scripts/8.plot_ensemble_dist.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## ❓ 8. Quantify network uncertainity (optional)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You are now familiar with how the CarveMe can build ensemble models, next we will inspect the generated output. This notebook loads the ensembles directly from the `/ensembles/` folder. You may follow along passively on GitHub or launch the jupyter notebook from your terminal to interactively run through each code chunk. To do so, launch jupyter from your terminal:\n",
    "```bash\n",
    "$ cd $ROOT\n",
    "$ jupyter notebook --browser firefox\n",
    "```\n",
    "This should launch a browser window where you can navigate to the `SymbNET/scripts/` folder and click on the appropriate file to launch this notebook interactively.\n",
    "\n",
    "Note: this script is forked from Daniel's [carveme_paper repo](https://github.com/cdanielmachado/carveme_paper/blob/master/notebooks/Ensemble%20distances.ipynb).\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "from scipy.spatial.distance import jaccard\n",
    "from framed.cobra.ensemble import load_ensemble\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "sns.set_style('white')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 🍎 Gut normal models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:306: UserWarning: Dataset has 0 variance; skipping density estimate.\n",
      "  warnings.warn(msg, UserWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x252 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# gut normal species\n",
    "organisms = ['B. uniformis', 'R. bromii', 'B. wexlerae', 'E. rectale', 'F. saccharivorans']\n",
    "\n",
    "models = {\n",
    "    'B. uniformis': '../ensembles/ERR260255_bin.24.s.xml',\n",
    "    'R. bromii': '../ensembles/ERR260255_bin.7.p.xml',\n",
    "    'B. wexlerae': '../ensembles/ERR260255_bin.9.s.xml',\n",
    "    'E. rectale': '../ensembles/ERR260255_bin.19.p.xml',\n",
    "    'F. saccharivorans': '../ensembles/ERR260255_bin.14.p.xml'\n",
    "}\n",
    "\n",
    "for label, model_file in models.items():\n",
    "    models[label] =  load_ensemble(model_file, flavor='cobra')\n",
    "    \n",
    "# calculate distances\n",
    "def ensemble_distance(ensemble):\n",
    "    bin_matrix = [[int(ensemble.reaction_states[r_id][i])\n",
    "                   for r_id in ensemble.reaction_states]\n",
    "                  for i in range(ensemble.size)]\n",
    "    \n",
    "    distances = [jaccard(bin_matrix[i], bin_matrix[j])\n",
    "                 for i in range(ensemble.size)\n",
    "                 for j in range(i+1, ensemble.size)]\n",
    "    \n",
    "    return distances\n",
    "dists = {label: ensemble_distance(model) for label, model in models.items()}\n",
    "\n",
    "# plot\n",
    "_, ax = plt.subplots(figsize=(8,3.5))\n",
    "colors = sns.color_palette(n_colors=5)\n",
    "for i, label in enumerate(organisms):\n",
    "    sns.distplot(dists[label], ax=ax, bins=20, kde=True, norm_hist=True, color=colors[i],\n",
    "                  hist_kws={'alpha': 0.2}, kde_kws={'shade': True, 'alpha':0.9})\n",
    "\n",
    "ax.set_xlabel('Jaccard distance')\n",
    "ax.set_ylabel('Frequency')\n",
    "ax.legend(organisms)\n",
    "ax.set_xlim(0,0.6)\n",
    "ax.set_ylim(0,70)\n",
    "plt.tight_layout()\n",
    "plt.savefig('../plots/ensemble_dist/gut_normal.png',dpi=300)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Compare with [plot generated for CarveMe paper](https://raw.githubusercontent.com/cdanielmachado/carveme_paper/master/results/other/ensemble_distances.png). Note that ensembles with no variance in Jaccard distance are not plotted and have a shaded out color in the legend."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 🍰 Gut impaired models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:306: UserWarning: Dataset has 0 variance; skipping density estimate.\n",
      "  warnings.warn(msg, UserWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:306: UserWarning: Dataset has 0 variance; skipping density estimate.\n",
      "  warnings.warn(msg, UserWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x252 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# gut impaired species\n",
    "organisms = ['B. uniformis', 'R. bromii', 'B. wexlerae', 'E. rectale', 'F. saccharivorans']\n",
    "\n",
    "models = {\n",
    "    'B. uniformis': '../ensembles/ERR260172_bin.10.p.xml',\n",
    "    'R. bromii': '../ensembles/ERR260172_bin.44.p.xml',\n",
