Overview of lines

library(readr)
library(dplyr)
library(scran)
library(scater)
library(viridis)
library(ggplot2)
library(ggforce)
library(ggridges)
library(SingleCellExperiment)
library(edgeR)
library(limma)
library(org.Hs.eg.db)
library(cowplot)
library(gplots)
library(ggrepel)
library(sigfit)
library(Rcpp)
library(deconstructSigs)
options(stringsAsFactors = FALSE)

donor_info <- as.data.frame(read_csv("data/donor_info_core.csv"))
# merge age bins
donor_info$age_decade <- ""
for (i in 1:nrow(donor_info)) {
  if (donor_info$age[i] %in% c("30-34", "35-39")) 
    donor_info$age_decade[i] <- "30-39"
  if (donor_info$age[i] %in% c("40-44", "45-49")) 
    donor_info$age_decade[i] <- "40-49"
  if (donor_info$age[i] %in% c("50-54", "55-59")) 
    donor_info$age_decade[i] <- "50-59"
  if (donor_info$age[i] %in% c("60-64", "65-69")) 
    donor_info$age_decade[i] <- "60-69"
  if (donor_info$age[i] %in% c("70-74", "75-79")) 
    donor_info$age_decade[i] <- "70-79"
}

exome_sites <- read_tsv("data/exome-point-mutations/high-vs-low-exomes.v62.ft.filt_lenient-alldonors.txt.gz",
                        col_types = "ciccdcciiiiccccccccddcdcll", comment = "#",
                        col_names = TRUE)
exome_sites <- dplyr::mutate(
  exome_sites,
  chrom = paste0("chr", gsub("chr", "", chrom)),
  var_id = paste0(chrom, ":", pos, "_", ref, "_", alt),
  chr_pos = paste0(chrom, "_", pos))

exome_sites <- as.data.frame(exome_sites)
## deduplicate sites list
exome_sites <- exome_sites[!duplicated(exome_sites[["var_id"]]),]
## calculate coverage at sites for each donor
donor_vars_coverage <- list()
for (i in unique(exome_sites$donor_short_id)) {
  exome_sites_subset <- exome_sites[exome_sites$donor_short_id == i, ]
  donor_vars_coverage[[i]] <- exome_sites_subset$nREF_fibro + exome_sites_subset$nALT_fibro
}

vep_best <- read_tsv("data/exome-point-mutations/high-vs-low-exomes.v62.ft.alldonors-filt_lenient.all_filt_sites.vep_most_severe_csq.txt")
colnames(vep_best)[1] <- "Uploaded_variation"
## deduplicate dataframe
vep_best <- as.data.frame(vep_best[!duplicated(vep_best[["Uploaded_variation"]]),])
as.data.frame(table(vep_best[["Consequence"]]))

vep_best[["var_id"]] <- paste0("chr", vep_best[["Uploaded_variation"]])
exome_sites <- inner_join(exome_sites, 
                          vep_best[, c("var_id", "Location", "Consequence")], 
                          by = "var_id")

impactful_csq <- c("stop_lost", "start_lost", "stop_gained",
                   "splice_donor_variant", "splice_acceptor_variant",
                   "splice_region_variant", "missense_variant")
donor_info$num_mutations <- NA
donor_info$num_synonymous <- NA
donor_info$num_missense <- NA
donor_info$num_splice_region <- NA
donor_info$num_splice_acceptor <- NA
donor_info$num_splice_donor <- NA
donor_info$num_stop_gained <- NA
donor_info$num_start_lost <- NA
donor_info$num_stop_lost <- NA
for (i in unique(donor_info$donor_short)) {
  if (i %in% unique(exome_sites$donor_short_id)) {
    exome_sites_subset <- exome_sites[exome_sites$donor_short_id == i, ]
    donor_info$num_mutations[donor_info$donor_short == i] <- length(exome_sites_subset$Consequence)
    donor_info$num_synonymous[donor_info$donor_short == i] <- sum(exome_sites_subset$Consequence == "synonymous_variant")
    donor_info$num_missense[donor_info$donor_short == i] <- sum(exome_sites_subset$Consequence == "missense_variant")
    donor_info$num_splice_region[donor_info$donor_short == i] <- sum(exome_sites_subset$Consequence == "splice_region_variant")
    donor_info$num_splice_acceptor[donor_info$donor_short == i] <- sum(exome_sites_subset$Consequence == "splice_acceptor_variant")
    donor_info$num_splice_donor[donor_info$donor_short == i] <- sum(exome_sites_subset$Consequence == "splice_donor_variant")
    donor_info$num_stop_gained[donor_info$donor_short == i] <- sum(exome_sites_subset$Consequence == "stop_gained")
    donor_info$num_start_lost[donor_info$donor_short == i] <- sum(exome_sites_subset$Consequence == "start_lost")
    donor_info$num_stop_lost[donor_info$donor_short == i] <- sum(exome_sites_subset$Consequence == "stop_lost")
  }
}

