2.2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "63372ade",
   "metadata": {},
   "source": [
    "# Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "911843c5",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.preprocessing import *\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import *\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from scipy.cluster import hierarchy\n",
    "from scipy.cluster.hierarchy import dendrogram, linkage, fcluster\n",
    "from sklearn.feature_selection import *"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b86b0fa6",
   "metadata": {},
   "source": [
    "# Importing rpy2 and dependencies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "c22fbe78",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Unable to determine R home: [WinError 2] The system cannot find the file specified\n"
     ]
    }
   ],
   "source": [
    "import rpy2\n",
    "import rpy2.robjects as robjects\n",
    "from rpy2.robjects.packages import importr\n",
    "from rpy2.robjects import pandas2ri\n",
    "from rpy2.robjects.conversion import localconverter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "1a9330f8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3.5.6\n"
     ]
    }
   ],
   "source": [
    "print(rpy2.__version__)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "2b5c7cfd",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<rpy2.rinterface_lib.sexp.NULLType object at 0x0000022AAF8A9340> [RTYPES.NILSXP]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# import rpy2's package module\n",
    "import rpy2.robjects.packages as rpackages\n",
    "\n",
    "# import R's utility package\n",
    "utils = rpackages.importr('utils')\n",
    "\n",
    "# select a mirror for R packages\n",
    "utils.chooseCRANmirror(ind=1) # select the first mirror in the list"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8a083381",
   "metadata": {},
   "source": [
    "# Installing R packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "4da8df78",
   "metadata": {},
   "outputs": [],
   "source": [
    "# R package names\n",
    "packnames = (\"zCompositions\", \"robCompositions\")\n",
    "\n",
    "# R vector of strings\n",
    "from rpy2.robjects.vectors import StrVector\n",
    "\n",
    "# Selectively install what needs to be install.\n",
    "names_to_install = [x for x in packnames if not rpackages.isinstalled(x)]\n",
    "if len(names_to_install) > 0:\n",
    "    utils.install_packages(StrVector(names_to_install))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "9bb785fd",
   "metadata": {},
   "outputs": [],
   "source": [
    "zCompositions = importr('zCompositions')\n",
    "robCompositions = importr(\"robCompositions\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "52a45540",
   "metadata": {},
   "source": [
    "# Import data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "458cb2d6",
   "metadata": {},
   "outputs": [],
   "source": [
    "path = 'CRC/'\n",
    "df = pd.read_csv(path + \"/all_data.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "3dd0d308",
   "metadata": {},
   "outputs": [],
   "source": [
    "lbl=LabelEncoder()\n",
    "df['condition'] = lbl.fit_transform(df['condition'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "57ef2350",
   "metadata": {},
   "outputs": [],
   "source": [
    "groups = df[\"Country\"].to_numpy()\n",
    "X = df.drop([\"condition\",\"Unnamed: 0\", \"Country\"], axis=1)\n",
    "#X = df.drop([\"condition\"], axis=1)\n",
    "y = df[\"condition\"].to_numpy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "1fe4d4bc",
   "metadata": {
    "scrolled": true
   },
   "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>Streptococcus_anginosus___ref_mOTU_v2_0004__</th>\n",
       "      <th>Enterobacteriaceae_sp.___ref_mOTU_v2_0036__</th>\n",
       "      <th>Citrobacter_sp.___ref_mOTU_v2_0076__</th>\n",
       "      <th>Klebsiella_michiganensis/oxytoca___ref_mOTU_v2_0079__</th>\n",
       "      <th>Enterococcus_faecalis___ref_mOTU_v2_0116__</th>\n",
       "      <th>Lactobacillus_salivarius___ref_mOTU_v2_0125__</th>\n",
       "      <th>Dielma_fastidiosa___ref_mOTU_v2_0138__</th>\n",
       "      <th>Streptococcus_constellatus/intermedius___ref_mOTU_v2_0143__</th>\n",
       "      <th>Streptococcus_parasanguinis___ref_mOTU_v2_0144__</th>\n",
       "      <th>Streptococcus_sp._HSISM1___ref_mOTU_v2_0145__</th>\n",
       "      <th>...</th>\n",
       "      <th>unknown_Porphyromonas___meta_mOTU_v2_7777__</th>\n",
       "      <th>unknown_Clostridiales___meta_mOTU_v2_7778__</th>\n",
       "      <th>unknown_Clostridiales___meta_mOTU_v2_7781__</th>\n",
       "      <th>unknown_Clostridiales___meta_mOTU_v2_7782__</th>\n",
       "      <th>unknown_Clostridiales___meta_mOTU_v2_7784__</th>\n",
       "      <th>Clostridium_sp._CAG__230___meta_mOTU_v2_7788__</th>\n",
       "      <th>Clostridium_sp._CAG__1193___meta_mOTU_v2_7789__</th>\n",
       "      <th>unknown_Erysipelotrichaceae___meta_mOTU_v2_7790__</th>\n",
       "      <th>unknown_Clostridiales___meta_mOTU_v2_7795__</th>\n",
       "      <th>unknown_Clostridiales___meta_mOTU_v2_7800__</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000083</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000249</td>\n",
       "      <td>0.000249</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000083</td>\n",
       "      <td>0.000083</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.000591</td>\n",
       "      <td>0.000066</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000066</td>\n",
       "      <td>0.002102</td>\n",
       "      <td>0.000197</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000788</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.000840</td>\n",
       "      <td>0.000065</td>\n",
       "      <td>0.002454</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000065</td>\n",
       "      <td>0.013111</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000065</td>\n",
       "      <td>0.012013</td>\n",
       "      <td>0.002777</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000194</td>\n",
