Merge branch 'main' into main

DESCRIPTION

@@ -25,6 +25,9 @@ Imports:
     rlang,
     ggplot2,
     dplyr,
+    purrr,
+    stringr,
     methods,
     utils,
+    RColorBrewer,
     magrittr

NAMESPACE

@@ -2,8 +2,16 @@
 
 export("%>%")
 export(get_study)
+export(plot_reads_group_abundance)
+export(plot_clonal_prop_per_sample)
+export(plot_clonal_prop_per_group)
+export(plot_clonotype_abundance)
+export(get_bottom_n_clonotypes)
+export(get_top_n_clonotypes)
+export(plot_diversity_index)
 export(plot_motif_counts)
 exportClasses(Basic)
+exportClasses(Clonality)
 exportClasses(Diversity)
 exportClasses(Motif)
 exportClasses(reTCRProj)

R/get_study.R

@@ -1,5 +1,7 @@
-source("R/utils-basic.R")
 source("R/utils-diversity.R")
+source("R/utils-clonality.R")
+source("R/utils-basic.R")
+source("R/utils-hill.R")
 
 #' Get pyTCR data of a project
 #'
@@ -12,9 +14,8 @@ source("R/utils-diversity.R")
 #'
 #' @examples
 #' \dontrun{
-#' proj <- reTCR::get_study(id="PRJNA473147", attr_col="cmv_status")
+#' proj <- reTCR::get_study(id = "PRJNA473147", attr_col = "cmv_status")
 #' }
-
 get_study <- function(id, attr_col) {
   stopifnot(is.character(id) && nchar(id) > 0)
   filename <- paste0(id, "_mixcr_metadata_file.csv")
@@ -23,11 +24,15 @@ get_study <- function(id, attr_col) {
 
   basic <- .get_basic(data = data, attr_col = attr_col)
   diversity <- .get_diversity(data = data, attr_col = attr_col)
+  clonality <- .get_clonality(data = data, attr_col = attr_col)
+  hill <- .get_hill_numbers(df = data)
 
   return(methods::new(
     "reTCRProj",
     data = data,
     basic = basic,
-    diversity = diversity
+    diversity = diversity,
+    clonality = clonality,
+    hill = hill
   ))
 }

R/plot_clonal_prop_per_group.R

@@ -0,0 +1,47 @@
+#' Plot clonal proportion per group
+#'
+#' This function plots the clonal proportion per group.
+#'
+#' @param data A data frame of clonal proportion per group.
+#' @param attr_col Character. Attribute column name.
+#'
+#' @return A ggplot2 plot of clonal proportion per group.
+#' @export
+#'
+#' @examples
+#' \dontrun{
+#' proj <- reTCR::get_study(id = "PRJNA473147", attr_col = "cmv_status")
+#' clonal_prop <- proj@clonality@clonal_prop
+#' reTCR::plot_clonal_prop_per_group(clonal_prop, "cmv_status")
+#' }
+plot_clonal_prop_per_group <- function(data, attr_col) {
+  data[[attr_col]] <- as.factor(data[[attr_col]])
+  ggplot2::ggplot(
+    data,
+    ggplot2::aes_string(
+      x = attr_col,
+      y = "clonality_portion"
+    )
+  ) +
+    ggplot2::geom_violin(trim = FALSE, fill = "skyblue", alpha = 0.5) +
+    ggplot2::geom_boxplot(outlier.shape = NA, width = 0.1, alpha = 0.7) +
+    ggplot2::geom_jitter(width = 0, height = 0.1, size = 1.5, alpha = 0.6) +
+    ggplot2::theme_classic() +
+    ggplot2::labs(x = attr_col, y = "number of clonotypes") +
+    ggplot2::theme(
+      axis.text.x = ggplot2::element_text(
+        angle = 90, vjust = 0.5, hjust = 1, size = 10
+      ),
+      axis.text.y = ggplot2::element_text(
+        size = 12,
+        margin = ggplot2::margin(l = 10)
+      ),
+      axis.title.x = ggplot2::element_text(
+        size = 14,
+        margin = ggplot2::margin(t = 10)
+      ),
+      axis.title.y = ggplot2::element_text(size = 14),
+      legend.title = ggplot2::element_text(size = 14),
+      legend.text = ggplot2::element_text(size = 12)
+    )
+}