    "    'B. wexlerae': '../ensembles/ERR260172_bin.7.s.xml',\n",
    "    'E. rectale': '../ensembles/ERR260172_bin.31.s.xml',\n",
    "    'F. saccharivorans': '../ensembles/ERR260172_bin.34.p.xml'\n",
    "}\n",
    "\n",
    "for label, model_file in models.items():\n",
    "    models[label] =  load_ensemble(model_file, flavor='cobra')\n",
    "\n",
    "# calculate distances\n",
    "def ensemble_distance(ensemble):\n",
    "    bin_matrix = [[int(ensemble.reaction_states[r_id][i])\n",
    "                   for r_id in ensemble.reaction_states]\n",
    "                  for i in range(ensemble.size)]\n",
    "    \n",
    "    distances = [jaccard(bin_matrix[i], bin_matrix[j])\n",
    "                 for i in range(ensemble.size)\n",
    "                 for j in range(i+1, ensemble.size)]\n",
    "    \n",
    "    return distances\n",
    "dists = {label: ensemble_distance(model) for label, model in models.items()}\n",
    "\n",
    "# plot\n",
    "_, ax = plt.subplots(figsize=(8,3.5))\n",
    "colors = sns.color_palette(n_colors=5)\n",
    "for i, label in enumerate(organisms):\n",
    "    sns.distplot(dists[label], ax=ax, bins=20, kde=True, norm_hist=True, color=colors[i],\n",
    "                  hist_kws={'alpha': 0.2}, kde_kws={'shade': True, 'alpha':0.9})\n",
    "\n",
    "ax.set_xlabel('Jaccard distance')\n",
    "ax.set_ylabel('Frequency')\n",
    "ax.legend(organisms)\n",
    "ax.set_xlim(0,0.6)\n",
    "ax.set_ylim(0,70)\n",
    "plt.tight_layout()\n",
    "plt.savefig('../plots/ensemble_dist/gut_impaired.png',dpi=300)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 🏥 Gut T2D models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:306: UserWarning: Dataset has 0 variance; skipping density estimate.\n",
      "  warnings.warn(msg, UserWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjgAAAD0CAYAAACWwFwQAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAA7I0lEQVR4nO3dd3xUdb7/8deZlt5I6KACUmRdVlEBRVGRtqKyFqpExVWviw0sCyrNlSKCooYFwV2VCxZAkSurrg1/soKiIi6KSmgqPZSEJJNk2jm/PyYJICQZIJPJhPfz8cgjyZkz53zmECbvfM/3nI9hWZaFiIiISB1ii3QBIiIiItVNAUdERETqHAUcERERqXMUcERERKTOUcARERGROkcBR0REROocRzg2umTJEt566y0APB4PP/74I6+++iqTJ0/GMAxat27N+PHjsdmUr0RERKT6GeG+D85jjz1Gu3bt+OSTTxg2bBidO3dm3LhxXHLJJfTs2TOcuxYREZFTVFiHUL777js2bdrEwIEDWb9+PZ06dQKgW7durFq1Kpy7FhERkVNYWE5RlZkzZw533XUXAJZlYRgGAAkJCRQUFFT63E6dOpPesFFI+3HZbdhtxskVKyIiIrXSjh07WL169XE9J2wBJz8/ny1bttClSxeAI+bbuN1ukpOTK31+4yZNmDjz5ZD21SQ1lvTEmBOuVURERGqv66677rifE7ZTVF999RUXXXRR+fft27cvT18rVqzg/PPPD9euRURE5BQXtoCzdetWmjVrVv79qFGjyMrKYuDAgfh8Pnr37h2uXYuIiMgpLmynqG677bYjvm/RogULFiwI1+5EREREyoV1krGIiEhN8Pl8bN++nZKSkkiXIichNjaWZs2a4XQ6T3pbCjgiIhL1tm/fTlJSEmeccUb5FbsSXSzLYv/+/Wzfvp0WLVqc9PZ0K2EREYl6JSUlpKenK9xEMcMwSE9Pr7ZROAUcERGpExRuol91/hvqFJWIiNQ5F0z6iL0FnmrbXv2kGL56tEe1bU/CTwFHRETqnL0FHhokVd8NYHOqCEurV69mxIgRnHnmmViWhd/vZ9KkSbRq1apa9n/33Xczc+ZM1q1bx8MPP0z37t154IEHTnq7S5YsISUlhSuuuKIaqqxdFHBERESqQZcuXZgxYwYAn332GU8++SRz5syplm3PnDmzfLuDBg0iMzOzWrZ7IncIjhYKOCIiItUsPz+fpk2bVvj46tWref3118sDUdeuXVm5ciWjR4/G5XKxY8cOcnJyeOKJJ/jd735H165dmT17Nm+88QZOp5NGjRoRHx/PM888Q0xMDKmpqUyePJkff/yR6dOn43Q6GTBgAP/85z85//zzyc7OpkWLFqSnp/P111/jcrmYO3cuzz//PBkZGfTu3ZsRI0ZgWRY+n4/HHnuMtm3b1tThCgsFHBERkWrwxRdfkJmZidfrZcOGDSc8etOkSRP+9re/sWjRIhYuXMjf/vY3ADp06MC1115LRkYGPXr04IorruC1116jYcOGzJs3j9mzZ3PZZZfh8XhYvHgxAM899xxXXXUV5513Hn369OHhhx9m5MiRDB06lE2bNpXvc926dSQlJfHUU0+xadMmCgsLT/6ARJiuohIREakGXbp0Yf78+SxcuJAlS5Zwzz33hHzJs2VZ5V+fddZZADRq1Aiv13vM9XNzc0lMTKRhw4YAXHDBBWzcuBHgqHvI/O53vwMgOTm5fE5QcnIyHs+heUXdunXjggsuYPjw4Tz33HNHNMiOVtH/CkRERGqZjIyMSh+PiYlh7969AOzYsYODBw+WPxbKpdJpaWkUFhaSk5MDwJdffskZZ5wBcFQ4CWV7q1evpkGDBrz44ov85S9/4emnn67yObWdTlGJiEidUz8ppsorn453e1UpO0Vls9lwu92MHj2a2NhYli1bRlFREQMHDixf9+yzzyYpKYn+/fvTqlWrI5pTh8IwDCZOnMg999yDYRikpKQwZcqU8lGc49WuXTtGjhzJvHnzsNls3HXXXSe0ndrEsA4fF6tF+v3pWibOfDmkdZukxpKeWH2XA4qISHT58ccfy0/tSHQ71r/lddddx5IlS45rOzpFJSIiInWOAo6IiIjUOQo4IiIiUuco4IiIiEido4AjIiIidY4uExcRkbpnehso3FN920tsCA9mV9/2JOwUcEREpO4p3AOJjapxe7srffjwbuIAbrebZs2aMX36dFwu1zGfk5mZyYQJE6qt4/jhRo4cydSpU3n55Zfp0qULHTp0qPZ91HZhCzhz5sxh+fLl+Hw+Bg8eTKdOnRg9ejSGYdC6dWvGjx9fJ24FLSIiAkd2Ewd44IEHWL58OX369KnxWsrquOOOO2p837VFWALO6tWrWbt2La+99hrFxcW8+OKLTJkyhREjRtC5c2fGjRvHxx9/TM+ePcOxexERkYjyer3k5OSQkpJS6XrPPfccubm5uFwunnzySTZu3HhEN/D69esfs2P43LlzcTqd7N69m0GDBvHFF1/w008/cdNNNzFkyBC6d+/Oe++9x/jx47nyyivp1q1bDb3y2iMsAeezzz6jTZs23HXXXRQWFvLXv/6VRYsW0alTJyDY1GvlypUKOCIiUmeUtWrYv38/NpuNAQMGcOGFF1b6nF69etG3b19eeeUV5syZQ/fu3cu7gVuWVWHH8N27d7N06VLWr1/Pfffdx4cffsiePXu4++67GTJkSA294totLOeIcnNz+f7773n22Wd57LHHePDBB7Esq7zhV0JCAgUFBeHYtYiISESUdRN/5ZVXcDqdIfWXOv/88wHo2LEjW7duBQ51A6+sY3jr1q1xOp0kJSVx2mmn4XK5SElJOaJD+KkuLAEnNTWViy++GJfLRcuWLYmJiTki0LjdbpKTkyvdxgG3l9FL1rF2W244ShQREQmLtLQ0pk2bxpgxY8q7fVfku+++A+Drr7+mdevWwKFu4JV1DA+lQ/ipLiwB57zzzuM///kPlmWxZ88eiouLufDCC1m9ejUAK1asKE+tFfH6/GT43ez8eSex7vxKP4yDufj2H8C3/wCBoqJwvCQREYkmiQ2DVz5V10diw+Pa/ZlnnklmZiYTJ04E4NZbb8Xr9R613kcffURmZiYrV648akLw4R3DBw0axOeff87w4cNP/JicYsLWTfzJJ59k9erVWJbFyJEjadasGWPHjsXn89GyZUsmTpyI3W6v8PldLu1Fgw4DSE1w8lCvdpXuq36Si5S44GV4jsaNcabXq9bXIiIitZu6idcd1dVNPGyXif/1r389atmCBQtCfr5hgM2AA4VHJ14RERGRytTqG9EYBvgCJsU+f6RLERERkShSewOOFTz/GOOws69As8JFREQkdLU34BiHPu8tVMARERGR0NXegFPKNC32HCyJdBkiIiISRaKi2eaOvOJIlyAiIlEk++JLCOzbV23bs2dk0Oaz/1Tb9iT8am3AKTtDZbMZ7NUcHBEROQ6Bfftw1K9fbdvz791b6eOHdxO3LAu/38+kSZPC0in8cFlZWWRkZDB48OCw7ica1dqA47DZaNu0AaZlYRhgd8ZVuG7AcOKxgi/FCoCzpooUEREpdXg38c8++4wnn3ySOXPmRLiqU1etDTgGBju35WFZFiV+k7wWGRWvm+DEjAnGmtR6usmfiIhEVn5+Pk2bNq3w8eHDh/OXv/yF3//+9/Tu3ZsHH3yQnj17cuuttzJlyhS++eYbXn75ZWw2G+eddx4PPvggU6dOxel0MmLECIYNG8awYcOO2OZTTz3FV199hWVZ3HLLLfzxj38kMzOTtLQ08vPzycrKYsyYMRQUFJCbm0v//v0ZMmQIGzZsKL/jclnH8qSkpLAen5pQawPO4SwL/KaJw1br50SLiMgpqqybuNfrZcOGDZWO3vTq1YsVK1aQmppKTEwMK1eupEuXLng8HmJiYsjKyuLNN98kLi6Ohx56iJUrV3L//fdz4403MmrUKDp06MBll11W3svq008/Zfv27bz++ut4PB4GDBhA165dAbj66qvp2bMn69evp2/fvvTq1Ys9e/aQmZnJkCFDGDt2LJMnT+bMM89k8eLF/OMf/2DkyJE1cszCqdYHHMMwsNsMir0BkmIVcEREpHY6/BTVli1bGDRoECtWrCA2NvaodS+//HKGDx9OWloat99+Oy+99BIrVqzg8ssv59dff+XAgQPlvancbjfbtm2ja9eu3HzzzYwaNYpPPvnkiO1lZ2ezfv16MjMzAfD7/ezcuRM41J08IyODefPm8cEHH5CYmIjfH7yJ7ubNm3nssccA8Pl85etHu1ofcAAcNoNiX4CkWM2uERGR2i8jo+JpFQApKSnExsby3nvvkZWVxfvvv8+8efOYPn06iYmJNG7cmBdffBGn08mSJUs466yzOHjwIM8//zyjR49m7NixPP/88+Xba9myJZ07d+bxxx/HNE1mzZpFs2bNgEOdx1988UXOOecchgwZwhdffMGnn34KBAPQ1KlTadKkCWvWrGFvFROqo0VUBByAIm8g0iWIiEiUsGdkVHnl0/Furyplp6hsNhtut5vRo0cTGxvLsmXLKCoqYuDAgUesf8UVV7BkyRJSU1O5+OKLefXVVznttNMAuOWWW8jMzCQQCNC0aVP++Mc/8tBDD3HbbbfRr18/vv/+e/73f/+3fFvdu3fnyy+/ZMiQIRQVFdGjRw8SExOP2N/ll1/OhAkTWLZsGampqdjtdrxeLxMmTGDUqFEEAsHfs5MmTTrZw1UrhK2b+Mm6tPsfOePsIQAETIvL2zXgrEbJx1w3LcFJYtkk47bNSWqaXmN1iohI5KmbeN1RXd3Eo2RSi0WxRnBEREQkRFERcPwWuL3qKC4iIiKhiYqAYwAFJQo4IiIiEpqoCDg2oMijgCMiIiKhiYqAYxgGRT7NwREREZHQRM1l4iUKOCIiEqKX/voZRfneattefLKLYU9eXG3bk/CLioBjA0p8ZqTLEBGRKFGU7yU+xVV92ztYeVg6vJt4mbS0NJ577rlqqwHA4/Hw9ttv079//wrX6d69O++99x4xMTHVuu9oExUBBwP8poUV/FJERKTWObxVQ7js3buXxYsXVxpwJChsAedPf/pTeTfSZs2aceeddzJ69GgMw6B169aMHz8eW4jNMw3DwG5AiT9AnMMerpJFRERqxOrVq5k+fTpOp5MBAwbQpEkTZsyYgd1up3nz5vztb38jEAjw8MMPs3PnTnw+H2PHjuXNN99k06ZNzJw5kxtuuIEJEybg8XjIy8vjrrvuokePHuX72LVrF2PHji1v4Pn444/TuHHjCL7qmhWWgOPxeACYP39++bI777yTESNG0LlzZ8aNG8fHH39Mz549Q96mzWbg8SngiIhI7VTWqqHMpZdeym233Vbh+h6Ph8WLF2NZFn369OHVV18lPT2dZ555hrfeeouioiKaNm3KjBkzyM7OZtWqVdx5551kZ2dz9913s2rVKoYNG0bnzp355ptvyMrKOiLgTJ06lczMTC699FI+//xzpk+fzlNPPRXWY1CbhCXg/PTTTxQXF3Prrbfi9/u5//77Wb9+PZ06dQKgW7durFy58rgCjt0w8Po1D0dERGqn4z1FVda1+8CBA+Tk5DBixAgASkpK6Nq1KwcOHKBbt24AtGnThjZt2rB9+/by59evX5/Zs2fzxhtvYBhGeXfwMtnZ2cyZM4d//OMfWJaF03lqNawOS8CJjY3lz3/+M/379+fnn3/m9ttvx7Ks8o6mCQkJFBQUHPd2PQo4IiJSR5RN00hLS6NRo0bMmjWLpKQkPv74Y+Lj48nOzua7776jR48ebNu2jWeeeYYHHngA0wz+Lnz22Wfp378/l156KW+++SZvvfXWEdtv2bIlt956Kx07dmTz5s189dVXNf4aIyksAadFixacfvrpGIZBixYtSE1NZf369eWPu91ukpOP3TizMrpUXEREQhGf7Kryyqfj3V5VfnuKCuCFF17gww8/PGY38TI2m41HH32UO+64A8uySEhI4Mknn6Rjx4488sgjDB06lEAgwCOPPEJ6ejo+n49p06bRp08fJk2axJw5c2jcuDG5ublHbHfUqFHlc3RKSkp49NFHT/wARKGwdBN/9dVXyc7OZsKECezZs4ebb76ZZs2acfvtt5fPwenSpQtXXnllhds4vJs4BDuKd2udwdlNU49aV93ERURObeomXndUVzfxsIzg3HDDDTz88MMMHjwYwzCYPHkyaWlpjB07lqeffpqWLVvSu3fv49qmZVl4AjpFJSIiIlULS8BxuVzHnKm9YMGCE95mwAKPV6eoREREpGpR0YsKwDCgSJOMRUREJATRE3CAYo3giIiISAiiJuDYgGJdRSUiIiIhiI5eVABG8E7GIiIiVSny+qt11D/OZSfeFT2/MiWaAg7g9Vf7Fe0iIlIHFXsD7MwrqbbtNUmNrTTgbN++nWuuuYbf/e535cs6d+7M3XffXW01VLbv+++/n0WLFh33c+fOnUuXLl3o0KFDGCqLrKgJOAbg9WsER0REaqczzzzziB6M0eCOO+6IdAlhEz0BxwCv7oMjIiJRbsaMGXzxxReYpknfvn255ZZb+PLLL5k5cyYQ7EU1depUWrRowaxZs/joo48IBAIMHjyYQYMGHbXs4osv5sCBAwwfPpy9e/fStm1bJk6cSHZ2Nk888QSmaZKfn8+YMWPo2LEjl19+OS1btqRly5YUFBRw5ZVXsmjRIm666SY6derEunXrmD17Ns899xyPPPII27ZtIxAIMGzYMK688koyMzNJS0sjPz+frKwsxowZQ0FBAbm5ufTv358hQ4aQmZlJu3bt2LhxI4WFhTz77LNkZGRw3333UVhYSElJCQ899BCdO3cO23GOmoAD4DfBIjiaIyIiUpts2rTpiFYN06dPp2HDhkett3TpUhYsWEDDhg3L7867ceNGpk2bRsOGDXn++ef597//zaWXXsqKFStYvHgxXq+Xp556ih9++OGoZV27dqWwsJApU6aQlJREz5492b9/P5s2bWLUqFG0bduWZcuWsWTJEjp27MiuXbtYsmQJaWlpjB49GoD+/fvz1ltv0alTJ9566y0GDBjAwoULSUtLY9q0aRQWFnLdddfRpUsXAK6++mp69uzJ+vXr6du3L7169WLPnj1kZmYyZEiwC0GHDh149NFHmTFjBu+88w6XX345+/bt4+WXX2b//v38/PPPYf33iJqAYxgGNoKnqWIc9kiXIyIicoRQT1E9/fTTPP300+zbt49LLrkEgIYNGzJp0iTi4+PZs2cPHTt2ZOvWrXTo0AG73U5cXBxjxozhnXfeOWrZ9u3bad