DGE-ontology.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Created by Michal Bukowski (michal.bukowski@tuta.io, m.bukowski@uj.edu.pl) under GPL-3.0 license as a part of the [dge-ontology](https://github.com/michalbukowski/dge-ontology) repository<br>\n",
    "Department of Analytical Biochemistry, Biophysics and Biotechnology, Jagiellonian University, Krakow, Poland"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "# DGE-ontology\n",
    "This Jupyter notebook aims at ontology analysis and visualisation for DGE results obtained using [Salmon](https://doi.org/10.1186/s13059-014-0550-8) and [DESeq2](https://doi.org/10.1038/nmeth.4197). However, the approach can be used for any kind of differential analysis (e.g. in proteomics). The exemplary input contains metadata on _Staphylococcus aureus_ RN4220 transcriptome and output from [rnaseq-pipeline-2](https://github.com/michalbukowski/rnaseq-pipeline-2) executed for two strains sampled in the logarithmic growth phase: a wild type (`wt51e`) vs. $\\Delta$_saoBC_ mutant (`mt51e`) (NCBI BioProject accession number PRJNA798259). Execution of subsequent cells in many instances relies on the output of preceding cells, thus cells should be executed in the order as presented in this notebook.\n",
    "\n",
    "<u>Notebook cell output has been saved in the original (repository) version of the notebook</u>. If rerun for the exemplary data, it should give the same results. All visualisations are saved to `output` directory as high-resolution PNG files.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Do the necessary imports and environment setup.\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from matplotlib import pyplot as plt\n",
    "from matplotlib.patches import Circle, Wedge\n",
    "from collections import defaultdict\n",
    "from scipy.stats import hypergeom\n",
    "from PIL import Image\n",
    "\n",
    "# Configure the environments to make sure the way\n",
    "# figures are displayed is the same as they are saved.\n",
    "%config InlineBackend.print_figure_kwargs = {'bbox_inches': None}\n",
    "\n",
    "# Attempt to create a directory for output files,\n",
    "# in case it does not exist.\n",
    "! mkdir -p \"output\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "## 1. Setting up required parameters regarding the input files and the output chart\n",
    "Required parameters are paths to two tabular input files that are required as well as provide names of obligatory columns:\n",
    "1. The first of the input files is a table that contains output of a differential analysis (`dge_file`). It contains results for the <u>complete population<u> (e.g. all possible/potentially expected transcripts) described in three obligatory columns:\n",
    "    - an ID (`id_col`)\n",
    "    - fold change (`fold_col`)\n",
    "    - <i>p</i>&#8209;value/false discovery rate (FDR) (`pval_col`)\n",
    "The <u>complete population</u> encompasses also entities, for which it is not possible to estimate the fold change (e.g. non-transcribed transcripts). In such cases fold change values and <i>p</i>&#8209;values should be empty.\n",
    "1. The other is a file providing metadata that contains two obligatory columns: an ID (`id_col`) and ontology labels (`onts_col`), ond three optional columns:\n",
    "    - sequence type (`type_col`) that is solely used in respect to `ncRNA` and `tRNA` values (any other ontology labels provided in remaining columns are ignored for rows described as such)\n",
    "    - a column that is used as source of an additional ontology label (`bont_col`) that is merged with labels provided in `onts_col`\n",
    "\n",
    "The values of `bont_col` column are treated as binary (true or false, whether the values are empty/NA or any non-empty value) and described with the given label (`bnt_label`). The ID from the first table corresponds to the id from the second table.\n",
    "\n",
    "The output chart will ba saved as `chart_file`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Results table parameters. All obligatory.\n",
    "dge_file   = 'input/mt51e_wt51e_DGE.tsv'\n",
    "id_col     = 'locus_tag'\n",
    "fold_col   = 'log2FoldChange'\n",
    "pval_col   = 'padj'\n",
    "\n",
    "# Metadata table parameters. Columns type_col\n",
    "# and bont_col as well as the label bont_lable\n",
    "# are not obligatory. Set them all to None if you\n",
    "# do not need/intend to use them or if corresponding\n",
    "# columns do not exist in the metadata input file.\n",
    "meta_file  = 'input/rn.tsv'\n",
    "onts_col   = 'cog'\n",
    "type_col   = 'type'\n",
    "bont_col   = 'vfdbont'\n",
    "bont_label = 'Virulence factors'\n",
    "\n",
    "# Output chart file path. A PNG file.\n",
    "chart_file = 'output/ontology_chart.png'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "## 2. Setting up filtering thresholds for input results\n",
    "Filtering threshold `fold_th` for the absolute fold change (`fold_col` column) and `fdr_th` for the <i>p</i>&#8209;value/FDR (`pval_col` column). The DGE results, for which |fold change| &ge; `fold_th` and <i>p</i>&#8209;value &le; `fdr_th`, are <u>interpreted as differentiating</u>."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Fold change value and p-value (FDR) thresholds.\n",
    "fold_th = 0.58\n",
    "fdr_th  = 0.05"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Setting up optional chart parameters\n",
    "Optional parameters help to modify the final look of the chart. It is recommended to first run the code with default parameters, and then modify then gradually to obtain the desired outcome. The meaning of the parameters is explaned in comments."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.1. Parameters related to differentiating cases\n",
    "These parameters help to adjust the color scale range, select manually ontology labels that are to be depicted in the chart or filter out ontology labels according to a minimal number of cases. Please read the comments carefully."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Max fold for the fold scale. If None, it is set\n",
    "# automatically to the highest absolute fold value.\n",
    "# Set it manually if you want to generate charts\n",
    "# depicting results in a fixed scale.\n",
    "max_fold = None\n",
    "\n",
    "# By default all ontology labels are depicted in\n",
    "# the chart in the order of descending number of\n",
    "# differentiating cases described by a given label.\n",
    "# If you want to set a custom order or select certain\n",
    "# groups, set the list of visualised labels manually:\n",
    "# sel_onts = ['First label', 'Second label', ...]\n",
    "# you can also indicate which labels to skip, e.g.:\n",
    "# skip_onts = ['Function unknown', 'ncRNA']\n",
    "sel_onts  = None\n",
    "skip_onts = None\n",
    "\n",
    "# By default all ontology labels are depicted in\n",
    "# the chart, for which the number of differentiating\n",
    "# cases is higher than 0. Change this value to\n",
    "# increase the threshold and skip less numerous groups.\n",
    "min_size = 0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.2. Parameters related to graphical formatting\n",
    "These parameters decide only about the graphical aspects of the chart. Please read the comments carefully."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# The output figure resolution in DPI (dots per inch).\n",
    "# The figure dimensions in pixels are fig_w*dpi x fig_h*dpi.\n",
    "dpi = 300\n",
    "\n",
    "# Set the figure dimensions in inches. The following figure\n",
    "# dimensions and their ratio work well in most cases.\n",
    "# It is not recommended to modify it. Adjust dimensions\n",
    "# only when a lot of blank space is left after all elements\n",
    "# of the figure are fit into the range of axes (see below).\n",
    "fig_w = 10.0\n",
    "fig_h =  3.0\n",
    "\n",
    "# The x and y axes fill up the whole figure. The range of the y-axis\n",
    "# is set automatically based on figure dimensions (see above)\n",
    "# and their ratio in order to keep the chart circular (not eliptical).\n",
    "# The scaffold circle of the chart is centred around the point (0, 0).\n",
    "# If ontology labels go horizontally beyond the figure, the range\n",
    "# of x-axis should be first modified:\n",
    "# xmin is the minimum (left most) x-coordinate of x-axis\n",
    "# xmax is the maximum (right most) x-coordinate of x-axis\n",
    "xmin = -2.2\n",
    "xmax =  1.8\n",
    "\n",
    "# General geometric measures of the pie chart:\n",
    "# pie_r        -- the radius of the scaffold circle, change if\n",
    "#                 the chart is too big or too small in respect\n",
    "#                 to labels\n",
    "# angle_offset -- the angle offset for placing wedges on the scaffold circle\n",
    "#                 expressed in degrees (0.0 - 360.0) - by default the third\n",
    "#                 wedge/slice starts at 12:00 o'clock, top centre,\n",
    "#                 which seems to be optimal for size-ordered slices\n",
    "# margin       -- margin between each wedge and the number of cases\n",
    "#                 as well as that number and the terminal parts of connectors\n",
    "#                 that join wedges and labels\n",
    "# label_at     -- x at which ontology (wedge) labels are left-aligned\n",
    "#                 on the right side of the pie chart, or\n",
    "#                 -x at which ontology (wedge) labels are right-aligned\n",
    "#                 on the left side of the pie chart,\n",
    "#                 change to bring labels closer or move further\n",
    "#                 from the chart\n",
    "# label_height -- vertical span a label is assumed to occupy,\n",
