#' Epithelial-Mesenchymal Transition Signature
#'
#' @description This signature is computed accordingly to the reference paper,
#' to have more details explore the function
#' \code{\link[signifinder]{availableSignatures}}.
#'
#' @param dataset Normalized expression values. A data frame or a matrix
#' where rows correspond to genes and columns correspond to samples.
#' Alternatively, an object of type \linkS4class{SummarizedExperiment},
#' \code{\link[SingleCellExperiment]{SingleCellExperiment}} or
#' \code{\link[SpatialExperiment]{SpatialExperiment}} where the normalized
#' expression values should be in an assay called 'norm_expr'.
#' @param nametype character string saying the type of gene name ID (row names
#' in dataset). Either one of "SYMBOL", "ENTREZID" or "ENSEMBL".
#' @param inputType character string saying the type of data you are using.
#' Either one of "microarray" or "rnaseq".
#' @param author character string saying the first author of the signature
#' publication. Check it in \code{\link[signifinder]{availableSignatures}}.
#' @param ... other arguments passed on to the \code{\link[GSVA]{gsva}}
#' function.
#'
#' @return If dataset is a \linkS4class{SummarizedExperiment} object, then
#' scores are added in the \code{\link[SummarizedExperiment]{colData}} section.
#' If dataset is a data frame or a matrix, then a
#' \linkS4class{SummarizedExperiment} object is created in which scores are
#' added in the \code{\link[SummarizedExperiment]{colData}} section.
#'
#' @importFrom GSVA gsva
#' @importFrom stats prcomp
#'
#' @examples
#' data(ovse)
#' EMTSign(dataset = ovse, inputType = "rnaseq")
#'
#' @export
EMTSign <- function(
dataset, nametype = "SYMBOL", inputType = "microarray",
author = "Miow", ...) {
.consistencyCheck(nametype, "EMTSign", author)
datasetm <- .getMatrix(dataset)
if (author == "Miow") {
sign_df <- EMT_Miow
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
EL <- sign_df[grep("Epithelial-like", sign_df$class), ]
ML <- sign_df[grep("Mesenchymal-like", sign_df$class), ]
.percentageOfGenesUsed(
"EMTSign", datasetm, EL$SYMBOL, "epithelial", author = author)
.percentageOfGenesUsed(
"EMTSign", datasetm, ML$SYMBOL, "mesenchymal", author = author)
gene_sets <- list( Epithelial = EL$SYMBOL, Mesenchymal = ML$SYMBOL)
names(gene_sets) <- paste0("EMT_Miow_", names(gene_sets))
dots <- list(...)
kcdftype <- ifelse(inputType == "microarray", "Gaussian", "Poisson")
args <- .matchArguments(dots, list(
expr = datasetm, gset.idx.list = gene_sets, method = "ssgsea",
kcdf = kcdftype, ssgsea.norm = FALSE, verbose = FALSE))
gsva_matrix <- do.call(gsva, args)
return(.returnAsInput(
userdata = dataset, result = gsva_matrix,
SignName = "", datasetm))
} else {
if (author == "Mak") {
sign_df <- EMT_Mak
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
Sign_E <- sign_df$SYMBOL[sign_df$class == "E"]
Sign_M <- sign_df$SYMBOL[sign_df$class == "M"]
.percentageOfGenesUsed(
"EMTSign", datasetm, Sign_E, "epithelial", author = author)
.percentageOfGenesUsed(
"EMTSign", datasetm, Sign_M, "mesenchymal", author = author)
columnNA <- .managena(datasetm, genes = sign_df$SYMBOL)
score <- colMeans(
datasetm[intersect(Sign_M, row.names(datasetm)), ]) - colMeans(
datasetm[intersect(Sign_E, row.names(datasetm)), ])
score[columnNA > 0.9] <- NA
} else if (author == "Cheng") {
sign_df <- EMT_Cheng
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
.percentageOfGenesUsed(
"EMTSign", datasetm, sign_df$SYMBOL, author = author)
datasetm_n <- datasetm[
intersect(row.names(datasetm), sign_df$SYMBOL), ]
datasetm_n <- if (inputType == "rnaseq") {
log2(datasetm_n + 1)
} else { datasetm_n }
columnNA <- .managena(datasetm = datasetm_n, genes = sign_df$SYMBOL)
score <- prcomp(t(datasetm_n))$x[, 1]
score[columnNA > 0.9] <- NA
}
return(.returnAsInput(
userdata = dataset, result = score,
SignName = paste0("EMT_", author), datasetm))
}
}
#' Pyroptosis Signature
#'
#' @inherit EMTSign description
#' @inheritParams EMTSign
#' @param hgReference character string saying the human reference genome.
#' Either one of "hg19" or "hg38".
