R/plotCNA.R

In MesKit: A tool kit for dissecting cancer evolution from multi-region derived tumor biopsies via somatic alterations

#' @title plotCNA
#'  
#' @param seg Object generated by \code{\link{readSegment}} function.
#' @param patient.id Select the specific patients. Default NULL, all patients are included.
#' @param sampleOrder A named list which contains the sample order used in plotting the final profile. Default NULL.
#' @param chrSilent Chromosomes excluded in the analysis. e.g, 1, 2, 3. Default NULL.
#' @param refBuild Human reference genome versions of hg18, hg19 or hg38 by UCSC. Default "hg19".
#' @param sample.text.size Fontsize of sample name. Default 11.
#' @param chrom.text.size Fontsize of chromosome text. Default 3.
#' @param legend.text.size Fontsize of legend text. Default 9.
#' @param legend.title.size Fontsize of legend title. Default 11.
#' @param annot.text.size Fontsize of cytoband or gene symbols. Default 3.
#' @param sample.bar.height Bar height of each sample. Default 0.5.
#' @param chrom.bar.height Bar height of each chromosome. Default 0.5.
#' @param showRownames Logical (Default: TRUE). Show sample names of rows.
#' @param removeEmptyChr Remove empty chromosomes that do not exist in all samples. Default TRUE. 
#' @param showCytoband Logical (Default: FALSE). Show cytobands on the plot. Only when the seg object is created with GISTIC results, this parameter can be TRUE.
#' @param showGene Logical (Default: FALSE). Show gene symbols on the plot. Only when the seg object is created with txdb, this parameter can be TRUE.
#' @param use.tumorSampleLabel Logical (Default: FALSE). Rename the 'Tumor_Sample_Barcode' with 'Tumor_Sample_Label'.
#' 
#' @examples
#' segFile <- system.file("extdata", "CRC_HZ.seg.txt", package = "MesKit")
#' seg <- readSegment(segFile = segFile)
#' plotCNA(seg)
#' 
#' ## showCytoband
#' gisticAmpGenesFile <- system.file("extdata", "COREAD_amp_genes.conf_99.txt", package = "MesKit")
#' gisticDelGenesFile <- system.file("extdata", "COREAD_del_genes.conf_99.txt", package = "MesKit")
#' gisticAllLesionsFile <- system.file("extdata", "COREAD_all_lesions.conf_99.txt", package = "MesKit")
#' seg <- readSegment(segFile = segFile,
#'                    gisticAmpGenesFile = gisticAmpGenesFile,
#'                     gisticDelGenesFile = gisticDelGenesFile,
#'                    gisticAllLesionsFile = gisticAllLesionsFile)
#' plotCNA(seg, showCytoband = TRUE)
#'
#' 
#' @return a heatmap plot of CNA profile
#' @import cowplot RColorBrewer
#' @importFrom Biostrings start end
#' @export plotCNA
#'


plotCNA <- function(seg,
                    patient.id = NULL,
                    sampleOrder = NULL,
                    chrSilent = NULL,
                    refBuild = "hg19",
                    sample.text.size = 11,
                    chrom.text.size = 3,
                    legend.text.size = 9,
                    legend.title.size = 11,
                    annot.text.size = 3,
                    sample.bar.height = 0.5,
                    chrom.bar.height = 0.5,
                    showRownames = TRUE,
                    removeEmptyChr = TRUE,
                    showCytoband = FALSE,
                    showGene = FALSE,
                    use.tumorSampleLabel = FALSE
){
    
    ## combine data frame
    if(is(seg, "list")){
        seg <- dplyr::bind_rows(seg) %>% as.data.table()
    }
    
    if(use.tumorSampleLabel){
        if(!"Tumor_Sample_Label" %in% colnames(seg)){
            stop("Tumor_Sample_Label was not found. Please check seg file or let use.tumorSampleLabel be 'FALSE'")
        }
        seg <- seg %>% 
            dplyr::mutate(Tumor_Sample_Barcode = .data$Tumor_Sample_Label)
    }
    
    if(!is.null(chrSilent)){    
        chrSilent <- gsub(pattern = 'chr', replacement = '', x = chrSilent, fixed = TRUE)
        seg <- seg[!seg$Chromosome %in% chrSilent]        
        chrSilent <- gsub(pattern = 'X', replacement = '23', x = chrSilent, fixed = TRUE)
        chrSilent <- gsub(pattern = 'Y', replacement = '24', x = chrSilent, fixed = TRUE)
    }
    
