WebGestaltR: The R interface for enrichment analysis with WebGestalt.

Main function for enrichment analysis

Usage

WebGestaltR(
  enrichMethod = "ORA",
  organism = "hsapiens",
  enrichDatabase = NULL,
  enrichDatabaseFile = NULL,
  enrichDatabaseType = NULL,
  enrichDatabaseDescriptionFile = NULL,
  interestGeneFile = NULL,
  interestGene = NULL,
  interestGeneType = NULL,
  interestGeneNames = NULL,
  collapseMethod = "mean",
  referenceGeneFile = NULL,
  referenceGene = NULL,
  referenceGeneType = NULL,
  referenceSet = NULL,
  minNum = 10,
  maxNum = 500,
  sigMethod = "fdr",
  fdrMethod = "BH",
  fdrThr = 0.05,
  topThr = 10,
  reportNum = 20,
  perNum = 1000,
  gseaP = 1,
  isOutput = TRUE,
  outputDirectory = getwd(),
  projectName = NULL,
  dagColor = "continuous",
  saveRawGseaResult = FALSE,
  gseaPlotFormat = c("png", "svg"),
  setCoverNum = 10,
  networkConstructionMethod = NULL,
  neighborNum = 10,
  highlightType = "Seeds",
  highlightSeedNum = 10,
  nThreads = 1,
  cache = NULL,
  hostName = "https://www.webgestalt.org/",
  useWeightedSetCover = FALSE,
  useAffinityPropagation = FALSE,
  usekMedoid = TRUE,
  kMedoid_k = 25,
  ...
)

WebGestaltRBatch(
  interestGeneFolder = NULL,
  enrichMethod = "ORA",
  isParallel = FALSE,
  nThreads = 3,
  ...
)

Arguments

enrichMethod

Enrichment methods: ORA, GSEA or NTA.

organism

Currently, WebGestaltR supports 12 organisms. Users can use the function listOrganism to check available organisms. Users can also input others to perform the enrichment analysis for other organisms not supported by WebGestaltR. For other organisms, users need to provide the functional categories, interesting list and reference list (for ORA method). Because WebGestaltR does not perform the ID mapping for the other organisms, the above data should have the same ID type.

enrichDatabase

The functional categories for the enrichment analysis. Users can use the function listGeneSet to check the available functional databases for the selected organism. Multiple databases in a vector are supported for ORA and GSEA.

enrichDatabaseFile

Users can provide one or more GMT files as the functional category for enrichment analysis. The extension of the file should be gmt and the first column of the file is the category ID, the second one is the external link for the category. Genes annotated to the category are from the third column. All columns are separated by tabs. The GMT files will be combined with enrichDatabase.

enrichDatabaseType

The ID type of the genes in the enrichDatabaseFile. If users set organism as others, users do not need to set this ID type because WebGestaltR will not perform ID mapping for other organisms. The supported ID types of WebGestaltR for the selected organism can be found by the function listIdType.

enrichDatabaseDescriptionFile

Users can also provide description files for the custom enrichDatabaseFile. The extension of the description file should be des. The description file contains two columns: the first column is the category ID that should be exactly the same as the category ID in the custom enrichDatabaseFile and the second column is the description of the category. All columns are separated by tabs.

interestGeneFile

If enrichMethod is ORA or NTA, the extension of the interestGeneFile should be txt and the file can only contain one column: the interesting gene list. If enrichMethod is GSEA, the extension of the interestGeneFile should be rnk and the file should contain two columns separated by tab: the gene list and the corresponding scores.

interestGene

Users can also use an R object as the input. If enrichMethod is ORA or NTA, interestGene should be an R vector object containing the interesting gene list. If enrichMethod is GSEA, interestGene should be an R data.frame object containing two columns: the gene list and the corresponding scores.

interestGeneType

The ID type of the interesting gene list. The supported ID types of WebGestaltR for the selected organism can be found by the function listIdType. If the organism is others, users do not need to set this parameter.

interestGeneNames

The names of the id lists for multiomics data.

collapseMethod

The method to collapse duplicate IDs with scores. mean, median, min and max represent the mean, median, minimum and maximum of scores for the duplicate IDs.

referenceGeneFile

For the ORA method, the users need to upload the reference gene list. The extension of the referenceGeneFile should be txt and the file can only contain one column: the reference gene list.

