R/pdat_prostate.R

In canprot: Chemical Metrics of Differentially Expressed Proteins

# canprot/R/pdat_prostate.R
# retrieve protein IDs for prostate cancer studies
# 20160420-20191230 assemble data for 2020 compilation

pdat_prostate <- function(dataset = 2020) {
  if(identical(dataset, 2020)) {
    return(c(
             "GTR+08",
             "KPB+10_localized",
             "HZH+12_Protein",
             "JHZ+13",
             "LCS+14",
             "CZL+16", "IWT+16",
             "GLZ+18_acinar", "GLZ+18_ductal", "LAJ+18_PC", "LAJ+18_CRPC", "MAN+18",
             "KRN+19_G1", "KRN+19_G2", "KRN+19_G3", "KRN+19_G4", "KRN+19_G5",
             "MMF+20_GS6", "MMF+20", "TOT+19", "ZYW+19_LG", "ZYW+19_HG",
             "KHN+20", "LDM+20", "SHC+20", "ZKL+20=mouse", "ZZX+20_CiRT"
             ))
  }
  # remove tags
  dataset <- strsplit(dataset, "=")[[1]][1]
  # get study and stage/condition
  study <- strsplit(dataset, "_")[[1]][1]
  stage <- paste(strsplit(dataset, "_")[[1]][-1], collapse="_")
  extdatadir <- system.file("extdata", package="canprot")
  datadir <- paste0(extdatadir, "/expression/prostate/")
  if(study=="KPB+10") {
    # 20160420 localized and metastatic prostate cancer, Khan et al., 2010
    # KPB+10_localized, KPB+10_metastatic
    dat <- read.csv(paste0(datadir, "KPB+10.csv.xz"), as.is=TRUE)
    if(stage=="localized") description <- "PCa / adjacent benign"
    if(stage=="metastatic") description <- "metastatic / localized"
    #print(paste0("pdat_metastasis: ", description, " [", dataset, "]"))
    # choose localized or metastatic
    dat <- dat[dat[, stage] != "", ]
    up2 <- dat[, stage] == "up"
    dat <- cleanup(dat, "Entry", up2)
    pcomp <- protcomp(dat$Entry)
    names <- dat$Gene.Symbol
  } else if(study=="IWT+16") {
    # 20170115 prostate cancer, Iglesias-Gato et al., 2016
    dat <- read.csv(paste0(datadir, "IWT+16.csv.xz"), as.is=TRUE)
    description <- "FFPE tumor / adjacent benign"
    dat <- check_IDs(dat, "Protein.IDs")
    pcomp <- protcomp(dat$Protein.IDs)
    up2 <- dat$Tmean.Cmean > 0
  } else if(study=="MAN+18") {
    # 20190319 PCa / BPH, Martiny et al., 2019
    dat <- read.csv(paste0(datadir, "MAN+18.csv.xz"), as.is=TRUE)
    description <- "PCa / BPH"
    dat <- check_IDs(dat, "ID.Protein")
    pcomp <- protcomp(dat$ID.Protein)
    up2 <- dat$Regulation == "Up"
  } else if(study=="LCS+14") {
    # 20190320 prostate glycoproteins, Liu et al., 2014
    dat <- read.csv(paste0(datadir, "LCS+14.csv.xz"), as.is=TRUE)
    description <- "glycoproteins PCa / normal"
    pcomp <- protcomp(dat$Uniprot)
    up2 <- dat$Cancer..Normal > 1
  } else if(study=="TOT+19") {
    # 20190321 tissue cancer / healthy
    dat <- read.csv(paste0(datadir, "TOT+19.csv.xz"), as.is=TRUE)
    description <- "FFPE TMA PCa / normal"
    pcomp <- protcomp(dat$Entry)
    up2 <- dat$Change == "increase"
  } else if(study=="LAJ+18") {
    # 20191202 PC / BPH and CRPC / BPH, Latonen et al., 2018
    # LAJ+18_PC, LAJ+18_CRPC
    dat <- read.csv(paste0(datadir, "LAJ+18.csv.xz"), as.is=TRUE)
    description <- paste(stage, "/ BPH")
    # get at least 2-fold differentially expressed proteins
    icol <- match(stage, colnames(dat))
    idiff <- dat[, icol] / dat$BPH > 2 | dat[, icol] / dat$BPH < 0.5
    dat <- dat[idiff, ]
    dat <- check_IDs(dat, "Protein")
    up2 <- dat[, icol] / dat$BPH > 2
    dat <- cleanup(dat, "Protein", up2)
    pcomp <- protcomp(dat$Protein)
  } else if(study=="GTR+08") {
    # 20191202 PCa / BPH, Garbis et al., 2008
    dat <- read.csv(paste0(datadir, "GTR+08.csv.xz"), as.is=TRUE)
    description <- "PCa / BPH"
    dat <- check_IDs(dat, "primary.accession.no.")
