R/textORA.R

In noriakis/wcGeneSummary: visualization and interpretation of textual information from omics summary data through network analysis

#' textORA
#' 
#' @description Return text p-values from hypergeometric distribution
#' @details Returns the p-values for the words using hypergeometric distribution (experimental).
#' By default uses the RefSeq background, but you can use the favorite backgrounds.
#' 
#' @param queries gene list (Entrez ID)
#' @param notGene boolean specifying type of input, default to gene IDs
#' @param bg default to NULL
#' @return p-values for the words
#' @import GeneSummary tm
#' @return p-values for the words
#' @examples textORA(c("2067","2068","2071","2072"))
#' @export
#' 
textORA <- function(queries, notGene=FALSE, bg=NULL) {
    ##TODO Options to use other backgrounds
    tb <- loadGeneSummary()
    if (!notGene) {
        fil <- tb |> dplyr::filter(tb$Gene_ID %in% queries)
        fil <- fil[!duplicated(fil$Gene_ID),]

        ## Make corpus for queried genes
        docs <- VCorpus(VectorSource(fil$Gene_summary))

        ## The same filtering as the `allFreqGeneSummary`
        docs <- docs |>
            tm_map(FUN=content_transformer(tolower)) |> 
            tm_map(FUN=removeNumbers) |>
            tm_map(removeWords, stopwords::stopwords("english", "stopwords-iso")) |>
            tm_map(FUN=removePunctuation) |>
            tm_map(FUN=stripWhitespace)
    } else {
        docs <- VCorpus(VectorSource(queries))
    }

    docs <- TermDocumentMatrix(docs)
    mat <- as.matrix(docs)
    matSorted <- sort(rowSums(mat), decreasing=TRUE)

    if (is.null(bg)) {
        data_env <- new.env(parent = emptyenv())
        load(system.file("extdata", "sysdata.rda", package = "biotextgraph"),
            envir=data_env)
        bg <- data_env[["allFreqGeneSummary"]]
    }

    pvs <- vapply(names(matSorted),
        function (x) returnP(x, matSorted, bg),
        FUN.VALUE=1)
    pvs
}

#' returnP
#' 
#' Calculate p-values
#' @param name name of a gene
#' @param matSorted words frequency in the cluster
#' @param allFreqGeneSummary words frequency in the whole description
#' @return p-values for the words
#' 
returnP <- function(name, matSorted, allFreqGeneSummary){
    query <- as.numeric(matSorted[name])
    noquery <- sum(matSorted) - query
    queryAll <- as.numeric(allFreqGeneSummary[name, "freq"])
    allwords <- sum(allFreqGeneSummary$freq) - queryAll
    
    ## p-value
    return(sum(dhyper(query:sum(matSorted), queryAll,
        allwords, sum(matSorted))))
}