usage

runGOEnrich(
  result,
  group = NULL,
  useBg = TRUE,
  orderBy = "padj",
  logFCThresh = 1,
  padjThresh = 0.05,
  splitReg = FALSE,
  ...
)

Arguments

result: Data frame of unfiltered output from runMarkerDEG or runPairwiseDEG.
group: Selection of one group available from result$group. Default NULL uses all groups involved in DE result table.
useBg: Logical, whether to set all genes involved in DE analysis (before threshold filtering) as a domain background of GO analysis. Default TRUE.
orderBy: Name of DE statistics metric to order the gene list for each group. Choose from "logFC" (default), "pval" or "padj". Or set NULL to turn off ranked mode.
logFCThresh: The log2FC threshold above which the genes will be used. Default 1.
padjThresh: The adjusted p-value threshold less than which the genes will be used. Default 0.05.
splitReg: Whether to have queries of both up-regulated and down-regulated genes for each group. Default FALSE only queries up-regulated genes and should be preferred when result comes from marker detection test. When result comes from group-to-group DE test, it is recommended to set splitReg = TRUE.
...: Additional arguments passed to gprofiler2::gost(). Arguments query, custom_bg, domain_scope, and ordered_query are pre-specified by this wrapper function. Users must set organism = "mmusculus" when working on mouse data.

Value

A list object where each element is a result list for a group. Each result list contains two elements:

result: data.frame of main GO analysis result.
meta: Meta information for the query.

See gprofiler2::gost(). for detailed explanation.

References

Kolberg, L. et al, 2020 and Raudvere, U. et al, 2019

Examples

defaultCluster(pbmc) <- pbmcPlot$leiden_cluster
#> ℹ Storing given cluster labels to `cellMeta(x)` field: "defaultCluster".
# Test the DEG between "stim" and "ctrl", within each cluster
result <- runPairwiseDEG(
    pbmc,
    groupTest = "stim",
    groupCtrl = "ctrl",
    variable1 = "dataset",
    splitBy = "defaultCluster",
    nPsdRep = 3,
    minCellPerRep = 3
)
#> ℹ Running DEG within: "0"
#> ℹ Calling pairwise DESeq2 Wald test
#> ✔ Calling pairwise DESeq2 Wald test ... done
#> 
#> ℹ Running DEG within: "1"
#> ℹ Calling pairwise DESeq2 Wald test
#> -- note: fitType='parametric', but the dispersion trend was not well captured by the
#>    function: y = a/x + b, and a local regression fit was automatically substituted.
#>    specify fitType='local' or 'mean' to avoid this message next time.
#> ✔ Calling pairwise DESeq2 Wald test ... done
#> 
#> ℹ Running DEG within: "2"
#> ℹ Calling pairwise DESeq2 Wald test
#> ✔ Calling pairwise DESeq2 Wald test ... done
#> 
#> ℹ Running DEG within: "3"
#> ℹ Calling pairwise DESeq2 Wald test
#> ✔ Calling pairwise DESeq2 Wald test ... done
#> 
#> ℹ Running DEG within: "4"
#> ℹ Calling pairwise DESeq2 Wald test
#> ✔ Calling pairwise DESeq2 Wald test ... done
#> 
#> ℹ Running DEG within: "5"
#> ℹ Calling pairwise DESeq2 Wald test
#> ✔ Calling pairwise DESeq2 Wald test ... done
#> 
#> ℹ Running DEG within: "6"
#> ℹ Calling pairwise DESeq2 Wald test
#> ✔ Calling pairwise DESeq2 Wald test ... done
#> 
#> ℹ Running DEG within: "7"
#> ℹ Calling pairwise DESeq2 Wald test
#> ! Ignoring replicates (size in bracket) with too few cells: "others.ctrl.rep1 (2)"
#> ℹ Calling pairwise DESeq2 Wald test

#> ℹ Consider decrease minCellPerRep to exclude less replicates or/and lower nPsdRep to generate larger pseudo-replicates.
#> ℹ Calling pairwise DESeq2 Wald test

#> ✖ Error when computing on "7.stim": Too few replicates with more than 3 cells (`minCellPerRep`) for condition "others".
#> ℹ Calling pairwise DESeq2 Wald test

#> ! Empty result returned for this test.
#> ℹ Calling pairwise DESeq2 Wald test

#> ✖ Calling pairwise DESeq2 Wald test ... failed
#> 
# Setting `significant = FALSE` because it's hard for a gene list obtained
# from small test dataset to represent real-life biology.
# \donttest{
if (requireNamespace("gprofiler2", quietly = TRUE)) {
    go <- runGOEnrich(result, group = "0.stim", significant = FALSE)
}
# }