    "#                 change to increase or decrease the vertical spacing\n",
    "#                 between adjacent labels, especially in case of overlap\n",
    "pie_r        = 0.30\n",
    "angle_offset = 0.00\n",
    "margin       = 0.03\n",
    "label_at     = 0.70\n",
    "label_height = 0.08\n",
    "\n",
    "# General scale factor. Change to increase or decrease the relative\n",
    "# wedge radial sizes, especially if inner parts pass through the\n",
    "# middle of the chart.\n",
    "scale = 0.03\n",
    "\n",
    "# The label located over the scale bar. LaTeX expression\n",
    "# are allowed to be used and must be enclosed with $ sign.\n",
    "scale_bar_label = 'Log$_{2}$ fold change'\n",
    "\n",
    "# The font size values for labels, numbers\n",
    "# of results and the scale bar label.\n",
    "label_font = 8.5\n",
    "num_font   = 6.0\n",
    "sbar_font  = 8.0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "## 4. Functions, declarations\n",
    "Four helper functions used for data analysis. Details on each are provided in doc-strings and comments in the code. Do not modify the code unless you known what you are doing."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_sign(pval: float, trend: str) -> tuple[str, str]:\n",
    "    '''For a p-value falling into a given bin and a trend\n",
    "       returns a string of asterisks and a hex colour. Arguments:\n",
    "       pval  -- p-value\n",
    "       trend -- trend (under/normal/over)\n",
    "       Returns:\n",
    "       sign  -- a string of asterisk or an empty string\n",
    "       color -- color encoded in a hex string\n",
    "    '''\n",
    "    sign  = ''\n",
    "    color = 'black'\n",
    "    if trend == 'over':\n",
    "        if pval <= 0.001:\n",
    "            sign = '***'\n",
    "        elif pval <= 0.01:\n",
    "            sign = '**'\n",
    "        elif pval <= 0.05:\n",
    "            sign = '*'\n",
    "        if pval <= 0.05:\n",
    "            color = '#990000' #if trend == 'over' else '#0044aa'\n",
    "    return sign, color"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def hypgeoprob(total_size: int, sample_size: int,\n",
    "               total_cases: int, sample_cases: int) -> float:\n",
    "    '''Provide a cumulative probability from hypergeometric distribution for\n",
    "       a given total_size of a population, total_cases in the whole population,\n",
    "       a sample_size sample from the population that contains sample_cases.\n",
    "       That is a probability of drawing sample_cases or less when having given\n",
    "       sample_size, total_size and total_cases. Or the probability of\n",
    "       sample_cases or more,  whichever smaller. Arguments:\n",
    "       total_size   -- number of elements in the total set\n",
    "       sample_size  -- number of element in a sample from the total set\n",
    "       total_cases  -- number of elements of interest in the total set\n",
    "       sample_cases -- number of elements of interest in the from the total set\n",
    "       Returns:\n",
    "       prob_1 or prob_2 -- left tail or right tail cumulative probability,\n",
    "                           whichever smaller, the left tail cumulative\n",
    "                           probability is given a minus sign that symbolises\n",
    "                           under-representation in the sample\n",
    "    '''\n",
    "    # Left tail cumulative probability of sample_cases or less.\n",
    "    prob_1 = hypergeom.cdf(sample_cases, total_size, total_cases, sample_size)\n",
    "    # Right tail cumulative probability of sample_cases or more.\n",
    "    prob_2 = hypergeom.sf(sample_cases-1, total_size, total_cases, sample_size)\n",
    "    if prob_1 <= prob_2: \n",
    "        return -1*prob_1\n",
    "    else: \n",
    "        return prob_2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def count_onts(fin_df: pd.DataFrame) -> defaultdict[int]:\n",
    "    '''Given a DataFrame with DGE results that are merged on id_col with semicolon-separated\n",
    "       ontology labels from the onts_col column, 'ncRNA' and 'tRNA' types\n",
    "       in the type_col column as well as not-NA values in the bont_col column\n",
    "       (labeled as bont_label), counts occurrences of all ontology labels\n",
    "       across all rows and puts the results in a dictionary ontology label -> count. Arguments:\n",
    "       fin_df -- DataFrame that contains DGE results merged with the onts_col\n",
    "                 and type_col columns from metadata on id_col\n",
    "       Returns:\n",
    "       ont_counts -- ontology label -> count dictionary (defaultdict with 0 as the default value)\n",
    "    '''\n",
    "    # Create a copy of the DataFrame being processed\n",
    "    # and a default dictionary with ontology label counts\n",
    "    fin_df = fin_df.copy()\n",
    "    ont_counts = defaultdict(int)\n",
    "\n",
    "    # Update onts_col column using bont_col column\n",
    "    # if this optional column is present.\n",
    "    if bont_col is not None:\n",
    "        bonts_nona = fin_df[bont_col].notna()\n",
    "        onts_na    = fin_df[onts_col].isna()\n",
    "        fin_df.loc[bonts_nona & onts_na,  onts_col] =  bont_label\n",
    "        fin_df.loc[bonts_nona & ~onts_na, onts_col] += '; ' + bont_label\n",
    "\n",
    "    # Update onts_col column using type_col column\n",
    "    # if this optional column is present.\n",
    "    if type_col is not None:\n",
    "        for gb_key in['ncRNA', 'tRNA']:\n",
    "            type_onts = fin_df[type_col] == gb_key\n",
    "            onts_na   = fin_df[onts_col].isna()\n",
    "            fin_df.loc[type_onts & onts_na,  onts_col] =  gb_key\n",
    "            fin_df.loc[type_onts & ~onts_na, onts_col] += '; ' + gb_key\n",
    "\n",
    "    # Assign 'Function unknown' label to remaining empty onts_col values \n",
    "    fin_df[onts_col] = fin_df[onts_col].fillna('Function unknown')\n",
    "    \n",
    "    # Fetch semicolon-separated ontologies from the onts_col column.\n",
    "    for onts in fin_df[onts_col]:\n",
    "        for item in onts.split(';'):\n",
    "            ont_counts[item.strip()] += 1\n",
    "    \n",
    "    return ont_counts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "def collect_folds(filt_df: pd.DataFrame) -> defaultdict[ list[float] ]:\n",
    "    '''Given a DataFrame with filtrated (considered as differentiating)\n",
    "       DGE results merged on id_col with semicolon-separated ontologies\n",
    "       in the onts_col column, 'ncRNA' and 'tRNA' types in the type_col column\n",
    "       as well as not-NA values in the bont_col column (labelled as bont_label),\n",
    "       collects fold_col values for occurrences of all ontology labels across\n",
    "       all rows and puts the results in a dictionary ontology label ->\n",
    "       [fold_1, fold_2, ...]. Arguments:\n",
    "       filt_df -- DataFrame that contains filtrated DESeq2 DGE results merged\n",
    "                  with the 'cog' and 'type' columns from metadata on locus_tag\n",
    "       Returns:\n",
    "       ont_counts -- ontology label -> [fold_1, fold_2, ...] dictionary\n",
    "                     (defaultdict with an empt list [] as the default value)\n",
    "    '''\n",
    "    # Create a copy of the DataFrame being processed and a default dictionary\n",
    "    # with ontology label fold change values\n",
    "    filt_df = filt_df.copy()\n",
    "    folds = defaultdict(list)\n",
    "\n",
    "    # Update onts_col column using bont_col column\n",
    "    # if this optional column is present.\n",
    "    if bont_col is not None:\n",
    "        bonts_nona = filt_df[bont_col].notna()\n",
    "        onts_na    = filt_df[onts_col].isna()\n",
    "        filt_df.loc[bonts_nona & onts_na, onts_col] = bont_label\n",
    "        filt_df.loc[bonts_nona & ~onts_na, onts_col] += '; ' + bont_label\n",
    "\n",
    "    # Update onts_col column using type_col column\n",
    "    # if this optional column is present.\n",
    "    if type_col is not None:\n",
    "        for gb_key in['ncRNA', 'tRNA']:\n",
    "            type_onts = filt_df[type_col] == gb_key\n",
    "            onts_na   = filt_df[onts_col].isna()\n",
    "            filt_df.loc[type_onts & onts_na,  onts_col] =  gb_key\n",
    "            filt_df.loc[type_onts & ~onts_na, onts_col] += '; ' + gb_key\n",
    "\n",
    "    # Assign 'Function unknown' label to remaining empty onts_col values \n",
    "    filt_df[onts_col] = filt_df[onts_col].fillna('Function unknown')\n",
    "\n",
    "    # Collect fold change values from fold_col column for each of\n",
    "    # semicolon-separated ontology label from the onts_col column.\n",
    "    for _, (fold, onts) in filt_df[ [fold_col, onts_col] ].iterrows():\n",
    "        for item in onts.split(';'):\n",
    "            folds[item.strip()].append(fold)\n",
    "    folds = { ont : np.array(values) for ont, values in folds.items() }\n",
    "\n",
    "    return folds"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "## 5. Data processing\n",
    "In this section input data is read into Pandas DataFrame objects, and then pre-processed for visualisation purposes. Details are provided in comments in the code. Do not modify the code unless you known what you are doing."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 3335 entries, SAOUHSC_03055 to srn_1040_sRNA92\n",
      "Data columns (total 5 columns):\n",
      " #   Column          Non-Null Count  Dtype  \n",
      "---  ------          --------------  -----  \n",
      " 0   baseMean        2503 non-null   float64\n",
      " 1   log2FoldChange  2503 non-null   float64\n",