    pcomp <- protcomp(dat$primary.accession.no.)
    up2 <- dat$mean.ratio > 1
  } else if(study=="JHZ+13") {
    # 20191202 PCa / benign, Jiang et al., 2013
    dat <- read.csv(paste0(datadir, "JHZ+13.csv.xz"), as.is=TRUE)
    description <- "PCa / adjacent benign"
    up2 <- dat$Ratio.fold. > 0
    dat <- cleanup(dat, "Protein_ID", up2)
    pcomp <- protcomp(dat$Protein_ID)
  } else if(study=="HZH+12") {
    # 20191202 PCa / benign (proteome and transcriptome), Han et al., 2012
    # HZH+12_Gene, HZH+12_Protein
    dat <- read.csv(paste0(datadir, "HZH+12.csv.xz"), as.is=TRUE)
    description <- paste("PCa / adjacent benign", stage)
    icol <- grep(paste0(stage, "_Fold_Change"), colnames(dat))
    dat <- dat[!is.na(dat[, icol]), ]
    dat <- check_IDs(dat, "Entry")
    up2 <- dat[, icol]  > 0
    dat <- cleanup(dat, "Entry", up2)
    pcomp <- protcomp(dat$Entry)
  } else if(study=="ZYW+19") {
    # 20191202 Low-grade or high-grade vs normal, Zhou et al., 2019
    # ZYW+19_LG, ZYW+19_HG
    dat <- read.csv(paste0(datadir, "ZYW+19.csv.xz"), as.is=TRUE)
    description <- paste("OCT", stage, "PCa / adjacent normal")
    icol <- grep(stage, colnames(dat))
    dat <- dat[dat[, icol] != "", ]
    dat <- check_IDs(dat, "Majority.protein.IDs")
    up2 <- dat[, icol] == "Up"
    dat <- cleanup(dat, "Majority.protein.IDs", up2)
    pcomp <- protcomp(dat$Majority.protein.IDs)
  } else if(study=="MMF+20") {
    # 20191202 PCa / benign (GS=6 or GS=6 and GS>=8), Mantsiou et al., 2019
    # MMF+20_GS6, MMF+20
    dat <- read.csv(paste0(datadir, "MMF+20.csv.xz"), as.is=TRUE)
    if(stage=="GS6") description <- "FFPE PCa / adjacent benign GS=6"
    else description <- "FFPE PCa / adjacent benign GS=6 and GS>=8"
    if(stage == "GS6") icol <- grep("ratio.GS6.cancer.Vs.GS6.benign", colnames(dat)) else icol <- grep("ratio.cancer.Vs.benign", colnames(dat))
    dat <- dat[!is.na(dat[, icol]), ]
    up2 <- dat[, icol] > 1
    pcomp <- protcomp(dat$Accession)
  } else if(study=="KRN+19") {
    # 20191212 cancer grades vs benign prostatic hyperplasia, Kawahara et al., 2019
    # KRN+19_G1, KRN+19_G2, KRN+19_G3, KRN+19_G4, KRN+19_G5
    dat <- read.csv(paste0(datadir, "KRN+19.csv.xz"), as.is=TRUE)
    description <- paste("PCa", stage, "/ BPH")
    icol <- grep(stage, colnames(dat))
    dat <- dat[!is.na(dat[, icol]), ]
    # ratios are given as BPH / GX
    up2 <- dat[, icol] == "Down"
    pcomp <- protcomp(dat$Majority.protein.IDs)
  } else if(study=="CZL+16") {
    # 20191230 prostate cancer bioinformatics analysis, Chen et al., 2016
    dat <- read.csv(paste0(datadir, "CZL+16.csv.xz"), as.is=TRUE)