#'
#' @inherit EMTSign return
#'
#' @importFrom SummarizedExperiment assays
#'
#' @examples
#' data(ovse)
#' pyroptosisSign(dataset = ovse)
#'
#' @export
pyroptosisSign <- function(
dataset, nametype = "SYMBOL", inputType = "rnaseq",
author = "Ye", hgReference = "hg38") {
.consistencyCheck(nametype, "pyroptosisSign", author)
sign_df <- get(paste0("Pyroptosis_", author))
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
datasetm <- .getMatrix(dataset)
if (author == "Ye") {
dataset <- .dataTransformation(
dataset, datasetm, "FPKM", hgReference, nametype)
datasetm_n <- as.matrix(assays(dataset)[["FPKM"]])
datasetm_n <- scale(datasetm_n)
} else if (author == "Shao") {
if (inputType == "rnaseq") {
dataset <- .dataTransformation(
dataset, datasetm, "FPKM", hgReference, nametype)
datasetm_n <- as.matrix(assays(dataset)[["FPKM"]])
} else { datasetm_n <- datasetm }
} else if (author == "Lin") {
dataset <- .dataTransformation(
dataset, datasetm, "TPM", hgReference, nametype)
datasetm_n <- as.matrix(assays(dataset)[["TPM"]])
} else { datasetm_n <- datasetm}
score <- .coeffScore(sign_df, datasetm_n, "pyroptosisSign", author = author)
return(.returnAsInput(
userdata = dataset, result = score,
SignName = paste0("Pyroptosis_", author), datasetm))
}
#' Ferroptosis Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @importFrom SummarizedExperiment assays
#'
#' @examples
#' data(ovse)
#' ferroptosisSign(dataset = ovse)
#'
#' @export
ferroptosisSign <- function(
dataset, nametype = "SYMBOL", inputType = "rnaseq",
author = "Ye", hgReference = "hg38") {
.consistencyCheck(nametype, "ferroptosisSign", author)
sign_df <- get(paste0("Ferroptosis_", author))
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
datasetm <- .getMatrix(dataset)
if (author %in% c("Liu", "Ye") & inputType == "rnaseq") {
dataset <- .dataTransformation(
dataset, datasetm, "FPKM", hgReference, nametype)
datasetm_n <- as.matrix(assays(dataset)[["FPKM"]])
datasetm_n <- datasetm_n[rownames(datasetm_n) %in% sign_df$SYMBOL, ]
} else {
datasetm_n <- datasetm[rownames(datasetm) %in% sign_df$SYMBOL, ]
}
score <- .coeffScore(
sign_df, datasetm_n, "ferroptosisSign", author = author)
if (author == "Liang") { score <- exp(score) }
return(.returnAsInput(
userdata = dataset, result = score,
SignName = paste0("Ferroptosis_", author), datasetm))
}
#' Lipid Metabolism Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @examples
#' data(ovse)
#' lipidMetabolismSign(dataset = ovse)
#'
#' @export
lipidMetabolismSign <- function(dataset, nametype = "SYMBOL") {
.consistencyCheck(nametype, "lipidMetabolismSign")
sign_df <- LipidMetabolism_Zheng
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
datasetm <- .getMatrix(dataset)
score <- .coeffScore(sign_df, datasetm, "lipidMetabolismSign")
return(.returnAsInput(
userdata = dataset, result = score,
SignName = "LipidMetabolism_Zheng", datasetm))
}
#' Hypoxia Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @examples
#' data(ovse)
#' hypoxiaSign(dataset = ovse, inputType = "rnaseq")
#'
#' @export
hypoxiaSign <- function(
dataset, nametype = "SYMBOL", inputType = "microarray") {
.consistencyCheck(nametype, "hypoxiaSign")
datasetm <- .getMatrix(dataset)
datasetm_n <- if (inputType == "rnaseq") {
log2(datasetm + 1)
} else { datasetm }
score <- .statScore(
Hypoxia_Buffa$SYMBOL, datasetm = abs(datasetm_n), nametype = nametype,
typeofstat = "median", namesignature = "hypoxiaSign")
return(.returnAsInput(
userdata = dataset, result = as.vector(scale(score)),
SignName = "Hypoxia_Buffa", datasetm))
}
#' Immunogenic Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @importFrom SummarizedExperiment assays
#'
#' @examples
#' data(ovse)
#' immunoScoreSign(dataset = ovse)
#'
#' @export
immunoScoreSign <- function(
dataset, nametype = "SYMBOL", author = "Hao",
inputType = "rnaseq", hgReference = "hg38") {
.consistencyCheck(nametype, "immunoScoreSign", author)
datasetm <- .getMatrix(dataset)
if (author == "Hao") {
sign_df <- ImmunoScore_Hao
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
g <- intersect(row.names(datasetm), sign_df$SYMBOL)
.percentageOfGenesUsed(
"immunoScoreSign", datasetm, sign_df$SYMBOL, author = author)
if (inputType == "rnaseq") {
dataset <- .dataTransformation(
dataset, datasetm, "FPKM", hgReference, nametype)
datasetm_n <- log2(as.matrix(assays(dataset)[["FPKM"]]) + 0.01)
datasetm_n <- datasetm_n[g, ]
} else { datasetm_n <- datasetm[g, ] }
sign_df <- sign_df[sign_df$SYMBOL %in% g, ]
columnNA <- .managena(datasetm_n, g)
SE <- (sign_df$HR - sign_df$`95CI_L`) / 1.96
k <- (1 - sign_df$HR) / SE
score <- unlist(lapply(seq_len(ncol(datasetm_n)), function(p) {
sum(k * datasetm_n[, p], na.rm = TRUE) }))
score[columnNA > 0.9] <- NA
} else if (author == "Roh") {
datasetm_n <- log2(datasetm + 1)
score <- .statScore(
ImmunoScore_Roh$SYMBOL, datasetm = datasetm_n,
nametype = nametype, typeofstat = ".meang",
namesignature = "immunoScoreSign", author = author)
}
return(.returnAsInput(
userdata = dataset, result = score,
SignName = paste0("ImmunoScore_", author), datasetm))
}
#' ConsensusOV Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#' @param ... optional parameters to be passed to
#' \code{\link[consensusOV]{get.subtypes}}.