ntrtestrPetr <- read.table("data/neutralitytestr-petr.tsv", 
                           stringsAsFactors = FALSE, header = TRUE)
negbinfitPetr = read.table("data/neg-bin-rsquared-petr.csv", 
                           stringsAsFactors = FALSE, header = TRUE, sep = ",")
negbinfitPetr$sampleID <- negbinfitPetr$fname
negbinfitPetr$sampleID <- gsub("petr-AF-", "", negbinfitPetr$sampleID)
negbinfitPetr$sampleID <- gsub(".tsv", "", negbinfitPetr$sampleID)
rownames(negbinfitPetr) <- negbinfitPetr$sampleID

dfrsq <- data.frame(sampleID = ntrtestrPetr$sampleID,
                    rsq_ntrtestr = ntrtestrPetr$rsq,
                    rsq_negbinfit = negbinfitPetr[ntrtestrPetr$sampleID, "rsq"])

cutoff_selection_cummut <- 0.85
cutoff_selection_negbin <- 0.25
cutoff_neutral_cummut <- 0.9
cutoff_neutral_negbin <- 0.55

dfrsq$candidatelabel <- NA
dfrsq$candidatelabel[dfrsq$sampleID == "joxm"] <- "joxm"

filter_selection <- (dfrsq$rsq_ntrtestr < cutoff_selection_cummut) &
  (dfrsq$rsq_negbinfit < cutoff_selection_negbin)
filter_neutral <- (dfrsq$rsq_ntrtestr > cutoff_neutral_cummut) &
  (dfrsq$rsq_negbinfit > cutoff_neutral_negbin)

dfrsq$selection <- "undetermined"
dfrsq$selection[filter_selection] <- "selected"
dfrsq$selection[filter_neutral] <- "neutral"

colnames(dfrsq)[1] <- "donor_short"
donor_info <- merge(donor_info, dfrsq, by = "donor_short")
donor_info$selection_colour <- "#CCCCCC"
for (i in 1:nrow(donor_info)) {
  if (donor_info$selection[i] == "neutral") 
    donor_info$selection_colour[i] <- "dodgerblue"
  if (donor_info$selection[i] == "selected") 
    donor_info$selection_colour[i] <- "dodgerblue4"
}

donors <- c("euts", "fawm", "feec", "fikt", "garx", "gesg", "heja", "hipn", 
            "ieki", "joxm", "kuco", "laey", "lexy", "naju", "nusw", "oaaz", 
            "oilg", "pipw", "puie", "qayj", "qolg", "qonc", "rozh", "sehl", 
            "ualf", "vass", "vils", "vuna", "wahn", "wetu", "xugn", "zoxy")

dfrsq_filt <- dfrsq[(dfrsq$donor_short %in% donors),]