      " 2   lfcSE           2503 non-null   float64\n",
      " 3   pvalue          2496 non-null   float64\n",
      " 4   padj            2496 non-null   float64\n",
      "dtypes: float64(5)\n",
      "memory usage: 156.3+ KB\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "None"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>baseMean</th>\n",
       "      <th>log2FoldChange</th>\n",
       "      <th>lfcSE</th>\n",
       "      <th>pvalue</th>\n",
       "      <th>padj</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>locus_tag</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_03055</th>\n",
       "      <td>900.082113</td>\n",
       "      <td>-0.190838</td>\n",
       "      <td>0.253922</td>\n",
       "      <td>0.107434</td>\n",
       "      <td>0.681069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_03054</th>\n",
       "      <td>12.453537</td>\n",
       "      <td>-0.023077</td>\n",
       "      <td>0.201086</td>\n",
       "      <td>0.677658</td>\n",
       "      <td>0.943881</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_03053</th>\n",
       "      <td>167.816277</td>\n",
       "      <td>0.057208</td>\n",
       "      <td>0.196688</td>\n",
       "      <td>0.491597</td>\n",
       "      <td>0.892770</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_03051</th>\n",
       "      <td>223.827558</td>\n",
       "      <td>0.025922</td>\n",
       "      <td>0.177333</td>\n",
       "      <td>0.784860</td>\n",
       "      <td>0.962877</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_03049</th>\n",
       "      <td>479.204817</td>\n",
       "      <td>-0.012935</td>\n",
       "      <td>0.165215</td>\n",
       "      <td>0.895382</td>\n",
       "      <td>0.981500</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 baseMean  log2FoldChange     lfcSE    pvalue      padj\n",
       "locus_tag                                                              \n",
       "SAOUHSC_03055  900.082113       -0.190838  0.253922  0.107434  0.681069\n",
       "SAOUHSC_03054   12.453537       -0.023077  0.201086  0.677658  0.943881\n",
       "SAOUHSC_03053  167.816277        0.057208  0.196688  0.491597  0.892770\n",
       "SAOUHSC_03051  223.827558        0.025922  0.177333  0.784860  0.962877\n",
       "SAOUHSC_03049  479.204817       -0.012935  0.165215  0.895382  0.981500"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load DGE results for a complete population,\n",
    "# i.e. even for those entities (genes, proteins)\n",
    "# that could be present in the sample (bacterial transcriptome,\n",
    "# proteome, etc.) but were not (e.g. non-transcribed genes).\n",
    "# In this example two groups were compared:\n",
    "# wt51e_lg - wild type, mt51e_lg - mutant.\n",
    "dge_df = pd.read_csv(\n",
    "    dge_file,\n",
    "    index_col = id_col,\n",
    "    sep = '\\t'\n",
    ")\n",
    "display(dge_df.info())\n",
    "dge_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 3466 entries, SAOUHSC_00001 to srn_5130_Sau6728\n",
      "Data columns (total 5 columns):\n",
      " #   Column           Non-Null Count  Dtype \n",
      "---  ------           --------------  ----- \n",
      " 0   type             3466 non-null   object\n",
      " 1   pan_gene_symbol  1522 non-null   object\n",
      " 2   product          3457 non-null   object\n",
      " 3   vfdbont          136 non-null    object\n",
      " 4   cog              1774 non-null   object\n",
      "dtypes: object(5)\n",
      "memory usage: 162.5+ KB\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "None"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>type</th>\n",
       "      <th>pan_gene_symbol</th>\n",
       "      <th>product</th>\n",
       "      <th>vfdbont</th>\n",
       "      <th>cog</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>locus_tag</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_00001</th>\n",
       "      <td>CDS</td>\n",
       "      <td>dnaA</td>\n",
       "      <td>chromosomal replication initiation protein</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Replication, recombination and repair</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_00002</th>\n",
       "      <td>CDS</td>\n",
       "      <td>dnaN</td>\n",
       "      <td>DNA polymerase III subunit beta</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Replication, recombination and repair</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_00003</th>\n",
       "      <td>CDS</td>\n",
       "      <td>NaN</td>\n",
       "      <td>hypothetical protein</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Translation, ribosomal structure and biogenesis</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_00004</th>\n",
       "      <td>CDS</td>\n",
       "      <td>recF</td>\n",
       "      <td>recombination protein F</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Replication, recombination and repair</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_00005</th>\n",
       "      <td>CDS</td>\n",
       "      <td>gyrB</td>\n",
       "      <td>DNA gyrase subunit B</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Replication, recombination and repair</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              type pan_gene_symbol  \\\n",
       "locus_tag                            \n",
       "SAOUHSC_00001  CDS            dnaA   \n",
       "SAOUHSC_00002  CDS            dnaN   \n",
       "SAOUHSC_00003  CDS             NaN   \n",
       "SAOUHSC_00004  CDS            recF   \n",
       "SAOUHSC_00005  CDS            gyrB   \n",
       "\n",
       "                                                  product vfdbont  \\\n",
       "locus_tag                                                           \n",
       "SAOUHSC_00001  chromosomal replication initiation protein     NaN   \n",
       "SAOUHSC_00002             DNA polymerase III subunit beta     NaN   \n",
       "SAOUHSC_00003                        hypothetical protein     NaN   \n",
       "SAOUHSC_00004                     recombination protein F     NaN   \n",
       "SAOUHSC_00005                        DNA gyrase subunit B     NaN   \n",
       "\n",
       "                                                           cog  \n",
       "locus_tag                                                       \n",
       "SAOUHSC_00001            Replication, recombination and repair  \n",
       "SAOUHSC_00002            Replication, recombination and repair  \n",
       "SAOUHSC_00003  Translation, ribosomal structure and biogenesis  \n",
       "SAOUHSC_00004            Replication, recombination and repair  \n",
       "SAOUHSC_00005            Replication, recombination and repair  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load metadata for all entities (genes, transcripts, proteins, etc.).\n",
    "meta_df = pd.read_csv(\n",
    "    meta_file,\n",
    "    index_col = id_col,\n",
    "    sep = '\\t'\n",
    ")\n",
    "display(meta_df.info())\n",
    "meta_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 3335 entries, SAOUHSC_03055 to srn_1040_sRNA92\n",
      "Data columns (total 10 columns):\n",
      " #   Column           Non-Null Count  Dtype  \n",
      "---  ------           --------------  -----  \n",
      " 0   baseMean         2503 non-null   float64\n",
      " 1   log2FoldChange   2503 non-null   float64\n",
      " 2   lfcSE            2503 non-null   float64\n",
      " 3   pvalue           2496 non-null   float64\n",
      " 4   padj             2496 non-null   float64\n",
      " 5   type             3335 non-null   object \n",
      " 6   pan_gene_symbol  1487 non-null   object \n",
      " 7   product          3326 non-null   object \n",
      " 8   vfdbont          135 non-null    object \n",
      " 9   cog              1773 non-null   object \n",
      "dtypes: float64(5), object(5)\n",
      "memory usage: 286.6+ KB\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "None"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>baseMean</th>\n",
       "      <th>log2FoldChange</th>\n",
       "      <th>lfcSE</th>\n",
       "      <th>pvalue</th>\n",
       "      <th>padj</th>\n",
       "      <th>type</th>\n",
       "      <th>pan_gene_symbol</th>\n",
       "      <th>product</th>\n",
       "      <th>vfdbont</th>\n",
       "      <th>cog</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>locus_tag</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_03055</th>\n",
       "      <td>900.082113</td>\n",
       "      <td>-0.190838</td>\n",
       "      <td>0.253922</td>\n",
       "      <td>0.107434</td>\n",
       "      <td>0.681069</td>\n",
       "      <td>CDS</td>\n",
       "      <td>rpmH</td>\n",
       "      <td>50S ribosomal protein L34</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Translation, ribosomal structure and biogenesis</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_03054</th>\n",
       "      <td>12.453537</td>\n",
       "      <td>-0.023077</td>\n",
       "      <td>0.201086</td>\n",
       "      <td>0.677658</td>\n",
       "      <td>0.943881</td>\n",
       "      <td>CDS</td>\n",
       "      <td>rnpA</td>\n",
       "      <td>ribonuclease P</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Translation, ribosomal structure and biogenesis</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_03053</th>\n",
       "      <td>167.816277</td>\n",
       "      <td>0.057208</td>\n",
       "      <td>0.196688</td>\n",
       "      <td>0.491597</td>\n",
       "      <td>0.892770</td>\n",
       "      <td>CDS</td>\n",
       "      <td>trmE</td>\n",
       "      <td>tRNA modification GTPase TrmE</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Translation, ribosomal structure and biogenesis</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_03051</th>\n",
