@@ -7,7 +7,7 @@ Authors@R: c(person("Justin", "Guinney", role=c("aut", "cre"), email="justin.gui

              person("Pablo Sebastian", "Rodriguez", role="ctb", email="pablosebastian.rodriguez@upf.edu"))

 Depends: R (>= 3.5.0)

 Imports: methods, stats, utils, graphics, BiocGenerics, S4Vectors,

-         Biobase, SummarizedExperiment, GSEABase,

+         Biobase, SummarizedExperiment, GSEABase, Matrix,

          parallel, BiocParallel, fastmatch, SingleCellExperiment

 Suggests: RUnit, limma, RColorBrewer, genefilter, edgeR, GSVAdata,

           shiny, shinythemes, ggplot2, data.table, plotly

@@ -7,6 +7,7 @@ importClassesFrom(Biobase, ExpressionSet)

 importClassesFrom(SummarizedExperiment, SummarizedExperiment)

 importClassesFrom(GSEABase, GeneSetCollection)

 importClassesFrom(SingleCellExperiment, SingleCellExperiment)

+importClassesFrom(Matrix, dgCMatrix)

 importMethodsFrom(Biobase, featureNames,

                            phenoData,

@@ -81,6 +81,61 @@ setMethod("gsva", signature(expr="SingleCellExperiment", gset.idx.list="list"),

   rval

           })

+          

+setMethod("gsva", signature(expr="dgCMatrix", gset.idx.list="list"),

+          function(expr, gset.idx.list, annotation,

+  method=c("gsva", "ssgsea", "zscore", "plage"),

+  kcdf=c("Gaussian", "Poisson", "none"),

+  abs.ranking=FALSE,

+  min.sz=1,

+  max.sz=Inf,

+  parallel.sz=1L,

+  mx.diff=TRUE,

+  tau=switch(method, gsva=1, ssgsea=0.25, NA),

+  ssgsea.norm=TRUE,

+  verbose=TRUE,

+  BPPARAM=SerialParam(progressbar=verbose))

+{

+  method <- match.arg(method)

+  kcdf <- match.arg(kcdf)

+  

+  ## filter genes according to verious criteria,

+  ## e.g., constant expression

+  expr <- .filterFeaturesSparse(expr, method)

+  

+  if (nrow(expr) < 2)

+    stop("Less than two genes in the input assay object\n")

+  

+  ## map to the actual features for which expression data is available

+  mapped.gset.idx.list <- lapply(gset.idx.list,

+                                 function(x, y) na.omit(fastmatch::fmatch(x, y)),

+                                 rownames(expr))

+  

+  if (length(unlist(mapped.gset.idx.list, use.names=FALSE)) == 0)

+    stop("No identifiers in the gene sets could be matched to the identifiers in the expression data.")

+  

+  ## remove gene sets from the analysis for which no features are available

+  ## and meet the minimum and maximum gene-set size specified by the user

+  mapped.gset.idx.list <- filterGeneSets(mapped.gset.idx.list,

+                                         min.sz=max(1, min.sz),

+                                         max.sz=max.sz)

+  

+  if (!missing(kcdf)) {

+    if (kcdf == "Gaussian") {

+      rnaseq <- FALSE

+      kernel <- TRUE

+    } else if (kcdf == "Poisson") {

+      rnaseq <- TRUE

+      kernel <- TRUE

+    } else

+      kernel <- FALSE

+  }

+  

+  rval <- .gsva(expr, mapped.gset.idx.list, method, kcdf, rnaseq, abs.ranking,

+                parallel.sz, mx.diff, tau, kernel, ssgsea.norm, verbose, BPPARAM)

+  

+  rval

+})

 setMethod("gsva", signature(expr="SummarizedExperiment", gset.idx.list="GeneSetCollection"),

           function(expr, gset.idx.list, annotation,

@@ -13,3 +13,42 @@

   expr

 }

+

+## transforms a dgCMatrix into a list of its

+## non-zero values by MARGIN (1 for row, 2 for column)

+.sparseToList <-function(dgCMat, MARGIN){

+  MARGIN <- as.integer(MARGIN)

+  J <- rep(1:ncol(dgCMat), diff(dgCMat@p))

+  I <- dgCMat@i + 1

+  x <- dgCMat@x

+  if (MARGIN == 1L) {

+    result <- split(x, I)

+    names(result) <- rownames(dgCMat)[as.numeric(names(result))]

+  } else if (MARGIN == 2L) {

+    result <- split(x, J)

+    names(result) <- colnames(dgCMat)[as.numeric(names(result))]

+  }

+  else {

+    warning("invalid MARGIN; return NULL")

+    result <- NULL

+  }

+  result

+}

+

+## filter out genes which non-zero values have

+## constant expression values

+.filterFeaturesSparse <- function(expr, method) {

+  

+  sdGenes <- sapply(.sparseToList(expr, 1), sd)

+  

+  if (any(sdGenes == 0) || any(is.na(sdGenes))) {

+    warning(sum(sdGenes == 0 | is.na(sdGenes)),

+            " genes with constant expression values throuhgout the samples.")