plt_scatter <- ggplot(dfrsq_filt, aes(x = rsq_negbinfit, y = rsq_ntrtestr)) +
  scale_colour_manual(values = c("neutral" = "dodgerblue",
                                 "selected" = "dodgerblue4",
                                 "undetermined" = "#CCCCCC")) +
  geom_point(aes(colour = selection)) +
  geom_label_repel(aes(label = candidatelabel), color = "white", size = 2.5,
                   fill = "black", box.padding = 0.35, point.padding = 0.5,
                   segment.color = "grey50") +
  theme_bw() +
  theme(text = element_text(size = 9,face = "bold"), 
        axis.text = element_text(size = 9, face = "bold"),
        axis.title = element_text(size = 10, face = "bold"),
        plot.title = element_text(size = 9, hjust = 0.5)) +
  labs(x = "Goodness of Fit - Negative Binomial Distribution", 
       y = "Goodness of Fit - Cumulative Mutations") +
  theme(strip.background = element_blank()) +
  labs(title = "") +
  theme(legend.justification = c(1,0), legend.position = c(1,0)) +
  theme(legend.background = element_rect(fill = "white", linetype = 1, 
                                         colour = "black", size =0.2),
        legend.key.size = unit(0.25, "cm")) +
  labs(colour = "Selection") #+
# coord_fixed()

ggsave("figures/overview_lines/neutral_selection_models_gof_scatter.png", 
       plot = plt_scatter, width = 12, height = 12, dpi = 300, units = "cm")

plt_scatter

data("cosmic_signatures", package = "sigfit")

##new input 
mutation_list <- read.table("data/exome-point-mutations/high-vs-low-exomes.v62.ft.filt_lenient-alldonors.txt.gz", header = TRUE)
mutation_list$chr_pos <- paste0("chr", mutation_list$chr, "_", mutation_list$pos)

mutation_donors <- unique(mutation_list$donor_short_id)
mutation_list_donors <- list()
for (i in mutation_donors) {
  cat("....reading ", i, "\n")
  mutation_list_donors[[i]] <- mutation_list[which(mutation_list$donor_short_id == i),]
  mutation_list_donors[[i]]$chr <- paste0("chr", mutation_list_donors[[i]]$chrom)
  mutation_list_donors[[i]]$chr_pos = paste0(mutation_list_donors[[i]]$chr, "_", mutation_list_donors[[i]]$pos)
}

## Calculate triNucleotide contexts for mutations using deconstructSigs command
mut_triNs <- list()
for (i in mutation_donors) {
  cat("....processing ", i, "\n")
  mut_triNs[[i]] <- mut.to.sigs.input(mutation_list_donors[[i]], sample.id = "donor_short_id", 
                                      chr = "chr", pos = "pos", ref = "ref", alt = "alt")
}

## Convert to correct format
sig_triNs <- character()
for (j in 1:96) {
  c1 <- substr(colnames(mut_triNs[[1]])[j], 1,1)
  ref <- substr(colnames(mut_triNs[[1]])[j], 3,3)
  alt <- substr(colnames(mut_triNs[[1]])[j], 5,5)
  c3 <- substr(colnames(mut_triNs[[1]])[j], 7,7)
  triN_sigfit <- paste0(c1,ref,c3,">",c1,alt,c3)
  sig_triNs[j] <- triN_sigfit
}

for(i in mutation_donors) {
  colnames(mut_triNs[[i]]) <- sig_triNs
}

## Fit signatures using sigfit
mcmc_samples_fit <- list()
set.seed(1234)
for (i in mutation_donors) {
  mcmc_samples_fit[[i]] <- sigfit::fit_signatures(
    counts = mut_triNs[[i]], signatures = cosmic_signatures,
    iter = 2000, warmup = 1000, chains = 1, seed = 1)
}

## Estimate exposures using sigfit
exposures <- list()
for (i in mutation_donors) {
  exposures[[i]] <- sigfit::retrieve_pars(
    mcmc_samples_fit[[i]], par = "exposures", hpd_prob = 0.90)
}