       "      <td>223.827558</td>\n",
       "      <td>0.025922</td>\n",
       "      <td>0.177333</td>\n",
       "      <td>0.784860</td>\n",
       "      <td>0.962877</td>\n",
       "      <td>CDS</td>\n",
       "      <td>gidB</td>\n",
       "      <td>16S rRNA methyltransferase GidB</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Translation, ribosomal structure and biogenesis</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_03049</th>\n",
       "      <td>479.204817</td>\n",
       "      <td>-0.012935</td>\n",
       "      <td>0.165215</td>\n",
       "      <td>0.895382</td>\n",
       "      <td>0.981500</td>\n",
       "      <td>CDS</td>\n",
       "      <td>noc</td>\n",
       "      <td>hypothetical protein</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Cell cycle control, cell division, chromosome ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 baseMean  log2FoldChange     lfcSE    pvalue      padj type  \\\n",
       "locus_tag                                                                      \n",
       "SAOUHSC_03055  900.082113       -0.190838  0.253922  0.107434  0.681069  CDS   \n",
       "SAOUHSC_03054   12.453537       -0.023077  0.201086  0.677658  0.943881  CDS   \n",
       "SAOUHSC_03053  167.816277        0.057208  0.196688  0.491597  0.892770  CDS   \n",
       "SAOUHSC_03051  223.827558        0.025922  0.177333  0.784860  0.962877  CDS   \n",
       "SAOUHSC_03049  479.204817       -0.012935  0.165215  0.895382  0.981500  CDS   \n",
       "\n",
       "              pan_gene_symbol                          product vfdbont  \\\n",
       "locus_tag                                                                \n",
       "SAOUHSC_03055            rpmH        50S ribosomal protein L34     NaN   \n",
       "SAOUHSC_03054            rnpA                   ribonuclease P     NaN   \n",
       "SAOUHSC_03053            trmE    tRNA modification GTPase TrmE     NaN   \n",
       "SAOUHSC_03051            gidB  16S rRNA methyltransferase GidB     NaN   \n",
       "SAOUHSC_03049             noc             hypothetical protein     NaN   \n",
       "\n",
       "                                                             cog  \n",
       "locus_tag                                                         \n",
       "SAOUHSC_03055    Translation, ribosomal structure and biogenesis  \n",
       "SAOUHSC_03054    Translation, ribosomal structure and biogenesis  \n",
       "SAOUHSC_03053    Translation, ribosomal structure and biogenesis  \n",
       "SAOUHSC_03051    Translation, ribosomal structure and biogenesis  \n",
       "SAOUHSC_03049  Cell cycle control, cell division, chromosome ...  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Merge DGE results with metadata on id_col column.\n",
    "fin_df = dge_df.merge(meta_df, on=id_col, how='left')\n",
    "display(fin_df.info())\n",
    "fin_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Index: 72 entries, SAOUHSC_03025 to SAOUHSC_01818\n",
      "Data columns (total 10 columns):\n",
      " #   Column           Non-Null Count  Dtype  \n",
      "---  ------           --------------  -----  \n",
      " 0   baseMean         72 non-null     float64\n",
      " 1   log2FoldChange   72 non-null     float64\n",
      " 2   lfcSE            72 non-null     float64\n",
      " 3   pvalue           72 non-null     float64\n",
      " 4   padj             72 non-null     float64\n",
      " 5   type             72 non-null     object \n",
      " 6   pan_gene_symbol  67 non-null     object \n",
      " 7   product          72 non-null     object \n",
      " 8   vfdbont          10 non-null     object \n",
      " 9   cog              61 non-null     object \n",
      "dtypes: float64(5), object(5)\n",
      "memory usage: 6.2+ KB\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>baseMean</th>\n",
       "      <th>log2FoldChange</th>\n",
       "      <th>lfcSE</th>\n",
       "      <th>pvalue</th>\n",
       "      <th>padj</th>\n",
       "      <th>type</th>\n",
       "      <th>pan_gene_symbol</th>\n",
       "      <th>product</th>\n",
       "      <th>vfdbont</th>\n",
       "      <th>cog</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>locus_tag</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_03025</th>\n",
       "      <td>65.012526</td>\n",
       "      <td>1.248228</td>\n",
       "      <td>0.402613</td>\n",
       "      <td>8.729000e-05</td>\n",
       "      <td>0.005067</td>\n",
       "      <td>CDS</td>\n",
       "      <td>pcp</td>\n",
       "      <td>pyrrolidone-carboxylate peptidase</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Posttranslational modification, protein turnov...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_03013</th>\n",
       "      <td>24.618089</td>\n",
       "      <td>1.269422</td>\n",
       "      <td>0.587313</td>\n",
       "      <td>1.111299e-03</td>\n",
       "      <td>0.038525</td>\n",
       "      <td>CDS</td>\n",
       "      <td>hisD</td>\n",
       "      <td>histidinol dehydrogenase</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Amino acid transport and metabolism</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_03010</th>\n",
       "      <td>12.871730</td>\n",
       "      <td>2.384308</td>\n",
       "      <td>0.899698</td>\n",
       "      <td>3.365842e-04</td>\n",
       "      <td>0.015851</td>\n",
       "      <td>CDS</td>\n",
       "      <td>hisH</td>\n",
       "      <td>imidazole glycerol phosphate synthase subunit ...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Amino acid transport and metabolism</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_02706</th>\n",
       "      <td>421.621567</td>\n",
       "      <td>2.348295</td>\n",
       "      <td>0.470815</td>\n",
       "      <td>2.628841e-08</td>\n",
       "      <td>0.000007</td>\n",
       "      <td>CDS</td>\n",
       "      <td>sbi</td>\n",
       "      <td>immunoglobulin G-binding protein Sbi</td>\n",
       "      <td>Sbi</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SAOUHSC_02681</th>\n",
       "      <td>65.388397</td>\n",
       "      <td>-1.432015</td>\n",
       "      <td>0.331830</td>\n",
       "      <td>7.925336e-07</td>\n",
       "      <td>0.000110</td>\n",
       "      <td>CDS</td>\n",
       "      <td>narG</td>\n",
       "      <td>nitrate reductase subunit alpha</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Energy production and conversion; Inorganic io...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 baseMean  log2FoldChange     lfcSE        pvalue      padj  \\\n",
       "locus_tag                                                                     \n",
       "SAOUHSC_03025   65.012526        1.248228  0.402613  8.729000e-05  0.005067   \n",
       "SAOUHSC_03013   24.618089        1.269422  0.587313  1.111299e-03  0.038525   \n",
       "SAOUHSC_03010   12.871730        2.384308  0.899698  3.365842e-04  0.015851   \n",
       "SAOUHSC_02706  421.621567        2.348295  0.470815  2.628841e-08  0.000007   \n",
       "SAOUHSC_02681   65.388397       -1.432015  0.331830  7.925336e-07  0.000110   \n",
       "\n",
       "              type pan_gene_symbol  \\\n",
       "locus_tag                            \n",
       "SAOUHSC_03025  CDS             pcp   \n",
       "SAOUHSC_03013  CDS            hisD   \n",
       "SAOUHSC_03010  CDS            hisH   \n",
       "SAOUHSC_02706  CDS             sbi   \n",
       "SAOUHSC_02681  CDS            narG   \n",
       "\n",
       "                                                         product vfdbont  \\\n",
       "locus_tag                                                                  \n",
       "SAOUHSC_03025                  pyrrolidone-carboxylate peptidase     NaN   \n",
       "SAOUHSC_03013                           histidinol dehydrogenase     NaN   \n",
       "SAOUHSC_03010  imidazole glycerol phosphate synthase subunit ...     NaN   \n",
       "SAOUHSC_02706               immunoglobulin G-binding protein Sbi     Sbi   \n",
       "SAOUHSC_02681                    nitrate reductase subunit alpha     NaN   \n",
       "\n",
       "                                                             cog  \n",
       "locus_tag                                                         \n",
       "SAOUHSC_03025  Posttranslational modification, protein turnov...  \n",
       "SAOUHSC_03013                Amino acid transport and metabolism  \n",
       "SAOUHSC_03010                Amino acid transport and metabolism  \n",
       "SAOUHSC_02706                                                NaN  \n",
       "SAOUHSC_02681  Energy production and conversion; Inorganic io...  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Filter fin_df (all DGE results merged with metadata) based on\n",
    "# fold_col column against fold_th and pval_col against fdr_th.\n",
    "filt_df = fin_df[\n",
    "    (fin_df[fold_col].abs() >= fold_th) &\n",
    "    (fin_df[pval_col] <= fdr_th)\n",
    "]\n",
    "filt_df.info()\n",
    "filt_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows are understood as results/cases related to given entities\n",
      "Total rows: 3335\n",
      "Non-empty (quantified) rows: 2503\n",
      "Rows that passed filtering (differentiating): 72\n",
      "Up:   45\n",
      "Down: 27\n"
     ]
    }
   ],
   "source": [
    "# Show general statistics on the differential analysis\n",
    "# (on raw data and filtered ones).\n",
    "print('Rows are understood as results/cases related to given entities')\n",
    "print('Total rows:', fin_df.shape[0])\n",
    "print('Non-empty (quantified) rows:', fin_df[fold_col].notna().sum())\n",
