05_Tutorial_Downstream_Analysis_R.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Downstream Analysis - R"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Note**: This tutorial was originally written using R Markdown. If you would like to run or edit the R Markdown document, it can be found [here](https://github.com/morriso1/clonedetective/blob/master/rcode/rmd/downstream_analysis_R.Rmd)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Recap - background and experimental aims\n",
    "\n",
    "Our original demo dataset consisted of 67 images of the Drosophila intestine from a MARCM genetic cell lineage experiment. GFP (C1 channel) marks genetic recombination events, and so clusters of GFP positive cells define \"clones\" of daughter cells that over time have grown out from a single parent cell. The remaining images include DAPI (C0 channel), which ubiquitously marks all cell nuclei; PDM1 (C2 channel), which labels EnteroCyte (EC) cells; and Prospero (C3 channel), which labels EnteroEndocrine cells.\n",
    "\n",
    "We aimed to investigate: 1) Does gene of interest regulate cell proliferation? 2) Does our gene of interest control cell differentiation? 3) Does gene of interest effect local tissue cell arrangement and structure?\n",
    "\n",
    "We will tackle each of these experimental aims using data previously generated by clonedetective.\n",
    "\n",
    "## Loading in data and libraries\n",
    "\n",
    "Now we wrangle our data and plot results! Below, we do this using R tidyverse, cowplot and emmeans packages."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#all_slow\n",
    "#hide_output\n",
    "\n",
    "knitr::opts_chunk$set(echo = TRUE)\n",
    "# stats\n",
    "library(MASS)\n",
    "library(emmeans)\n",
    "library(lme4)\n",
    "library(glmmTMB)\n",
    "\n",
    "# tidying up stat outputs\n",
    "library(broom.mixed)\n",
    "\n",
    "# data wrangling\n",
    "library(tidyverse)\n",
    "\n",
    "# ehlp with plotting\n",
    "library(cowplot)\n",
    "library(ggpubr)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "\u001b[36m──\u001b[39m \u001b[1m\u001b[1mColumn specification\u001b[1m\u001b[22m \u001b[36m──────────────────────────────────────────────────────────────────────────────\u001b[39m\n",
      "cols(\n",
      "  .default = col_double(),\n",
      "  int_img = \u001b[31mcol_character()\u001b[39m,\n",
      "  seg_ch = \u001b[31mcol_character()\u001b[39m,\n",
      "  int_img_ch = \u001b[31mcol_character()\u001b[39m,\n",
      "  seg_img = \u001b[31mcol_character()\u001b[39m,\n",
      "  GFP_clonepos = \u001b[33mcol_logical()\u001b[39m,\n",
      "  EC_pos = \u001b[33mcol_logical()\u001b[39m,\n",
      "  EE_pos = \u001b[33mcol_logical()\u001b[39m,\n",
      "  ECEE_pos = \u001b[33mcol_logical()\u001b[39m,\n",
      "  ISCorEB_pos = \u001b[33mcol_logical()\u001b[39m,\n",
      "  preEC_pos = \u001b[33mcol_logical()\u001b[39m\n",
      ")\n",
      "\u001b[36mℹ\u001b[39m Use \u001b[30m\u001b[47m\u001b[30m\u001b[47m`spec()`\u001b[47m\u001b[30m\u001b[49m\u001b[39m for the full column specifications.\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# hide_output\n",
    "# absolute path to data\n",
    "path_to_data <- \"/Users/ottomorris/Documents/my_python_packages/clonedetective/data/example_results.csv\"\n",
    "\n",
    "count_df <- read_csv(path_to_data)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"dataframe\">\n",
       "<caption>A tibble: 6 × 32</caption>\n",
       "<thead>\n",
       "\t<tr><th scope=col>int_img</th><th scope=col>label</th><th scope=col>GFP_clone</th><th scope=col>GFPneg_nc</th><th scope=col>GFPpos_nc</th><th scope=col>tot_nc</th><th scope=col>ECneg_nc</th><th scope=col>ECpos_nc</th><th scope=col>EEneg_nc</th><th scope=col>EEpos_nc</th><th scope=col>⋯</th><th scope=col>centroid-1</th><th scope=col>eccentricity</th><th scope=col>area_um2</th><th scope=col>GFP_clonepos</th><th scope=col>total_intensity</th><th scope=col>EC_pos</th><th scope=col>EE_pos</th><th scope=col>ECEE_pos</th><th scope=col>ISCorEB_pos</th><th scope=col>preEC_pos</th></tr>\n",
       "\t<tr><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;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>⋯</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;lgl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;lgl&gt;</th><th scope=col>&lt;lgl&gt;</th><th scope=col>&lt;lgl&gt;</th><th scope=col>&lt;lgl&gt;</th><th scope=col>&lt;lgl&gt;</th></tr>\n",
       "</thead>\n",
       "<tbody>\n",
       "\t<tr><td>a1g01p1</td><td>11</td><td>0</td><td>2</td><td>2</td><td>4</td><td>1</td><td>3</td><td>4</td><td>0</td><td>⋯</td><td> 89.55212</td><td>0.4408815</td><td>73.28063</td><td>FALSE</td><td>2196457</td><td>TRUE</td><td>FALSE</td><td>FALSE</td><td>FALSE</td><td>FALSE</td></tr>\n",
       "\t<tr><td>a1g01p1</td><td>11</td><td>0</td><td>2</td><td>2</td><td>4</td><td>1</td><td>3</td><td>4</td><td>0</td><td>⋯</td><td> 89.55212</td><td>0.4408815</td><td>73.28063</td><td>FALSE</td><td>  52757</td><td>TRUE</td><td>FALSE</td><td>FALSE</td><td>FALSE</td><td>FALSE</td></tr>\n",