68OSkpKQCkp6dTXFxMgwYNmDVrFrGxsbjdbhITE4Fgc8+0tLQjarrkkkuYNm0aeXl5fP3114wZM4aJEydy0UUXAZCYmEirVq3Ytm0bcKgTekZGBvPmzeODDz4gMTHxiI7m7du3B6BRo0bs27eP1q1bc+ONN3L//ffj9/uP6ttV3aLmMnEAm83Aq5v9iYhIlPJ6vfz73//m6aefZt68ebz11lvs2LGDMWPGMHnyZJ544gkaNGiAZVm0bNmSH374AdM08fl8DBs2jGbNmh21zOv1YhhHn9uYNGkS9957L1OnTqVNmzaUtZ4s62J+OJvNRp8+fZgwYQI9evTAbrfTqlUrvv76awAKCwvJzs6mWbNmAOX7e/HFFznnnHOYPn06ffr0obL2lhs2bMDtdjN37lyeeOIJHn/88ZM+npWJmhEcAIcNPP4ASTgjXYqIiNRicS47TVJjq3V7xysvL48xY8aUz62BYCujlJQU+vXrR0pKCl27dqVJkyb069ePAQMGkJycTEZGBjk5OZx11llccsklDB48GNM0GTx4MH/4wx+OWuZyHbvT+TXXXMPw4cNJT0+nUaNGR3Ub/63rr7+eHj168P777wMwYMAAxo4dy+DBg/F4PNx9992kpx/ZzPryyy9nwoQJLFu2jNTUVOx2O17vsbu4n3HGGfz9739n6dKlOJ1O7r333uM5nMctLN3Eq8Nvu4lDsKN4v3Oa0jQ17ojl6iYuInJqUzfxuqO6uolH1SkqgJLDzu+JiIiIHEvUBRzd7E9ERESqEnUBx6NJxiIicgy1dMaFHIfq/DcMaZLxvn37yMjIqLadhsJug9YpR75Qn2mRZi8m1jxykrHN66Q8q3ndUGiCKx5cCTVUrYiIRFJsbCz79+8nPT39mFcUSe1nWRb79+8nNrZ6JoeHFHDuuece6tWrxw033MCll156zEvMqpuBRWHe3iOW+QMmha5izKTfzMMxHRAoneHuSYP8/ZDcTAFHROQU0axZM7Zv387evXurXllqrdjY2PJL0U9WSAHntddeY/PmzbzxxhvMnj2bCy+8kBtuuIHmzZtXSxEhMwxK1K5BRER+w+l0lt98TgSOYw5OgwYNaN68ObGxsWRnZzNp0iSeffbZcNZ2FAPw+NRwU0RERCoX0gjOfffdx8aNG7nmmmvKe2VA8Lr0++67L6wFHs4AfKZGcERERKRyIQWcAQMGcM4555CQkEBOTk758tdee63C5+zfv5/rrruOF198EYfDwejRozEMg9atWzN+/PgTm8djGPh0mbiIiIhUIaSUsXbtWrKysgCYOHEic+fOBSAmJuaY6/t8PsaNG1c+E3rKlCmMGDGCV199Fcuy+Pjjj0+oWAPwaQ6OiIiIVCGkgLN8+fLylurPPfccy5cvr3T9qVOnMmjQIBo0aADA+vXr6dSpEwDdunVj1apVJ1ywL6ARHBEREalcSAHHMIzy5lk+n6/SG/EsWbKEevXqlbeAh+C17WX3JUhISKCgoOCEijUAv6mAIyIiIpULaQ7OoEGDuPrqq2nTpg1btmzhtttuq3DdN998E8Mw+Pzzz/nxxx8ZNWoUBw4cKH/c7XaTnJx8wgWbAV1FJSIiIpULKeD079+fK664gm3bttG8eXPq1atX4bqvvPJK+deZmZlMmDCBadOmsXr1ajp37syKFSvo0qXLCRdsEuwqbrfpTpUiIiJybCEFnB9//JGFCxfi8XjKl02ZMiXknYwaNYqxY8fy9NNP07JlS3r37n38lQIYYDMM/AETu81+YtsQERGROi+kgDN69GiGDh1Ko0aNjmvj8+fPL/96wYIFx1dZBWw2G76ARYyz6nVFRETk1BRSwMnIyKB///7hriUkNsPAq0vFRUREpBIhBZymTZsyd+5czjrrrPKroS6++OKwFlYRA/D5FXBERESkYiEFHJ/Px9atW9m6dWv5skgFHFC7BhEREalcSAFnypQpbN26lV9//ZW2bduW38AvUvw6RSUiIiKVCCngLFiwgA8//JCDBw9y7bXX8ssvvzBu3Lhw13ZMFuBRPyoRERGpREh3Mn7nnXd4+eWXSUpK4uabb+a///1vuOuqkGVZ+HWKSkRERCoRUsApa81QNsHY5XKFr6IQavFqBEdEREQqEdIpqquuuoobb7yRnTt3cvvtt9OjR49w11Uxw8DjV7sGERERqVhIAWfo0KFceOGFZGdn06JFC9q1axfuuipkAF5dJi4iIiKVCCngzJw5s/zrzZs389FHH3H33XeHrajKGIBXDTdFRESkEiHfyRiC819++OEHzEhO8jUMzcERERGRSoUUcAYNGnTE97fddltYigmFAfh0HxwRERGpREgB5/A7GO/du5ddu3aFraBQ+AIawREREZGKhRRwDr+pX0xMDH/961/DVlBVDCBgKuCIiIhIxUIKOPPnzw93HcclYGqSsYiIiFQspIBzzTXX4Ha7iYmJwePxAMEJx4Zh8PHHH4e1wGMxreAojt1m1Pi+RUREpPYLKeCce+65/OlPf+Lcc89lw4YN/POf/2TixInhru3YDLAZBv6Aid1mj0wNIiIiUquFFHA2b97MueeeC0Dbtm3ZtWtXRNs12Gw2fAGLGGfEShAREZFaLKSAk5SUxDPPPEOHDh1Ys2YNTZo0CXddlbIZBp6ASWJEqxAREZHaKqRmm0899RSJiYn85z//oXnz5kyaNCncdVXKAPxq1yAiIiIVCGkEJyYmhpSUFIqKimjRogX5+fnUq1evwvUDgQBjxoxh69at2O12pkyZgmVZjB49GsMwaN26NePHj8dmCylfHZMvkndTFhERkVotpIQxbtw4du7cycqVK3G73YwaNarS9T/55BMAXn/9de69916mTJnClClTGDFiBK+++iqWZZ301Vd+3c1YREREKhBSwPn111+57777cLlcdO/enYKCgkrX79GjB48//jgAO3fuJCMjg/Xr19OpUycAunXrxqpVq064aAvwqB+ViIiIVCCkgBMIBDhw4ACGYVBYWBjSqSWHw8GoUaN4/PHH6d27d/l9cwASEhKqDEmVsSxLIzgiIiJSoZDm4IwcOZLBgwezd+9eBg4cyKOPPhrSxqdOncqDDz7IgAEDym8QCOB2u0lOTj6xigkGHHUUFxERkYqEFHB27drF+++/z4EDB0hLSysfianI0qVL2bNnD//zP/9DXFwchmFw9tlns3r1ajp37syKFSvo0qXLiVdtGHgCatcgIiIixxbSKapFixYBUK9evSrDDUCvXr344YcfuPHGG/nzn//MI488wrhx48jKymLgwIH4fD569+5deWGWSaP8HAzr6JEaA/D4dIpKREREji2kERyv18uf/vQnWrRoUT7/5qmnnqpw/fj4eJ599tmjli9YsCDkwlwBH233bKZxfg7fNW2H3zhUqgH41HBTREREKlBpwJk1axbDhw/nwQcfZM+ePTRs2LCm6gIgYLeRWFLA73f8yNpmvz/0gGFoDo6IiIhUqNJTVF988QUAnTp1YvHixXTq1Kn8o6ZYNhvJxQWkFuWVLzMAr66iEhERkQpUGnCsw+a/WMeYC1NTLAxa7/05eAOcUn6N4IiIiEgFKg04h08oDmVycbiYdhvx3mLqFecFawH8lkZwRERE5NgqnYOzfv16Bg0ahGVZbNq0qfxrwzB4/fXXa6pGIDh40zx3BwfiUwG1ahAREZGKVRpw3n777Zqqo0qW3aCeOw+H6cdvc2ADAgETu/3EG3aKiIhI3VRpwGnatGlN1VElCwO/zUZG4X52JzfEZrPhCZjEK+CIiIjIb0RXOrCgad5uIDgnyKfTVCIiInIMURVwLLuN5JJCYgJebIaB16+AIyIiIkeLqoAD4LfbSXPnAeAN6FJxEREROVrUBRzDtKhfuA8An1/tGkRERORoURdwLFvwaiosE49u9iciIiLHEH0BxzCwMEgoceMLaARHREREjhZ1AQfAMqC+e79GcEREROSYojLgADRwH8Dj0wiOiIiIHK3WBhzT5mJ/xrm4E5pjYT/iMctmI9FbhNfri1B1IiIiUpvV2oATbKlpoySuAQcy/oBpizniUb/NTkx+XkQqExERkdqtFgccALP0wyAvtR3mYSM5hgUJ+fsjVpmIiIjUXrU84JQxsWwOClLPLF9iGZBWmBvBmkRERKS2ipKAA2DidyTicaUCwXk48R43lt8f2bJERESk1omigBPkTmqBhQGA37Dj36fTVCIiInIkR3Vv0Ofz8cgjj7Bjxw68Xi9/+ctfOPPMMxk9ejSGYdC6dWvGjx+PzXYi2crEMmwUxzUivngXAN49u+GMptX7IkRERCSqVXvAefvtt0lNTWXatGnk5uZy7bXX0q5dO0aMGEHnzp0ZN24cH3/8MT179jzhfRQnNCGueDcG4Nu1p/qKFxERkTqh2gNOnz596N27d/n3drud9evX06lTJwC6devGypUrqw44NoipH1fBgwb+pNa4Sg5iEqAYP5YZfCTGMvGZfvAVQrG9gudXr1hHLPHO+BrZl4iIiFSt2gNOQkICAIWFhdx7772MGDGCqVOnYhhG+eMFBQUhbWvvwZxjLrewYbNMkg/uxAnY920jITUFAKc3n5LCvWB5wSw6+RcUgsbxjRVwREREapGwTDLetWsXN910E/369ePqq68+Yr6N2+0mOTn5pLZvYGIZdnyuJCwDzLyDJ1uyiIiI1CHVHnD27dvHrbfeykMPPcQNN9wAQPv27Vm9ejUAK1as4Pzzzz/p/VhASUwGWBb+3LyT3p6IiIjUHdUecJ5//nny8/OZNWsWmZmZZGZmMmLECLKyshg4cCA+n++IOTonyiBAwBFPwObCPJh38oWLiIhInVHtc3DGjBnDmDFjjlq+YMGC6t4VWBYlsWlYZgGBoiLs8ZoHIyIiIlF4o7/DGViUuNIxbDb8e/dFuhwRERGpJaI64JQ14vQ64vHn6H44IiIiEhTlASc42bjYlYZ/97EvKRcREZFTT9QHHAMTnyMef0kJphpvioiICHUg4IAFFpTE1yewX403RUREpE4EnCB3bAP8e/dGugwRERGpBepIwDHxOhIo2a0rqURERKSOBBwDMCyLwhInmFakyxEREZEIqxMBp0xRfAPMwtAaeYqIiEjdVWcCjg0Tnz2Wkn1qvCkiInKqqzMBB0pPU+WVRLoMERERibA6FXDAwu2PDd79T0RERE5ZdSLglOUZAxPL5uTgr7pcXERE5FRWJwJOGQMDMNn13e5IlyIiIiIRVKsDjq30wyj9CIllsXuvGbaaREREpPar1QGnTFnACS3sWAQCBoW5mmwsIiJyqoqKgHM4g8pDTnC5ye5s9aUSERE5VUVdwClz+IjObx8xLIucbYVYOlMlIiJySoragFPmWCM5NsvECpjs+1V3NRYRETkVRX3AgdKQY1lHLLBMky3rdLm4iIjIqcgRrg3/97//Zfr06cyfP59ffvmF0aNHYxgGrVu3Zvz48dhs1Z2trCNu8GeYAXb8dADzyhbYHCFfgyUi4eZ1g7foxJ/vigdXQvXVIyJ1UlgCzgsvvMDbb79NXFwcAFOmTGHEiBF07tyZcePG8fHHH9OzZ88w7NkqHckxSufomOzclEezdmlh2JeInBBvEeRvP/HnJzdTwBGRKoXlFNVpp51GVlZW+ffr16+nU6dOAHTr1o1Vq1aFY7eHlJ6uMv0BNn+TE959iYiISK0TloDTu3dvHI5Dg0OWZWEYwdNECQkJFBSEe/KvBVgYPg/7thdSUugP8/5ERESkNqmRScaHz7dxu90kJyfXwF4tDMBuWPz8/b4a2J+IiIjUFjUScNq3b8/q1asBWLFiBeeff35N7BawwFfC5q9z1GFcRETkFFIjAWfUqFFkZWUxcOBAfD4fvXv3rondBvl9eN1F7N9RWHP7FBERkYgK22XizZo1Y9GiRQC0aNGCBQsWhGtXlTPBMH1sfP8L0v/cIzI1iIiISI2qEzf6q5RRek+cnHi8a5ZGuhoRERGpAXU/4ABWwMKOny0rv4NNn0S6HBEREQmzUyLgYAJYbCy6GOv/TYVtX0a6IhEREQmjUyPgGGCYJgHLwU7f7+GDsbD7+0hXJSIiImFyagQcwDKDn78vvBxsLnj3r7BvU2SLEhERkbA4ZQIOAbDhp8ifzD5/KzBssOw+yP0l0pWJiIhINTt1Ao4BVgAM4PuDl4AjJrjw/+6BvJNo/CciIiK1zqkTcAgGHBs+cn2N2FfSLBhyLBOWDlfIERERqUNOqYBTxsBkbV7PYPuG8pDzF52uEhERqSNOuYBj+oNzcdz+VHaVtAoudMSAZQVHcvZuiGyBIiIictJOuYBD6TwcMFlz4I8ETHtwedmcnGX3wfZvIlefiIiInLRTL+AYwVEcuxHAbzn4If/iQ485YsBwwvsPw0/vRK5GEREROSmnXsAhONkYCwwCbCw8n4PejEMPOlzgiIVVWbByJgQCEatTRERETkzYuonXagYE/GB3WlhWgJX7bqBXo3/gsPmDj9scwfvkZL8HOeuh92SIT4tszSJ1jbcQCvZAYQ6U5IEnH3zF4CsJTvy3O4Ojqs4EiEmGhHoQXx8SGka6chGJAqdmwAEIgOUITjj2mrGsOfBHOmcsO/S4YQN7DORtg4VD4fIxcMaFkatXJFqZJuzfCDvWwLavoDgXNn4Afg84Y8FWOg/ODAQ/jmCAzXZonYAf6reB3J+hXitocg406QiN/wD12wZDkYgIp3LAMcD0BjOMzfKzs+RMfsq/gHbJXx22jhH8CzLgg+WPwekXQ9d7ITY5cnWL1HZmAHZ+C1tXBIPMrm/BsAfDh+mDRn8Ifm2PCf4fK2N3gL2KbTsc4EoJbm//puBVj98vCU6s85dAems442I4/SJo3gmSm4TxhYpIbXbqBpxSAV/wvdZm+fkh/xJibSWckfjdkSvZncG/IH/9An79HC68C9r0Cf5lKXKK83n8+Pdth13rYNtq2PcTGGVvLSnQuHswkJRJbQ5mCg6rEOf+0JreBlLbYlqlozNxZ0DqMZKQZQaDzi/r4dcfwT8Lm8uOPTEFWnUPBp6GZweDlIjUefqfbgb/4LTZLbB8fJPXC68ZS5vDR3IgeMrKERMcIv/87/DNfLjobji9a2TqFomkkoOw9T+w4V38ufkU/LwZ7K7g6Si788hA81tFMbB/F0lNGxPqCSXTcuLfvD74TboH8n6u+kmWhaPlWdh/WQnbvggGIF8x1G8HZ3SFZp2gybmQdsaRI0mV8brBWxRi1b/higdXwok9V0SOmwIOYPnApCzkeFmffzEHfRmcm/bhoYnHZewOsOzgKYBPJkNMElw+DtpeHZxPIFJLFHn9FHtDuwowzmUn3lXJ24GvBLZ/BZuXQ/Z7mPu2UGg7nYPeerjrd2OnuzXuQCrFZiKeQDw+K5aA6cC07FilF2saWBiGSUPLJLD3ZxrGJxLIaU6co5B4x0Hi7AXE2QtJsB8kzpGP0/BXXE8oDCMYtJzxh5aVzatbuwDWLQqe1rJMSG8TDDtN/gAZbSGjDSRkHB18vEWQf4JtXZKbKeCI1CAFHAg24vQF//i0OS0My8+O4rbsKWnJBen/omHsb1o4lM3NsSzwuGHN/8L/3QtnXwfnZkKzCw5NihSJkGJvgJ15Jcd8rKDEx7rtB9mwp4DdB0to3SCBpsmxnJmWSPsmSaTaPLAvG//unyj89WcKDwbINxtx0NeeQn9XPI4EUtNsOAwvsSkJ5O7JAcCORTwWwT4owVADh0KWHztYFn7LhZHWBG8gEQ+QB9iM4PMsy0bAcmDDj8sowYWbVJsDK749TrOEuPRWBHLXYceDYQFG8Oadhg2wB//r2ZzBiyFdSafjdyYC4LDcOA98X3pTz5hgQS5nMODk/QoHtsCP/xfcoK8o+BqSm0LqacHAk3o6pDSFknyIrwdxqeBMLLtzqIjUMjUWcEzTZMKECWzYsAGXy8XEiRM5/fTTa2r3VTOC73MBTzDk2G1+/JaTVfuuI8mxn9+lrKRR7BYMwzrsOUbwvjl2Z/BS1h/ehh+XBTd0Zk9o0zs44TG5aehD4CJh4vEH+GLLft5Zt4uNewuJc9rxB0zSjYOcFfCTuM3NgV+2s8bvwO1vREGgER6rKU6jAQYmZulIjI0iDNw4ih0U/bqT0zq1oGTbtqN3aB322Qp+ijutCQRMLL8JpsH+rfsOrXcUA8swsLCRkhFHyS9pGFg0b3k2329PxsKG3fJiD5Tg8Jfg8Bfh9BfhMIuxBzw4/B7impv4fvYR6yimQbtU4nMMYlItXCnBucr2WDDKTj+XhR4Inm6zTCg+CEXfwvY1weTU8Pew+7/B89oBT/CvorJTTzFJ4EqEmJTghQixqRCbGHxvcMVDww7B5TFJwY/Y5OD6+mNIJCxqLOB89NFHeL1eFi5cyLfffssTTzzB7Nmza2r3x8X0BT8bdj82GxT601m9/2oAGsdtomncJuq5dhBvLzj015vNDrbSoXDTDxs/hC3/L3jO3+EKvjE27QgN2kO9lpB2OiQ00ERlCasir48fNv/CfzdsZd/2PTTARxfLR6+Aid0dh9dMoSiQxmlx9YmzfOwrbMwB7ARTh4mTYmwE87kNM7jRssBiErzdggmWPzigiVX6ucxvc71Z+rzDw0+FLAzLwgDsph/DCpRu0sJuekufaiNgj8Vvj4PYeqU7NDFKi0ho3IS9zoswDTsZTdLY5xsSDET7SnDuKsLpdxNj5RNvyyc+ppD42EIS4tzBj4RCYh1FxNrd2B2lo1Dlw0PO4A1BreBxwldSOjdnd3BZ2cEw7GAzgsNLDTsE76tlBYLvEQFf8MPhCp5GcyYcCj6xqRCXBvHpwXtwlYWmmOQjQ5Kr9LMjRn9EifxGjQWcNWvWcMkllwBwzjnn8P33VV89YYY49msBphEMCpZhYBo2TOyYhnnC9ZaPrAegfIjdZmNbYTu2uc/CKrtKxDJp60nH2pRDrJWPyyrCQQl2vNjxYcePzfJjy/ZjWDsx2I5BIPjKLKt0CB9MmwPT5jz0YdixsGPZgnMYLCP4EfyrtvS1Yiv9BWKUnxAo/41icNiy46E3ydBV9pv88HWM33xfun5ZErAqWKs8BBhYWMHPloFF6WfLwLRsWARP6fiJwW/F4Dfi