       "      <td>0.000065</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000840</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.001808</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000071</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000213</td>\n",
       "      <td>0.000071</td>\n",
       "      <td>0.000142</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.000086</td>\n",
       "      <td>0.019236</td>\n",
       "      <td>0.001111</td>\n",
       "      <td>0.000256</td>\n",
       "      <td>0.000086</td>\n",
       "      <td>0.000086</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000513</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000171</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>763</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000118</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000059</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000118</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000059</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.002590</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.001118</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>764</th>\n",
       "      <td>0.000038</td>\n",
       "      <td>0.000038</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000075</td>\n",
       "      <td>0.000038</td>\n",
       "      <td>0.000113</td>\n",
       "      <td>0.000113</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000151</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000151</td>\n",
       "      <td>0.000038</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>765</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000348</td>\n",
       "      <td>0.000044</td>\n",
       "      <td>0.001741</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000435</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000044</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>766</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000312</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000208</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000208</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000156</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>767</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000784</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000087</td>\n",
       "      <td>0.000087</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>768 rows × 849 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     Streptococcus_anginosus___ref_mOTU_v2_0004__  \\\n",
       "0                                        0.000000   \n",
       "1                                        0.000591   \n",
       "2                                        0.000840   \n",
       "3                                        0.000000   \n",
       "4                                        0.000086   \n",
       "..                                            ...   \n",
       "763                                      0.000000   \n",
       "764                                      0.000038   \n",
       "765                                      0.000000   \n",
       "766                                      0.000000   \n",
       "767                                      0.000000   \n",
       "\n",
       "     Enterobacteriaceae_sp.___ref_mOTU_v2_0036__  \\\n",
       "0                                       0.000000   \n",
       "1                                       0.000066   \n",
       "2                                       0.000065   \n",
       "3                                       0.000071   \n",
       "4                                       0.019236   \n",
       "..                                           ...   \n",
       "763                                     0.000118   \n",
       "764                                     0.000038   \n",
       "765                                     0.000000   \n",
       "766                                     0.000312   \n",
       "767                                     0.000000   \n",
       "\n",
       "     Citrobacter_sp.___ref_mOTU_v2_0076__  \\\n",
       "0                                0.000000   \n",
       "1                                0.000000   \n",
       "2                                0.002454   \n",
       "3                                0.000000   \n",
       "4                                0.001111   \n",
       "..                                    ...   \n",
       "763                              0.000000   \n",
       "764                              0.000000   \n",
       "765                              0.000000   \n",
       "766                              0.000000   \n",
       "767                              0.000784   \n",
       "\n",
       "     Klebsiella_michiganensis/oxytoca___ref_mOTU_v2_0079__  \\\n",
       "0                                             0.000000       \n",
       "1                                             0.000000       \n",
       "2                                             0.000000       \n",
       "3                                             0.000000       \n",
       "4                                             0.000256       \n",
       "..                                                 ...       \n",
       "763                                           0.000000       \n",
       "764                                           0.000000       \n",
       "765                                           0.000000       \n",
       "766                                           0.000000       \n",
       "767                                           0.000000       \n",
       "\n",
       "     Enterococcus_faecalis___ref_mOTU_v2_0116__  \\\n",
       "0                                      0.000000   \n",
       "1                                      0.000000   \n",
       "2                                      0.000065   \n",
       "3                                      0.000000   \n",
       "4                                      0.000086   \n",