       "\t<tr><td>a1g01p1</td><td>11</td><td>0</td><td>2</td><td>2</td><td>4</td><td>1</td><td>3</td><td>4</td><td>0</td><td>⋯</td><td> 89.55212</td><td>0.4408815</td><td>73.28063</td><td>FALSE</td><td>2262648</td><td>TRUE</td><td>FALSE</td><td>FALSE</td><td>FALSE</td><td>FALSE</td></tr>\n",
       "\t<tr><td>a1g01p1</td><td>11</td><td>0</td><td>2</td><td>2</td><td>4</td><td>1</td><td>3</td><td>4</td><td>0</td><td>⋯</td><td> 89.55212</td><td>0.4408815</td><td>73.28063</td><td>FALSE</td><td>  67331</td><td>TRUE</td><td>FALSE</td><td>FALSE</td><td>FALSE</td><td>FALSE</td></tr>\n",
       "\t<tr><td>a1g01p1</td><td>12</td><td>0</td><td>4</td><td>0</td><td>4</td><td>1</td><td>3</td><td>4</td><td>0</td><td>⋯</td><td>486.79847</td><td>0.4652274</td><td>98.69063</td><td>FALSE</td><td>4113516</td><td>TRUE</td><td>FALSE</td><td>FALSE</td><td>FALSE</td><td>FALSE</td></tr>\n",
       "\t<tr><td>a1g01p1</td><td>12</td><td>0</td><td>4</td><td>0</td><td>4</td><td>1</td><td>3</td><td>4</td><td>0</td><td>⋯</td><td>486.79847</td><td>0.4652274</td><td>98.69063</td><td>FALSE</td><td>  99772</td><td>TRUE</td><td>FALSE</td><td>FALSE</td><td>FALSE</td><td>FALSE</td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "A tibble: 6 × 32\n",
       "\\begin{tabular}{lllllllllllllllllllll}\n",
       " int\\_img & label & GFP\\_clone & GFPneg\\_nc & GFPpos\\_nc & tot\\_nc & ECneg\\_nc & ECpos\\_nc & EEneg\\_nc & EEpos\\_nc & ⋯ & centroid-1 & eccentricity & area\\_um2 & GFP\\_clonepos & total\\_intensity & EC\\_pos & EE\\_pos & ECEE\\_pos & ISCorEB\\_pos & preEC\\_pos\\\\\n",
       " <chr> & <dbl> & <dbl> & <dbl> & <dbl> & <dbl> & <dbl> & <dbl> & <dbl> & <dbl> & ⋯ & <dbl> & <dbl> & <dbl> & <lgl> & <dbl> & <lgl> & <lgl> & <lgl> & <lgl> & <lgl>\\\\\n",
       "\\hline\n",
       "\t a1g01p1 & 11 & 0 & 2 & 2 & 4 & 1 & 3 & 4 & 0 & ⋯ &  89.55212 & 0.4408815 & 73.28063 & FALSE & 2196457 & TRUE & FALSE & FALSE & FALSE & FALSE\\\\\n",
       "\t a1g01p1 & 11 & 0 & 2 & 2 & 4 & 1 & 3 & 4 & 0 & ⋯ &  89.55212 & 0.4408815 & 73.28063 & FALSE &   52757 & TRUE & FALSE & FALSE & FALSE & FALSE\\\\\n",
       "\t a1g01p1 & 11 & 0 & 2 & 2 & 4 & 1 & 3 & 4 & 0 & ⋯ &  89.55212 & 0.4408815 & 73.28063 & FALSE & 2262648 & TRUE & FALSE & FALSE & FALSE & FALSE\\\\\n",
       "\t a1g01p1 & 11 & 0 & 2 & 2 & 4 & 1 & 3 & 4 & 0 & ⋯ &  89.55212 & 0.4408815 & 73.28063 & FALSE &   67331 & TRUE & FALSE & FALSE & FALSE & FALSE\\\\\n",
       "\t a1g01p1 & 12 & 0 & 4 & 0 & 4 & 1 & 3 & 4 & 0 & ⋯ & 486.79847 & 0.4652274 & 98.69063 & FALSE & 4113516 & TRUE & FALSE & FALSE & FALSE & FALSE\\\\\n",
       "\t a1g01p1 & 12 & 0 & 4 & 0 & 4 & 1 & 3 & 4 & 0 & ⋯ & 486.79847 & 0.4652274 & 98.69063 & FALSE &   99772 & TRUE & FALSE & FALSE & FALSE & FALSE\\\\\n",
       "\\end{tabular}\n"
      ],
      "text/markdown": [
       "\n",
       "A tibble: 6 × 32\n",
       "\n",
       "| int_img &lt;chr&gt; | label &lt;dbl&gt; | GFP_clone &lt;dbl&gt; | GFPneg_nc &lt;dbl&gt; | GFPpos_nc &lt;dbl&gt; | tot_nc &lt;dbl&gt; | ECneg_nc &lt;dbl&gt; | ECpos_nc &lt;dbl&gt; | EEneg_nc &lt;dbl&gt; | EEpos_nc &lt;dbl&gt; | ⋯ ⋯ | centroid-1 &lt;dbl&gt; | eccentricity &lt;dbl&gt; | area_um2 &lt;dbl&gt; | GFP_clonepos &lt;lgl&gt; | total_intensity &lt;dbl&gt; | EC_pos &lt;lgl&gt; | EE_pos &lt;lgl&gt; | ECEE_pos &lt;lgl&gt; | ISCorEB_pos &lt;lgl&gt; | preEC_pos &lt;lgl&gt; |\n",
       "|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n",
       "| a1g01p1 | 11 | 0 | 2 | 2 | 4 | 1 | 3 | 4 | 0 | ⋯ |  89.55212 | 0.4408815 | 73.28063 | FALSE | 2196457 | TRUE | FALSE | FALSE | FALSE | FALSE |\n",
       "| a1g01p1 | 11 | 0 | 2 | 2 | 4 | 1 | 3 | 4 | 0 | ⋯ |  89.55212 | 0.4408815 | 73.28063 | FALSE |   52757 | TRUE | FALSE | FALSE | FALSE | FALSE |\n",
       "| a1g01p1 | 11 | 0 | 2 | 2 | 4 | 1 | 3 | 4 | 0 | ⋯ |  89.55212 | 0.4408815 | 73.28063 | FALSE | 2262648 | TRUE | FALSE | FALSE | FALSE | FALSE |\n",
       "| a1g01p1 | 11 | 0 | 2 | 2 | 4 | 1 | 3 | 4 | 0 | ⋯ |  89.55212 | 0.4408815 | 73.28063 | FALSE |   67331 | TRUE | FALSE | FALSE | FALSE | FALSE |\n",
       "| a1g01p1 | 12 | 0 | 4 | 0 | 4 | 1 | 3 | 4 | 0 | ⋯ | 486.79847 | 0.4652274 | 98.69063 | FALSE | 4113516 | TRUE | FALSE | FALSE | FALSE | FALSE |\n",
       "| a1g01p1 | 12 | 0 | 4 | 0 | 4 | 1 | 3 | 4 | 0 | ⋯ | 486.79847 | 0.4652274 | 98.69063 | FALSE |   99772 | TRUE | FALSE | FALSE | FALSE | FALSE |\n",
       "\n"
      ],
      "text/plain": [
       "  int_img label GFP_clone GFPneg_nc GFPpos_nc tot_nc ECneg_nc ECpos_nc EEneg_nc\n",
       "1 a1g01p1 11    0         2         2         4      1        3        4       \n",
       "2 a1g01p1 11    0         2         2         4      1        3        4       \n",
       "3 a1g01p1 11    0         2         2         4      1        3        4       \n",
       "4 a1g01p1 11    0         2         2         4      1        3        4       \n",
       "5 a1g01p1 12    0         4         0         4      1        3        4       \n",
       "6 a1g01p1 12    0         4         0         4      1        3        4       \n",
       "  EEpos_nc ⋯ centroid-1 eccentricity area_um2 GFP_clonepos total_intensity\n",
       "1 0        ⋯  89.55212  0.4408815    73.28063 FALSE        2196457        \n",