8BlxuM5owObt+QAkkUoSqRhWABOTPCgNAcGkYg94gtv0BwNCcN8GJgYV/S8yAwb+gA3TtBEwbYcOQyU/QhYGJvbgz/NhP8tVMct//ku3cdTzjkxWh/JT8PXZLBNboAR7oAQw8Dni8TkSsIwGlM4OCpZtglVkYBXbIffo2pr7TDzbN2K3vDgsb3AECS8OyxP8/274seEP/r8ngA0TwxYgJslPTEJbCvJOx8DEVvreZGBieIOJL9G+nxaxa4KjxJZ5KCQd/tNhlV4VYVU1t6r0NLozDhxxwfmBzrhgkHLEBD/bXYd9OEuvFnWUBrjSc32GvXSZ/dCcJsNW+rWt9LygURquDvsMwa/j0qD9n3TlmkREjf3UFRYWkpiYWP693W7H7/fjcBy7hF279vDGgVXHv6NvTuA5J8jCwLLZ+bjG9ihyHNYAllk6ClLpUAmffHkS+9lwHOv+dNjXJ7rP9e+f2P7WnuD+KP2/bhjwVdkv8RORU+UaGbHJVf5bHVtZKDo8GJXxlH4cPIHtVoP6L5WeAhQ5cTt27Dju59RYwElMTMTtdpd/b5pmheEG4Nv1J/FuJCIiIqe0GpsA0rFjR1asWAHAt99+S5s2bWpq1yIiInKKMSzLOpHx0ONWdhVVdnY2lmUxefJkWrVqVRO7FhERkVNMjQUcERERkZqia5RFRESkzlHAERERkTpHAUdERETqnIgHHNM0GTduHAMHDiQzM5Nffjmy79Py5cu5/vrrGThwIIsWLYpQldGnquMKUFxczKBBg9i8eXMEKoxOVR3Xf/3rX/Tv359BgwYxbtw4TPMkbjZ5CqnquL7//vtcf/313HDDDSxevDhCVUafUN4HAMaOHcv06dNruLroVdVxfemll+jbty+ZmZlkZmayZcuWCFUaXao6ruvWrWPIkCEMHjyYe++9F4/HU/kGrQh7//33rVGjRlmWZVlr16617rzzzvLHvF6v1aNHDysvL8/yeDzWddddZ+Xk5ESq1KhS2XG1LMtat26dde2111oXXXSRtWnTpkiUGJUqO67FxcXWFVdcYRUVFVmWZVkjR460Pvroo4jUGW0qO65+v9/q2bOnlZ+fb/n9fqtXr17W/v37I1VqVKnqfcCyLOu1116zBgwYYE2bNq2my4taVR3XBx54wPruu+8iUVpUq+y4mqZpXXPNNdbPP/9sWZZlLVq0yNq8eXOl24v4CE5lLRw2b97MaaedRkpKCi6Xi/POO4+vv/46UqVGlapaY3i9Xv7+97/TsmXLSJQXtSo7ri6Xi9dff524uDgA/H4/MTG6g2soKjuudrudd999l6SkJPLy8gBISEiIRJlRp6r3gbVr1/Lf//6XgQMHRqK8qFXVcV2/fj1z585l8ODBzJkzJxIlRqXKjuvWrVtJTU1l3rx5DB06lLy8vCp/f0U84FTUwqHssaSkpPLHEhISKCwsrPEao1FlxxXgvPPOo3HjxpEoLapVdlxtNhsZGRkAzJ8/n6KiIrp27RqROqNNVT+vDoeDDz74gH79+nH++edXehd0OaSy45qTk8PMmTMZN25cpMqLWlX9vPbt25cJEyYwb9481qxZwyeffBKJMqNOZcc1NzeXtWvXMmTIEF566SW++OILPv/880q3F/GAU1kLh98+5na7jwg8UrHjbY0hoanquJqmydSpU1m5ciVZWVkY6vAcklB+Xnv16sWKFSvw+XwsXbq0hiuMTpUd13//+9/k5uZyxx13MHfuXP71r3+xZMmSSJUaVSo7rpZlcfPNN1OvXj1cLheXXnopP/zwQ6RKjSqVHdfU1FROP/10zjzzTJxOJ5dcckmVTbsjHnAqa+HQqlUrfvnlF/Ly8vB6vXz99dece+65kSo1qqg1RnhUdVzHjRuHx+Nh1qxZ5aeqpGqVHdfCwkKGDh2K1+vFZrMRFxeHzRbxt66oUNlxvemmm1iyZAnz58/njjvu4KqrruK6666LVKlRpaqf16uuugq3241lWaxevZqzzz47UqVGlcqOa/PmzXG73eUTj7/++mtat25d6fYififjY7Vw+OGHHygqKmLgwIEsX76cv//971iWxfXXX8+NN94YyXKjRlXHtUxmZiYTJkxQ24wQVXZczz77bK6//nrOP//88pGbm266iZ49e0a46tqvqp/XhQsX8sYbb+BwOGjbti1jx47FbrdHuuxaL9T3gSVLlrBlyxYefPDBCFYbPao6rkuXLmX+/Pm4XC4uvPBC7r333kiXHBWqOq6ff/45Tz31FJZlce655zJmzJhKtxfxgCMiIiJS3TTOKyIiInWOAo6IiIjUOQo4IiIiUuco4IiIiEido4AjIiIidY4CjogcJTMzMyJNWKdPn17pzeY2b95MZmYmACNHjsTr9R5zvby8PJYtWxaWGkUkOijgiEhUmjFjBi6X65iPbdiwgeXLl9dwRSJSm+je/SJSod27dzNhwgQ8Hg95eXncdddd9OjRg08++YSZM2cC0L59ex577DE+/fTTo5Z98MEHvPLKK+Xbe/bZZ9m4cSPTp0/H6XQyYMAA4uLimD17NvXq1cPn8x3VQC8nJ4cHH3wQy7KoX79++fLu3bvz3nvv8emnn/LCCy/gcDho2rQpTz75JM8//zw//fQTCxcu5Nxzz+WJJ57ANE3y8/MZM2YMHTt2pFevXnTs2JGtW7eSnp5OVlYWPp+Phx9+mJ07d+Lz+Rg7dixnn30248eP55dffsE0TUaMGEHnzp1r4OiLyMlQwBGRCm3ZsoVhw4bRuXNnvvnmG7Kysrjssst4/PHHWbx4Menp6cycOZPdu3cfc9nPP//M3LlziYuLY9y4cXz22Wc0bNgQj8fD4sWLAejRoweLFy8mNTWVO+6446gaXnrpJa666ioGDBjAu+++y2uvvXbE4//617+45ZZb6Nu3L0uXLqWwsJA777yT119/nYEDB/Luu+8yatQo2rZty7Jly1iyZAkdO3Zk27ZtzJs3j8aNGzNo0CC+++47vv32W5o2bcqMGTPIzs5m1apV/Pjjj6SlpTF58mRyc3MZOnQo77zzTo0cfxE5cQo4IgIEm9m6XC6cTicAhmFQv359Zs+ezRtvvIFhGPj9fnJzc0lOTiY9PR2Au+++m7179x61DCA9PZ1Ro0aRkJDAli1bOOeccwBo0aIFAPv27SMxMZG0tDSAY/aa27hxI/369QOCvWp+G3Aefvhh5syZw2uvvUbLli3p0aPHEY83aNCAWbNmERsbi9vtLu9WnJaWRuPGjQFo3LgxHo+HLVu20K1bNwDatGlDmzZtmDBhAmvWrGHdunUA5cegrGYRqZ00B0dEABg9ejRr1qzBNE32799PvXr1ePbZZ+nXrx/Tpk2jc+fOWJZFeno6+fn55OXlATBx4kR27dp11LIvv/yS5557jhkzZjBx4kRiYmIo6wxT1iwzNTWVgoICDhw4AMB33313VF0tW7Zk7dq1FT6+cOFC7rnnHhYsWADAhx9+iM1mwzRNACZNmsS9997L1KlTadOmTXkNx+r03qpVq/J9bNu2jQceeICWLVvSt29f5s+fzwsvvECfPn1ISUk5oWMsIjVHIzgiAsCwYcPKg8i1115Lamoqffr0YdKkScyZM4fGjRuTm5uLzWZj/Pjx/M///A82m4327dvz+9///qhlF1xwAR07duTaa68lPj6e5ORkcnJyaNasWfk+HQ4HU6ZM4c9//jMpKSk4HEe/Jd13332MHDmSd99994jnlunQoQPDhg0jNTWVhIQELrvsMrxeL9nZ2bz88stcc801DB8+nPT0dBo1akRubm6Fx2DQoEE88sgjDB06lEAgwCOPPELbtm0ZM2YMQ4cOpbCwkCFDhqibuUgUULNNERERqXP0Z4iIiIjUOQo4IiIiUuco4IiIiEido4AjIiIidY4CjoiIiNQ5CjgiIiJS5yjgiIiISJ3z/wG5RxEkVSQjpgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 576x252 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# gut t2d species\n",
    "organisms = ['B. uniformis', 'R. bromii', 'B. wexlerae', 'E. rectale', 'F. saccharivorans']\n",
    "\n",
    "models = {\n",
    "    'B. uniformis': '../ensembles/ERR260140_bin.3.s.xml',\n",
    "    'R. bromii': '../ensembles/ERR260140_bin.19.p.xml',\n",
    "    'B. wexlerae': '../ensembles/ERR260140_bin.13.s.xml',\n",
    "    'E. rectale': '../ensembles/ERR260140_bin.11.s.xml',\n",
    "    'F. saccharivorans': '../ensembles/ERR260140_bin.10.p.xml'\n",
    "}\n",