#'
#' @inherit EMTSign return
#'
#' @importFrom consensusOV get.subtypes
#' @importFrom AnnotationDbi mapIds
#' @importFrom org.Hs.eg.db org.Hs.eg.db
#'
#' @examples
#' data(ovse)
#' consensusOVSign(dataset = ovse)
#'
#' @export
consensusOVSign <- function(dataset, nametype = "SYMBOL", ...) {
.consistencyCheck(nametype, "consensusOVSign")
datasetm <- .getMatrix(dataset)
if (nametype != "ENTREZID") {
genename <- mapIds(
org.Hs.eg.db,
keys = row.names(datasetm),
column = "ENTREZID",
keytype = nametype,
multiVals = "first")
datasetm_n <- datasetm[!is.na(genename), ]
genename <- genename[!is.na(genename)]
datasetm_n <- datasetm_n[!duplicated(genename), ]
genename <- genename[!duplicated(genename)]
} else {
genename <- row.names(datasetm)
datasetm_n <- datasetm
}
consensus_subtypes <- get.subtypes(
expression.dataset = datasetm_n, entrez.ids = genename, ...)
scores <- consensus_subtypes$rf.probs
colnames(scores) <- paste0(
"ConsensusOV_Chen_", substring(colnames(scores), 1, 3))
return(.returnAsInput(
userdata = dataset, result = t(scores), SignName = "", datasetm))
}
#' ImmunoPhenoScore Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @importFrom stats sd
#' @importFrom SummarizedExperiment assays
#'
#' @examples
#' data(ovse)
#' IPSSign(dataset = ovse)
#'
#' @export
IPSSign <- function(dataset, nametype = "SYMBOL", hgReference = "hg38") {
## code adapted from https://github.com/icbi-lab/Immunophenogram
.consistencyCheck(nametype, "IPSSign")
sign_df <- IPS_Charoentong
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
datasetm <- .getMatrix(dataset)
sample_names <- colnames(datasetm)
.percentageOfGenesUsed("IPSSign", datasetm, sign_df$SYMBOL)
dataset <- .dataTransformation(
dataset, datasetm, "TPM", hgReference, nametype)
datasetm_n <- log2(as.matrix(assays(dataset)[["TPM"]]) + 1)
datasetm_n <- datasetm_n[rownames(datasetm_n) %in% sign_df$SYMBOL, ]
IPS <- NULL
MHC <- NULL
CP <- NULL
EC <- NULL
SC <- NULL
AZ <- NULL
for (i in seq_len(length(sample_names))) {
GE <- datasetm_n[, i]
Z1 <- (datasetm_n[match(sign_df$SYMBOL, row.names(datasetm_n)), i] -
mean(GE)) / sd(GE)
WEIGHT <- NULL
MIG <- NULL
k <- 1
for (gen in unique(sign_df$NAME)) {
MIG[k] <- mean(Z1[which(sign_df$NAME == gen)], na.rm = TRUE)
WEIGHT[k] <- mean(sign_df$WEIGHT[which(sign_df$NAME == gen)])
k <- k + 1}
WG <- MIG * WEIGHT
MHC[i] <- mean(WG[which(
unique(sign_df$NAME) %in% unique(
sign_df$NAME[sign_df$class == "MHC"]))], na.rm = TRUE)
CP[i] <- mean(WG[which(
unique(sign_df$NAME) %in% unique(
sign_df$NAME[sign_df$class == "CP"]))], na.rm = TRUE)
EC[i] <- mean(WG[which(
unique(sign_df$NAME) %in% unique(
sign_df$NAME[sign_df$class == "EC"]))], na.rm = TRUE)
SC[i] <- mean(WG[which(
unique(sign_df$NAME) %in% unique(
sign_df$NAME[sign_df$class == "SC"]))], na.rm = TRUE)
AZ[i] <- sum(MHC[i], CP[i], EC[i], SC[i])
IPS[i] <- .ipsmap(AZ[i])
}
ipsres <- data.frame(IPS, MHC, CP, EC, SC)
rownames(ipsres) <- sample_names
colnames(ipsres) <- c(
"IPS_Charoentong", "IPS_Charoentong_MHC", "IPS_Charoentong_CP",
"IPS_Charoentong_EC", "IPS_Charoentong_SC")
return(.returnAsInput(
userdata = dataset, result = t(ipsres), SignName = "", datasetm))
}
#' Core Matrisome Gene signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @examples
#' data(ovse)
#' matrisomeSign(dataset = ovse)
#'
#' @export
matrisomeSign <- function(dataset, nametype = "SYMBOL") {
.consistencyCheck(nametype, "matrisomeSign")
datasetm <- .getMatrix(dataset)
score <- .statScore(
Matrisome_Yuzhalin$SYMBOL, datasetm,
nametype, "median", "matrisomeSign")
return(.returnAsInput(
userdata = dataset, result = score,
SignName = "Matrisome_Yuzhalin", datasetm))
}
#' Mitotic Index
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @examples
#' data(ovse)
#' mitoticIndexSign(dataset = ovse)
#'
#' @export
mitoticIndexSign <- function(dataset, nametype = "SYMBOL") {
.consistencyCheck(nametype, "mitoticIndexSign")
datasetm <- .getMatrix(dataset)
score <- .statScore(
MitoticIndex_Yang$SYMBOL, datasetm,
nametype, "mean", "mitoticIndexSign")