    

    seg$Chromosome = gsub(pattern = 'X', replacement = '23', x = seg$Chromosome, fixed = TRUE)
    seg$Chromosome = gsub(pattern = 'Y', replacement = '24', x = seg$Chromosome, fixed = TRUE)
    ## select patients in patient.id 
    if(is.null(patient.id)|length(patient.id) > 1){
        if(length(patient.id) > 1){
            patient.setdiff <- setdiff(patient.id, unique(seg$Patient_ID))
            if(length(patient.setdiff) > 0){
                stop(paste0(patient.setdiff, " cannot be found in your data"))
            }
            seg <- seg[seg$Patient_ID %in% patient.id, ]
            seg$patient <- factor(seg$Patient_ID, levels = rev(patient.id))
        }else{
            seg$patient <- seg$Patient_ID
        }
        patient_num <- length(unique(seg$Patient_ID)) 
        seg <- dplyr::select(seg, -"Patient_ID")
        seg$Patient_ID <- "ALL"
    }else if(length(patient.id) == 1){
        patient.setdiff <- setdiff(patient.id, unique(seg$Patient_ID))
        if(length(patient.setdiff) > 0){
            stop(paste0(patient.setdiff, " cannot be found in your data"))
        }
        patient_num <- 1
        seg <- seg[seg$Patient_ID %in% patient.id, ]
        seg$patient <- seg$Patient_ID
        seg <- dplyr::select(seg, -"Patient_ID")
    }
    
    
    
    # if(show.GISTIC.gene){
    #     if(!"Gistic_type" %in% colnames(seg)){
    #         stop("Error: GISTIC genes information were not found. Please check readSegment")
    #     }
    # }
    # if("Start" %in% colnames(seg)){
    #   seg <- dplyr::rename(seg,Start_Position = Start, End_Position = End)
    # }
    ref.options <- c("hg19","hg18","hg38")
    if(!refBuild %in% ref.options){
        stop("refBuild should be either 'hg18','hg19' or 'hg38'.")
    }
    if(refBuild == 'hg19'){
        chrLens = c(249250621, 243199373, 198022430, 191154276, 180915260, 171115067, 159138663,
                    146364022, 141213431, 135534747, 135006516, 133851895, 115169878, 107349540,
                    102531392, 90354753, 81195210, 78077248, 59128983, 63025520, 48129895, 51304566,
                    155270560, 59373566)
    } else if(refBuild == 'hg18'){
        chrLens = c(247249719, 242951149, 199501827, 191273063, 180857866, 170899992,
                    158821424, 146274826, 140273252, 135374737, 134452384, 132349534,
                    114142980, 106368585, 100338915, 88827254, 78774742, 76117153,
                    63811651, 62435964, 46944323, 49691432, 154913754, 57772954)
    } else if(refBuild == 'hg38'){ 
        chrLens = c(248956422, 242193529, 198295559, 190214555, 181538259, 170805979,
                    159345973, 145138636, 138394717, 133797422, 135086622, 133275309,
                    114364328, 107043718, 101991189, 90338345, 83257441, 80373285,
                    58617616, 64444167, 46709983, 50818468, 156040895, 57227415)
    }
    
    chrLabels <- seq_len(24)
    
    chrTable <- data.table::data.table(chr = as.character(chrLabels),
                                       start = as.numeric(chrLens),
                                       end = as.numeric(chrLens))
    if(removeEmptyChr){
        chr_seg <- unique(as.character(seg$Chromosome)) 
        chrTable <- dplyr::filter(chrTable, .data$chr %in% chr_seg)
    }
    chrTable <- chrTable %>% 
        dplyr::filter(!.data$chr %in% chrSilent)
    seg <- seg %>% 
        dplyr::filter(!.data$Chromosome %in% chrSilent)
    
    
    chrLens <- append(cumsum(chrTable$start),1,after = 0)
    chrTable <- chrTable %>% 
        dplyr::mutate(
            start = chrLens[seq_len(length(chrLens)-1)],
            end = chrLens[2:length(chrLens)]
        )
    
    chrLabels <- chrTable$chr
    
    # chr_bar_color <- c()
    # chr_text_color <- c()
    
    chr_bar_color <- lapply(seq_len(length(as.numeric(chrLabels) )), function(i){
        if((i %% 2) == 0){
            return("white")
        }else{
            return("black")
        }
    }) %>% unlist()
    
