7.figures/SupplementaryFigure_tsne_figures.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "f53b8f8c-2979-4f55-8160-300c4b1898ee",
   "metadata": {},
   "source": [
    "## t-SNE visualization\n",
    "\n",
    "Generate a stand-alone figure for each perplexity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "fd4b9fec-a0f3-4674-90f3-c86be2d1aa60",
   "metadata": {},
   "outputs": [],
   "source": [
    "suppressPackageStartupMessages(library(dplyr))\n",
    "suppressPackageStartupMessages(library(ggplot2))\n",
    "suppressPackageStartupMessages(library(patchwork))\n",
    "\n",
    "# Load variables important for plotting (e.g., themes, phenotypes, etc.)\n",
    "source(\"themes.r\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "ae4e03b9-6364-41ea-97bf-51bf295b17cb",
   "metadata": {},
   "outputs": [],
   "source": [
    "figure_dir <- \"figures\"\n",
    "\n",
    "output_sup_fig_tsne <- file.path(\n",
    "    figure_dir,\n",
    "    \"supplementary\",\n",
    "    \"supplementary_tsne_figure.png\"\n",
    ")\n",
    "\n",
    "list_of_perplexities <- c(2, 10, 15, 30, 40, 60, 80, 100, 150, 300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "13d4f63f-ac8a-45d8-a9a7-fca82ce98bdc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1] 85860     7\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<table class=\"dataframe\">\n",
       "<caption>A tibble: 6 × 7</caption>\n",
       "<thead>\n",
       "\t<tr><th scope=col>Mitocheck_Phenotypic_Class</th><th scope=col>Feature_Type</th><th scope=col>Cell_UUID</th><th scope=col>tsne_x</th><th scope=col>tsne_y</th><th scope=col>Mitocheck_Plot_Label</th><th scope=col>perplexity</th></tr>\n",
       "\t<tr><th scope=col>&lt;chr&gt;</th><th scope=col>&lt;fct&gt;</th><th scope=col>&lt;chr&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;fct&gt;</th><th scope=col>&lt;dbl&gt;</th></tr>\n",
       "</thead>\n",
       "<tbody>\n",
       "\t<tr><td>Large</td><td>CellProfiler</td><td>21da27ab-873a-41f4-ab98-49170cae9a2d</td><td>15.23036</td><td>-30.34632</td><td>Large</td><td>2</td></tr>\n",
       "\t<tr><td>Large</td><td>CellProfiler</td><td>82f7949b-4ea2-45c8-8dd9-7854caf49077</td><td>39.67961</td><td>-47.58160</td><td>Large</td><td>2</td></tr>\n",
       "\t<tr><td>Large</td><td>CellProfiler</td><td>cec7234f-fe35-4411-aded-f8112bb31219</td><td>14.80523</td><td>-30.46914</td><td>Large</td><td>2</td></tr>\n",
       "\t<tr><td>Large</td><td>CellProfiler</td><td>43d9e7c9-c9ec-45ce-8820-048bfb896989</td><td>29.56186</td><td>-54.32789</td><td>Large</td><td>2</td></tr>\n",
       "\t<tr><td>Large</td><td>CellProfiler</td><td>63ce6652-338e-4afd-9c77-dbc0e903bf92</td><td>21.11539</td><td>-55.36263</td><td>Large</td><td>2</td></tr>\n",
       "\t<tr><td>Large</td><td>CellProfiler</td><td>08c1dd64-a50a-470a-8fac-3df28fd3b825</td><td>40.62339</td><td>-48.02852</td><td>Large</td><td>2</td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "A tibble: 6 × 7\n",
       "\\begin{tabular}{lllllll}\n",
       " Mitocheck\\_Phenotypic\\_Class & Feature\\_Type & Cell\\_UUID & tsne\\_x & tsne\\_y & Mitocheck\\_Plot\\_Label & perplexity\\\\\n",
       " <chr> & <fct> & <chr> & <dbl> & <dbl> & <fct> & <dbl>\\\\\n",
       "\\hline\n",
       "\t Large & CellProfiler & 21da27ab-873a-41f4-ab98-49170cae9a2d & 15.23036 & -30.34632 & Large & 2\\\\\n",
       "\t Large & CellProfiler & 82f7949b-4ea2-45c8-8dd9-7854caf49077 & 39.67961 & -47.58160 & Large & 2\\\\\n",
       "\t Large & CellProfiler & cec7234f-fe35-4411-aded-f8112bb31219 & 14.80523 & -30.46914 & Large & 2\\\\\n",
       "\t Large & CellProfiler & 43d9e7c9-c9ec-45ce-8820-048bfb896989 & 29.56186 & -54.32789 & Large & 2\\\\\n",
       "\t Large & CellProfiler & 63ce6652-338e-4afd-9c77-dbc0e903bf92 & 21.11539 & -55.36263 & Large & 2\\\\\n",
       "\t Large & CellProfiler & 08c1dd64-a50a-470a-8fac-3df28fd3b825 & 40.62339 & -48.02852 & Large & 2\\\\\n",
       "\\end{tabular}\n"
      ],
      "text/markdown": [
       "\n",
       "A tibble: 6 × 7\n",
       "\n",