    "print('Rows that passed filtering (differentiating):', filt_df.shape[0])\n",
    "print('Up:  ', (filt_df[fold_col] > 0.0).sum())\n",
    "print('Down:', (filt_df[fold_col] < 0.0).sum())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "## 6. Gen Set Enrichment Analysis (GSEA) (ontology enrichment analysis)\n",
    "Firstly, an analysis of over-representation of ontology labels among differentiating cases (filtered results) is performed in relation to the distribution of the labels in the complete population (raw input results). The analysis is based on hypergeometric distribution. Important aspects are explained in comments in the code. Do not modify the code unless you known what you are doing."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>pval</th>\n",
       "      <th>trend</th>\n",
       "      <th>total_size</th>\n",
       "      <th>sample_size</th>\n",
       "      <th>total_cases</th>\n",
       "      <th>sample_cases</th>\n",
       "      <th>downreg_cases</th>\n",
       "      <th>upreg_cases</th>\n",
       "      <th>expected</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ont</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Function unknown</th>\n",
       "      <td>4.415900e-07</td>\n",
       "      <td>under</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>974</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ncRNA</th>\n",
       "      <td>1.182506e-05</td>\n",
       "      <td>under</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>481</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Amino acid transport and metabolism</th>\n",
       "      <td>1.617569e-19</td>\n",
       "      <td>over</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>228</td>\n",
       "      <td>32</td>\n",
       "      <td>5</td>\n",
       "      <td>27</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Translation, ribosomal structure and biogenesis</th>\n",
       "      <td>9.417760e-03</td>\n",
       "      <td>under</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>207</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Inorganic ion transport and metabolism</th>\n",
       "      <td>5.237931e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>164</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Carbohydrate transport and metabolism</th>\n",
       "      <td>5.514674e-02</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>160</td>\n",
       "      <td>7</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Transcription</th>\n",
       "      <td>3.843791e-02</td>\n",
       "      <td>under</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>146</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Coenzyme transport and metabolism</th>\n",
       "      <td>4.158144e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>139</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Virulence factors</th>\n",
       "      <td>5.304979e-04</td>\n",
       "      <td>over</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>135</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>10</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Cell wall, membrane, envelope biogenesis</th>\n",
       "      <td>3.129979e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>131</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Energy production and conversion</th>\n",
       "      <td>2.886463e-02</td>\n",
       "      <td>over</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>109</td>\n",
       "      <td>6</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Replication, recombination and repair</th>\n",
       "      <td>3.306023e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>105</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Posttranslational modification, protein turnover, chaperones</th>\n",
       "      <td>6.146056e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>95</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Lipid transport and metabolism</th>\n",
       "      <td>4.152321e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>90</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Nucleotide transport and metabolism</th>\n",
       "      <td>9.014071e-07</td>\n",
       "      <td>over</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>82</td>\n",
       "      <td>11</td>\n",
       "      <td>11</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Signal transduction mechanisms</th>\n",
       "      <td>1.746552e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>79</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Defense mechanisms</th>\n",
       "      <td>2.608986e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>61</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Secondary metabolites biosynthesis, transport and catabolism</th>\n",
       "      <td>3.975547e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>42</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tRNA</th>\n",
       "      <td>3.975547e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>42</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Cell cycle control, cell division, chromosome partitioning</th>\n",
       "      <td>4.247978e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>39</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Mobilome: prophages, transposons</th>\n",
       "      <td>4.538801e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>36</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Intracellular trafficking, secretion, and vesicular transport</th>\n",
       "      <td>6.043077e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>23</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Cell motility</th>\n",
       "      <td>8.036793e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Extracellular structures</th>\n",
       "      <td>9.163633e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Intracellular trafficking and secretion</th>\n",
       "      <td>9.784108e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Chromatin structure and dynamics</th>\n",
       "      <td>9.784108e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RNA processing and modification</th>\n",
       "      <td>9.784108e-01</td>\n",
       "      <td>normal</td>\n",
       "      <td>3335</td>\n",
       "      <td>72</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                            pval   trend  \\\n",
       "ont                                                                        \n",
       "Function unknown                                    4.415900e-07   under   \n",
       "ncRNA                                               1.182506e-05   under   \n",
       "Amino acid transport and metabolism                 1.617569e-19    over   \n",
       "Translation, ribosomal structure and biogenesis     9.417760e-03   under   \n",
       "Inorganic ion transport and metabolism              5.237931e-01  normal   \n",
       "Carbohydrate transport and metabolism               5.514674e-02  normal   \n",
       "Transcription                                       3.843791e-02   under   \n",
       "Coenzyme transport and metabolism                   4.158144e-01  normal   \n",
       "Virulence factors                                   5.304979e-04    over   \n",
       "Cell wall, membrane, envelope biogenesis            3.129979e-01  normal   \n",
       "Energy production and conversion                    2.886463e-02    over   \n",
       "Replication, recombination and repair               3.306023e-01  normal   \n",
       "Posttranslational modification, protein turnove...  6.146056e-01  normal   \n",
       "Lipid transport and metabolism                      4.152321e-01  normal   \n",
       "Nucleotide transport and metabolism                 9.014071e-07    over   \n",
       "Signal transduction mechanisms                      1.746552e-01  normal   \n",
       "Defense mechanisms                                  2.608986e-01  normal   \n",
       "Secondary metabolites biosynthesis, transport a...  3.975547e-01  normal   \n",
       "tRNA                                                3.975547e-01  normal   \n",
       "Cell cycle control, cell division, chromosome p...  4.247978e-01  normal   \n",
       "Mobilome: prophages, transposons                    4.538801e-01  normal   \n",
       "Intracellular trafficking, secretion, and vesic...  6.043077e-01  normal   \n",
       "Cell motility                                       8.036793e-01  normal   \n",
       "Extracellular structures                            9.163633e-01  normal   \n",
       "Intracellular trafficking and secretion             9.784108e-01  normal   \n",
       "Chromatin structure and dynamics                    9.784108e-01  normal   \n",
       "RNA processing and modification                     9.784108e-01  normal   \n",
       "\n",
       "                                                    total_size  sample_size  \\\n",
       "ont                                                                           \n",
       "Function unknown                                          3335           72   \n",
       "ncRNA                                                     3335           72   \n",
       "Amino acid transport and metabolism                       3335           72   \n",
       "Translation, ribosomal structure and biogenesis           3335           72   \n",
       "Inorganic ion transport and metabolism                    3335           72   \n",
       "Carbohydrate transport and metabolism                     3335           72   \n",
       "Transcription                                             3335           72   \n",