R/plot_clonal_prop_per_sample.R

@@ -0,0 +1,47 @@
+#' Plot clonal proportion per sample
+#'
+#' This function plots the clonal proportion per sample.
+#'
+#' @param data A data frame of clonal proportion per sample.
+#' @param attr_col Character. Attribute column name.
+#'
+#' @return A ggplot2 plot of clonal proportion per sample.
+#' @export
+#'
+#' @examples
+#' \dontrun{
+#' proj <- reTCR::get_study(id="PRJNA473147", attr_col="cmv_status")
+#' clonal_prop <- proj@clonality@clonal_prop
+#' reTCR::plot_clonal_prop_per_sample(clonal_prop, "cmv_status")
+#' }
+plot_clonal_prop_per_sample <- function(data, attr_col) {
+  ggplot2::ggplot(
+    data,
+    ggplot2::aes(
+      x = sample,
+      y = clonality_portion,
+      fill = !!rlang::sym(attr_col)
+    )
+  ) +
+    ggplot2::geom_bar(stat = "identity", position = ggplot2::position_dodge()) +
+    ggplot2::labs(x = "sample", y = "number of clonotypes") +
+    ggplot2::theme_classic() +
+    ggplot2::theme(
+      axis.text.x = ggplot2::element_text(
+        angle = 90, vjust = 0.5, hjust = 1, size = 10
+      ),
+      axis.text.y = ggplot2::element_text(
+        size = 12,
+        margin = ggplot2::margin(l = 10)
+      ),
+      axis.title.x = ggplot2::element_text(
+        size = 14,
+        margin = ggplot2::margin(t = 10)
+      ),
+      axis.title.y = ggplot2::element_text(size = 14),
+      legend.title = ggplot2::element_text(size = 14),
+      legend.text = ggplot2::element_text(size = 12)
+    ) +
+    ggplot2::scale_y_continuous(expand = c(0, 0)) +
+    ggplot2::scale_fill_brewer(palette = "Set1", name = attr_col)
+}

R/plot_clonotype_abundance.R

@@ -0,0 +1,60 @@
+#' Plot relative abundance for all clonotypes
+#'
+#' This function plots the relative abundance for all clonotypes.
+#'
+#' @param data A data frame of relative abundance for all clonotypes.
+#'
+#' @return A ggplot2 plot of relative abundance for all clonotypes.
+#' @export
+#'
+#' @examples
+#' \dontrun{
+#' proj <- reTCR::get_study(id="PRJNA473147", attr_col="cmv_status")
+#' reTCR::plot_clonotype_abundance(proj@clonality@abundance)
+#' }
+plot_clonotype_abundance <- function(data) {
+  label_order <- unique(data$clonotype_group)
+  colors <- RColorBrewer::brewer.pal(n = length(label_order), name = "Set1")
+
+  ggplot2::ggplot(
+    data,
+    ggplot2::aes(x = sample, y = relative_abundance, fill = clonotype_group)
+  ) +
+    ggplot2::geom_bar(
+      stat = "identity",
+      width = 0.5,
+      color = NA
+    ) +
+    ggplot2::scale_fill_manual(
+      values = colors,
+      breaks = label_order
+    ) +
+    ggplot2::labs(x = "sample", y = "clonotype frequency") +
+    ggplot2::theme_minimal() +
+    ggplot2::theme(
+      axis.text.x = ggplot2::element_text(
+        angle = 90,
+        vjust = 0.5,
+        hjust = 1,
+        size = 10
+      ),
+      axis.text.y = ggplot2::element_text(size = 10),
+      axis.title = ggplot2::element_text(size = 14),
+      legend.title = ggplot2::element_blank(),
+      legend.text = ggplot2::element_text(size = 12),
+      legend.position = "right",
+      panel.grid.major = ggplot2::element_blank(),
+      panel.grid.minor = ggplot2::element_blank(),
+      axis.line = ggplot2::element_line(color = "black"),
+      axis.text = ggplot2::element_text(margin = ggplot2::margin(t = 10)),
+      axis.ticks = ggplot2::element_line(color = "black", linewidth = 0.3),
+      strip.background = ggplot2::element_blank(),
+      strip.text = ggplot2::element_text(size = 12)
+    ) +
+    ggplot2::scale_x_discrete(
+      labels = function(x) {
+        stringr::str_wrap(x, width = 10)
+      }
+    ) +
+    ggplot2::scale_y_continuous(expand = c(0, 0))
+}