referenceGene

For the ORA method, users can also use an R object as the reference gene list. referenceGene should be an R vector object containing the reference gene list.

referenceGeneType

The ID type of the reference gene list. The supported ID types of WebGestaltR for the selected organism can be found by the function listIdType. If the organism is others, users do not need to set this parameter.

referenceSet

Users can directly select the reference set from existing platforms in WebGestaltR and do not need to provide the reference set through referenceGeneFile. All existing platforms supported in WebGestaltR can be found by the function listReferenceSet. If referenceGeneFile and refereneceGene are NULL, WebGestaltR will use the referenceSet as the reference gene set. Otherwise, WebGestaltR will use the user supplied reference set for enrichment analysis.

minNum

WebGestaltR will exclude the categories with the number of annotated genes less than minNum for enrichment analysis. The default is 10.

maxNum

WebGestaltR will exclude the categories with the number of annotated genes larger than maxNum for enrichment analysis. The default is 500.

sigMethod

Two methods of significance are available in WebGestaltR: fdr and top. fdr means the enriched categories are identified based on the FDR and top means all categories are ranked based on FDR and then select top categories as the enriched categories. The default is fdr.

fdrMethod

For the ORA method, WebGestaltR supports five FDR methods: holm, hochberg, hommel, bonferroni, BH and BY. The default is BH.

fdrThr

The significant threshold for the fdr method. The default is 0.05.

topThr

The threshold for the top method. The default is 10.

reportNum

The number of enriched categories visualized in the final report. The default is 20. A larger reportNum may be slow to render in the report.

perNum

The number of permutations for the GSEA method. The default is 1000.

gseaP

The exponential scaling factor of the phenotype score. The default is 1. When p=0, ES reduces to standard K-S statistics (See original paper for more details).

isOutput

If isOutput is TRUE, WebGestaltR will create a folder named by the projectName and save the results in the folder. Otherwise, WebGestaltR will only return an R data.frame object containing the enrichment results. If hundreds of gene list need to be analyzed simultaneously, it is better to set isOutput to FALSE. The default is TRUE.

outputDirectory

The output directory for the results.

projectName

The name of the project. If projectName is NULL, WebGestaltR will use time stamp as the project name.

dagColor

If dagColor is binary, the significant terms in the DAG structure will be colored by steel blue for ORA method or steel blue (positive related) and dark orange (negative related) for GSEA method. If dagColor is continous, the significant terms in the DAG structure will be colored by the color gradient based on corresponding FDRs.

saveRawGseaResult

Whether the raw result from GSEA is saved as a RDS file, which can be used for plotting. Defaults to FALSE. The list includes

Enrichment_Results: A data frame of GSEA results with statistics
Running_Sums: A matrix of running sum of scores for each gene set
Items_in_Set: A list with ranks of genes for each gene set

gseaPlotFormat

The graphic format of GSEA enrichment plots. Either svg, png, or c("png", "svg") (default).

setCoverNum

The number of expected gene sets after set cover to reduce redundancy. It could get fewer sets if the coverage reaches 100%. The default is 10.

networkConstructionMethod

Netowrk construction method for NTA. Either Network_Retrieval_Prioritization or Network_Expansion. Network Retrieval & Prioritization first uses random walk analysis to calculate random walk probabilities for the input seeds, then identifies the relationships among the seeds in the selected network and returns a retrieval sub-network. The seeds with the top random walk probabilities are highlighted in the sub-network. Network Expansion first uses random walk analysis to rank all genes in the selected network based on their network proximity to the input seeds and then return an expanded sub-network in which nodes are the input seeds and their top ranking neighbors and edges represent their relationships.

neighborNum

The number of neighbors to include in NTA Network Expansion method.

highlightType

The type of nodes to highlight in the NTA Network Expansion method, either Seeds or Neighbors.

highlightSeedNum

The number of top input seeds to highlight in NTA Network Retrieval & Prioritizaiton method.

nThreads

The number of cores to use for GSEA and set cover, and in batch function.

cache

A directory to save data cache for reuse. Defaults to NULL and disabled.

hostName

The server URL for accessing data. Mostly for development purposes.

useWeightedSetCover

Use weighted set cover for ORA. Defaults to TRUE.

useAffinityPropagation

Use affinity propagation for ORA. Defaults to FALSE.

usekMedoid

Use k-medoid for ORA. Defaults to TRUE.

kMedoid_k

The number of clusters for k-medoid. Defaults to 25.

