Add platemap visualization

5_analyze_data/nbconverted/visualize_platemaps.r

@@ -0,0 +1,139 @@
+suppressPackageStartupMessages(library(dplyr))
+suppressPackageStartupMessages(library(ggplot2))
+suppressPackageStartupMessages(library(platetools))
+
+# Set directories
+output_dir <- "platemap_figures"
+
+input_platemap_dir <- file.path(
+    "..", 
+    "CellProfiler_pipelines",
+    "Metadata"
+)
+
+input_feature_dir <- file.path(
+    "..",
+    "4_processing_features",
+    "data"
+)
+
+plates <- c(
+    "plate_1" = "platemap_NF1_CP.csv",
+    "plate_2" = "platemap_NF1_CP_Plate2.csv"
+)
+
+# Corresponding to feature directory per plate
+plate_feature_dir <- c(
+    "plate_1" = "Plate1",
+    "plate_2" = "Plate2"
+)
+
+plates
+
+platemap_dfs <- list()
+for (platemap in names(plates)) {
+    platemap_file <- file.path(
+        input_platemap_dir,
+        plates[[platemap]]
+    )
+
+    platemap_dfs[[platemap]] <- readr::read_csv(
+        platemap_file,
+        col_types = readr::cols(
+            .default = "c"
+        )
+    )
+
+    print(platemap)
+    print(dim(platemap_dfs[[platemap]]))
+}
+
+cell_count_platemaps <- list()
+for (plate in names(plates)) {
+    plate_dir <- plate_feature_dir[[plate]]
+
+    feature_file <- file.path(
+        input_feature_dir,
+        plate_dir,
+        "CellProfiler",
+        "nf1_sc_norm_fs_cellprofiler.csv.gz"
+    )
+
+    # Currently, plate 2 cell count does not exist in main
+    # Skip for now
+    if (plate == "plate_2") {
+        cell_count_platemaps[[plate]] <- platemap_dfs[[plate]]
+        break
+    }
+
+    # Update the platemap to contain cell counts
+    cell_count_platemaps[[plate]] <- readr::read_csv(
+        feature_file,
+        col_types = readr::cols(
+            .default = "c"
+        )
+    ) %>%
+        dplyr::group_by(Metadata_Well) %>%
+        dplyr::count() %>%
+        dplyr::full_join(platemap_dfs[[plate]], by = c("Metadata_Well" = "well_position")) %>%
+        dplyr::rename(cell_count = n, well_position = Metadata_Well) %>%
+        dplyr::ungroup()
+}
+
+for (plate in names(plates)) {
+    output_fig_genotype <- file.path(
+        output_dir,
+        paste0("genotype_platemap_", plate, ".png")
+    )
+
+    plate_replicate_gg <- (
+        platetools::raw_map(
+            data = cell_count_platemaps[[plate]]$genotype,
+            well = cell_count_platemaps[[plate]]$well_position,
+            plate = 96,
+            size = 8
+        )
+        + scale_fill_discrete(name="Genotype")
+        + ggtitle(paste("NF1 Schwann cell", plate))
+    )
+
+
+    ggsave(
+        output_fig_genotype,
+        plate_replicate_gg,
+        dpi = 500,
+        height = 3.5,
+        width = 6
+    )
+
+    print(plate_replicate_gg)
+
+    # Plate 1 contains cell counts, add here while only 1 plate has counts
+    if (plate == "plate_1") {
+        output_fig_genotype <- file.path(
+            output_dir,
+            paste0("cell_count_platemap_", plate, ".png")
+        )
+
+        cell_count_gg <- (
+            platetools::raw_map(
+                data = cell_count_platemaps[[plate]]$cell_count,
+                well = cell_count_platemaps[[plate]]$well_position,
+                plate = 96,
+                size = 8
+            )
+            + scale_fill_continuous(name="Cell Count")
+            + ggtitle(paste("NF1 Schwann cell", plate))
+        )
+
+        ggsave(
+            output_fig_genotype,
+            cell_count_gg,
+            dpi = 500,
+            height = 3.5,
+            width = 6
+        )
+
+        print(cell_count_gg)
+    }
+}

5_analyze_data/visualize_env.yml

@@ -24,5 +24,6 @@ dependencies:
   - conda-forge::r-reshape2
   - conda-forge::r-cowplot
   - conda-forge::r-patchwork
+  - conda-forge::r-platetools
   - conda-forge::r-pwr
   - bioconda::bioconductor-complexheatmap