       "Coenzyme transport and metabolism                         3335           72   \n",
       "Virulence factors                                         3335           72   \n",
       "Cell wall, membrane, envelope biogenesis                  3335           72   \n",
       "Energy production and conversion                          3335           72   \n",
       "Replication, recombination and repair                     3335           72   \n",
       "Posttranslational modification, protein turnove...        3335           72   \n",
       "Lipid transport and metabolism                            3335           72   \n",
       "Nucleotide transport and metabolism                       3335           72   \n",
       "Signal transduction mechanisms                            3335           72   \n",
       "Defense mechanisms                                        3335           72   \n",
       "Secondary metabolites biosynthesis, transport a...        3335           72   \n",
       "tRNA                                                      3335           72   \n",
       "Cell cycle control, cell division, chromosome p...        3335           72   \n",
       "Mobilome: prophages, transposons                          3335           72   \n",
       "Intracellular trafficking, secretion, and vesic...        3335           72   \n",
       "Cell motility                                             3335           72   \n",
       "Extracellular structures                                  3335           72   \n",
       "Intracellular trafficking and secretion                   3335           72   \n",
       "Chromatin structure and dynamics                          3335           72   \n",
       "RNA processing and modification                           3335           72   \n",
       "\n",
       "                                                    total_cases  sample_cases  \\\n",
       "ont                                                                             \n",
       "Function unknown                                            974             4   \n",
       "ncRNA                                                       481             0   \n",
       "Amino acid transport and metabolism                         228            32   \n",
       "Translation, ribosomal structure and biogenesis             207             0   \n",
       "Inorganic ion transport and metabolism                      164             3   \n",
       "Carbohydrate transport and metabolism                       160             7   \n",
       "Transcription                                               146             0   \n",
       "Coenzyme transport and metabolism                           139             2   \n",
       "Virulence factors                                           135            10   \n",
       "Cell wall, membrane, envelope biogenesis                    131             4   \n",
       "Energy production and conversion                            109             6   \n",
       "Replication, recombination and repair                       105             1   \n",
       "Posttranslational modification, protein turnove...           95             2   \n",
       "Lipid transport and metabolism                               90             1   \n",
       "Nucleotide transport and metabolism                          82            11   \n",
       "Signal transduction mechanisms                               79             0   \n",
       "Defense mechanisms                                           61             0   \n",
       "Secondary metabolites biosynthesis, transport a...           42             0   \n",
       "tRNA                                                         42             0   \n",
       "Cell cycle control, cell division, chromosome p...           39             0   \n",
       "Mobilome: prophages, transposons                             36             0   \n",
       "Intracellular trafficking, secretion, and vesic...           23             0   \n",
       "Cell motility                                                10             0   \n",
       "Extracellular structures                                      4             0   \n",
       "Intracellular trafficking and secretion                       1             0   \n",
       "Chromatin structure and dynamics                              1             0   \n",
       "RNA processing and modification                               1             0   \n",
       "\n",
       "                                                    downreg_cases  \\\n",
       "ont                                                                 \n",
       "Function unknown                                                1   \n",
       "ncRNA                                                           0   \n",
       "Amino acid transport and metabolism                             5   \n",
       "Translation, ribosomal structure and biogenesis                 0   \n",
       "Inorganic ion transport and metabolism                          1   \n",
       "Carbohydrate transport and metabolism                           6   \n",
       "Transcription                                                   0   \n",
       "Coenzyme transport and metabolism                               0   \n",
       "Virulence factors                                               0   \n",
       "Cell wall, membrane, envelope biogenesis                        1   \n",
       "Energy production and conversion                                5   \n",
       "Replication, recombination and repair                           0   \n",
       "Posttranslational modification, protein turnove...              1   \n",
       "Lipid transport and metabolism                                  0   \n",
       "Nucleotide transport and metabolism                            11   \n",
       "Signal transduction mechanisms                                  0   \n",
       "Defense mechanisms                                              0   \n",
       "Secondary metabolites biosynthesis, transport a...              0   \n",
       "tRNA                                                            0   \n",
       "Cell cycle control, cell division, chromosome p...              0   \n",
       "Mobilome: prophages, transposons                                0   \n",
       "Intracellular trafficking, secretion, and vesic...              0   \n",
       "Cell motility                                                   0   \n",
       "Extracellular structures                                        0   \n",
       "Intracellular trafficking and secretion                         0   \n",
       "Chromatin structure and dynamics                                0   \n",
       "RNA processing and modification                                 0   \n",
       "\n",
       "                                                    upreg_cases  expected  \n",
       "ont                                                                        \n",
       "Function unknown                                              3        21  \n",
       "ncRNA                                                         0        10  \n",
       "Amino acid transport and metabolism                          27         5  \n",
       "Translation, ribosomal structure and biogenesis               0         4  \n",
       "Inorganic ion transport and metabolism                        2         4  \n",
       "Carbohydrate transport and metabolism                         1         3  \n",
       "Transcription                                                 0         3  \n",
       "Coenzyme transport and metabolism                             2         3  \n",
       "Virulence factors                                            10         3  \n",
       "Cell wall, membrane, envelope biogenesis                      3         3  \n",
       "Energy production and conversion                              1         2  \n",
       "Replication, recombination and repair                         1         2  \n",
       "Posttranslational modification, protein turnove...            1         2  \n",
       "Lipid transport and metabolism                                1         2  \n",
       "Nucleotide transport and metabolism                           0         2  \n",
       "Signal transduction mechanisms                                0         2  \n",
       "Defense mechanisms                                            0         1  \n",
       "Secondary metabolites biosynthesis, transport a...            0         1  \n",
       "tRNA                                                          0         1  \n",
       "Cell cycle control, cell division, chromosome p...            0         1  \n",
       "Mobilome: prophages, transposons                              0         1  \n",
       "Intracellular trafficking, secretion, and vesic...            0         0  \n",
       "Cell motility                                                 0         0  \n",
       "Extracellular structures                                      0         0  \n",
       "Intracellular trafficking and secretion                       0         0  \n",
       "Chromatin structure and dynamics                              0         0  \n",
       "RNA processing and modification                               0         0  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Total size is the number of all analysed entities,\n",
    "# sample size, the number of differentiating cases\n",
    "# (filtered results).\n",
    "total_size  = fin_df.shape[0]\n",
    "sample_size = filt_df.shape[0]\n",
    "# Create a new DataFrame for over-representation analysis.\n",
    "columns = ('ont pval trend total_size sample_size ' +\n",