    chr_text_color <- lapply(seq_len(length(as.numeric(chrLabels))), function(i){
        if((i %% 2) == 0){
            return("black")
        }else{
            return("white")
        }
    }) %>% unlist()
    
    
    chrTable$color = chr_bar_color
    chr_start_pos <- as.numeric(chrTable$start) 
    names(chr_start_pos) <- paste0("chr",chrTable$chr)

    seg <- seg %>% 
        dplyr::mutate(
            Update_Start = as.numeric(.data$Start_Position) + chr_start_pos[paste0("chr",.data$Chromosome)],
            Update_End = as.numeric(.data$End_Position) + chr_start_pos[paste0("chr",.data$Chromosome)]
        )
    
    
    seg$hmin <- 0
    seg$hmax <- 0
    h <- chrom.bar.height
    
    if("patient" %in% colnames(seg)& length(unique(seg$patient)) > 1){
        
        ## sort sampleid 
        seg <- seg %>% 
            dplyr::arrange(dplyr::desc(.data$patient) ,
                           dplyr::desc(.data$Tumor_Sample_Barcode),
                           .data$Chromosome,
                           .data$Start_Position) %>%
            as.data.table()
        
        seg$Sample_ID <- paste(seg$patient, seg$Tumor_Sample_Barcode,sep = "&")
        sampleids <- unique(seg$Sample_ID)
        ## reorder sample by sampleOrder
        if(!is.null(sampleOrder)){
            patient.setdiff <- setdiff(names(sampleOrder), unique(seg$patient))
            if(length(patient.setdiff) > 0){
                stop(paste0(patient.setdiff,
                            " cannot be found in your data. Please check sampleOrder!"))
            }
            
            slist <- lapply(unique(seg$patient) ,function(p){
                o1 <- sampleOrder[[p]]
                if(is.null(o1)){
                    return(sampleids[grepl(p, sampleids)])
                }
                o2 <- unique(seg[seg$patient == p]$Tumor_Sample_Barcode)
                s1 <- paste(p, o1, sep = "&") 
                sample.setdiff <- setdiff(o1, o2)
                if(length(sample.setdiff) > 0){
                    stop(paste0(paste(sample.setdiff, collapse = ","),
                                " cannot be found in ", p, ". Please check sampleOrder!"))
                }else{
                    return(s1)
                }
            }) %>% unlist() 
            sampleids <- slist
        }
        
        
        patient.rect.hmin <- c(h)
        patient.rect.hmax <- c()
        sample_patient_list <- (seg %>% dplyr::distinct(.data$Sample_ID, .keep_all = TRUE))$patient
        
        hlist1 <- lapply(seq_len(length(sampleids)), function(i){
            p <- sample_patient_list[i]
            h1 <- h + (sample.bar.height + sample.bar.height/10) * (i-1)
            p_idx <- which(unique(seg$patient) == p)
            h1 <- h1 + sample.bar.height*2/5 * (p_idx - 1)
            return(h1)
        }) %>% unlist()
        seg_sample_list <- lapply(seq_len(length(sampleids)), function(i){
            h1 <- hlist1[i]
            sampleid <- sampleids[i]
            dat <- seg[seg$Sample_ID == sampleid,]
            dat$hmin <- h1
            dat$hmax <- h1 + sample.bar.height
            return(dat)
        })
        seg <- dplyr::bind_rows(seg_sample_list)
        
        patient.rect.hmin <- lapply(unique(seg$patient), function(p){
            return(min(seg[seg$patient == p,]$hmin))
        }) %>% unlist()
        
        patient.rect.hmax <- lapply(unique(seg$patient), function(p){
            return(max(seg[seg$patient == p,]$hmax))
        }) %>% unlist()
        
        patient.rect.table <- data.frame(hmin = patient.rect.hmin,
                                         hmax = patient.rect.hmax,
                                         patient = as.character(unique(seg$patient)))

    }else{
        seg <- seg %>% 
            dplyr::arrange(.data$patient,
                           .data$Tumor_Sample_Barcode,
                           .data$Chromosome,
                           .data$Start_Position) %>%
            as.data.table()
        
        seg$Sample_ID <- paste(seg$patient,seg$Tumor_Sample_Barcode,sep = "&")
        
        sampleids <- sort(unique(seg$Sample_ID))
        