       "..                                          ...   \n",
       "763                                    0.000059   \n",
       "764                                    0.000000   \n",
       "765                                    0.000000   \n",
       "766                                    0.000208   \n",
       "767                                    0.000000   \n",
       "\n",
       "     Lactobacillus_salivarius___ref_mOTU_v2_0125__  \\\n",
       "0                                         0.000000   \n",
       "1                                         0.000000   \n",
       "2                                         0.013111   \n",
       "3                                         0.000000   \n",
       "4                                         0.000086   \n",
       "..                                             ...   \n",
       "763                                       0.000000   \n",
       "764                                       0.000000   \n",
       "765                                       0.000000   \n",
       "766                                       0.000000   \n",
       "767                                       0.000000   \n",
       "\n",
       "     Dielma_fastidiosa___ref_mOTU_v2_0138__  \\\n",
       "0                                  0.000083   \n",
       "1                                  0.000000   \n",
       "2                                  0.000000   \n",
       "3                                  0.000213   \n",
       "4                                  0.000000   \n",
       "..                                      ...   \n",
       "763                                0.000118   \n",
       "764                                0.000075   \n",
       "765                                0.000000   \n",
       "766                                0.000000   \n",
       "767                                0.000000   \n",
       "\n",
       "     Streptococcus_constellatus/intermedius___ref_mOTU_v2_0143__  \\\n",
       "0                                             0.000000             \n",
       "1                                             0.000066             \n",
       "2                                             0.000065             \n",
       "3                                             0.000071             \n",
       "4                                             0.000000             \n",
       "..                                                 ...             \n",
       "763                                           0.000000             \n",
       "764                                           0.000038             \n",
       "765                                           0.000000             \n",
       "766                                           0.000000             \n",
       "767                                           0.000000             \n",
       "\n",
       "     Streptococcus_parasanguinis___ref_mOTU_v2_0144__  \\\n",
       "0                                            0.000249   \n",
       "1                                            0.002102   \n",
       "2                                            0.012013   \n",
       "3                                            0.000142   \n",
       "4                                            0.000513   \n",
       "..                                                ...   \n",
       "763                                          0.000059   \n",
       "764                                          0.000113   \n",
       "765                                          0.000000   \n",
       "766                                          0.000000   \n",
       "767                                          0.000087   \n",
       "\n",
       "     Streptococcus_sp._HSISM1___ref_mOTU_v2_0145__  ...  \\\n",
       "0                                         0.000249  ...   \n",
       "1                                         0.000197  ...   \n",
       "2                                         0.002777  ...   \n",
       "3                                         0.000000  ...   \n",
       "4                                         0.000000  ...   \n",
       "..                                             ...  ...   \n",
       "763                                       0.000000  ...   \n",
       "764                                       0.000113  ...   \n",
       "765                                       0.000000  ...   \n",
       "766                                       0.000000  ...   \n",
       "767                                       0.000087  ...   \n",
       "\n",
       "     unknown_Porphyromonas___meta_mOTU_v2_7777__  \\\n",
       "0                                       0.000000   \n",
       "1                                       0.000000   \n",
       "2                                       0.000000   \n",
       "3                                       0.000000   \n",
       "4                                       0.000000   \n",
       "..                                           ...   \n",
       "763                                     0.002590   \n",
       "764                                     0.000151   \n",
       "765                                     0.000000   \n",
       "766                                     0.000000   \n",
       "767                                     0.000000   \n",
       "\n",
       "     unknown_Clostridiales___meta_mOTU_v2_7778__  \\\n",
       "0                                       0.000000   \n",
       "1                                       0.000000   \n",
       "2                                       0.000000   \n",
       "3                                       0.000000   \n",
       "4                                       0.000171   \n",
       "..                                           ...   \n",
       "763                                     0.000000   \n",
       "764                                     0.000000   \n",
       "765                                     0.000348   \n",
       "766                                     0.000000   \n",