       "2 0        ⋯  89.55212  0.4408815    73.28063 FALSE          52757        \n",
       "3 0        ⋯  89.55212  0.4408815    73.28063 FALSE        2262648        \n",
       "4 0        ⋯  89.55212  0.4408815    73.28063 FALSE          67331        \n",
       "5 0        ⋯ 486.79847  0.4652274    98.69063 FALSE        4113516        \n",
       "6 0        ⋯ 486.79847  0.4652274    98.69063 FALSE          99772        \n",
       "  EC_pos EE_pos ECEE_pos ISCorEB_pos preEC_pos\n",
       "1 TRUE   FALSE  FALSE    FALSE       FALSE    \n",
       "2 TRUE   FALSE  FALSE    FALSE       FALSE    \n",
       "3 TRUE   FALSE  FALSE    FALSE       FALSE    \n",
       "4 TRUE   FALSE  FALSE    FALSE       FALSE    \n",
       "5 TRUE   FALSE  FALSE    FALSE       FALSE    \n",
       "6 TRUE   FALSE  FALSE    FALSE       FALSE    "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "count_df %>% head()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## **Cell Proliferation - number of cells per clone**\n",
    "\n",
    "Our first aim is to determine whether there is a difference in the number of cells per clone in mutant vs control animals. `count_df` currently contains four rows for each cell - one for each of the four intensity image channels. Since we are presently interested in cell types and numbers, rather than the intensity values of individual cells, we can query count_df for a single intensity image channel. This ensures that `count_data` contains only a single row for each cell.\n",
    "\n",
    "For simplicity, we also filter out columns not required for counting the number and type of cells per clone.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "count_df <- count_df %>%\n",
    "  pivot_wider(\n",
    "    names_from = int_img_ch,\n",
    "    values_from = c(\"total_intensity\", \"mean_intensity\")\n",
    "  ) %>%\n",
    "  dplyr::select(matches(\"clone$|_pos$|int_img\"))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we groupby clones per image and sum to get the number of cells of each type per clone.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "`summarise()` has grouped output by 'int_img'. You can override using the `.groups` argument.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "count_df <- count_df %>%\n",
    "  group_by(int_img, GFP_clone) %>%\n",
    "  summarise(\n",
    "    EC = sum(EC_pos),\n",
    "    ECEE = sum(ECEE_pos),\n",
    "    preEC = sum(preEC_pos),\n",
    "    EE = sum(EE_pos),\n",
    "    ISCorEB = sum(ISCorEB_pos)\n",
    "  )\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To ease downstream wrangling and plotting, we update some of the column metadata.\n",
    "\n",
    "For example, when plotting, it will be helpful to label whether a cell is inside a clone (IC) or outside a clone (OC). This information is present within the GFP_clone column: 0 indicates if a cell is outside a clone, and \\>0 indicate if a cell is inside a clone.\n",
    "\n",
    "We also can extract metadata stored within the image names. We know that images are named 'a1' if from a wild-type (ctrl) animal, and 'a2' if from a mutant (mut) animal. The 'a\\\\dg\\\\d\\\\d' pattern of the image name identifies the gut (animal) from which an image is taken (e.g. 'a1g01' is ctrl animal gut 1) . Using this information, we can add 'genotype' and 'gut' columns.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "count_df <- count_df %>% mutate(\n",
    "  clone_status = as.factor(if_else(GFP_clone == 0, \"OC\", \"IC\")),\n",
    "  genotype = as.factor(if_else(str_detect(int_img, \"a1\"), \"ctrl\", \"mut\")),\n",
    "  gut = as_factor(str_extract(int_img, \"g\\\\d\\\\d\")),\n",
    "  total = EC + ECEE + preEC + EE + ISCorEB\n",
    ")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "count_df_tidy <-\n",
    "  count_df %>%\n",
    "  pivot_longer(cols = EC:ISCorEB, names_to = \"cell_type\", values_to = \"cell_num\") %>%\n",
    "  mutate(cell_type = factor(cell_type, levels = c(\"ECEE\", \"EE\", \"EC\", \"preEC\", \"ISCorEB\")))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We next generate a bar plot displaying cell numbers and types per clone. Here, we are only interested in plotting cells inside clones, and so we first filter on 'clone_status':\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWgAAAHgCAYAAACIBvdgAAAEDmlDQ1BrQ0dDb2xvclNwYWNl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",
      "text/plain": [
       "plot without title"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 240,
       "width": 180
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# colors for our bar plot\n",
    "col_pal <- c(\"#000000\", \"#95a5a6\", \"#F57171\", \"#539DC2\", \"#008b68\")\n",
    "\n",
    "stacked_bar_mean <- count_df_tidy %>%\n",
    "  filter(clone_status == \"IC\") %>%\n",
    "  ggplot(aes(fill = cell_type, y = cell_num, x = genotype)) +\n",
    "  geom_bar(\n",
    "    position = \"stack\",\n",
    "    stat = \"summary\",\n",
    "    fun = \"mean\"\n",
    "  ) +\n",
    "  theme_bw() +\n",
    "  scale_fill_manual(values = col_pal)\n",
    "\n",
    "options(repr.plot.width = 3, repr.plot.height = 4)\n",