return(.returnAsInput(
userdata = dataset, result = score,
SignName = "MitoticIndex_Yang", datasetm))
}
#' Immune Cytolytic Activity Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @importFrom SummarizedExperiment assays
#'
#' @examples
#' data(ovse)
#' immuneCytSign(dataset = ovse, inputType = "rnaseq")
#'
#' @export
immuneCytSign <- function(
dataset, nametype = "SYMBOL", inputType = "microarray",
author = "Rooney", hgReference = "hg38") {
.consistencyCheck(nametype, "immuneCytSign", author)
datasetm <- .getMatrix(dataset)
if (author == "Rooney") {
if (inputType == "rnaseq") {
dataset <- .dataTransformation(
dataset, datasetm, "TPM", hgReference, nametype)
datasetm_n <- as.matrix(assays(dataset)[["TPM"]]) + 0.01
} else { datasetm_n <- datasetm }
score <- .statScore(
ImmuneCyt_Rooney$SYMBOL, datasetm_n, nametype,
".meang", "ImmuneCytSign", author = author)
} else if (author == "Davoli") {
datasetm_n <- datasetm[
row.names(datasetm) %in% ImmuneCyt_Davoli$SYMBOL, ]
datasetm_r <- apply(datasetm_n, 1, rank)
datasetm_r <- (datasetm_r - 1) / (nrow(datasetm_r) - 1)
score <- .statScore(
ImmuneCyt_Davoli$SYMBOL, t(datasetm_r), nametype,
".meang", "ImmuneCytSign", author = author)
}
return(.returnAsInput(
userdata = dataset, result = score,
SignName = paste0("ImmuneCyt_", author), datasetm))
}
#' IFN-gamma Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @examples
#' data(ovse)
#' IFNSign(dataset = ovse)
#'
#' @export
IFNSign <- function(dataset, nametype = "SYMBOL") {
.consistencyCheck(nametype, "IFNSign")
datasetm <- .getMatrix(dataset)
datasetm_n <- log10(datasetm + 1)
score <- .statScore(
IFN_Ayers$SYMBOL, datasetm_n, nametype, "mean", "IFNSign")
return(.returnAsInput(
userdata = dataset, result = score, SignName = "IFN_Ayers", datasetm))
}
#' ExpandedImmune Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @examples
#' data(ovse)
#' expandedImmuneSign(dataset = ovse)
#'
#' @export
expandedImmuneSign <- function(dataset, nametype = "SYMBOL") {
.consistencyCheck(nametype, "expandedImmuneSign")
datasetm <- .getMatrix(dataset)
datasetm_n <- log10(datasetm + 1)
score <- .statScore(
ExpandedImmune_Ayers$SYMBOL, datasetm_n,
nametype, "mean", "expandedImmuneSign")
return(.returnAsInput(
userdata = dataset, result = score,
SignName = "ExpandedImmune_Ayers", datasetm))
}
#' TinflamSign Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @examples
#' data(ovse)
#' TinflamSign(dataset = ovse)
#'
#' @export
TinflamSign <- function(dataset, nametype = "SYMBOL") {
.consistencyCheck(nametype, "TinflamSign")
sign_df <- Tinflam_Ayers
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
datasetm <- .getMatrix(dataset)
datasetm_n <- log10(datasetm + 1)
housekeeping <- intersect(
row.names(datasetm_n), sign_df$SYMBOL[sign_df$class == "Housekeeping"])
genes_pred <- intersect(
row.names(datasetm_n), sign_df$SYMBOL[sign_df$class == "TInflam"])
housekeeping_m <- apply(datasetm_n[housekeeping, ], 2, mean)
datasetm_n <- sweep(datasetm_n[genes_pred, ], 2, housekeeping_m, FUN = "-")
score <- .coeffScore(
sign_df[sign_df$SYMBOL %in% genes_pred, ], datasetm_n, "TinflamSign")
return(.returnAsInput(
userdata = dataset, result = score,
SignName = "Tinflam_Ayers", datasetm))
}
#' Tertiary Lymphoid Structures (TLS) Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @examples
#' data(ovse)
#' TLSSign(dataset = ovse)
#'
#' @export
TLSSign <- function(dataset, nametype = "SYMBOL", inputType = "rnaseq") {
.consistencyCheck(nametype, "TLSSign")
datasetm <- .getMatrix(dataset)
datasetm_n <- if (inputType == "rnaseq") {
log2(datasetm + 1) } else { datasetm }
score <- .statScore(
TLS_Cabrita$SYMBOL, datasetm_n, nametype, "mean", "TLSSign")
return(.returnAsInput(
userdata = dataset, result = score,
SignName = "TLS_Cabrita", datasetm))
}
#' CD49fHi Basal Stem Cell Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @examples
#' data(ovse)
#' stemCellCD49fSign(dataset = ovse)
#'
#' @export
stemCellCD49fSign <- function(dataset, nametype = "SYMBOL") {
