R/Get_Percentplot.R

In Lian-Lab/LIANLAB: To integrate analysis methods and simplify analysis process

#' Get percentage of seurat.obj, which has more than one samples
#'
#' @param seurat_object seurat.obj, including cols could split
#' @param filename the name of the generated file
#' @param order the order of the samples in figure
#' @param split_by split by the seurat.obj by cols, default 'samples'
#' @param colorm the colors of the clusters
#'
#' @importFrom reshape2 melt
#' @importFrom ggplot2 theme_bw
#' @importFrom dplyr group_by mutate
#' @export
#'
#' @examples
#' \dontrun{
#' input.file <- system.file('extdata','pbmc_1k.RDS',package = 'LIANLABDATA')
#' pbmc_1k <- readRDS(input.file)
#' Get_percentplot_split(pbmc_1k)
#' }
Get_percentplot_split = function(seurat_object,filename=NULL,order='increasing',split_by=NULL,colorm=NULL){

  variable <- count <- value <- Sample <- Percent <- Cluster <- NULL

  num <- table(seurat_object@active.ident)
  if (is.null(split_by)) {
    split_by <- 'samples'
    sample_name=unique(seurat_object@meta.data$samples)
  }else{
    sample_name=unique(seurat_object@meta.data[,split_by])
  }
  frequency_matrix <- data.frame(num)
  colnames(frequency_matrix)=c('type','num')

  for (i in sample_name) {
    cell_name=row.names(seurat_object@meta.data[seurat_object@meta.data[,split_by]==i,])
    subset_data=subset(seurat_object, cells = cell_name)
    frequency=table(subset_data@active.ident)/ncol(subset_data)
    frequency1 = data.frame(frequency)
    colnames(frequency1)=c('type',i)
    frequency_matrix = merge(frequency_matrix,frequency1,by='type',all=T)
  }

  frequency_matrix[is.na(frequency_matrix)] <- 0

  o = order(frequency_matrix$type,decreasing = F)
  frequency_matrix = frequency_matrix[o,]

  rownames(frequency_matrix) = frequency_matrix$type
  colnames(frequency_matrix)[1] = 'cluster'
  frequency_matrix = frequency_matrix[,-2]

  frequency_matrix[] <- lapply(frequency_matrix, as.character)
  frequency_matrix[is.na(frequency_matrix)] <- 0
  frequency_matrix[,2:ncol(frequency_matrix)] <- as.data.frame(lapply(frequency_matrix[,2:ncol(frequency_matrix)],as.numeric))
  frequency_matrix$cluster = as.character(frequency_matrix$cluster)

  if(order=='increasing'){
    o = order(colnames(frequency_matrix),decreasing = F)
  }else{
    o = order(colnames(frequency_matrix),decreasing = T)
  }
  frequency_matrix = frequency_matrix[,o]
  write.csv(frequency_matrix,paste0(filename,"_frequency_matrix.csv"),row.names = T)

  data_rownames <- rownames(frequency_matrix)

  frequency_matrix_m <- melt(frequency_matrix, id.vars=c("cluster"))
  frequency_matrix_m <- frequency_matrix_m %>% group_by(variable) %>%
    mutate(count=sum(value)) %>% mutate(freq=round(100*value/count,2))
  colnames(frequency_matrix_m)=c("Cluster","Sample","Frequency","count","Percent")

  if (is.null(colorm)) {
    colorful <- LIANLAB::colorful
    colors = colorful[["colors"]]
    colorm <- data.frame(frequency_matrix$cluster,colors[1:length(rownames(frequency_matrix))])
    row.names(colorm) <- colorm[,1]
  }

  p <- ggplot(frequency_matrix_m, aes(x=Sample, y=Percent, group=Cluster)) +
    geom_bar(stat="identity", position="fill", aes(fill=Cluster))+
    scale_fill_manual(values=colorm[,2],breaks=colorm[,1])+ theme_bw()+
    theme(panel.border = element_blank(),panel.grid.major = element_blank(),
          panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),
          axis.text.x = element_text( vjust = 0.5, hjust = 0.5, angle = 45))

  pdf(paste0("percent_plot_",filename,".pdf"),width =ncol(frequency_matrix)*0.8+2, height =nrow(frequency_matrix)*0.25+1)
  print(p)
  dev.off()
}


#' Get percentage of seurat.obj
#'
#' @param seurat_object seurat.obj, including cols could split
#' @param filename the name of the generated file
#' @param colorm the colors of the clusters
#' @importFrom ggplot2 xlab
#' @export
#'
#' @examples
#' \dontrun{
#' input.file <- system.file('extdata','pbmc_1k.RDS',package = 'LIANLABDATA')
#' pbmc_1k <- readRDS(input.file)
#' Get_percentplot(pbmc_1k)
#' }
Get_percentplot = function(seurat_object,filename,colorm=NULL){

  Sample <- Percent <- Cluster <- NULL
  num <- table(seurat_object@active.ident)
  frequency_matrix=data.frame(num)
  frequency_matrix = frequency_matrix[which(frequency_matrix$Freq!=0),]

  frequency=table(seurat_object@active.ident)/ncol(seurat_object)
  frequency=as.data.frame(frequency)
  frequency=frequency[,-1]
  frequency_matrix=cbind(frequency_matrix,frequency)


  row.names(frequency_matrix)=frequency_matrix[,1]
  colnames(frequency_matrix)[1] = 'cluster'
  colnames(frequency_matrix)[2] = 'count'
  add = data.frame(rep('Sample',nrow(frequency_matrix)))
  colnames(add)[1] = 'Sample'
  frequency_matrix = cbind(frequency_matrix,add)

  write.csv(frequency_matrix,paste0(filename,"_frequency_matrix.csv"),row.names = F)

  data_rownames <- rownames(frequency_matrix)
  frequency_matrix$Percent <- frequency_matrix$frequency*100

  if (is.null(colorm)) {
    colorful <- LIANLAB::colorful
    colors = colorful[["colors"]]
    colorm <- data.frame(frequency_matrix$cluster,colors[1:length(rownames(frequency_matrix))])
    row.names(colorm) <- colorm[,1]
  }

    color_percent=as.character(colorm[,2])

  p <- ggplot(frequency_matrix, aes(x = Sample, y=Percent, group=cluster)) +
    geom_bar(stat="identity", position="fill", aes(fill=Cluster)) + xlab('') +
    scale_fill_manual(values=color_percent,breaks=data_rownames) + theme_bw() +
    theme(panel.border = element_blank(),panel.grid.major = element_blank(),
          panel.grid.minor = element_blank(),axis.line = element_line(colour = "black"),
          axis.text.x = element_text( vjust = 0.5, hjust = 0.5, angle = 45))

  pdf(paste0("percent_plot_",filename,".pdf"),width =2, height = 6)
  print(p)
  dev.off()
}