From chromosome-arm heatmaps to bar plots

Background

Chromosome-arm calls can be represented as a cell-by-arm heatmap. This workflow summarizes those calls as arm-level bar plots.

Goal

Convert the cell-by-arm heatmap into chromosome-arm bar plots. Both representations use chromosome arms on the x-axis.

Download original image: Cell-by-chromosome-arm heatmap for the 14-day dataset

14d.pdf

Convert heatmap dataframe to long format

final_mtx_filtered_14_long <- tidyr::pivot_longer(final_mtx_filtered_14d,
                                                  cols = colnames(final_mtx_filtered_14d),
                                                  names_to = "amplicon",
                                                  values_to = "ploidy")
final_mtx_filtered_14_long$ploidy <- factor(final_mtx_filtered_14_long$ploidy, levels = c(0,1,2))
final_mtx_filtered_14_long$amplicon <- factor(final_mtx_filtered_14_long$amplicon, levels = AmpliconList_for_39SNPs)

Visualization

pdf(file = "PDF/Barplot/14d.pdf", width = 30,height = 20)

final_mtx_filtered_14d_long %>%
  dplyr::count(amplicon, ploidy) %>% 
  tidyr::complete(amplicon, ploidy, fill = list(n = 0)) %>% # Complete missing combinations of amplicon and ploidy with 0 counts
  ungroup() %>% 
  group_by(amplicon) %>%
  mutate(pct = (n / sum(n)) * 100) %>%
  ungroup() %>%

  ggplot() +
  aes(amplicon, pct, fill = ploidy) +
  geom_bar(stat = "identity") +
  ylab("Cell number (%)") +
  xlab("Amplicons") +
  ggrepel::geom_text_repel(aes( label = paste0(sprintf("%1.2f", pct),"%")),
                           position = position_stack(vjust = 0.5),
                           direction="y", 
                           hjust=0.5, 
                           box.padding=0.1,
                           color = "white",
                           size = 8) +
  ggtitle("Ploidy composition of each amplicon in 14d data (511 consistent cells)") +
  theme(text = element_text(size = 35),
        axis.text.x = element_text(angle = 45, vjust = 1, hjust=1)) +
  scale_fill_manual(values = c("#000000","#0080FF","#FF9933"),
                    name = "Ploidy")

dev.off()

Compare one chromosome arm across three datasets

pdf(file = "PDF/Barplot/per_amplicon.pdf", width = 10,height = 8)


###

for (amplicon_subset in AmpliconList_for_39SNPs) {
    df <- data.frame(row.names = 1:3)
    
    ####
    for (data in c("2N","8d","14d")) {
    
      matrix <- get(paste0("final_mtx_filtered_",data,"_long"))
      
      matrix_subset_pct <-
            matrix %>%
            dplyr::count(amplicon, ploidy) %>% 
            tidyr::complete(amplicon, ploidy, fill = list(n = 0)) %>% # Complete missing combinations of amplicon and ploidy with 0 counts
            ungroup() %>% 
            group_by(amplicon) %>%
            mutate(pct = (n / sum(n)) * 100) %>%
            ungroup() %>%
            subset(amplicon == amplicon_subset) %>%
            select(pct)
      
      colnames(matrix_subset_pct) = c(paste0("pct_", data))
      df <- cbind(df, matrix_subset_pct)
    }
    ####
    
    
    # output df
    df$ploidy = as.factor(c(0,1,2))
    assign(paste0("pct_of_", amplicon_subset) , df,envir = .GlobalEnv)
    

    # visualization
    # preprocess the data format
    matrix <- get(paste0("pct_of_", amplicon_subset))
    long_matrix <- matrix %>% tidyr::pivot_longer(cols = c(-ploidy),
                                          names_to = "dataset",
                                          values_to = "percentage")
    long_matrix$dataset <- factor(long_matrix$dataset, levels = c("pct_2N","pct_8d","pct_14d"))

    plot <-
          ggplot(long_matrix) +
          aes(dataset, percentage, fill = ploidy) +
          geom_bar(stat = "identity") +
          ylab("Cell number (%)") +
          xlab("Datasets") +
          ggrepel::geom_text_repel(aes(label = paste0(sprintf("%1.2f", percentage),"%")),
                                   position = position_stack(vjust = 0.5),
                                   direction="y",
                                   hjust=0.5,
                                   box.padding=0.1,
                                   color = "white",
                                   size = 8) +
          ggtitle(paste0("Ploidy composition of ", amplicon_subset," of each dataset")) +
          theme(text = element_text(size = 20),
                axis.text.x = element_text(angle = 45, vjust = 1, hjust=1)) +
          scale_fill_manual(values = c("#000000","#0080FF","#FF9933"),
                            name = "Ploidy")
    print(plot)
    rm(df)
}

dev.off()

An example section of the resulting plot.