       "| Mitocheck_Phenotypic_Class &lt;chr&gt; | Feature_Type &lt;fct&gt; | Cell_UUID &lt;chr&gt; | tsne_x &lt;dbl&gt; | tsne_y &lt;dbl&gt; | Mitocheck_Plot_Label &lt;fct&gt; | perplexity &lt;dbl&gt; |\n",
       "|---|---|---|---|---|---|---|\n",
       "| Large | CellProfiler | 21da27ab-873a-41f4-ab98-49170cae9a2d | 15.23036 | -30.34632 | Large | 2 |\n",
       "| Large | CellProfiler | 82f7949b-4ea2-45c8-8dd9-7854caf49077 | 39.67961 | -47.58160 | Large | 2 |\n",
       "| Large | CellProfiler | cec7234f-fe35-4411-aded-f8112bb31219 | 14.80523 | -30.46914 | Large | 2 |\n",
       "| Large | CellProfiler | 43d9e7c9-c9ec-45ce-8820-048bfb896989 | 29.56186 | -54.32789 | Large | 2 |\n",
       "| Large | CellProfiler | 63ce6652-338e-4afd-9c77-dbc0e903bf92 | 21.11539 | -55.36263 | Large | 2 |\n",
       "| Large | CellProfiler | 08c1dd64-a50a-470a-8fac-3df28fd3b825 | 40.62339 | -48.02852 | Large | 2 |\n",
       "\n"
      ],
      "text/plain": [
       "  Mitocheck_Phenotypic_Class Feature_Type Cell_UUID                           \n",
       "1 Large                      CellProfiler 21da27ab-873a-41f4-ab98-49170cae9a2d\n",
       "2 Large                      CellProfiler 82f7949b-4ea2-45c8-8dd9-7854caf49077\n",
       "3 Large                      CellProfiler cec7234f-fe35-4411-aded-f8112bb31219\n",
       "4 Large                      CellProfiler 43d9e7c9-c9ec-45ce-8820-048bfb896989\n",
       "5 Large                      CellProfiler 63ce6652-338e-4afd-9c77-dbc0e903bf92\n",
       "6 Large                      CellProfiler 08c1dd64-a50a-470a-8fac-3df28fd3b825\n",
       "  tsne_x   tsne_y    Mitocheck_Plot_Label perplexity\n",
       "1 15.23036 -30.34632 Large                2         \n",
       "2 39.67961 -47.58160 Large                2         \n",
       "3 14.80523 -30.46914 Large                2         \n",
       "4 29.56186 -54.32789 Large                2         \n",
       "5 21.11539 -55.36263 Large                2         \n",
       "6 40.62339 -48.02852 Large                2         "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Load tsne coordinates and process\n",
    "tsne_dir <- file.path(\"../3.evaluate_model/evaluations/\")\n",
    "tsne_file <- file.path(tsne_dir, \"tsne_embeddings.tsv.gz\")\n",
    "\n",
    "tsne_df <- readr::read_tsv(\n",
    "    tsne_file,\n",
    "    col_types = readr::cols(\n",
    "        .default = \"c\",\n",
    "        tsne_x = \"d\",\n",
    "        tsne_y = \"d\",\n",
    "        perplexity = \"d\"\n",
    "    )\n",
    ") %>%\n",
    "    # Note, we define focus_phenotypes in themes.r\n",
    "    dplyr::mutate(Mitocheck_Plot_Label = if_else(\n",
    "        Mitocheck_Phenotypic_Class %in% focus_phenotypes,\n",
    "        Mitocheck_Phenotypic_Class,\n",
    "        \"Other\"\n",
    "    ))\n",
    "\n",
    "tsne_df$Mitocheck_Plot_Label <-\n",
    "    dplyr::recode_factor(tsne_df$Mitocheck_Plot_Label, !!!focus_phenotype_labels)\n",
    "\n",
    "tsne_df$Feature_Type <-\n",
    "    dplyr::recode_factor(tsne_df$feature_group, !!!facet_labels)\n",
    "\n",
    "# Select certain columns\n",
    "tsne_df <- tsne_df %>%\n",
    "    dplyr::select(\n",
    "        Mitocheck_Phenotypic_Class,\n",
    "        Feature_Type,\n",
    "        Cell_UUID,\n",
    "        tsne_x,\n",
    "        tsne_y,\n",
    "        Mitocheck_Plot_Label,\n",
    "        perplexity\n",
    "    )\n",
    "\n",
    "print(dim(tsne_df))\n",
    "head(tsne_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "ab165476-e8f0-4d80-ab3b-8ff6890f8867",
   "metadata": {},
   "outputs": [],
   "source": [
    "for(perplexity in list_of_perplexities){\n",
    "    output_tsne_figure_file <- file.path(\n",
    "        \"figures\",\n",
    "        \"tsne\",\n",
    "        paste0(\"tsne_figure_perplexity_\", perplexity, \".png\")\n",
    "    )\n",
    "\n",
    "    # Create a background dataset to show in greyed color across all facets\n",
    "    tsne_focus_df <- tsne_df %>%\n",
    "        dplyr::filter(perplexity == !!perplexity) %>%\n",