    "           'total_cases sample_cases ' +\n",
    "           'downreg_cases upreg_cases expected').split()\n",
    "ont_df = pd.DataFrame(columns=columns)\n",
    "\n",
    "# Calculate total number of each ontology occurrences, and\n",
    "# the number of occurrences of each ontology label among\n",
    "# differentiating cases.\n",
    "tot_ont_counts = count_onts(fin_df)\n",
    "filt_ont_folds = collect_folds(filt_df)\n",
    "\n",
    "# For each ontology label (sorted by the number of cases\n",
    "# in the complete population in a descending order)\n",
    "# add a row to ont_df with results of over-representation analysis.\n",
    "for ont in sorted(tot_ont_counts, key=lambda ont: tot_ont_counts[ont],\n",
    "                  reverse=True):\n",
    "    # Fetch the number of total cases of a given ontology label\n",
    "    # (the number among all entities).\n",
    "    total_cases  = tot_ont_counts[ont]\n",
    "    # Fetch the number of sample cases of a given ontology label\n",
    "    # (the number among differentiating cases).\n",
    "    if ont in filt_ont_folds:\n",
    "        # The number of occurences is equal to the number of\n",
    "        # fold change values in the item list.\n",
    "        sample_cases = len(filt_ont_folds[ont])\n",
    "        # Additionally count separately up and down-regulated cases.\n",
    "        downreg_cases = sum(filt_ont_folds[ont] <= 0)\n",
    "        upreg_cases   = sum(filt_ont_folds[ont] > 0)\n",
    "    else:\n",
    "        # If the key is not in filt_ont_folds, it means that\n",
    "        # no occurrences are observed.\n",
    "        sample_cases, downreg_cases, upreg_cases = 0, 0, 0\n",
    "    # Count the expected number of occurrences of a given ontology label\n",
    "    # among differentiating cases.\n",
    "    expected = int( round(total_cases/total_size * sample_size) )\n",
    "\n",
    "    # Calculate cumulative probability and describe every ontology label\n",
    "    # with a proper trend: under(-represented), over(-represented), normal.\n",
    "    pval = hypgeoprob(total_size, sample_size, total_cases, sample_cases)\n",
    "    if pval < -0.05 or pval > 0.05:\n",
    "        trend = 'normal'\n",
    "    else:\n",
    "        trend = 'under' if pval < 0.0 else 'over'\n",
    "    pval = abs(pval)\n",
    "\n",
    "    # Put the collected statistics into the ont_df DataFrame.\n",
    "    ont_df.loc[ ont_df.shape[0] ] = ont, pval, trend, total_size, sample_size, \\\n",
    "                                    total_cases, sample_cases, \\\n",
    "                                    downreg_cases, upreg_cases, expected\n",
    "\n",
    "# Replace the default numeric DataFrame index with the values\n",
    "# of ontology labels (the 'ont' column).\n",
    "ont_df.set_index('ont', drop=True, inplace=True)\n",
    "ont_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "## 7. Visulisation of GSEA results combined with fold change values\n",
    "Secondly, GSEA results combined with fold change values for all differentiating cases are visualised as a pie chart. Dimensions of wedges of the chart are proportional to the number of cases described by each ontology label (the numbers are also given in the chart, next to wedges). Pieces of wedges corresponding to down-regulated cases shoot inwards from the scaffold circle, whereas those corresponding to up-regulated ones, outwards. Fold change values are mapped to blues and reds, respectively.\n",
    "\n",
    "The chart is of very specific requirements, thus it is not rendered with any high-level Matplotlib plotting function. Instead Matplotlib patches such as Circle and Wedge have been utilised and the chart is plotted based on manually calculated coordinates. Consequently the code is a bit lengthy. Do not modify it unless you known what you are doing."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Collect ontology labels to which at least one differentiating\n",
    "# result is assigned. Sort them according to the manual list\n",
    "# of labels, if provided, or the number of transcripts\n",
    "# that fall in each ontology in a descending order (default).\n",
    "# Skip selected labels if such indicated.\n",
    "if sel_onts != None and len(sel_onts) > 0:\n",
    "    onts_list = [ [ontname, filt_ont_folds[ontname]] for ontname in sel_onts\n",
    "                  if ontname in filt_ont_folds ]\n",
    "else:\n",
    "    onts_list = [ [ontname, folds] for ontname, folds\n",
    "                  in filt_ont_folds.items() ]\n",
    "    onts_list = sorted(onts_list, key=lambda items: len(items[1]),\n",
    "                       reverse=True)\n",
    "if skip_onts is None:\n",
    "    skip_onts = []\n",
    "onts_list = [ [ontname, folds] for ontname, folds in onts_list\n",
    "              if len(folds) >= min_size and ontname not in skip_onts ]\n",
    "\n",
    "# Find the maximal absolute fold change value (for scaling the heatmap).\n",
    "if max_fold is None:\n",
    "    max_fold = np.max([ np.abs(folds).max() for _, folds in onts_list ])\n",
    "\n",
    "# Collect sizes (lengths, the number of cases) for each ontology.\n",
    "# Square-root sizes (scaling). Next calculate arc sizes (360° scaffold\n",
    "# circle pieces) proportional to scaled ontology sizes. Base on the arc\n",
    "# sizes assign to each ontology wedge its beginning radius (theta).\n",
    "# Shift each theta in order the third ontology starts at hour 12 (0°).\n",
    "sizes   = np.array( [ len(ont[1]) for ont in onts_list ] )\n",
    "scaled  = np.sqrt(sizes)\n",
    "arcs    = 360*scaled / scaled.sum()\n",
    "thetas  = np.append([0], arcs[:-1]).cumsum() + 90 - arcs[:2].sum() + angle_offset\n",
    "\n",
    "# Calculate the fraction of positive fold change values\n",
    "# for each ontology label.\n",
    "possizes = [ len([ value for value in ont[1] if value >= 0.0 ])\n",
    "             for ont in onts_list ]\n",
    "posfracs = possizes / sizes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x450 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Draw a pie chart depicting of fold change values\n",
    "# within ontology labels found among differentiating cases.\n",
    "# Make each wedge (ontology label) radial and arc dimensions\n",
    "# proportional to the number of cases falling into the label.\n",
    "# Draw each result as a subwedge of colour proportional to\n",
    "# its fold change value. Draw the part of each wedge that\n",
    "# corresponds to up-regulated cases outwards in respect\n",
    "# to the scaffold circle, and the down-regulated inwards.\n",
    "\n",
    "# Calculate the ratio of the figure dimensions.\n",
    "# That is necessary to calculate ranges of axes in order\n",
    "# to avoid the circular shape turning into elliptical.\n",
    "ratio = fig_h/fig_w\n",
    "\n",
    "# Create a figure and axes.\n",
    "fig = plt.figure(figsize=(fig_w,fig_h), facecolor='white', dpi=150)\n",
    "\n",
    "ax = fig.add_axes([0.0, 0.0, 1.0, 1.0])\n",
    "ax.set_axis_off()\n",
    "\n",
    "# Create a new DataFrame to which geometrical measures of each\n",
    "# ontology wedge will be collected. These measures will be\n",
    "# used to position connectors and labels:\n",
    "# theta -- start radius\n",
    "# r_in  -- radial width of the wedge portion directed inwards\n",
    "#          the scaffold circle (log2FoldChange < 0)\n",
    "# r_out -- radial width of the wedge portion directed outwards\n",
    "#          the scaffold circle (log2FoldChange > 0)\n",
    "# text  -- wedge label (ontology name)\n",
    "# sign  -- asterisks denoting significance thresholds for\n",
    "#          up-regulated transcripts\n",
    "# color -- label colour (red if an ontology is over-represented)\n",
    "label_cols = 'theta r_in r_out text sign color'.split()\n",
    "labels_df = pd.DataFrame(columns=label_cols)\n",
    "\n",
    "# Iterate over calculated thetas, arc sizes, scaled sizes, positive fractions\n",
    "# and ontology list to fill up the labels_df DataFrame, i.e. measures of wedges\n",
    "# corresponding to ontologies, as well as to draw subwedges corresponding\n",
    "# to cases within each ontology.\n",
    "for theta, arc, size, posfrac, rawdata in zip(\n",
    "    thetas, arcs, scaled, posfracs, onts_list):\n",
    "    # Calculate the \"real\" radial width of a wedge, i.e. the actual\n",
    "    # radial width of a wedge based on the scaled number of cases falling\n",
    "    # into an ontology and an arbitrarily assumed scale.\n",
    "    real_w = size*scale\n",
    "    \n",
    "    # Collect the ontology name as well as the positive and negative\n",
    "    # fold change values\n",
    "    ontname  = rawdata[0]\n",
    "    negfolds = sorted([ abs(fold) for fold in rawdata[1] if fold <  0.0 ])\n",
    "    posfolds = sorted([ fold      for fold in rawdata[1] if fold >= 0.0 ])\n",
    "\n",
    "    # Draw subwedges. Calculate their radial span as a fraction of the real\n",
    "    # radial width of a wedge. Use colours proportional to fold change values\n",
    "    # scaled against the maximal value (max_fold). Use scale of reds for\n",
    "    # up-regulated transcripts and of blues for down-regulated ones.\n",
    "    for i, fold in enumerate(negfolds):\n",
    "        span = real_w*(1.0-posfrac) / len(negfolds)\n",
    "        color = (1.0-abs(fold)/max_fold, 1.0-abs(fold)/max_fold, 1.0)\n",
    "        ax.add_patch( Wedge((0.0, 0.0), pie_r-span*i, theta, theta+arc,\n",
    "                            width=span, linewidth=0.5, edgecolor=color,\n",
    "                            facecolor=color) )\n",
    "\n",
    "    for i, fold in enumerate(posfolds):\n",
    "        span = real_w*posfrac / len(posfolds)\n",