    "stacked_bar_mean\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From this plot we can see that 'mutant' clones are on average bigger than 'ctrl' clones. This suggests that mutation of our gene of interest may affect cell proliferation. Since we have count data, let's fit a negative binomial regression using the 'MASS' package to test if this difference is significant.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"dataframe\">\n",
       "<caption>A tibble: 1 × 8</caption>\n",
       "<thead>\n",
       "\t<tr><th scope=col>null.deviance</th><th scope=col>df.null</th><th scope=col>logLik</th><th scope=col>AIC</th><th scope=col>BIC</th><th scope=col>deviance</th><th scope=col>df.residual</th><th scope=col>nobs</th></tr>\n",
       "\t<tr><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;int&gt;</th><th scope=col>&lt;logLik&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;int&gt;</th><th scope=col>&lt;int&gt;</th></tr>\n",
       "</thead>\n",
       "<tbody>\n",
       "\t<tr><td>1849.033</td><td>495</td><td>-1723.41</td><td>3456.821</td><td>3477.854</td><td>518.8748</td><td>492</td><td>496</td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "A tibble: 1 × 8\n",
       "\\begin{tabular}{llllllll}\n",
       " null.deviance & df.null & logLik & AIC & BIC & deviance & df.residual & nobs\\\\\n",
       " <dbl> & <int> & <logLik> & <dbl> & <dbl> & <dbl> & <int> & <int>\\\\\n",
       "\\hline\n",
       "\t 1849.033 & 495 & -1723.41 & 3456.821 & 3477.854 & 518.8748 & 492 & 496\\\\\n",
       "\\end{tabular}\n"
      ],
      "text/markdown": [
       "\n",
       "A tibble: 1 × 8\n",
       "\n",
       "| null.deviance &lt;dbl&gt; | df.null &lt;int&gt; | logLik &lt;logLik&gt; | AIC &lt;dbl&gt; | BIC &lt;dbl&gt; | deviance &lt;dbl&gt; | df.residual &lt;int&gt; | nobs &lt;int&gt; |\n",
       "|---|---|---|---|---|---|---|---|\n",
       "| 1849.033 | 495 | -1723.41 | 3456.821 | 3477.854 | 518.8748 | 492 | 496 |\n",
       "\n"
      ],
      "text/plain": [
       "  null.deviance df.null logLik   AIC      BIC      deviance df.residual nobs\n",
       "1 1849.033      495     -1723.41 3456.821 3477.854 518.8748 492         496 "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "m1 <- glm.nb(total ~ genotype * clone_status, count_df)\n",
    "glance(m1)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "However, this simple model ignores the fact that each clone is 'nested' inside a gut. Errors may, therefore, correlate within guts, and so we should try to take account of this using a mixed-effect model:\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"dataframe\">\n",
       "<caption>A tibble: 1 × 6</caption>\n",
       "<thead>\n",
       "\t<tr><th scope=col>sigma</th><th scope=col>logLik</th><th scope=col>AIC</th><th scope=col>BIC</th><th scope=col>deviance</th><th scope=col>df.residual</th></tr>\n",
       "\t<tr><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th></tr>\n",
       "</thead>\n",
       "<tbody>\n",
       "\t<tr><td>1</td><td>-1715.031</td><td>3442.062</td><td>3467.301</td><td>500.5485</td><td>490</td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "A tibble: 1 × 6\n",
       "\\begin{tabular}{llllll}\n",
       " sigma & logLik & AIC & BIC & deviance & df.residual\\\\\n",
       " <dbl> & <dbl> & <dbl> & <dbl> & <dbl> & <dbl>\\\\\n",
       "\\hline\n",
       "\t 1 & -1715.031 & 3442.062 & 3467.301 & 500.5485 & 490\\\\\n",
       "\\end{tabular}\n"
      ],
      "text/markdown": [
       "\n",
       "A tibble: 1 × 6\n",
       "\n",
       "| sigma &lt;dbl&gt; | logLik &lt;dbl&gt; | AIC &lt;dbl&gt; | BIC &lt;dbl&gt; | deviance &lt;dbl&gt; | df.residual &lt;dbl&gt; |\n",
       "|---|---|---|---|---|---|\n",
       "| 1 | -1715.031 | 3442.062 | 3467.301 | 500.5485 | 490 |\n",
       "\n"
      ],
      "text/plain": [
       "  sigma logLik    AIC      BIC      deviance df.residual\n",
       "1 1     -1715.031 3442.062 3467.301 500.5485 490        "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "m2 <- glmer.nb(total ~ genotype * clone_status + (1 | gut), count_df)\n",
    "glance(m2)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This model shows shows an improved (note the improved AIC score). However, the current negative binomial distribution does not account for the fact that clone cannot contain 0 cells! We, therefore, adjust our model to use a truncated negative binomial distribution.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"dataframe\">\n",
       "<caption>A tibble: 1 × 5</caption>\n",
       "<thead>\n",
       "\t<tr><th scope=col>sigma</th><th scope=col>logLik</th><th scope=col>AIC</th><th scope=col>BIC</th><th scope=col>df.residual</th></tr>\n",
       "\t<tr><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;int&gt;</th></tr>\n",
       "</thead>\n",
       "<tbody>\n",
       "\t<tr><td>0.608627</td><td>-1630.508</td><td>3273.017</td><td>3298.256</td><td>490</td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "A tibble: 1 × 5\n",
       "\\begin{tabular}{lllll}\n",
       " sigma & logLik & AIC & BIC & df.residual\\\\\n",