.consistencyCheck(nametype, "StemCellCD49fSign")
sign_df <- StemCellCD49f_Smith
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
datasetm <- .getMatrix(dataset)
datasetm_n <- log2(datasetm + 1)
score <- .coeffScore(sign_df, datasetm_n, "StemCellCD49fSign")
return(.returnAsInput(
userdata = dataset, result = score,
SignName = "StemCellCD49f_Smith", datasetm))
}
#' Chromosomal instability Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @examples
#' data(ovse)
#' CINSign(dataset = ovse, inputType = "rnaseq")
#'
#' @export
CINSign <- function(dataset, nametype = "SYMBOL", inputType = "microarray") {
.consistencyCheck(nametype, "CINSign")
datasetm <- .getMatrix(dataset)
datasetm_n <- if (inputType == "rnaseq") {
log2(datasetm + 1)
} else { datasetm }
datasetm_n <- scale(datasetm_n, center = TRUE, scale = FALSE)
score25 <- .statScore(
CIN_Carter$SYMBOL[CIN_Carter$class == "CIN25"],
datasetm_n, nametype, "sum", "CINSign")
score70 <- .statScore(
CIN_Carter$SYMBOL, datasetm_n, nametype, "sum", "CINSign")
score <- data.frame(score25, score70)
colnames(score) <- c("CIN_Carter_25", "CIN_Carter_70")
return(.returnAsInput(
userdata = dataset, result = t(score), SignName = "", datasetm))
}
#' Cell-cycle Signature classifier
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @examples
#' data(ovse)
#' cellCycleSign(dataset = ovse, inputType = "rnaseq")
#'
#' @export
cellCycleSign <- function(
dataset, nametype = "SYMBOL",
author = "Lundberg", inputType = "microarray") {
.consistencyCheck(nametype, "cellCycleSign", author)
datasetm <- .getMatrix(dataset)
if (author == "Lundberg") {
datasetm_n <- if (inputType == "rnaseq") {
log2(datasetm + 1)
} else { datasetm }
score <- .statScore(
CellCycle_Lundberg$SYMBOL, datasetm_n, nametype,
"sum", "cellCycleSign", author = author)
} else if (author == "Davoli") {
datasetm_n <- datasetm[
row.names(datasetm) %in% CellCycle_Davoli$SYMBOL, ]
datasetm_r <- apply(datasetm_n, 1, rank)
datasetm_r <- (datasetm_r - 1) / (nrow(datasetm_r) - 1)
score <- .statScore(
CellCycle_Davoli$SYMBOL, t(datasetm_r), nametype,
".meang", "cellCycleSign", author = author)
}
return(.returnAsInput(
userdata = dataset, result = score,
SignName = paste0("CellCycle_", author), datasetm))
}
#' Chemokine Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @importFrom stats prcomp
#'
#' @examples
#' data(ovse)
#' chemokineSign(dataset = ovse, inputType = "rnaseq")
#'
#' @export
chemokineSign <- function(
dataset, nametype = "SYMBOL", inputType = "microarray") {
.consistencyCheck(nametype, "chemokineSign")
sign_df <- Chemokines_Messina
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
datasetm <- .getMatrix(dataset)
.percentageOfGenesUsed("ChemokineSign", datasetm, sign_df$SYMBOL)
datasetm_n <- datasetm[intersect(row.names(datasetm), sign_df$SYMBOL), ]
datasetm_n <- if (inputType == "rnaseq") {
log2(datasetm_n + 1)
} else { datasetm_n }
columnNA <- .managena(datasetm_n, sign_df$SYMBOL)
score <- prcomp(t(datasetm_n), center = TRUE, scale = TRUE)$x[, 1]
score[columnNA > 0.9] <- NA
return(.returnAsInput(
userdata = dataset, result = score,
SignName = "Chemokines_Messina", datasetm))
}
#' Adult Stem Cell Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @examples
#' data(ovse)
#' ASCSign(dataset = ovse)
#'
#' @export
ASCSign <- function(dataset, nametype = "SYMBOL") {
.consistencyCheck(nametype, "ASCSign")
sign_df <- ASC_Smith
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
datasetm <- .getMatrix(dataset)
.percentageOfGenesUsed("ASCSign", datasetm, sign_df$SYMBOL)
datasetm_n <- log2(datasetm[row.names(datasetm) %in% sign_df$SYMBOL, ] + 1)
columnNA <- .managena(datasetm_n, sign_df$SYMBOL)
score <- rowSums(
scale(t(datasetm_n), center = TRUE, scale = TRUE), na.rm = TRUE)
score[columnNA > 0.9] <- NA
return(.returnAsInput(
userdata = dataset, result = score, SignName = "ASC_Smith", datasetm))
}
#' passON Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#' @param ... other arguments passed on to the \code{\link[GSVA]{gsva}}
#' function.