+    if (method != "ssgsea") {

+      warning("Since argument method!=\"ssgsea\", genes with constant expression values are discarded.")

+      expr <- expr[sdGenes > 0 & !is.na(sdGenes), ]

+    }

+  }

+  

+  expr

+}

\ No newline at end of file

@@ -1,6 +1,7 @@

 \name{gsva}

 \alias{gsva}

 \alias{gsva,SingleCellExperiment,list-method}

+\alias{gsva,dgCMatrix,list-method}

 \alias{gsva,SummarizedExperiment,list-method}

 \alias{gsva,SummarizedExperiment,GeneSetCollection-method}

 \alias{gsva,ExpressionSet,list-method}

@@ -29,6 +30,18 @@ Estimates GSVA enrichment scores.

     ssgsea.norm=TRUE,

     verbose=TRUE,

     BPPARAM=SerialParam(progressbar=verbose))

+\S4method{gsva}{dgCMatrix,list}(expr, gset.idx.list, annotation,

+    method=c("gsva", "ssgsea", "zscore", "plage"),

+    kcdf=c("Gaussian", "Poisson", "none"),

+    abs.ranking=FALSE,

+    min.sz=1,

+    max.sz=Inf,

+    parallel.sz=1L,

+    mx.diff=TRUE,

+    tau=switch(method, gsva=1, ssgsea=0.25, NA),

+    ssgsea.norm=TRUE,

+    verbose=TRUE,

+    BPPARAM=SerialParam(progressbar=verbose))

 \S4method{gsva}{SummarizedExperiment,GeneSetCollection}(expr, gset.idx.list, annotation,

     method=c("gsva", "ssgsea", "zscore", "plage"),

     kcdf=c("Gaussian", "Poisson", "none"),

@@ -106,7 +119,8 @@ Estimates GSVA enrichment scores.

   \item{expr}{Gene expression data which can be given either as a

               \code{SummarizedExperiment}, \code{SingleCellExperiment}

               \code{ExpressionSet} object, or as a matrix of expression

-              values where rows correspond to genes and columns correspond to samples.}

+              values where rows correspond to genes and columns correspond to samples.

+              This matrix can be also in a sparse format, as a \code{dgCMatrix}.}

   \item{gset.idx.list}{Gene sets provided either as a \code{list} object or as a

                        \code{GeneSetCollection} object.}

   \item{annotation}{In the case of calling \code{gsva()} on a

@@ -190,10 +204,11 @@ However, then the input gene sets in \code{gset.idx.list} is provided as a

 those identifiers to the type of identifier in the input expression data \code{expr}.

 Such an automatic conversion, however, will only occur in three scenarios: 1. when

 \code{expr} is an \code{ExpressionSet} object with an appropriately set

-\code{annotation} slot; 2. when \code{expr} is a \code{SummarizedExperiment} object

-with an appropriately set \code{annotation} slot in the metadata of \code{expr};

-3. when \code{expr} is a \code{matrix} and the \code{annotation} argument of the

-\code{gsva()} function is set to the name of the annotation package that provides

+\code{annotation} slot; 2. when \code{expr} is a \code{SummarizedExperiment} or a

+\code{SingleCellExperiment} object with an appropriately set \code{annotation} slot

+in the metadata of \code{expr}; 3. when \code{expr} is a \code{matrix} or a 

+\code{dgCMatrix} and the \code{annotation} argument of the \code{gsva()} function

+is set to the name of the annotation package that provides

 the relationships between the type of identifiers in \code{expr} and \code{gset.idx.list}.

 The collection of gene sets resulting from the previous identifier matching,

@@ -201,7 +216,8 @@ can be further filtered to require a minimun and/or maximum size by using the

 arguments \code{min.sz} and \code{max.sz}.

 }

 \value{

-A gene-set by sample matrix of GSVA enrichment scores.

+A gene-set by sample matrix (of \code{matrix} or \code{dgCMatrix} type, 

+depending on the input) of GSVA enrichment scores.

 }

 \references{

 Barbie, D.A. et al. Systematic RNA interference reveals that oncogenic KRAS-driven