       " <dbl> & <dbl> & <dbl> & <dbl> & <int>\\\\\n",
       "\\hline\n",
       "\t 0.608627 & -1630.508 & 3273.017 & 3298.256 & 490\\\\\n",
       "\\end{tabular}\n"
      ],
      "text/markdown": [
       "\n",
       "A tibble: 1 × 5\n",
       "\n",
       "| sigma &lt;dbl&gt; | logLik &lt;dbl&gt; | AIC &lt;dbl&gt; | BIC &lt;dbl&gt; | df.residual &lt;int&gt; |\n",
       "|---|---|---|---|---|\n",
       "| 0.608627 | -1630.508 | 3273.017 | 3298.256 | 490 |\n",
       "\n"
      ],
      "text/plain": [
       "  sigma    logLik    AIC      BIC      df.residual\n",
       "1 0.608627 -1630.508 3273.017 3298.256 490        "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "m3 <- glmmTMB(total ~ genotype * clone_status + (1 | gut), count_df, family = \"truncated_nbinom2\")\n",
    "glance(m3)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This further improves the fit of our model.\n",
    "\n",
    "We next move to test our desired contrast - do mutant clones possess fewer cells than control clones?\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"dataframe\">\n",
       "<caption>A tibble: 2 × 9</caption>\n",
       "<thead>\n",
       "\t<tr><th scope=col>clone_status</th><th scope=col>term</th><th scope=col>contrast</th><th scope=col>null.value</th><th scope=col>estimate</th><th scope=col>std.error</th><th scope=col>df</th><th scope=col>statistic</th><th scope=col>p.value</th></tr>\n",
       "\t<tr><th scope=col>&lt;chr&gt;</th><th scope=col>&lt;chr&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;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th></tr>\n",
       "</thead>\n",
       "<tbody>\n",
       "\t<tr><td>IC</td><td>genotype</td><td>ctrl - mut</td><td>0</td><td>1.55072248</td><td>0.1372035</td><td>490</td><td>11.3023514</td><td>1.760469e-26</td></tr>\n",
       "\t<tr><td>OC</td><td>genotype</td><td>ctrl - mut</td><td>0</td><td>0.09263063</td><td>0.3181798</td><td>490</td><td> 0.2911267</td><td>7.710776e-01</td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "A tibble: 2 × 9\n",
       "\\begin{tabular}{lllllllll}\n",
       " clone\\_status & term & contrast & null.value & estimate & std.error & df & statistic & p.value\\\\\n",
       " <chr> & <chr> & <chr> & <dbl> & <dbl> & <dbl> & <dbl> & <dbl> & <dbl>\\\\\n",
       "\\hline\n",
       "\t IC & genotype & ctrl - mut & 0 & 1.55072248 & 0.1372035 & 490 & 11.3023514 & 1.760469e-26\\\\\n",
       "\t OC & genotype & ctrl - mut & 0 & 0.09263063 & 0.3181798 & 490 &  0.2911267 & 7.710776e-01\\\\\n",
       "\\end{tabular}\n"
      ],
      "text/markdown": [
       "\n",
       "A tibble: 2 × 9\n",
       "\n",
       "| clone_status &lt;chr&gt; | term &lt;chr&gt; | contrast &lt;chr&gt; | null.value &lt;dbl&gt; | estimate &lt;dbl&gt; | std.error &lt;dbl&gt; | df &lt;dbl&gt; | statistic &lt;dbl&gt; | p.value &lt;dbl&gt; |\n",
       "|---|---|---|---|---|---|---|---|---|\n",
       "| IC | genotype | ctrl - mut | 0 | 1.55072248 | 0.1372035 | 490 | 11.3023514 | 1.760469e-26 |\n",
       "| OC | genotype | ctrl - mut | 0 | 0.09263063 | 0.3181798 | 490 |  0.2911267 | 7.710776e-01 |\n",
       "\n"
      ],
      "text/plain": [
       "  clone_status term     contrast   null.value estimate   std.error df \n",
       "1 IC           genotype ctrl - mut 0          1.55072248 0.1372035 490\n",
       "2 OC           genotype ctrl - mut 0          0.09263063 0.3181798 490\n",
       "  statistic  p.value     \n",
       "1 11.3023514 1.760469e-26\n",
       "2  0.2911267 7.710776e-01"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "em_m3 <- emmeans(m3, ~ genotype * clone_status)\n",
    "\n",
    "contrast(em_m3, \"pairwise\", by = \"clone_status\", adjust = \"fdr\") %>% tidy()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From this we see that the total number of cells per clone is significantly different in \"ctrl\" vs \"mutant\" animals.\n",
    "\n",
    "**Conclusion**:\\\n",
    "These data are consistent with a role for our gene of interest in controlling cell proliferation. However, complementary studies are be required to test this hypothesis further and eliminate alternative theories, e.g. does our gene of interest enhance cell death, leading to few cells per clone.\n",
    "\n",
    "## Cell Differentiation - percentage cell type inside and outside clones\n",
    "\n",
    "Next, let's investigate whether the types of cells inside and outside clones are different in ctrl and mutant animals. This information will help us determine whether our gene of interest affects cell differentiation.\n",
    "\n",
    "First, since we already know that 'mutant' clones possess fewer cells than 'ctrl' clones, it is helpful to summarise cell types as proportions and percentages. We also create a composite\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "prop_df_tidy <- count_df_tidy %>% mutate(\n",
    "  pro_cell_num = cell_num / total,\n",
    "  perc_cell_num = pro_cell_num * 100,\n",
    "  genotype_clone_status = interaction(genotype, clone_status, sep = \"\\n\")\n",
    ")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"dataframe\">\n",
       "<caption>A grouped_df: 6 × 11</caption>\n",