## Plot exposure bar charts
donors <- c("euts", "fawm", "feec", "fikt", "garx", "gesg", "heja", "hipn", 
            "ieki", "joxm", "kuco", "laey", "lexy", "naju", "nusw", "oaaz", 
            "oilg", "pipw", "puie", "qayj", "qolg", "qonc", "rozh", "sehl", 
            "ualf", "vass", "vils", "vuna", "wahn", "wetu", "xugn", "zoxy")

signature_names <- c("1", "2", "3", "4", "5", "6", "7", "8", "9", "10", "11", 
                     "12", "13", "14", "15", "16", "17", "18", "19", "20", "21", 
                     "22", "23", "24", "25", "26", "27", "28", "29", "30")

for (j in donors) {
  cat("....plotting ", j, "\n")
   sigfit::plot_exposures(mcmc_samples_fit[[j]], 
                          signature_names = signature_names)
   png(paste0("figures/overview_lines/mutational_signatures/exposure_barchart_", 
             j, ".png"),
      units = "in", width = 12, height = 10, res = 500)
  sigfit::plot_exposures(mcmc_samples_fit[[j]], 
                         signature_names = signature_names)
  dev.off()
 
}

get_signature_df <- function(exposures, samples, signature) {
  signature_mat_mean <- matrix(NA, nrow = length(samples), 30)
  for (i in 1:length(samples)) {
    signature_mat_mean[i,] <- as.numeric(exposures[[i]]$mean)
  }
  rownames(signature_mat_mean) <- samples
  colnames(signature_mat_mean) <- colnames(exposures[[i]]$mean)
  signature_mat_lower90 <- matrix(NA, nrow = length(samples), 30)
  for (i in 1:length(samples)) {
    signature_mat_lower90[i,] <- as.numeric(exposures[[i]]$lower_90)
  }
  rownames(signature_mat_lower90) <- samples
  colnames(signature_mat_lower90) <- colnames(exposures[[i]]$lower_90)
  signature_mat_upper90 <- matrix(NA, nrow = length(samples), 30)
  for (i in 1:length(samples)) {
    signature_mat_upper90[i,] <- as.numeric(exposures[[i]]$upper_90)
  }
  rownames(signature_mat_upper90) <- samples
  colnames(signature_mat_upper90) <- colnames(exposures[[i]]$upper_90)
  signature_df <- cbind(
    as.data.frame(signature_mat_mean[,signature]),
    as.data.frame(signature_mat_lower90[,signature]),
    as.data.frame(signature_mat_upper90[,signature]))
  colnames(signature_df) <- c(paste0("Sig",signature,"_mean"),
                              paste0("Sig",signature,"_lower"),
                              paste0("Sig",signature,"_upper"))
  signature_df$donor <- rownames(signature_df)
  signature_df
}

## Get lower, mean and upper values for signatures 7 & 11
sig7_df <- get_signature_df(exposures, mutation_donors, 7)
sig11_df <- get_signature_df(exposures, mutation_donors, 11)

## Add Sigs 7/11 to donor info table
sig_subset_df <- merge(sig7_df,sig11_df, by = "donor")
sig_subset_df_means <- sig_subset_df[,c(1,grep("mean", colnames(sig_subset_df)))]
sig_subset_df_means_melt <- reshape2::melt(sig_subset_df_means)
sig_subset_df_means_melt$signature <- substr(sig_subset_df_means_melt$variable, 1, 5)
sig_subset_df_means_melt <- sig_subset_df_means_melt[,c("donor", "signature", "value")]

colnames(sig_subset_df)[1] <- "donor_short" 
donor_info <- merge(donor_info, sig_subset_df, by = "donor_short")

params <- list()
params$callset <- "filt_lenient.cell_coverage_sites"
fls <- list.files("data/sces")
fls <- fls[grepl(params$callset, fls)]
donors <- gsub(".*ce_([a-z]+)_.*", "\\1", fls)
sce_unst_list <- list()
for (don in donors) {
    sce_unst_list[[don]] <- readRDS(file.path("data/sces", 
        paste0("sce_", don, "_with_clone_assignments.", params$callset, ".rds")))
    cat(paste("reading", don, ":   ", ncol(sce_unst_list[[don]]), "cells.\n"))
}