R/plot_diversity_index.R

@@ -0,0 +1,76 @@
+#' Plot diversity index metrics
+#'
+#' This function plots the diversity index metrics.
+#'
+#' @param data A data frame of diversity index metrics.
+#' @param index Character. Diversity index metric.
+#' @param attr_col Character. Attribute column name.
+#'
+#' @return A ggplot2 plot of diversity index metrics.
+#' @export
+#'
+#' @examples
+#' \dontrun{
+#' proj <- reTCR::get_study(id = "PRJNA473147", attr_col = "cmv_status")
+#' reTCR::plot_diversity_index(
+#'   proj@diversity@shannon,
+#'   "shannon_wiener_index",
+#'   "cmv_status"
+#' )
+#' }
+plot_diversity_index <- function(data, index, attr_col) {
+  label_order <- unique(data[[attr_col]])
+  colors <- RColorBrewer::brewer.pal(n = 3, name = "Set1")[c(2, 1)]
+
+  df_summary <- data %>%
+    dplyr::group_by(.data[[attr_col]]) %>%
+    dplyr::summarize(
+      mean_index = mean(.data[[index]], na.rm = TRUE),
+      sem_index = stats::sd(.data[[index]], na.rm = TRUE) / sqrt(dplyr::n())
+    )
+
+  ggplot2::ggplot(
+    df_summary,
+    ggplot2::aes(
+      x = .data[[attr_col]],
+      y = .data$mean_index,
+      fill = .data[[attr_col]]
+    )
+  ) +
+    ggplot2::geom_bar(position = "dodge", stat = "identity") +
+    ggplot2::geom_errorbar(
+      ggplot2::aes(
+        ymin = .data$mean_index - .data$sem_index,
+        ymax = .data$mean_index + .data$sem_index
+      ),
+      position = ggplot2::position_dodge(0.9),
+      width = 0.2
+    ) +
+    ggplot2::scale_fill_manual(
+      values = colors,
+      breaks = label_order,
+      name = attr_col
+    ) +
+    ggplot2::labs(x = attr_col, y = index) +
+    ggplot2::theme_minimal() +
+    ggplot2::theme(
+      axis.text.x = ggplot2::element_text(
+        angle = 90, vjust = 0.5, hjust = 1, size = 10
+      ),
+      axis.text.y = ggplot2::element_text(size = 10),
+      axis.title = ggplot2::element_text(size = 14),
+      legend.title = ggplot2::element_blank(),
+      legend.text = ggplot2::element_text(size = 12),
+      legend.position = "right",
+      panel.grid.major = ggplot2::element_blank(),
+      panel.grid.minor = ggplot2::element_blank(),
+      axis.line = ggplot2::element_line(color = "black"),
+      axis.text = ggplot2::element_text(margin = ggplot2::margin(t = 10)),
+      axis.ticks = ggplot2::element_line(color = "black", size = 0.3),
+      strip.background = ggplot2::element_blank(),
+      strip.text = ggplot2::element_text(size = 12)
+    ) +
+    ggplot2::scale_x_discrete(
+      labels = function(x) stringr::str_wrap(x, width = 10)
+    )
+}