    description <- "protein expression bioinformatics"
    dat <- check_IDs(dat, "Entry")
    up2 <- dat$Differentially.expressed == "Up-regulation"
    pcomp <- protcomp(dat$Entry)
  } else if(study=="GLZ+18") {
    # 20191230 prostate acinar and ductal adenocarcinoma, Guo et al., 2018
    # GLZ+18_acinar, GLZ+18_ductal
    dat <- read.csv(paste0(datadir, "GLZ+18.csv.xz"), as.is=TRUE)
    description <- paste(stage, "PCa / matched BPH")
    icol <- grep(stage, colnames(dat))
    # kep highly differential proteins
    dat <- dat[dat[, icol] > 1.5 | dat[, icol] < 2/3, ]
    up2 <- dat[, icol] > 1.5
    pcomp <- protcomp(dat$Entry)
  } else if(study=="SHC+20") {
    # 20200329 prostate cancer, Sun et al., 2020
    dat <- read.csv(paste0(datadir, "SHC+20.csv.xz"), as.is=TRUE)
    description <- "FFPE PCa / BPH"
    # keep proteins that have more than 2-fold change
    dat <- dat[abs(dat$logFC) > log10(2), ]
    up2 <- dat$logFC > 0
    dat <- check_IDs(dat, "UniprotID")
    pcomp <- protcomp(dat$UniprotID)
  } else if(study=="KHN+20") {
    # 20200404 PCa / control, Kwon et al., 2020
    dat <- read.csv(paste0(datadir, "KHN+20.csv.xz"), as.is=TRUE)
    description <- "PCa / control"
    up2 <- up2 <- dat$log2FC > 1
    dat <- check_IDs(dat, "Uniprot.ID")
    pcomp <- protcomp(dat$Uniprot.ID)
  } else if(study=="ZZX+20") {
    # 20200420 Zhu et al., 2020
    # ZZX+20_CiRT, ZZX+20_SiRT
    dat <- read.csv(paste0(datadir, "ZZX+20.csv.xz"), as.is=TRUE)
    description <- paste("tumor / normal", stage)
    icol <- grep(stage, colnames(dat))
    dat <- dat[!is.na(dat[, icol]), ]
    dat <- check_IDs(dat, "Protein")
    up2 <- dat[, icol] > 0
    pcomp <- protcomp(dat$Protein)
  } else if(study=="LDM+20") {
    # 20201016 PCa / BPH, Latosinska et al., 2020
    dat <- read.csv(paste0(datadir, "LDM+20.csv.xz"), as.is=TRUE)
    description <- "PCa / BPH"
    up2 <- dat$Fold.change..PCa.BPH. >= 1
    pcomp <- protcomp(dat$UniprotID)
  } else if(study=="ZKL+20") {
    # 20201016 mouse Pten gene-knockout prostate cancer, Zhang et al., 2020
    dat <- read.csv(paste0(datadir, "ZKL+20.csv.xz"), as.is=TRUE)
    description <- "mouse Pten-KO / WT"
    up2 <- dat$Average > 1
    dat <- cleanup(dat, "Accession", up2)
    pcomp <- protcomp(dat$Accession, aa_file = paste0(extdatadir, "/aa/mouse/ZKL+20_aa.csv.xz"))
  } else stop(paste("prostate dataset", dataset, "not available"))
  print(paste0("pdat_prostate: ", description, " [", dataset, "]"))
  # use the up2 from the cleaned-up data, if it exists 20190429
  if("up2" %in% colnames(dat)) up2 <- dat$up2
  return(list(dataset=dataset, pcomp=pcomp, up2=up2, names=names, description=description))
}