## Calculate single cell data metrics
sc_metrics_summary <- list()
sc_metrics_summary_df <- data.frame()
for (don in donors) {
  sc_metrics_summary[[don]]$num_unst_cells <- ncol(sce_unst_list[[don]])
  sc_metrics_summary[[don]]$num_assignable <- sum(sce_unst_list[[don]]$assignable)
  sc_metrics_summary[[don]]$num_unassignable <- 
    (sc_metrics_summary[[don]]$num_unst_cells -
       sc_metrics_summary[[don]]$num_assignable)
  sc_metrics_summary[[don]]$num_clones_with_cells <-
    length(unique(sce_unst_list[[don]]$assigned[
      which(sce_unst_list[[don]]$assigned != "unassigned")]))
  sc_metrics_summary[[don]]$donor <- don
  sc_metrics_summary_df <-
    rbind(sc_metrics_summary_df, as.data.frame(sc_metrics_summary[[don]]))
}

colnames(sc_metrics_summary_df) <- c("total_unst_cells", "assigned_unst_cells",
                                     "unassigned_unst_cells", 
                                     "num_clones_with_cells", "donor_short")

## Merge with donor info table
donor_info <- merge(donor_info, sc_metrics_summary_df, by = "donor_short", 
                    all.x = TRUE)
donor_info$percent_assigned_cells <-
  donor_info$assigned_unst_cells / donor_info$total_unst_cells

canopy_files <- list.files("data/canopy")
canopy_files <- canopy_files[grepl(params$callset, canopy_files)]
canopy_list <- list()
for (don in donors) {
  canopy_list[[don]] <- readRDS(
    file.path("data/canopy", 
              paste0("canopy_results.", don, ".", params$callset, ".rds")))
}

clone_mut_list <- list()
for (don in donors) {
  cat(paste("summarising clones and mutations for", don, '\n'))
  clone_mut_list[[don]] <- colSums(canopy_list[[don]]$tree$Z)
  cat(colSums(canopy_list[[don]]$tree$Z))
  cat('\n')
}

clone_mut_df <- data.frame(clone1 = numeric(), clone2 = numeric(), 
                           clone3 = numeric(), clone4 = numeric())
for (i in 1:length(donors)) {
  num_clones <- length(clone_mut_list[[i]])
  num_NAs <- 4 - num_clones
  temp_row  <- c(clone_mut_list[[i]], rep(NA, num_NAs))
  clone_mut_df[i,] <- temp_row
}
rownames(clone_mut_df) <- donors
clone_mut_df$donor_short <- donors

clone_mut_unique_list <- list()
for (don in donors) {
        cat(paste("summarising clones and mutations for", don, '\n'))
        clone_mut_unique_list[[don]] <-
          colSums(canopy_list[[don]]$tree$Z[
            (rowSums(canopy_list[[don]]$tree$Z) == 1),])
        cat(colSums(canopy_list[[don]]$tree$Z[
          (rowSums(canopy_list[[don]]$tree$Z) == 1),]))
        cat('\n')
}

clone_mut_unique_df <- data.frame(clone1 = numeric(), clone2 = numeric(), 
                                  clone3 = numeric(),
                           clone4 = numeric(), min_unique_muts = numeric())
for (i in 1:length(donors)) {
  num_clones <- length(clone_mut_unique_list[[i]])
  num_NAs <- 4 - num_clones
  temp_row  <- c(clone_mut_unique_list[[i]], rep(NA, num_NAs), 
                 min(clone_mut_unique_list[[i]][2:num_clones]))
  clone_mut_unique_df[i,] <- temp_row
}
rownames(clone_mut_df) <- donors
clone_mut_df$donor_short <- donors
rownames(clone_mut_unique_df) <- donors
clone_mut_unique_df$donor_short <- donors
donor_info <- merge(donor_info, clone_mut_unique_df, by = "donor_short", all.x = T)