       "<thead>\n",
       "\t<tr><th scope=col>int_img</th><th scope=col>GFP_clone</th><th scope=col>clone_status</th><th scope=col>genotype</th><th scope=col>gut</th><th scope=col>total</th><th scope=col>cell_type</th><th scope=col>cell_num</th><th scope=col>pro_cell_num</th><th scope=col>perc_cell_num</th><th scope=col>genotype_clone_status</th></tr>\n",
       "\t<tr><th scope=col>&lt;chr&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;fct&gt;</th><th scope=col>&lt;fct&gt;</th><th scope=col>&lt;fct&gt;</th><th scope=col>&lt;int&gt;</th><th scope=col>&lt;fct&gt;</th><th scope=col>&lt;int&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;fct&gt;</th></tr>\n",
       "</thead>\n",
       "<tbody>\n",
       "\t<tr><td>a1g01p1</td><td>0</td><td>OC</td><td>ctrl</td><td>g01</td><td>175</td><td>EC     </td><td>70</td><td>0.400000000</td><td>40.0000000</td><td>ctrl\n",
       "OC</td></tr>\n",
       "\t<tr><td>a1g01p1</td><td>0</td><td>OC</td><td>ctrl</td><td>g01</td><td>175</td><td>ECEE   </td><td> 4</td><td>0.022857143</td><td> 2.2857143</td><td>ctrl\n",
       "OC</td></tr>\n",
       "\t<tr><td>a1g01p1</td><td>0</td><td>OC</td><td>ctrl</td><td>g01</td><td>175</td><td>preEC  </td><td> 1</td><td>0.005714286</td><td> 0.5714286</td><td>ctrl\n",
       "OC</td></tr>\n",
       "\t<tr><td>a1g01p1</td><td>0</td><td>OC</td><td>ctrl</td><td>g01</td><td>175</td><td>EE     </td><td> 2</td><td>0.011428571</td><td> 1.1428571</td><td>ctrl\n",
       "OC</td></tr>\n",
       "\t<tr><td>a1g01p1</td><td>0</td><td>OC</td><td>ctrl</td><td>g01</td><td>175</td><td>ISCorEB</td><td>98</td><td>0.560000000</td><td>56.0000000</td><td>ctrl\n",
       "OC</td></tr>\n",
       "\t<tr><td>a1g01p1</td><td>1</td><td>IC</td><td>ctrl</td><td>g01</td><td> 14</td><td>EC     </td><td> 9</td><td>0.642857143</td><td>64.2857143</td><td>ctrl\n",
       "IC</td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "A grouped\\_df: 6 × 11\n",
       "\\begin{tabular}{lllllllllll}\n",
       " int\\_img & GFP\\_clone & clone\\_status & genotype & gut & total & cell\\_type & cell\\_num & pro\\_cell\\_num & perc\\_cell\\_num & genotype\\_clone\\_status\\\\\n",
       " <chr> & <dbl> & <fct> & <fct> & <fct> & <int> & <fct> & <int> & <dbl> & <dbl> & <fct>\\\\\n",
       "\\hline\n",
       "\t a1g01p1 & 0 & OC & ctrl & g01 & 175 & EC      & 70 & 0.400000000 & 40.0000000 & ctrl\n",
       "OC\\\\\n",
       "\t a1g01p1 & 0 & OC & ctrl & g01 & 175 & ECEE    &  4 & 0.022857143 &  2.2857143 & ctrl\n",
       "OC\\\\\n",
       "\t a1g01p1 & 0 & OC & ctrl & g01 & 175 & preEC   &  1 & 0.005714286 &  0.5714286 & ctrl\n",
       "OC\\\\\n",
       "\t a1g01p1 & 0 & OC & ctrl & g01 & 175 & EE      &  2 & 0.011428571 &  1.1428571 & ctrl\n",
       "OC\\\\\n",
       "\t a1g01p1 & 0 & OC & ctrl & g01 & 175 & ISCorEB & 98 & 0.560000000 & 56.0000000 & ctrl\n",
       "OC\\\\\n",
       "\t a1g01p1 & 1 & IC & ctrl & g01 &  14 & EC      &  9 & 0.642857143 & 64.2857143 & ctrl\n",
       "IC\\\\\n",
       "\\end{tabular}\n"
      ],
      "text/markdown": [
       "\n",
       "A grouped_df: 6 × 11\n",
       "\n",
       "| int_img &lt;chr&gt; | GFP_clone &lt;dbl&gt; | clone_status &lt;fct&gt; | genotype &lt;fct&gt; | gut &lt;fct&gt; | total &lt;int&gt; | cell_type &lt;fct&gt; | cell_num &lt;int&gt; | pro_cell_num &lt;dbl&gt; | perc_cell_num &lt;dbl&gt; | genotype_clone_status &lt;fct&gt; |\n",
       "|---|---|---|---|---|---|---|---|---|---|---|\n",
       "| a1g01p1 | 0 | OC | ctrl | g01 | 175 | EC      | 70 | 0.400000000 | 40.0000000 | ctrl\n",
       "OC |\n",
       "| a1g01p1 | 0 | OC | ctrl | g01 | 175 | ECEE    |  4 | 0.022857143 |  2.2857143 | ctrl\n",
       "OC |\n",
       "| a1g01p1 | 0 | OC | ctrl | g01 | 175 | preEC   |  1 | 0.005714286 |  0.5714286 | ctrl\n",
       "OC |\n",
       "| a1g01p1 | 0 | OC | ctrl | g01 | 175 | EE      |  2 | 0.011428571 |  1.1428571 | ctrl\n",
       "OC |\n",
       "| a1g01p1 | 0 | OC | ctrl | g01 | 175 | ISCorEB | 98 | 0.560000000 | 56.0000000 | ctrl\n",
       "OC |\n",
       "| a1g01p1 | 1 | IC | ctrl | g01 |  14 | EC      |  9 | 0.642857143 | 64.2857143 | ctrl\n",
       "IC |\n",
       "\n"
      ],
      "text/plain": [
       "  int_img GFP_clone clone_status genotype gut total cell_type cell_num\n",
       "1 a1g01p1 0         OC           ctrl     g01 175   EC        70      \n",
       "2 a1g01p1 0         OC           ctrl     g01 175   ECEE       4      \n",
       "3 a1g01p1 0         OC           ctrl     g01 175   preEC      1      \n",
       "4 a1g01p1 0         OC           ctrl     g01 175   EE         2      \n",
       "5 a1g01p1 0         OC           ctrl     g01 175   ISCorEB   98      \n",
       "6 a1g01p1 1         IC           ctrl     g01  14   EC         9      \n",
       "  pro_cell_num perc_cell_num genotype_clone_status\n",
       "1 0.400000000  40.0000000    ctrl\\nOC             \n",
       "2 0.022857143   2.2857143    ctrl\\nOC             \n",
       "3 0.005714286   0.5714286    ctrl\\nOC             \n",
       "4 0.011428571   1.1428571    ctrl\\nOC             \n",
       "5 0.560000000  56.0000000    ctrl\\nOC             \n",