    "\n",
    "for label, model_file in models.items():\n",
    "    models[label] =  load_ensemble(model_file, flavor='cobra')\n",
    "\n",
    "# calculate distances\n",
    "def ensemble_distance(ensemble):\n",
    "    bin_matrix = [[int(ensemble.reaction_states[r_id][i])\n",
    "                   for r_id in ensemble.reaction_states]\n",
    "                  for i in range(ensemble.size)]\n",
    "    \n",
    "    distances = [jaccard(bin_matrix[i], bin_matrix[j])\n",
    "                 for i in range(ensemble.size)\n",
    "                 for j in range(i+1, ensemble.size)]\n",
    "    \n",
    "    return distances\n",
    "dists = {label: ensemble_distance(model) for label, model in models.items()}\n",
    "\n",
    "# plot\n",
    "_, ax = plt.subplots(figsize=(8,3.5))\n",
    "colors = sns.color_palette(n_colors=5)\n",
    "for i, label in enumerate(organisms):\n",
    "    sns.distplot(dists[label], ax=ax, bins=20, kde=True, norm_hist=True, color=colors[i],\n",
    "                  hist_kws={'alpha': 0.2}, kde_kws={'shade': True, 'alpha':0.9})\n",
    "\n",
    "ax.set_xlabel('Jaccard distance')\n",
    "ax.set_ylabel('Frequency')\n",
    "ax.legend(organisms)\n",
    "ax.set_xlim(0,0.6)\n",
    "ax.set_ylim(0,70)\n",
    "plt.tight_layout()\n",
    "plt.savefig('../plots/ensemble_dist/gut_t2d.png',dpi=300)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 🤖 RefSeq models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:306: UserWarning: Dataset has 0 variance; skipping density estimate.\n",
      "  warnings.warn(msg, UserWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:306: UserWarning: Dataset has 0 variance; skipping density estimate.\n",
      "  warnings.warn(msg, UserWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x252 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# gut refseq species\n",
    "organisms = ['B. uniformis', 'R. bromii', 'B. wexlerae', 'E. rectale', 'F. saccharivorans']\n",
    "\n",
    "models = {\n",
    "    'B. uniformis': '../ensembles/GCF_018292165.1_ASM1829216v1_protein.xml',\n",
    "    'R. bromii': '../ensembles/GCF_002834225.1_ASM283422v1_protein.xml',\n",
    "    'B. wexlerae': '../ensembles/GCF_018785545.1_ASM1878554v1_protein.xml',\n",
    "    'E. rectale': '../ensembles/GCF_000020605.1_ASM2060v1_protein.xml',\n",
    "    'F. saccharivorans': '../ensembles/GCF_015557635.1_ASM1555763v1_protein.xml'\n",
    "}\n",
    "\n",
    "for label, model_file in models.items():\n",
    "    models[label] =  load_ensemble(model_file, flavor='cobra')\n",
    "\n",
    "# calculate distances\n",
    "def ensemble_distance(ensemble):\n",
    "    bin_matrix = [[int(ensemble.reaction_states[r_id][i])\n",
    "                   for r_id in ensemble.reaction_states]\n",
    "                  for i in range(ensemble.size)]\n",
    "    \n",
    "    distances = [jaccard(bin_matrix[i], bin_matrix[j])\n",
    "                 for i in range(ensemble.size)\n",
    "                 for j in range(i+1, ensemble.size)]\n",
    "    \n",
    "    return distances\n",
    "dists = {label: ensemble_distance(model) for label, model in models.items()}\n",
    "\n",
    "# plot\n",
    "_, ax = plt.subplots(figsize=(8,3.5))\n",
    "colors = sns.color_palette(n_colors=5)\n",
    "for i, label in enumerate(organisms):\n",
    "    sns.distplot(dists[label], ax=ax, bins=20, kde=True, norm_hist=True, color=colors[i],\n",
    "                  hist_kws={'alpha': 0.2}, kde_kws={'shade': True, 'alpha':0.9})\n",
    "\n",
    "ax.set_xlabel('Jaccard distance')\n",
    "ax.set_ylabel('Frequency')\n",
    "ax.legend(organisms)\n",
    "ax.set_xlim(0,0.6)\n",
    "ax.set_ylim(0,70)\n",
    "plt.tight_layout()\n",
    "plt.savefig('../plots/ensemble_dist/gut_refseq.png',dpi=300)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 🥛 Kefir models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x252 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# kefir species\n",
    "organisms = ['L. mesenteroides', 'L. lactis', 'A. fabarum', 'L. kefiranofaciens', 'L. kefiri']\n",
    "\n",
    "models = {\n",
    "    'L. mesenteroides': '../ensembles/230a_03182016.xml',\n",
    "    'L. lactis': '../ensembles/261_03182016.xml',\n",
    "    'A. fabarum': '../ensembles/290_03182016.xml',\n",
    "    'L. kefiranofaciens': '../ensembles/322b_03182016.xml',\n",
    "    'L. kefiri': '../ensembles/376_03182016.xml'\n",
    "}\n",
    "\n",
    "for label, model_file in models.items():\n",