#'
#' @inherit EMTSign return
#'
#' @importFrom GSVA gsva
#' @importFrom SummarizedExperiment assays
#' @importFrom stats weighted.mean
#'
#' @examples
#' data(ovse)
#' PassONSign(dataset = ovse)
#'
#' @export
PassONSign <- function(
dataset, nametype = "SYMBOL", hgReference = "hg38", ...) {
.consistencyCheck(nametype, "PassONSign")
datasetm <- .getMatrix(dataset)
dataset <- .dataTransformation(
dataset, datasetm, "TPM", hgReference, nametype)
datasetm_n <- as.matrix(assays(dataset)[["TPM"]])
sign_df <- PassON_Du
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
sign_df <- sign_df[sign_df$SYMBOL %in% rownames(datasetm_n), ]
sign_list <- split(sign_df$SYMBOL, sign_df$class)
.percentageOfGenesUsed("PassONSign", datasetm_n, sign_df$SYMBOL)
dots <- list(...)
args <- .matchArguments(dots, list(
expr = datasetm_n, gset.idx.list = sign_list, method = "ssgsea",
kcdf = "Poisson", ssgsea.norm = TRUE, verbose = FALSE))
gsva_matrix <- do.call(gsva, args)
gsva_mean <- vapply(seq_len(ncol(gsva_matrix)), function(x) {
weighted.mean(gsva_matrix[, x], lengths(sign_list))},
double(1))
return(.returnAsInput(
userdata = dataset, result = gsva_mean,
SignName = "PassON_Du", datasetm))
}
#' IPRES Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#' @param ... other arguments passed on to the \code{\link[GSVA]{gsva}}
#' function.
#'
#' @inherit EMTSign return
#'
#' @importFrom GSVA gsva
#' @importFrom SummarizedExperiment assays
#'
#' @examples
#' data(ovse)
#' IPRESSign(dataset = ovse)
#'
#' @export
IPRESSign <- function(
dataset, nametype = "SYMBOL", hgReference = "hg38", ...) {
.consistencyCheck(nametype, "IPRESSign")
datasetm <- .getMatrix(dataset)
dataset <- .dataTransformation(
dataset, datasetm, "CPM", hgReference, nametype)
datasetm_n <- log2(as.matrix(assays(dataset)[["CPM"]]))
sign_df <- IPRES_Hugo
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
sign_df <- sign_df[sign_df$SYMBOL %in% rownames(datasetm_n), ]
sign_list <- split(sign_df$SYMBOL, sign_df$class)
.percentageOfGenesUsed("IPRESSign", datasetm_n, sign_df$SYMBOL)
dots <- list(...)
args <- .matchArguments(dots, list(
expr = datasetm_n, gset.idx.list = sign_list, method = "ssgsea",
kcdf = "Gaussian", ssgsea.norm = TRUE, verbose = FALSE))
gsva_matrix <- do.call(gsva, args)
score <- rowMeans(vapply(as.data.frame(t(gsva_matrix)), scale,
double(ncol(gsva_matrix))))
return(.returnAsInput(
userdata = dataset, result = score,
SignName = "IPRES_Hugo", datasetm))
}
#' CIS (carcinoma-in situ) Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @importFrom matrixStats rowMedians
#'
#' @examples
#' data(ovse)
#' CISSign(dataset = ovse)
#'