       "6 0.642857143  64.2857143    ctrl\\nIC             "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "prop_df_tidy %>% head()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we plot a stacked barplot as before. However, this time we plot percentage cell types:\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWgAAAHgCAYAAACIBvdgAAAEDmlDQ1BrQ0dDb2xvclNwYWNl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\ngAI6U0aK7SQBEsg5Av8Hf6vsCIbXrPQAAAAASUVORK5CYII=",
      "text/plain": [
       "plot without title"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 240,
       "width": 180
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "stacked_bar_prop <- prop_df_tidy %>%\n",
    "  ggplot(aes(fill = cell_type, y = pro_cell_num, x = genotype_clone_status)) +\n",
    "  geom_bar(\n",
    "    position = \"fill\",\n",
    "    stat = \"summary\",\n",
    "    fun = \"mean\"\n",
    "  ) +\n",
    "  theme_bw() +\n",
    "  scale_fill_manual(values = col_pal)\n",
    "\n",
    "options(repr.plot.width = 3, repr.plot.height = 4)\n",
    "stacked_bar_prop\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This plot suggests there might be a difference in percentage cell-types of mutant clones vs ctrl. Let's investigate if these differences are statistically significant. Since we have proportional data, we construct a quasibinomal regression model:\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "ml_qb <- glm(pro_cell_num ~ genotype * clone_status * cell_type, family = quasibinomial(), prop_df_tidy)\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Unfortunately, quasibinomal regression models do not currently support mixed-effects (via lme4 package). We, therefore, move onto our desired contrasts.\n",
    "\n",
    "We wish to compare:  \n",
    "* the percentage of each cell type inside mutant vs ctrl clones  \n",
    "* the percentage of each cell type outside mutant vs ctrl clones  \n",
    "* the percentage of each cell type outside mutant vs inside mutant clones\n",
    "\n",
    "To do this, we setup pairwise contrasts by cell type and adjust for multiple comparisons using Benjamini/Hochberg false-discovery rate procedure:\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table class=\"dataframe\">\n",
       "<caption>A tibble: 6 × 9</caption>\n",
       "<thead>\n",
       "\t<tr><th scope=col>cell_type</th><th scope=col>term</th><th scope=col>contrast</th><th scope=col>null.value</th><th scope=col>estimate</th><th scope=col>std.error</th><th scope=col>df</th><th scope=col>z.ratio</th><th scope=col>adj.p.value</th></tr>\n",
       "\t<tr><th scope=col>&lt;chr&gt;</th><th scope=col>&lt;chr&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;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th><th scope=col>&lt;dbl&gt;</th></tr>\n",
       "</thead>\n",
       "<tbody>\n",
       "\t<tr><td>ECEE</td><td>genotype*clone_status</td><td>ctrl IC - mut IC </td><td>0</td><td> 0.02604153</td><td>0.741427</td><td>Inf</td><td> 0.03512353</td><td>0.9719812</td></tr>\n",
       "\t<tr><td>ECEE</td><td>genotype*clone_status</td><td>ctrl IC - ctrl OC</td><td>0</td><td>-0.34499874</td><td>1.227151</td><td>Inf</td><td>-0.28113790</td><td>0.9719812</td></tr>\n",
       "\t<tr><td>ECEE</td><td>genotype*clone_status</td><td>ctrl IC - mut OC </td><td>0</td><td>-0.12599085</td><td>1.357413</td><td>Inf</td><td>-0.09281688</td><td>0.9719812</td></tr>\n",
       "\t<tr><td>ECEE</td><td>genotype*clone_status</td><td>mut IC - ctrl OC </td><td>0</td><td>-0.37104027</td><td>1.209524</td><td>Inf</td><td>-0.30676542</td><td>0.9719812</td></tr>\n",
       "\t<tr><td>ECEE</td><td>genotype*clone_status</td><td>mut IC - mut OC  </td><td>0</td><td>-0.15203239</td><td>1.341499</td><td>Inf</td><td>-0.11333023</td><td>0.9719812</td></tr>\n",
       "\t<tr><td>ECEE</td><td>genotype*clone_status</td><td>ctrl OC - mut OC </td><td>0</td><td> 0.21900788</td><td>1.660062</td><td>Inf</td><td> 0.13192753</td><td>0.9719812</td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "A tibble: 6 × 9\n",
       "\\begin{tabular}{lllllllll}\n",
       " cell\\_type & term & contrast & null.value & estimate & std.error & df & z.ratio & adj.p.value\\\\\n",
       " <chr> & <chr> & <chr> & <dbl> & <dbl> & <dbl> & <dbl> & <dbl> & <dbl>\\\\\n",
       "\\hline\n",
       "\t ECEE & genotype*clone\\_status & ctrl IC - mut IC  & 0 &  0.02604153 & 0.741427 & Inf &  0.03512353 & 0.9719812\\\\\n",
       "\t ECEE & genotype*clone\\_status & ctrl IC - ctrl OC & 0 & -0.34499874 & 1.227151 & Inf & -0.28113790 & 0.9719812\\\\\n",
       "\t ECEE & genotype*clone\\_status & ctrl IC - mut OC  & 0 & -0.12599085 & 1.357413 & Inf & -0.09281688 & 0.9719812\\\\\n",
       "\t ECEE & genotype*clone\\_status & mut IC - ctrl OC  & 0 & -0.37104027 & 1.209524 & Inf & -0.30676542 & 0.9719812\\\\\n",
       "\t ECEE & genotype*clone\\_status & mut IC - mut OC   & 0 & -0.15203239 & 1.341499 & Inf & -0.11333023 & 0.9719812\\\\\n",
       "\t ECEE & genotype*clone\\_status & ctrl OC - mut OC  & 0 &  0.21900788 & 1.660062 & Inf &  0.13192753 & 0.9719812\\\\\n",
       "\\end{tabular}\n"
      ],