        ## sort sampleid 
        p <- unique(seg$patient)
        if(!is.null(sampleOrder)){
            patient.setdiff <- setdiff(names(sampleOrder), p)
            if(length(patient.setdiff) > 0){
                stop(paste0(patient.setdiff,
                            " cannot be found in your data. Please check sampleOrder!"))
            }
            o1 <- sampleOrder[[p]]
            o2 <- unique(seg[seg$patient == p]$Tumor_Sample_Barcode)
            s1 <- paste(p, sampleOrder[[p]], sep = "&") 
            s2 <- paste(p, unique(seg[seg$patient == p]$Tumor_Sample_Barcode), sep = "&") 
            sample.setdiff <- setdiff(o1, o2)
            if(length(sample.setdiff) > 0){
                stop(paste0(paste(sample.setdiff, collapse = ","),
                            " cannot be found in ", p, ". Please check sampleOrder!"))
            }else if(length(sample.setdiff) == 0 & length(s1)< length(s2)){
                s3 <- setdiff(s2, s1)
                sampleids <-  sampleids[!sampleids %in% s3]
                sampleids[grepl(p, sampleids)] <- s1
            }else if(length(sample.setdiff) > 0){
                sampleids[grepl(p, sampleids)] <- s1
            }
            
        }
        
        seg$Sample_ID <- factor(seg$Sample_ID, levels = sampleids)
        seg <- dplyr::arrange(seg, .data$Sample_ID)
        
        hlist1 <- seq(h,
                  h + (sample.bar.height + sample.bar.height/10) * (length(sampleids)-1),
                  (sample.bar.height + sample.bar.height/10))
        hlist2 <- hlist1 + sample.bar.height
        
        slist <- lapply(seq_len(length(sampleids)), function(i){
            h <- hlist1[i]
            sampleid <- sampleids[i]
            dat <- seg[seg$Sample_ID == sampleid,]
            dat$hmin <- h
            dat$hmax <- h + sample.bar.height
            return(dat)
        })
        seg <- dplyr::bind_rows(slist)
        
        # for(sampleid in sampleids){
            # seg[seg$Sample_ID == sampleid,]$hmin <- h
            # seg[seg$Sample_ID == sampleid,]$hmax <- h + sample.bar.height
            # h <- h + sample.bar.height + sample.bar.height/10
        # }
    }
    
    seg <- seg[seg$Sample_ID %in% sampleids]
    
    min <- sort(unique(seg$hmin))
    max <- sort(unique(seg$hmax))
    CNADat <- seg[seg$Type != "Neutral",]
    patient.type <- unique(CNADat$Type)
    
    ## set color for CNA type
    type.all.levels <- c("Loss", "Deletion",  "Gain",  "Amplification")
    type.all.colors <- c("#6baed6", "#084594", "#f4a582", "#d73027")
    type.level <- type.all.levels[type.all.levels %in% patient.type]
    type.color <- type.all.colors[type.all.levels %in% patient.type]
    
    CNADat$Type <- factor(CNADat$Type, levels = type.level)
    segmentTable <- data.table::data.table(y = c(min(min),max(max)), yend = c(min(min),max(max)))
    seg.add <- 20000000
    text.add <- -200000000
    segmentTable$x <- -seg.add
    segmentTable$xend <- max(chrTable$end)+seg.add
    segmentTable <- rbind(segmentTable, list(min(min), max(max), -seg.add,-seg.add))
    segmentTable <- rbind(segmentTable, list(min(min), max(max), max(chrTable$end)+seg.add,max(chrTable$end)+seg.add))
    
    ## get tumor sample barcode of patient
    patient.tsbs <- unlist(lapply(sampleids,
                                  function(x){return(strsplit(x,"&")[[1]][2])}))

    ## label on axis Y
    Y.text.table <- data.table::data.table(Pos = min + (max-min)/2, 
                                           Tumor_Sample_Barcode = patient.tsbs)
    
    ## set background for the graph
    backgroundTable <- data.table::data.table(
        xmin = min(chrTable$start), 
        xmax = max(chrTable$end), 
        ymin = min, 
        ymax = max)
    
    ## set color for patient bar
    allScales <- type.color
    names(allScales) <- type.level
    if("patient" %in% colnames(seg)&
       length(unique(seg$patient)) > 1){
        # set.seed(1234)
        # patient_colors <- sample(colors(),length(patient.rect.table$patient),replace = FALSE)
        # names(patient_colors) <- patient.rect.table$patient
        qual_col_pals <- RColorBrewer::brewer.pal.info[RColorBrewer::brewer.pal.info$category == 'qual',]
        col_vector <- unlist(mapply(RColorBrewer::brewer.pal, qual_col_pals$maxcolors, rownames(qual_col_pals)))
        
        patient_colors <- col_vector[seq_len(length(patient.rect.table$patient))]
        names(patient_colors) <- patient.rect.table$patient
    }
    