    "    models[label] =  load_ensemble(model_file, flavor='cobra')\n",
    "    \n",
    "# calculate distances\n",
    "def ensemble_distance(ensemble):\n",
    "    bin_matrix = [[int(ensemble.reaction_states[r_id][i])\n",
    "                   for r_id in ensemble.reaction_states]\n",
    "                  for i in range(ensemble.size)]\n",
    "    \n",
    "    distances = [jaccard(bin_matrix[i], bin_matrix[j])\n",
    "                 for i in range(ensemble.size)\n",
    "                 for j in range(i+1, ensemble.size)]\n",
    "    \n",
    "    return distances\n",
    "dists = {label: ensemble_distance(model) for label, model in models.items()}\n",
    "\n",
    "# plot\n",
    "_, ax = plt.subplots(figsize=(8,3.5))\n",
    "colors = sns.color_palette(n_colors=5)\n",
    "for i, label in enumerate(organisms):\n",
    "    sns.distplot(dists[label], ax=ax, bins=20, kde=True, norm_hist=True, color=colors[i],\n",
    "                  hist_kws={'alpha': 0.2}, kde_kws={'shade': True, 'alpha':0.9})\n",
    "\n",
    "ax.set_xlabel('Jaccard distance')\n",
    "ax.set_ylabel('Frequency')\n",
    "ax.legend(organisms)\n",
    "ax.set_xlim(0,0.6)\n",
    "ax.set_ylim(0,70)\n",
    "plt.tight_layout()\n",
    "plt.savefig('../plots/ensemble_dist/kefir.png', dpi=300)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 🪱 Soil models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:306: UserWarning: Dataset has 0 variance; skipping density estimate.\n",
      "  warnings.warn(msg, UserWarning)\n",
      "/opt/anaconda3/lib/python3.8/site-packages/seaborn/distributions.py:2557: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x252 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# soil species\n",
    "organisms = ['f__Thermoleophilaceae;g__', 'f__Herpetosiphonaceae;g__', 'f__Phormidiaceae;g__Microcoleus', 'f__Geodermatophilaceae;g__Geodermatophilus', 'f__Rubrobacteraceae;g__Rubrobacter_A']\n",
    "\n",
    "models = {\n",
    "    'f__Thermoleophilaceae;g__': '../ensembles/ERR671933_bin.1.o.xml',\n",
    "    'f__Herpetosiphonaceae;g__': '../ensembles/ERR671933_bin.2.o.xml',\n",
    "    'f__Phormidiaceae;g__Microcoleus': '../ensembles/ERR671933_bin.3.s.xml',\n",
    "    'f__Geodermatophilaceae;g__Geodermatophilus': '../ensembles/ERR671933_bin.4.o.xml',\n",
    "    'f__Rubrobacteraceae;g__Rubrobacter_A': '../ensembles/ERR671933_bin.5.s.xml'\n",
    "}\n",
    "\n",
    "for label, model_file in models.items():\n",
    "    models[label] =  load_ensemble(model_file, flavor='cobra')\n",
    "    \n",
    "# calculate distances\n",
    "def ensemble_distance(ensemble):\n",
    "    bin_matrix = [[int(ensemble.reaction_states[r_id][i])\n",
    "                   for r_id in ensemble.reaction_states]\n",
    "                  for i in range(ensemble.size)]\n",
    "    \n",
    "    distances = [jaccard(bin_matrix[i], bin_matrix[j])\n",
    "                 for i in range(ensemble.size)\n",
    "                 for j in range(i+1, ensemble.size)]\n",
    "    \n",
    "    return distances\n",
    "dists = {label: ensemble_distance(model) for label, model in models.items()}\n",
    "\n",
    "# plot\n",
    "_, ax = plt.subplots(figsize=(8,3.5))\n",
    "colors = sns.color_palette(n_colors=5)\n",
    "for i, label in enumerate(organisms):\n",
    "    sns.distplot(dists[label], ax=ax, bins=20, kde=True, norm_hist=True, color=colors[i],\n",
    "                  hist_kws={'alpha': 0.2}, kde_kws={'shade': True, 'alpha':0.9})\n",
    "\n",
    "ax.set_xlabel('Jaccard distance')\n",
    "ax.set_ylabel('Frequency')\n",
    "ax.legend(organisms)\n",
    "ax.set_xlim(0,0.6)\n",
    "ax.set_ylim(0,70)\n",
    "plt.tight_layout()\n",
    "plt.savefig('../plots/ensemble_dist/soil.png', dpi=300)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 🐈‍ Discussion questions\n",
    "\n",
    " - Compare your community's plot with the one [generated for CarveMe paper](https://raw.githubusercontent.com/cdanielmachado/carveme_paper/master/results/other/ensemble_distances.png), is there more or less uncertainity?\n",
    " - What does it mean if the average/median Jaccard distance between models is 1 for a given ensemble?\n",
    " - What does it mean if the average/median Jaccard distance between models is 0 for a given ensemble?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 🍾 You have completed this tutorial, move on to [the general discussion](https://github.com/franciscozorrilla/SymbNET/blob/main/scripts/9.wrap_up_discussion.md)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}