      "text/markdown": [
       "\n",
       "A tibble: 6 × 9\n",
       "\n",
       "| cell_type &lt;chr&gt; | term &lt;chr&gt; | contrast &lt;chr&gt; | null.value &lt;dbl&gt; | estimate &lt;dbl&gt; | std.error &lt;dbl&gt; | df &lt;dbl&gt; | z.ratio &lt;dbl&gt; | adj.p.value &lt;dbl&gt; |\n",
       "|---|---|---|---|---|---|---|---|---|\n",
       "| ECEE | genotype*clone_status | ctrl IC - mut IC  | 0 |  0.02604153 | 0.741427 | Inf |  0.03512353 | 0.9719812 |\n",
       "| ECEE | genotype*clone_status | ctrl IC - ctrl OC | 0 | -0.34499874 | 1.227151 | Inf | -0.28113790 | 0.9719812 |\n",
       "| ECEE | genotype*clone_status | ctrl IC - mut OC  | 0 | -0.12599085 | 1.357413 | Inf | -0.09281688 | 0.9719812 |\n",
       "| ECEE | genotype*clone_status | mut IC - ctrl OC  | 0 | -0.37104027 | 1.209524 | Inf | -0.30676542 | 0.9719812 |\n",
       "| ECEE | genotype*clone_status | mut IC - mut OC   | 0 | -0.15203239 | 1.341499 | Inf | -0.11333023 | 0.9719812 |\n",
       "| ECEE | genotype*clone_status | ctrl OC - mut OC  | 0 |  0.21900788 | 1.660062 | Inf |  0.13192753 | 0.9719812 |\n",
       "\n"
      ],
      "text/plain": [
       "  cell_type term                  contrast          null.value estimate   \n",
       "1 ECEE      genotype*clone_status ctrl IC - mut IC  0           0.02604153\n",
       "2 ECEE      genotype*clone_status ctrl IC - ctrl OC 0          -0.34499874\n",
       "3 ECEE      genotype*clone_status ctrl IC - mut OC  0          -0.12599085\n",
       "4 ECEE      genotype*clone_status mut IC - ctrl OC  0          -0.37104027\n",
       "5 ECEE      genotype*clone_status mut IC - mut OC   0          -0.15203239\n",
       "6 ECEE      genotype*clone_status ctrl OC - mut OC  0           0.21900788\n",
       "  std.error df  z.ratio     adj.p.value\n",
       "1 0.741427  Inf  0.03512353 0.9719812  \n",
       "2 1.227151  Inf -0.28113790 0.9719812  \n",
       "3 1.357413  Inf -0.09281688 0.9719812  \n",
       "4 1.209524  Inf -0.30676542 0.9719812  \n",
       "5 1.341499  Inf -0.11333023 0.9719812  \n",
       "6 1.660062  Inf  0.13192753 0.9719812  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "e_ml <- emmeans(ml_qb, ~ genotype * clone_status * cell_type)\n",
    "\n",
    "comp_results <- contrast(e_ml, \"pairwise\",\n",
    "  by = c(\"cell_type\"),\n",
    "  adjust = \"fdr\"\n",
    ")\n",
    "\n",
    "stat_results <- comp_results %>% tidy()\n",
    "stat_results %>% head()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To aid visualization, let's add these results to our stacked percentage barplot from earlier. We annotations only statistically significant comparisons:\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "stat_anots <- stat_results %>%\n",
    "  mutate(\n",
    "    p = case_when(\n",
    "      adj.p.value < 0.0001 ~ \"****\",\n",
    "      adj.p.value < 0.001 ~ \"***\",\n",
    "      adj.p.value < 0.01 ~ \"**\",\n",
    "      adj.p.value < 0.05 ~ \"*\",\n",
    "      adj.p.value < 1 ~ \"ns\"\n",
    "    ),\n",
    "  ) %>%\n",
    "  filter(p != \"ns\", contrast != \"mut IC - ctrl OC\") %>%\n",
    "  select(cell_type, contrast, p) %>%\n",
    "  mutate(contrast = str_replace_all(contrast, \"\\ \", \"\\n\"),\n",
    "         y.position = c(1.05, 1.10, 1.15, 1.20, 1.25), \n",
    "         color = case_when(cell_type == 'EE' ~ \"#95a5a6\",\n",
    "                           cell_type == 'EC' ~ \"#F57171\",\n",
    "                           cell_type == 'ISCorEB' ~ \"#008b68\")) %>% \n",
    "  separate(contrast, sep = \"\\n-\\n\", c(\"group1\", \"group2\"))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWgAAAHgCAYAAACIBvdgAAAEDmlDQ1BrQ0dDb2xvclNwYWNl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",
      "text/plain": [
       "plot without title"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 240,
       "width": 180
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "options(repr.plot.width = 3, repr.plot.height = 4)\n",
    "stacked_bar_prop + stat_pvalue_manual(stat_anots, color = rep(stat_anots$color, each=3), tip.length=0.01)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Summary of results**:  \n",
    "- comparing the percentage makeup of control and mutant clones, mutant clones possess significantly fewer ECs and significantly more progenitor cells (e.g. ISCs or EBs) and EEs.  \n",
    "- comparing the percentage makeup of cells inside mutant clones and outside mutant clones, cells inside clones similarly possess significantly fewer ECs and more progenitor cells.  \n",
    "- no significant difference is seen in the percentage makeup of cells outside clones in control vs mutant animals, suggesting our gene of interest may be haploinsufficient.\n",
    "\n",
    "**Conclusion**:  \n",
    "Together, these data suggest mutation of our gene of interest may hinder ISC -> EC differentiation. However, complementary studies are be required to further test this hypothesis and eliminate alternative theories, e.g. does our gene of interest lead to enhanced EC cell death, resulting in higher percentage progenitor and EE cells.\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}