    "        color = (1.0, 1.0-abs(fold)/max_fold, 1.0-abs(fold)/max_fold)\n",
    "        ax.add_patch( Wedge((0.0, 0.0), pie_r+span*(i+1), theta, theta+arc,\n",
    "                            width=span, linewidth=0.5, edgecolor=color,\n",
    "                            facecolor=color) )\n",
    "\n",
    "    # Sketch a wedge frame around already drawn subwedges.\n",
    "    ax.add_patch( Wedge((0.0, 0.0), pie_r+posfrac*real_w, theta, theta+arc,\n",
    "                        width=real_w, fill=False, linewidth=0.3) )\n",
    "    \n",
    "    # Assigned colour and optionally add asterisks to an ontology (wedge) label\n",
    "    # if the ontology happens to be over-represented.\n",
    "    sign, color = get_sign(*ont_df.loc[ontname, ['pval', 'trend']])\n",
    "\n",
    "    # Recalculate degrees into radians. Add all collected and\n",
    "    # calculated measures to labels_df DataFrame.\n",
    "    theta_rad = (theta+0.5*arc)/180*np.pi\n",
    "    labels_df.loc[ labels_df.shape[0] ] = theta_rad,      \\\n",
    "        pie_r-(1.0-posfrac)*real_w, pie_r+posfrac*real_w, \\\n",
    "        ontname, sign, color\n",
    "\n",
    "# Sketch the scaffold circle across already drawn wedges.\n",
    "ax.add_patch( Circle((0.0, 0.0), pie_r, fill=False, linewidth=0.2,\n",
    "                     color='black') )\n",
    "\n",
    "# Based on values from different columns from labels_df DataFrame, calculate\n",
    "# coordinates of connectors fragments, number of transcript labels and\n",
    "# ontology labels. Add the coordinates to labels_df DataFrame.\n",
    "# numup_       -- labels for number of up-regulated transcripts,\n",
    "#                 (outwards the wedges)\n",
    "# numupdown_   -- labels for number of down-regulated transcripts,\n",
    "#                 (inwards the wedges)\n",
    "# Coordinates of connector points starting at a wedge and finishing at\n",
    "# an ontology (wedge) label:\n",
    "# conn_term_   -- start points at the wedge\n",
    "# conn_proxy_  -- points on the outer scaffold circle\n",
    "# conn_middle_ -- x and y values shared by next two middle points\n",
    "# conn_init_   -- points at ontology (wedge) labels\n",
    "labels_df['numup_x']   = np.cos(labels_df['theta']) *  \\\n",
    "                         (labels_df['r_out'] + margin)\n",
    "labels_df['numup_y']   = np.sin(labels_df['theta']) *  \\\n",
    "                         (labels_df['r_out'] + margin)\n",
    "\n",
    "labels_df['numdown_x'] = np.cos(labels_df['theta']) *  \\\n",
    "                         (labels_df['r_in'] - margin)\n",
    "labels_df['numdown_y'] = np.sin(labels_df['theta']) *  \\\n",
    "                         (labels_df['r_in'] - margin)\n",
    "\n",
    "labels_df['conn_term_x'] = np.cos(labels_df['theta']) *      \\\n",
    "                           (labels_df['r_out'] + 2.0*margin)\n",
    "labels_df['conn_term_y'] = np.sin(labels_df['theta']) *      \\\n",
    "                           (labels_df['r_out'] + 2.0*margin)\n",
    "\n",
    "proxy_r = labels_df['r_out'].max() + 3.0*margin\n",
    "\n",
    "labels_df['conn_proxy_x'] = np.cos(labels_df['theta']) * (proxy_r)\n",
    "labels_df['conn_proxy_y'] = np.sin(labels_df['theta']) * (proxy_r)\n",
    "\n",
    "labels_df.sort_values('conn_proxy_y', ascending=False, inplace=True)\n",
    "sub_index_left = labels_df.loc[ labels_df['conn_proxy_x'] <= 0.0 ].index\n",
    "sub_index_right = labels_df.index.difference(sub_index_left, sort=False)\n",
    "\n",
    "for sub_index in sub_index_left, sub_index_right:\n",
    "    max_rank = sub_index.shape[0] // 2\n",
    "    labels_df.loc[sub_index, 'rank'] = pd.Series(\n",
    "        range(max_rank, max_rank-sub_index.shape[0], -1), index=sub_index)\n",
    "\n",
    "labels_df.loc[sub_index_left, 'rank_norm'] =              \\\n",
    "    (labels_df.loc[sub_index_left, 'rank'] +              \\\n",
    "    np.abs(labels_df.loc[sub_index_left,  'rank'].min()))\n",
    "labels_df.loc[sub_index_left, 'rank_norm'] =        \\\n",
    "    labels_df.loc[sub_index_left, 'rank_norm'] /    \\\n",
    "    labels_df.loc[sub_index_left,'rank_norm'].max()\n",
    "\n",
    "labels_df.loc[sub_index_right, 'rank_norm'] =              \\\n",
    "    (labels_df.loc[sub_index_right, 'rank'] +              \\\n",
    "    np.abs(labels_df.loc[sub_index_right,  'rank'].min()))\n",
    "labels_df.loc[sub_index_right, 'rank_norm'] =        \\\n",
    "    labels_df.loc[sub_index_right, 'rank_norm'] /    \\\n",
    "    labels_df.loc[sub_index_right,'rank_norm'].max()\n",
    "\n",
    "labels_df['rank_norm'] = labels_df['rank_norm'].fillna(0.0)\n",
    "\n",
    "labels_df['conn_init_y'] = label_height * labels_df['rank']\n",
    "labels_df.loc[labels_df['conn_term_x'] <  0, 'conn_init_x'] = label_at * -1\n",
    "labels_df.loc[labels_df['conn_term_x'] >= 0, 'conn_init_x'] = label_at\n",
    "\n",
    "conn_range = label_at - proxy_r\n",
    "\n",
    "# Upper right quarter labels (if x-axis is reversed, default).\n",
    "sub_index = labels_df[\n",
    "    (labels_df['conn_init_x'] < 0) &\n",
    "    (labels_df['conn_init_y'] < labels_df['conn_proxy_y'])\n",
    "].index\n",
    "if sub_index.shape[0] != 0:\n",
    "    step = conn_range / sub_index.shape[0]\n",
    "    labels_df.loc[sub_index, 'conn_middle_x'] = -1* proxy_r - \\\n",
    "        step * np.arange(sub_index.shape[0]-1, -1, -1)\n",
    "\n",
    "# Upper left quarter labels (if x-axis is reversed, default).\n",
    "sub_index = labels_df[\n",
    "    (labels_df['conn_init_x'] > 0) &\n",
    "    (labels_df['conn_init_y'] < labels_df['conn_proxy_y'])\n",
    "].index\n",
    "if sub_index.shape[0] != 0:\n",
    "    step = conn_range / sub_index.shape[0]\n",
    "    labels_df.loc[sub_index, 'conn_middle_x'] = proxy_r + \\\n",
    "        step * np.arange(sub_index.shape[0]-1, -1, -1)\n",
    "\n",
    "# Lower right quarter labels (if x-axis is reversed, default).\n",
    "sub_index = labels_df[\n",
    "    (labels_df['conn_init_x'] < 0) &\n",
    "    (labels_df['conn_init_y'] > labels_df['conn_proxy_y'])\n",
    "].index\n",
    "if sub_index.shape[0] != 0:\n",
    "    step = conn_range / sub_index.shape[0]\n",
    "    labels_df.loc[sub_index, 'conn_middle_x'] = -proxy_r - \\\n",
    "        step * np.arange(1, sub_index.shape[0]+1, 1)\n",
    "\n",
    "# Lower left quarter labels (if x-axis is reversed, default).\n",
    "sub_index = labels_df[\n",
    "    (labels_df['conn_init_x'] > 0) &\n",
    "    (labels_df['conn_init_y'] > labels_df['conn_proxy_y'])\n",
    "].index\n",
    "if sub_index.shape[0] != 0:\n",
    "    step = conn_range / sub_index.shape[0]\n",
    "    labels_df.loc[sub_index, 'conn_middle_x'] = proxy_r + \\\n",
    "        step * np.arange(1, sub_index.shape[0]+1, 1)\n",
    "\n",
    "# Iterate over labels_df DataFrame and based on already calculated\n",
    "# geometric measures draw connectors and labels.\n",
    "for i, row in labels_df.iterrows():\n",
    "    x, y = row['numup_x'], row['numup_y']\n",
    "    ax.text(x, y, ont_df.loc[row['text'], 'upreg_cases'],\n",
    "            va='center', ha='center', fontsize=num_font)\n",
    "\n",
    "    x, y = row['numdown_x'], row['numdown_y']\n",
    "    ax.text(x, y, ont_df.loc[row['text'], 'downreg_cases'],\n",
    "            va='center', ha='center', fontsize=num_font)\n",
    "    \n",
    "    X = row['conn_init_x conn_middle_x conn_middle_x'.split() +\n",
    "            'conn_proxy_x conn_term_x'.split()]\n",
    "    Y = row['conn_init_y conn_init_y conn_proxy_y'.split() +\n",
    "            'conn_proxy_y conn_term_y'.split()]\n",
    "    ax.plot(X, Y, linewidth=0.4, color='black', solid_joinstyle='round',\n",
    "            solid_capstyle='round')\n",
    "\n",
    "    x, y, text, sign, color = \\\n",
    "        row['conn_init_x conn_init_y text sign color'.split()]\n",
    "    if x < 0:\n",
    "        x -= margin\n",
    "        label = f'{text} {sign}'\n",
    "        ha    = 'left'\n",
    "    else:\n",
    "        x += margin\n",
    "        label = f'{sign} {text}'\n",
    "        ha    = 'right'\n",
    "    ax.text(x, y, label, va='center', ha=ha, color=color, fontsize=label_font)\n",
    "\n",
    "# Draw the log2FoldChange scale bar (heatmap) on separate axes, assuming\n",
    "# arbitrary number of cells (colour bands) and their widths (sizes).\n",
    "bar_cell_size  = 0.05\n",
    "bar_cell_count = 7\n",
    "ax_bar = fig.add_axes([0.88, 0.1, bar_cell_size*ratio*bar_cell_count,\n",
    "                       bar_cell_size])\n",
    "ax_bar.set_yticks([])\n",
    "for spine in ['left', 'top', 'right', 'bottom']:\n",
    "    ax_bar.spines[spine].set_linewidth(0.6)\n",
    "ax_bar.tick_params(width=0.6, labelsize=6.0)\n",
    "steps  = np.linspace(0.0, 1.0, 50)\n",
    "colors = np.stack([steps, steps, np.ones(steps.shape[0])], axis=1)\n",
    "steps  = np.linspace(1.0, 0.0, 50)[1:]\n",
    "colors = np.vstack([colors, np.stack([np.ones(steps.shape[0]), steps, steps],\n",
    "                    axis=1)])\n",
    "ax_bar.imshow(\n",
    "    colors.reshape((1,99,3)),\n",
    "    extent = (-max_fold, max_fold, 0, max_fold/(bar_cell_count + 1))\n",
    ")\n",
    "ax_bar.set_title(scale_bar_label, fontsize=sbar_font)\n",
    "\n",
    "# Set x-axis limits. Based on figure\n",
    "# fig_h/fig_w ratio, calculate y-axis limits.\n",
    "xrange = np.abs(xmax - xmin)\n",
    "xmin_p = -xrange - xmin\n",
    "xmax_p =  xrange - xmax\n",
    "ax.set_xlim(xmax_p, xmin_p)\n",
    "ax.set_ylim(-xrange/2*ratio, xrange/2*ratio)\n",
    "\n",
    "# Save the figure.\n",
    "fig.savefig(chart_file, transparent=False, dpi=dpi)\n",
    "Image.open(chart_file).convert('RGB').save(chart_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}