...

In batch function, passes parameters to WebGestaltR function. Also handles backward compatibility for some parameters in old versions.

interestGeneFolder

Run WebGestaltR for gene list files in the folder.

isParallel

If jobs are run parallelly in the batch.

Value

The WebGestaltR function returns a data frame containing the enrichment analysis result and also outputs an user-friendly HTML report if isOutput is TRUE. The columns in the data frame depend on the enrichMethod and they are the following:

geneSet: ID of the gene set.
description: Description of the gene set if available.
link: Link to the data source.
size: The number of genes in the set after filtering by minNum and maxNum.
overlap: The number of mapped input genes that are annotated in the gene set.
expect: Expected number of input genes that are annotated in the gene set.
enrichmentRatio: Enrichment ratio, overlap / expect.
enrichmentScore: Enrichment score, the maximum running sum of scores for the ranked list.
normalizedEnrichmentScore: Normalized enrichment score, normalized against the average enrichment score of all permutations.
leadingEdgeNum: Number of genes/phosphosites in the leading edge.
pValue: P-value from hypergeometric test for ORA. For GSEA, please refer to its original publication or online at https://software.broadinstitute.org/gsea/doc/GSEAUserGuideTEXT.htm.
FDR: Corrected P-value for mulilple testing with fdrMethod for ORA.
overlapId: The gene/phosphosite IDs of overlap for ORA (entrez gene IDs or phosphosite sequence).
leadingEdgeId: Genes/phosphosites in the leading edge in entrez gene ID or phosphosite sequence.
userId: The gene/phosphosite IDs of overlap for ORA or leadingEdgeId for GSEA in User input IDs.
plotPath: Path of the GSEA enrichment plot.
database: Name of the source database if multiple enrichment databases are given.
goId: In NTA, like geneSet, the enriched GO terms of genes in the returned subnetwork.
interestGene: In NTA, the gene IDs in the subnetwork with 0/1 annotations indicating if it is from user input.

The WebGestaltRBatch function returns a list of enrichment results.

Details

WebGestaltR function can perform three enrichment analyses: ORA (Over-Representation Analysis) and GSEA (Gene Set Enrichment Analysis).and NTA (Network Topology Analysis). Based on the user-uploaded gene list or gene list with scores, WebGestaltR function will first map the gene list to the entrez gene ids and then summarize the gene list based on the GO (Gene Ontology) Slim. After performing the enrichment analysis, WebGestaltR function also returns a user-friendly HTML report containing GO Slim summary and the enrichment analysis result. If functional categories have DAG (directed acyclic graph) structure or genes in the functional categories have network structure, those relationship can also be visualized in the report.

Examples

if (FALSE) {
####### ORA example #########
geneFile <- system.file("extdata", "interestingGenes.txt", package = "WebGestaltR")
refFile <- system.file("extdata", "referenceGenes.txt", package = "WebGestaltR")
outputDirectory <- getwd()
enrichResult <- WebGestaltR(
  enrichMethod = "ORA", organism = "hsapiens",
  enrichDatabase = "pathway_KEGG", interestGeneFile = geneFile,
  interestGeneType = "genesymbol", referenceGeneFile = refFile,
  referenceGeneType = "genesymbol", isOutput = TRUE,
  outputDirectory = outputDirectory, projectName = NULL
)

####### GSEA example #########
rankFile <- system.file("extdata", "GeneRankList.rnk", package = "WebGestaltR")
outputDirectory <- getwd()
enrichResult <- WebGestaltR(
  enrichMethod = "GSEA", organism = "hsapiens",
  enrichDatabase = "pathway_KEGG", interestGeneFile = rankFile,
  interestGeneType = "genesymbol", sigMethod = "top", topThr = 10, minNum = 5,
  outputDirectory = outputDirectory
)

####### NTA example #########
enrichResult <- WebGestaltR(
  enrichMethod = "NTA", organism = "hsapiens",
  enrichDatabase = "network_PPI_BIOGRID", interestGeneFile = geneFile,
  interestGeneType = "genesymbol", sigMethod = "top", topThr = 10,
  outputDirectory = getwd(), highlightSeedNum = 10,
  networkConstructionMethod = "Network_Retrieval_Prioritization"
)
}