    ## cytoband table
    # Cytoband_table <- CNADat %>% 
    #     distinct(Cytoband, .keep_all = TRUE)
    
    ## convert chromosome name
    chrTable$chr = gsub(pattern = '23', replacement = 'X', x = chrTable$chr, fixed = TRUE)
    chrTable$chr = gsub(pattern = '24', replacement = 'Y', x = chrTable$chr, fixed = TRUE)
    
    ## initialize 
    xmin <- NULL
    xmax <- NULL
    ymin <- NULL
    ymax <- NULL
    Update_Start <- NULL
    Update_End <- NULL
    hmin <- NULL
    hmax <- NULL
    Type <- NULL
    chr <- NULL
    sample_num <- length(unique(CNADat$Sample_ID))

    p <- ggplot()+
        ## background color
        geom_rect(data = backgroundTable, 
                  mapping = aes(xmin = xmin, xmax = xmax, ymin = ymin, ymax = ymax), 
                  fill = "#f0f0f0")+
        
        ## CNA regions
        geom_rect(data = CNADat,
                  mapping = aes(xmin = Update_Start, xmax = Update_End, ymin = hmin, ymax = hmax, fill = Type))+
        scale_fill_manual(values = type.color, name = "Type")+
        
        ## chr bar
        geom_rect(data = chrTable,
                  mapping = aes(xmin = start, xmax = end, ymin = 0, ymax = chrom.bar.height),
                  fill = chr_bar_color, color = "black")+
        geom_text(data = chrTable,
                  mapping = aes(x = start + (end - start)/2, y = chrom.bar.height/2, label = chr), 
                  size = chrom.text.size,
                  color = chr_text_color)+
        
        ## vertical line 
        geom_segment(aes(x = chrTable$end[seq_len(nrow(chrTable)-1)],
                         xend = chrTable$end[seq_len(nrow(chrTable)-1)],
                         y = chrom.bar.height,
                         yend = max(backgroundTable$ymax)),
                     linetype = "dotted",
                     color = "#252525",
                     size = 0.7) +
        
        theme(
            axis.text.x = element_blank(),
            # axis.text.y.left =  element_text(size = sample.text.size,
            #                                  margin = margin(l = 0),
            #                                  colour = "black"),
            axis.ticks = element_blank(),
            # axis.line.y = element_line(colour = "darkblue", size = 1, linetype = "solid"),
            panel.grid =element_blank(),
            panel.border = element_blank(),
            panel.background = element_blank(),
            axis.title.x = element_blank(),
            axis.title.y = element_blank(),
            # legend.key.width  = unit(1.1,'lines'),
            # legend.box.margin = margin(l = 0),
            # legend.title = element_text(size = legend.title.size),
            # legend.text = element_text(size = legend.text.size,margin = margin(b = 3))
        )+
        # coord_fixed(ratio = 1) + 
        scale_x_continuous(expand = c(0,0))+
        # scale_y_continuous(breaks = Y.text.table$Pos,
        #                    labels = Y.text.table$Tumor_Sample_Barcode)+
        #ggtitle("Copy number alteration profile")+
        theme(plot.title = element_text(size = 13.5, face = "bold", hjust = 0.5, vjust = -2))
    if(showRownames){
        p <- p + scale_y_continuous(breaks = Y.text.table$Pos,
                               labels = Y.text.table$Tumor_Sample_Barcode) + 
            theme(
                axis.text.y.left =  element_text(size = sample.text.size,
                                                 margin = margin(l = 0),
                                                 colour = "black")
            )
    }else{
        p <- p + theme(axis.text.y.left =  element_blank())
    }
    
    if(showCytoband){
        if(!"Cytoband_pos" %in% colnames(CNADat)){
            stop("Cannot find information of cytobands. Please check your seg object.")
        }
        ## cytoband
        CNADat$Cytoband_pos <- as.numeric(CNADat$Start_Position) + chr_start_pos[paste0("chr",CNADat$Chromosome)]
        Cytoband_table <- CNADat %>%
            distinct(Cytoband, .keep_all = TRUE)
        p <- p + 
            