#' @export
CISSign <- function(dataset, nametype = "SYMBOL") {
## code instruction from an e-mail by Jaegil Kim:
## 'We first compute "log2(fold change) = log2(expression) - log2(median
## expression across samples)" for each gene and define the signature
## score as "mean of log2(fold change) across CIS-up genes - mean of
## log2(fold change) across CIS-down genes".'
.consistencyCheck(nametype, "CISSign")
sign_df <- CIS_Robertson
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
datasetm <- .getMatrix(dataset)
sign_up <- sign_df[grep("up", sign_df$class), ]
sign_down <- sign_df[grep("down", sign_df$class), ]
.percentageOfGenesUsed("CISSign", datasetm, sign_up$SYMBOL, "up")
.percentageOfGenesUsed("CISSign", datasetm, sign_down$SYMBOL, "down")
datasetm_n <- datasetm[intersect(row.names(datasetm), sign_df$SYMBOL), ]
datasetm_n <- log2(datasetm_n + 1) - log2(rowMedians(datasetm_n) + 1)
score_up <- colMeans(
datasetm_n[intersect(row.names(datasetm_n), sign_up$SYMBOL), ])
score_down <- colMeans(
datasetm_n[intersect(row.names(datasetm_n), sign_down$SYMBOL), ])
score <- score_up - score_down
return(.returnAsInput(
userdata = dataset, result = score,
SignName = "CIS_Robertson", datasetm))
}
#' Glycolysis Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @examples
#' data(ovse)
#' glycolysisSign(dataset = ovse)
#'
#' @export
glycolysisSign <- function(dataset, nametype = "SYMBOL", author = "Zhang") {
.consistencyCheck(nametype, "glycolysisSign", author)
sign_df <- get(paste0("Glycolysis_", author))
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
datasetm <- .getMatrix(dataset)
datasetm_n <- log2(datasetm + 1)
score <- .coeffScore(sign_df, datasetm_n, "glycolysisSign", author = author)
return(.returnAsInput(
userdata = dataset, result = score,
SignName = paste0("Glycolysis_", author), datasetm))
}
#' Autophagy Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @importFrom SummarizedExperiment assays
#'
#' @examples
#' data(ovse)
#' autophagySign(dataset = ovse)
#'
#' @export
autophagySign <- function(
dataset, nametype = "SYMBOL", author = "Xu", hgReference = "hg38") {
.consistencyCheck(nametype, "autophagySign", author)
sign_df <- get(paste0("Autophagy_", author))
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
datasetm <- .getMatrix(dataset)
if (author == "ChenM") {
dataset <- .dataTransformation(
dataset, datasetm, "FPKM", hgReference, nametype)
datasetm_n <- as.matrix(assays(dataset)[["FPKM"]])
OSscore <- .coeffScore(
sign_df[sign_df$class == "OS", ], datasetm_n,
"autophagySign", "OS", author = author)
DFSscore <- .coeffScore(
sign_df[sign_df$class == "DFS", ], datasetm_n,
"autophagySign", "DFS", author = author)
score <- data.frame(OSscore, DFSscore)
colnames(score) <- c("Autophagy_ChenM_OS", "Autophagy_ChenM_DFS")
return(.returnAsInput(
userdata = dataset, result = t(score), SignName = "", datasetm))
} else {
score <- .coeffScore(
sign_df, datasetm, "autophagySign", author = author)
return(.returnAsInput(
userdata = dataset, result = score,
SignName = paste0("Autophagy_", author), datasetm))
}
}
#' Extracellular Matrix Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#' @param ... other arguments passed on to the \code{\link[GSVA]{gsva}}
#' function.
#'
#' @inherit EMTSign return
#'
#' @importFrom GSVA gsva
#'
#' @examples
#' data(ovse)
#' ECMSign(dataset = ovse)
#'
#' @export
ECMSign <- function(dataset, nametype = "SYMBOL", ...) {
.consistencyCheck(nametype, "ECMSign")
sign_df <- ECM_Chakravarthy
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
datasetm <- .getMatrix(dataset)
sign_up <- sign_df[grep("ECMup", sign_df$class), ]
sign_down <- sign_df[grep("ECMdown", sign_df$class), ]
.percentageOfGenesUsed("ECMSign", datasetm, sign_up$SYMBOL, "up")
.percentageOfGenesUsed("ECMSign", datasetm, sign_down$SYMBOL, "down")
gene_sets <- list(
ECM_Chakravarthy_up = sign_up$SYMBOL,
ECM_Chakravarthy_down = sign_down$SYMBOL)
dots <- list(...)
args <- .matchArguments(dots, list(
expr = datasetm, gset.idx.list = gene_sets, method = "ssgsea",
kcdf = "Poisson", ssgsea.norm = FALSE, verbose = FALSE))
gsva_count <- do.call(gsva, args)
return(.returnAsInput(
userdata = dataset, result = gsva_count, SignName = "", datasetm))
}
#' Homologous Recombination Deficiency Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @importFrom matrixStats rowMedians
#' @importFrom stats t.test
#'
#' @examples
#' data(ovse)
#' HRDSSign(dataset = ovse)
#'
#' @export
HRDSSign <- function(dataset, nametype = "SYMBOL") {
.consistencyCheck(nametype, "HRDSSign")
sign_df <- HRDS_Lu
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
datasetm <- .getMatrix(dataset)
datasetm_n <- datasetm - rowMedians(datasetm)
.percentageOfGenesUsed("HRDSSign", datasetm, sign_df$SYMBOL)