    "        dplyr::filter(Mitocheck_Phenotypic_Class %in% focus_phenotypes)\n",
    "    \n",
    "    tsne_other_df <- tsne_df %>%\n",
    "        dplyr::filter(perplexity == !!perplexity) %>%\n",
    "        dplyr::filter(!Mitocheck_Phenotypic_Class %in% focus_phenotypes)\n",
    "    \n",
    "    # Custom function for name repair\n",
    "    name_repair_function <- function(names) {\n",
    "      names[1] <- paste0(names[1], \"_original\")\n",
    "      return(names)\n",
    "    }\n",
    "    \n",
    "    df_focus_background <- tidyr::crossing(\n",
    "        tsne_df %>% dplyr::filter(perplexity == !!perplexity),\n",
    "        Mitocheck_Phenotypic_Class = unique(tsne_focus_df$Mitocheck_Phenotypic_Class),\n",
    "        .name_repair = name_repair_function\n",
    "    )\n",
    "    \n",
    "    df_background <- tidyr::crossing(\n",
    "        tsne_df %>% dplyr::filter(perplexity == !!perplexity),\n",
    "        Mitocheck_Phenotypic_Class = unique(tsne_other_df$Mitocheck_Phenotypic_Class),\n",
    "        .name_repair = name_repair_function\n",
    "    )\n",
    "    \n",
    "    tsne_fig_gg <- (\n",
    "        ggplot(\n",
    "            tsne_focus_df,\n",
    "            aes(x = tsne_x, y = tsne_y)\n",
    "        )\n",
    "        + geom_point(\n",
    "            data = df_focus_background,\n",
    "            color = \"lightgray\",\n",
    "            size = 0.05,\n",
    "            alpha = 0.4\n",
    "        )\n",
    "        + geom_point(\n",
    "            aes(color = Mitocheck_Phenotypic_Class),\n",
    "            size = 0.05\n",
    "        )\n",
    "        + facet_grid(\"Feature_Type~Mitocheck_Phenotypic_Class\")\n",
    "        + theme_bw()\n",
    "        + scale_color_manual(\n",
    "            \"Phenotype\",\n",
    "            values = focus_phenotype_colors,\n",
    "            labels = focus_phenotype_labels\n",
    "        )\n",
    "        + phenotypic_ggplot_theme\n",
    "        + guides(\n",
    "            color = guide_legend(\n",
    "                override.aes = list(size = 2)\n",
    "            )\n",
    "        )\n",
    "        + labs(x = \"tSNE 1\", y = \"tSNE 2\")\n",
    "    )\n",
    "    \n",
    "    tsne_fig_other_gg <- (\n",
    "        ggplot(\n",
    "            tsne_other_df,\n",
    "            aes(x = tsne_x, y = tsne_y)\n",
    "        )\n",
    "        + geom_point(\n",
    "            data = df_background,\n",
    "            color = \"lightgray\",\n",
    "            size = 0.1,\n",
    "            alpha = 0.4\n",
    "        )\n",
    "        + geom_point(\n",
    "            aes(color = Mitocheck_Phenotypic_Class),\n",
    "            size = 0.1\n",
    "        )\n",
    "        + facet_grid(\"Feature_Type~Mitocheck_Phenotypic_Class\")\n",
    "        + theme_bw()\n",
    "        + phenotypic_ggplot_theme\n",
    "        + guides(\n",
    "            color = guide_legend(\n",
    "                override.aes = list(size = 2)\n",
    "            )\n",
    "        )\n",
    "        + labs(x = \"tSNE 1\", y = \"tSNE 2\")\n",
    "        + theme(\n",
    "            legend.position = \"none\",\n",
    "            strip.text = element_text(size = 8.5),\n",
    "        )\n",
    "    )\n",
    "    \n",
    "    nested_plot <- (\n",
    "        tsne_fig_gg | plot_spacer()\n",
    "    ) + plot_layout(widths = c(3, 1.35))\n",
    "    \n",
    "    tsne_full_sup_fig <- (\n",
    "        nested_plot / tsne_fig_other_gg\n",
    "    ) + plot_layout(heights = c(1, 1))\n",
    "    \n",
    "    ggsave(output_tsne_figure_file, dpi = 500, height = 10, width = 13)\n",
    "    \n",
    "    tsne_full_sup_fig\n",
    "}"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "R [conda env:phenotype_figures] *",
   "language": "R",
   "name": "conda-env-phenotype_figures-r"
  },
  "language_info": {
   "codemirror_mode": "r",
   "file_extension": ".r",
   "mimetype": "text/x-r-source",
   "name": "R",
   "pygments_lexer": "r",
   "version": "4.4.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}