            ## blank bar
            geom_rect(aes(xmin = 0,
                         xmax = max(backgroundTable$xmax),
                         ymin = max(backgroundTable$ymax),
                         ymax = max(backgroundTable$ymax) + sample.bar.height*sample_num/5),
                     fill = "white") + 
            
            ggrepel::geom_text_repel(data = Cytoband_table, 
                                     aes(x = Cytoband_pos,
                                         y = max(backgroundTable$ymax),
                                         label = Cytoband),
                                     size = annot.text.size,
                                     max.overlaps = 200,
                                     angle = 90,
                                     nudge_y  = sample.bar.height*0.3*sample_num/8,
                                     direction = "x",
                                     vjust = 1)
                                                                           
    }
    
    if(showGene){
        if(showGene & showCytoband){
            stop("showGene and showCytoband cannot both be TRUE.")
        }
        if(!"Hugo_Symbol" %in% colnames(CNADat)){
            stop("Cannot find gene information. Please provide correct 'txdb' object in readSegment.")
        }
        CNADat$gene_pos <- as.numeric(CNADat$Update_Start) + (as.numeric(CNADat$Update_End) - as.numeric(CNADat$Update_Start ))/2
        CNADat$gene_id <- paste(CNADat$Hugo_Symbol)
        gene_table <- CNADat %>%
            distinct(gene_id, .keep_all = TRUE)
        p <- p + 
            geom_rect(aes(xmin = 0,
                          xmax = max(backgroundTable$xmax),
                          ymin = max(backgroundTable$ymax),
                          ymax = max(backgroundTable$ymax) + sample.bar.height*sample_num/5),
                      fill = "white") + 
            ggrepel::geom_text_repel(data = gene_table, 
                                     aes(x = gene_pos,
                                         y = max(backgroundTable$ymax) + sample.bar.height*sample_num/5,
                                         label = Hugo_Symbol),
                                     size = annot.text.size,
                                     angle = 90,
                                     # nudge_y  = sample.bar.height*0.3*patient_num,
                                     direction = "both",
                                     vjust = 1)
    }
    ## patient bar
    if("patient" %in% colnames(seg)&
       length(unique(seg$patient)) > 1){
        p <- p + 
            geom_rect(data = patient.rect.table,
                      mapping = aes(xmax = -20000000, xmin = -40000000,ymin = hmin, ymax = hmax),
                      fill =  patient_colors)
        # print(CNADat)
        ## multi legend
        patients <- unique(seg$patient)
        ## initialize
        patient <- NULL
        patient_legend <- (
            ggplot()+
                geom_rect(data = patient.rect.table,
                          mapping = aes(xmax = -20000000, xmin = -40000000,
                                        ymin = hmin, ymax = hmax,fill = patient)) +  
                scale_fill_manual(breaks = rev(patients),
                                  values = patient_colors,
                                  name = "Patient") + 
                theme(legend.background = element_blank(),
                      legend.title = element_text(size = legend.title.size),
                      legend.text = element_text(size = legend.text.size,margin = margin(b = 3)))
        )%>% ggplotGrob()
        patient_legend <- patient_legend$grobs[[which(unlist(lapply(patient_legend$grobs, function(x) {x$name})) == "guide-box")]]
        
        type_legend <- (
            ggplot()+
                geom_rect(data = CNADat,mapping = aes(xmin = Update_Start, xmax = Update_End, ymin = hmin, ymax = hmax, fill = Type))+
                scale_fill_manual(values = type.color, name = "Type")+
                theme(legend.background = element_blank(),
                      legend.title = element_text(size = legend.title.size),
                      legend.text = element_text(size = legend.text.size,margin = margin(b = 3)))
        ) %>% ggplotGrob()
        type_legend <- type_legend$grobs[[which(unlist(lapply(type_legend$grobs, function(x) {x$name})) == "guide-box")]]
        
        legends_column <-
            plot_grid(
                patient_legend + theme(legend.position = "none"),
                type_legend,
                patient_legend + theme(legend.position = "none"),
                patient_legend,
                patient_legend + theme(legend.position = "none"),
                ncol = 1,
                rel_heights = c(1,1,.1,1,1),
                align = "v")
        
        p <- plot_grid(p + theme(legend.position = "none"),
                       legends_column + theme(plot.margin = margin(r = 15)),
                       nrow = 1,
                       rel_widths = c(1,.2))
        
    }
        
    
    # if(length(CNAplot.list) == 1){
    #     return(CNAplot.list[[1]])
    # }
    
    return(p)
}