HRDS_P <- datasetm_n[intersect(
row.names(datasetm_n), sign_df[sign_df$coeff == 1, ]$SYMBOL), ]
HRDS_N <- datasetm_n[intersect(
row.names(datasetm_n), sign_df[sign_df$coeff == -1, ]$SYMBOL), ]
score <- unlist(lapply(seq_len(ncol(datasetm_n)), function(x) {
tmp <- t.test(HRDS_P[, x], HRDS_N[, x], alternative = "two.sided")
tmp[["statistic"]] }))
return(.returnAsInput(
userdata = dataset, result = score, SignName = "HRDS_Lu", datasetm))
}
#' Adult Intestinal Stem Cell Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @examples
#' data(ovse)
#' ISCSign(dataset = ovse, inputType = "rnaseq")
#'
#' @export
ISCSign <- function(dataset, nametype = "SYMBOL", inputType = "microarray") {
.consistencyCheck(nametype, "ISCSign")
sign_df <- ISC_MerlosSuarez
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
sign_list <- split(sign_df$SYMBOL, sign_df$class)
datasetm <- .getMatrix(dataset)
datasetm_n <- if (inputType == "rnaseq") {
log2(datasetm + 1)
} else { datasetm }
datasetm_n <- datasetm_n - rowMeans(datasetm_n)
.percentageOfGenesUsed(
"ISCSign", datasetm_n, sign_list$`ISCEphB2`, "ISCEphB2")
.percentageOfGenesUsed(
"ISCSign", datasetm_n, sign_list$`LateTA`, "LateTA")
.percentageOfGenesUsed(
"ISCSign", datasetm_n, sign_list$`ISCLgr5`, "ISCLgr5")
.percentageOfGenesUsed(
"ISCSign", datasetm_n, sign_list$`Prolif`, "Prolif")
scores <- vapply(sign_list, function(x) {
colMeans(datasetm_n[intersect(
x, row.names(datasetm_n - colMeans(datasetm_n))), ])
}, double(ncol(datasetm_n)))
colnames(scores) <- paste0("ISC_MerlosSuarez_", colnames(scores))
return(.returnAsInput(
userdata = dataset, result = t(scores), SignName = "", datasetm))
}
#' VEGF Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @examples
#' data(ovse)
#' VEGFSign(dataset = ovse)
#'
#' @export
VEGFSign <- function(dataset, nametype = "SYMBOL") {
.consistencyCheck(nametype, "VEGFSign")
datasetm <- .getMatrix(dataset)
datasetm_n <- log2(datasetm + 1)
score <- .statScore(
VEGF_Hu$SYMBOL, datasetm_n, nametype, "mean", "VEGFSign")
return(.returnAsInput(
userdata = dataset, result = score, SignName = "VEGF_Hu", datasetm))
}
#' DNA Repair Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#'
#' @inherit EMTSign return
#'
#' @examples
#' data(ovse)
#' DNArepSign(dataset = ovse, inputType = "rnaseq")
#'
#' @export
DNArepSign <- function(
dataset, nametype = "SYMBOL", inputType = "microarray") {
.consistencyCheck(nametype, "DNArepSign")
sign_df <- DNArep_Kang
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
datasetm <- .getMatrix(dataset)
.percentageOfGenesUsed("DNArepSign", datasetm, sign_df$SYMBOL)
datasetm_n <- if (inputType == "rnaseq") {
log2(datasetm + 1)
} else { datasetm }
datasetm_n <- datasetm_n[row.names(datasetm_n) %in% sign_df$SYMBOL, ]
datasetm_n <- scale(t(datasetm_n), center = TRUE, scale = FALSE)
medianexp <- apply(datasetm_n, 2, median)
genes_h <- intersect(
colnames(datasetm_n), sign_df[sign_df$class == "high", ]$SYMBOL)
genes_l <- intersect(
colnames(datasetm_n), sign_df[sign_df$class == "low", ]$SYMBOL)
score <- rowSums(datasetm_n[, genes_h] > medianexp[genes_h]) +
rowSums(datasetm_n[, genes_l] < medianexp[genes_l])
return(.returnAsInput(
userdata = dataset, result = score, SignName = "DNArep_Kang", datasetm))
}
#' IPSOV Signature
#'
#' @inherit EMTSign description
#' @inheritParams pyroptosisSign
#' @param ... other arguments passed on to the \code{\link[GSVA]{gsva}}
#' function.
#'
#' @inherit EMTSign return
#'
#' @examples
#' data(ovse)
#' IPSOVSign(dataset = ovse)
#'
#' @export
IPSOVSign <- function(
dataset, nametype = "SYMBOL", inputType = "microarray", ...) {
.consistencyCheck(nametype, "IPSOVSign")
datasetm <- .getMatrix(dataset)
datasetm_n <- if (inputType == "rnaseq") { log2(datasetm + 1)
} else { datasetm }
datasetm_n <- scale(datasetm, center = TRUE, scale = TRUE)
sign_df <- IPSOV_Shen
sign_df$SYMBOL <- .geneIDtrans(nametype, sign_df$SYMBOL)
sign_df <- sign_df[sign_df$SYMBOL %in% rownames(datasetm),]
sign_list <- split(sign_df$SYMBOL, sign_df$class)
.percentageOfGenesUsed("IPSOVSign", datasetm, sign_df$SYMBOL)
dots <- list(...)
args <- .matchArguments(dots, list(
expr = datasetm_n, gset.idx.list = sign_list,
method = "ssgsea", kcdf = "Gaussian",
ssgsea.norm = FALSE, verbose = FALSE))
gsva_matrix <- do.call(gsva, args)
sign_class <- unique(sign_df[,2:3])
sign_class <- sign_class[sign_class$class %in% row.names(gsva_matrix), ]
columnNA <- .managena(datasetm = gsva_matrix, genes = sign_class$class)
score <- colSums(
gsva_matrix[sign_class$class, ] * sign_class$coeff, na.rm = TRUE)
score[columnNA > 0.9] <- NA
return(.returnAsInput(
userdata = dataset, result = score, SignName = "IPSOV_Shen", datasetm))
return(dataset)
}
