TCGAbiolinks下载CNV数据

下载Gene水平的数据

library(TCGAbiolinks)
query <- GDCquery(
    project = "TCGA-PRAD",
    data.category = "Copy Number Variation",
    data.type = "Gene Level Copy Number",              
    access = "open"
)
GDCdownload(query)
data <- GDCprepare(query)
saveRDS(data, 'prad_cnv.rds')

下载Masked数据

query <- GDCquery(
    project = "TCGA-PRAD",
    data.category = "Copy Number Variation",
    data.type = "Masked Copy Number Segment",              
    access = "open"
)
GDCdownload(query)
data <- GDCprepare(query)
saveRDS(data, 'prad_cnv_masked.rds')

清洗数据

初步清洗

library(SummarizedExperiment)
data <- readRDS('prad_cnv.rds')
cnT <- data@assays@data$copy_number
cnTcol <- colnames(cnT)
type <- as.numeric(substr(cnTcol, 14, 15))
cnT <- cnT[, type<10]
colnames(cnT) <- substr(cnTcol,1, 12)
rownames(cnT) <-  data@rowRanges$gene_name
cnT <- na.omit(cnT)

精细清洗

f_dedup_IQR

cnT <- f_dedup_IQR(cnT, rownames(cnT))
cnT <- cnT[,f_rm_duplicated(colnames(cnT))]
group <- readRDS('../idea_2/fig3.2/fig5/tcga.predict.rds')
cnT <- cnT[,colnames(cnT) %in% rownames(group)]

构造cnTable

慢速

df <- NULL
for (i in 1:ncol(cnT)){
    colnames(cnT)[[i]]
    tmp_df <- data.frame(Hugo_Symbol = rownames(cnT), Tumor_Sample_Barcode = colnames(cnT)[[i]], Variant_Classification=cnT[,i])
    tmp_df <- subset(tmp_df, Variant_Classification != 2)
    df <- rbind(df, tmp_df)
}

快速

library(reshape2)
df <- melt(cnT)
colnames(df) = c('Hugo_Symbol', 'Tumor_Sample_Barcode', 'Variant_Classification')
df <- subset(df, Variant_Classification != 2)
df

贴标签

df$Variant_Classification[df$Variant_Classification > 2]  <- 'Amp'
df$Variant_Classification[df$Variant_Classification < 2]  <- 'Del'
table(df$Variant_Classification)

分组别

df_l <- subset(df, Tumor_Sample_Barcode %in% rownames(group)[group$group == 'Low Risk'])
df_h <- subset(df, Tumor_Sample_Barcode %in% rownames(group)[group$group == 'High Risk'])
saveRDS(df_l, 'cnT.l.rds')
saveRDS(df_h, 'cnT.h.rds')

导入maftools

TCGAbiolinks下载maf数据

清洗数据

cnv_l <- readRDS('cnT.l.rds')
cnv_h <- readRDS('cnT.h.rds')
prad_l$Tumor_Sample_Barcode <- prad_l$BarCode
prad_l <- subset(prad_l, Tumor_Sample_Barcode %in% cnv_l$Tumor_Sample_Barcode)
cnv_l <- subset(cnv_l, Tumor_Sample_Barcode %in% prad_l$Tumor_Sample_Barcode)
prad_h$Tumor_Sample_Barcode <- prad_h$BarCode
prad_h <- subset(prad_h, Tumor_Sample_Barcode %in% cnv_h$Tumor_Sample_Barcode)
cnv_h <- subset(cnv_h, Tumor_Sample_Barcode %in% prad_h$Tumor_Sample_Barcode)

读入maftools

maf_l <- read.maf(prad_l, cnTable = cnv_l)
maf_h <- read.maf(prad_h, cnTable = cnv_h)

绘制瀑布图

options(repr.plot.width=12, repr.plot.height=8)
genes <- subset(lvsh$results, pval < 0.05)$Hugo_Symbol
coOncoplot(m1=maf_l, m2=maf_h, m1Name="Low Risk", m2Name="High Risk",genes=genes)