R/plot_motif_counts.R

@@ -10,7 +10,7 @@
 #'
 #' @examples
 #' \dontrun{
-#' proj <- reTCR::get_study(id = "PRJNA473147")
+#' proj <- reTCR::get_study(id="PRJNA473147", attr_col="cmv_status")
 #' reTCR::plot_motif_counts(proj@motif@aa_motif_count, 20)
 #' }
 plot_motif_counts <- function(data, threshold = 60) {
@@ -29,7 +29,11 @@ plot_motif_counts <- function(data, threshold = 60) {
     )
   ) +
     ggplot2::geom_jitter(width = 0.2, size = 3) +
-    ggplot2::labs(x = "Amino acid motif", y = "Count", color = "CMV Status") +
+    ggplot2::labs(
+      x = "Amino acid motif",
+      y = "Count",
+      color = "CMV Status"
+    ) +
     ggplot2::theme_bw() +
     ggplot2::theme(
       axis.text.x = ggplot2::element_text(

R/plot_reads_group_abundance.R

@@ -0,0 +1,62 @@
+#' Plot reads group abundance for top 100 clonotypes
+#'
+#' This function plots reads group abundance for top 100 clonotypes.
+#'
+#' @param data A data frame of relative abundance for top 100 clonotypes.
+#'
+#' @return A ggplot2 plot of reads group abundance for top 100 clonotypes.
+#' @export
+#'
+#' @examples
+#' \dontrun{
+#' proj <- reTCR::get_study(id = "PRJNA473147", attr_col = "cmv_status")
+#' reTCR::plot_reads_group_abundance(proj@clonality@abundance_top)
+#' }
+plot_reads_group_abundance <- function(data) {
+  label_order <- unique(data$reads_group)
+  colors <- RColorBrewer::brewer.pal(n = length(label_order), name = "Set1")
+
+  df_counts <- data %>%
+    dplyr::filter(reads_group != "None") %>%
+    dplyr::group_by(sample, reads_group) %>%
+    dplyr::summarize(count = dplyr::n()) %>%
+    dplyr::ungroup()
+
+  ggplot2::ggplot(
+    df_counts,
+    ggplot2::aes(
+      x = sample,
+      y = count,
+      fill = reads_group
+    )
+  ) +
+    ggplot2::geom_bar(position = "stack", stat = "identity", width = 0.7) +
+    ggplot2::scale_fill_manual(
+      values = colors,
+      breaks = label_order,
+      name = "reads_group"
+    ) +
+    ggplot2::labs(x = "sample", y = "number of clonotypes") +
+    ggplot2::theme_minimal() +
+    ggplot2::theme(
+      axis.text.x = ggplot2::element_text(
+        angle = 90, vjust = 0.5, hjust = 1, size = 10
+      ),
+      axis.text.y = ggplot2::element_text(size = 10),
+      axis.title = ggplot2::element_text(size = 14),
+      legend.title = ggplot2::element_blank(),
+      legend.text = ggplot2::element_text(size = 12),
+      legend.position = "right",
+      panel.grid.major = ggplot2::element_blank(),
+      panel.grid.minor = ggplot2::element_blank(),
+      axis.line = ggplot2::element_line(color = "black"),
+      axis.text = ggplot2::element_text(margin = ggplot2::margin(t = 10)),
+      axis.ticks = ggplot2::element_line(color = "black", size = 0.3),
+      strip.background = ggplot2::element_blank(),
+      strip.text = ggplot2::element_text(size = 12)
+    ) +
+    ggplot2::scale_x_discrete(
+      labels = function(x) stringr::str_wrap(x, width = 10)
+    ) +
+    ggplot2::scale_y_continuous(expand = c(0, 0))
+}