## Calculate Hamming distance
clone_mut_list_hamming <- list()
for (don in donors) {
  Config <- canopy_list[[don]]$tree$Z
  unique_sites_paired <- c()
  for (i in seq_len(ncol(Config) - 1)) {
    for (j in seq(i + 1, ncol(Config))) {
      n_sites <- sum(rowSums(Config[, c(i,j)]) == 1)
      unique_sites_paired <- c(unique_sites_paired, n_sites)
    }
  }
  clone_mut_list_hamming[[don]] <- unique_sites_paired  
  cat("....hamming distances for ", don, ": ", unique_sites_paired, "\n") 
}

min_hamming_distance <- data.frame("donor_short" = donors,
                                   "min_hamming_dist" = 0)
for (i in 1:length(donors)) {
  min_hamming_distance$min_hamming_dist[i] <- min(clone_mut_list_hamming[[i]])
}

donor_info <- merge(donor_info, min_hamming_distance, by = "donor_short", 
                    all.x = TRUE)

## Number of clones
donor_info$num_clones_total <- 0
for (i in donors) {
  num_clones <- length(clone_mut_unique_list[[i]])
  donor_info$num_clones_total[which(donor_info$donor_short == i)] <- num_clones
}

donor_info_filt <- donor_info[which(donor_info$donor_short %in% donors), ]
donor_filt_order <- dplyr::arrange(donor_info_filt, desc(num_mutations))
donor_filt_order <- donor_filt_order$donor_short
rownames(donor_info_filt) <- donor_info_filt$donor_short

# Plot age
selected_vars <- c("donor_short", "age_decade", "selection_colour")
donor_info_filt_melt <- reshape2::melt(donor_info_filt[,selected_vars],
                                       "donor_short")
donor_info_filt_melt <- donor_info_filt[,selected_vars]

age_plot_filt <- ggplot(
  donor_info_filt_melt, 
  aes(x = "Age", y = factor(donor_short, levels = rev(donor_filt_order)), 
      fill = age_decade)) + 
  geom_tile() +
  labs(x = "", y = "Donor", fill = "") + 
  scale_fill_manual(values = rev(c(magma(8)[-c(1,8)])))  +
  guides(fill = guide_legend(nrow = 3,byrow = TRUE)) +
  theme(axis.text.x = element_text(size = 28, face = "bold"), 
        axis.line = element_blank(), 
        legend.position = "top", legend.direction = "horizontal", 
        legend.text = element_text(size = 24, face = "bold"), 
        legend.key.size = unit(0.3,"in"),
        axis.ticks.x = element_blank(),
        legend.margin = margin(unit(0.01, "cm")),
        axis.text.y = element_text(size = 28, face = "bold"), 
        axis.title.y = element_text(size = 28, face = "bold"), 
        axis.title.x = element_text(size = 28, face = "bold"))

# Plot number of mutations
selected_vars <- c("donor_short", "num_mutations")
donor_info_filt_melt <- reshape2::melt(donor_info_filt[,selected_vars],
                                       "donor_short")
num_mut_plot_filt <- ggplot(
  donor_info_filt_melt, 
  aes(x = value, y = factor(donor_short, levels = rev(donor_filt_order)))) + 
  geom_point(size = 4, alpha = 0.7) + 
  scale_shape_manual(values = c(4, 19)) + 
  scale_x_log10(breaks=c(50, 100, 500)) +
  theme_bw(16) + 
  ggtitle(" ") + 
  labs(x = "Number of mutations", y = "") + 
  theme(axis.text.y = element_blank(), 
        axis.ticks.y = element_blank(), 
        axis.line = element_blank(), 
        legend.position = "top", 
        axis.text.x = element_text(size = 28, face = "bold"), 
        title = element_text(colour="black",size = 26, face = "bold"), 
        axis.title.x = element_text(size = 28, face = "bold"))

# Plot signature 7 
donor_info_filt_melt <- reshape2::melt(
  donor_info_filt[,c(1,grep("Sig7_mean",colnames(donor_info_filt)))])
donor_info_filt_melt$signature <- substr(donor_info_filt_melt$variable, 1, 5)
donor_info_filt_melt <- 
  donor_info_filt_melt[, c("donor_short", "signature", "value")]
sig_decomp_plot_filt <- ggplot(
  donor_info_filt_melt, 
  aes(y = value, x = factor(donor_short, levels = rev(donor_filt_order)), 
      fill = factor(signature))) + 
  geom_bar(stat = "identity") + 
  coord_flip() + 
  scale_fill_manual(values = rev(c(magma(10)[-c(1,10)]))) +
  theme_bw(16) + 
  ggtitle(" ") + 
  labs(x = "",y = "Signature 7 exposure", fill = "") +
  theme(axis.text.y = element_blank(), 
        axis.ticks.y = element_blank(), 
        axis.line = element_blank(), 
        axis.text.x = element_text(size = 28, face = "bold"),
        title = element_text(colour = "black", size = 26, face = "bold"),
        axis.title.x = element_text(size = 28, face = "bold")) +
  guides(fill = FALSE)

# Plot total number of mutations per clone
donor_info_filt_subset <-
  donor_info_filt[,c(1,grep("clone",colnames(donor_info_filt)))]
# Remove columns that do not relate to number of mutations per clone: 
donor_info_filt_subset <- donor_info_filt_subset[,c(-2,-7)]
donor_info_filt_melt <- reshape2::melt(donor_info_filt_subset, "donor_short")
donor_info_filt_melt$variable <- substr(donor_info_filt_melt$variable, 6, 6)
num_clones_plot_filt <- ggplot(
  donor_info_filt_melt, 
  aes(variable, factor(donor_short, levels = rev(donor_filt_order)))) +
  geom_tile(aes(fill = value), colour = "white") + 
  scale_fill_distiller(palette = "BuPu", values = c(0,0.05,0.1,0.15,0.25,0.5,1),
                       na.value = "white", breaks = c(0,25,50,75), 
                       direction = 1) +
  labs(x = "Clone", y = "", fill = "Number of mutations per clone") +  
  theme(axis.ticks.y = element_blank(), 
        axis.line = element_blank(), 
        legend.position = "top",
        legend.key.size = unit(0.5,"in"), 
        axis.title.x = element_text(size = 28, face = "bold"),
        axis.text.x = element_text(size = 28, face = "bold"),
        axis.text.y = element_blank(), 
        legend.justification = "center", 
        legend.text = element_text(size = 26, face = "bold"), 
        legend.title = element_text(size = 26, face = "bold"))

## Combine above plots into Fig 2a
fig_2a <- cowplot::plot_grid(age_plot_filt, num_mut_plot_filt,
                             sig_decomp_plot_filt, num_clones_plot_filt, 
                             nrow = 1, rel_widths = c(3, 4, 4, 6), align = "h",
                             axis = "t", scale =  c(1, 0.988, 0.988, 0.988))
ggsave("figures/overview_lines/overview_lines_BuPu.png", 
       plot = fig_2a, width = 30, height = 14)

fig_2a

fig_2c_no_size <- ggplot(
  donor_info_filt, 
  aes(y = as.numeric(percent_assigned_cells), x = as.numeric(min_hamming_dist), 
      fill = total_unst_cells)) + 
  geom_point(pch = 21, colour = "gray50", size = 4) +
  scale_shape_manual(values = c(4, 19)) + 
  ylim(c(0, 1)) + 
  scale_x_log10() + 
  scale_fill_viridis(option = "magma") +
  ylab("Proportion cells assigned") + 
  xlab("Minimum number of variants distinguishing clones") + 
  labs(title = "") +
  labs(fill = "Number of cells") + 
  theme_bw() +
  theme(text = element_text(size = 9,face = "bold"), 
        axis.text = element_text(size = 9, face = "bold"),
        axis.title = element_text(size = 10, face = "bold"),
        plot.title = element_text(size = 9, hjust = 0.5)) +
  theme(strip.background = element_blank()) +
  theme(legend.justification = c(1,0), 
        legend.position = c(1,0), 
        legend.direction = "horizontal",
        legend.background = element_rect(fill = "white", linetype = 1, 
                                         colour = "black", size = 0.2),
        legend.key.size = unit(0.25, "cm"))

ggsave("figures/overview_lines/cell_assignment_vs_min_hamming_dist.png",
       plot = fig_2c_no_size, width = 12, height = 12, dpi = 300, units = "cm")
fig_2c_no_size

write_tsv(donor_info_filt, path = "output/line_info.tsv", col_names = TRUE)

devtools::session_info()