       "767                                     0.000000   \n",
       "\n",
       "     unknown_Clostridiales___meta_mOTU_v2_7781__  \\\n",
       "0                                       0.000000   \n",
       "1                                       0.000000   \n",
       "2                                       0.000194   \n",
       "3                                       0.000000   \n",
       "4                                       0.000000   \n",
       "..                                           ...   \n",
       "763                                     0.000000   \n",
       "764                                     0.000000   \n",
       "765                                     0.000044   \n",
       "766                                     0.000000   \n",
       "767                                     0.000000   \n",
       "\n",
       "     unknown_Clostridiales___meta_mOTU_v2_7782__  \\\n",
       "0                                       0.000000   \n",
       "1                                       0.000788   \n",
       "2                                       0.000065   \n",
       "3                                       0.000000   \n",
       "4                                       0.000000   \n",
       "..                                           ...   \n",
       "763                                     0.001118   \n",
       "764                                     0.000151   \n",
       "765                                     0.001741   \n",
       "766                                     0.000000   \n",
       "767                                     0.000000   \n",
       "\n",
       "     unknown_Clostridiales___meta_mOTU_v2_7784__  \\\n",
       "0                                       0.000000   \n",
       "1                                       0.000000   \n",
       "2                                       0.000000   \n",
       "3                                       0.000000   \n",
       "4                                       0.000000   \n",
       "..                                           ...   \n",
       "763                                     0.000000   \n",
       "764                                     0.000038   \n",
       "765                                     0.000000   \n",
       "766                                     0.000208   \n",
       "767                                     0.000000   \n",
       "\n",
       "     Clostridium_sp._CAG__230___meta_mOTU_v2_7788__  \\\n",
       "0                                          0.000000   \n",
       "1                                          0.000000   \n",
       "2                                          0.000840   \n",
       "3                                          0.000000   \n",
       "4                                          0.000000   \n",
       "..                                              ...   \n",
       "763                                        0.000000   \n",
       "764                                        0.000000   \n",
       "765                                        0.000435   \n",
       "766                                        0.000000   \n",
       "767                                        0.000000   \n",
       "\n",
       "     Clostridium_sp._CAG__1193___meta_mOTU_v2_7789__  \\\n",
       "0                                                0.0   \n",
       "1                                                0.0   \n",
       "2                                                0.0   \n",
       "3                                                0.0   \n",
       "4                                                0.0   \n",
       "..                                               ...   \n",
       "763                                              0.0   \n",
       "764                                              0.0   \n",
       "765                                              0.0   \n",
       "766                                              0.0   \n",
       "767                                              0.0   \n",
       "\n",
       "     unknown_Erysipelotrichaceae___meta_mOTU_v2_7790__  \\\n",
       "0                                                  0.0   \n",
       "1                                                  0.0   \n",
       "2                                                  0.0   \n",
       "3                                                  0.0   \n",
       "4                                                  0.0   \n",
       "..                                                 ...   \n",
       "763                                                0.0   \n",
       "764                                                0.0   \n",
       "765                                                0.0   \n",
       "766                                                0.0   \n",
       "767                                                0.0   \n",
       "\n",
       "     unknown_Clostridiales___meta_mOTU_v2_7795__  \\\n",
       "0                                       0.000083   \n",
       "1                                       0.000000   \n",
       "2                                       0.001808   \n",
       "3                                       0.000000   \n",
       "4                                       0.000000   \n",
       "..                                           ...   \n",
       "763                                     0.000000   \n",
       "764                                     0.000000   \n",
       "765                                     0.000000   \n",
       "766                                     0.000000   \n",
       "767                                     0.000000   \n",
       "\n",
       "     unknown_Clostridiales___meta_mOTU_v2_7800__  \n",
       "0                                       0.000083  \n",
       "1                                       0.000000  \n",
       "2                                       0.000000  \n",
       "3                                       0.000000  \n",
       "4                                       0.000000  \n",
       "..                                           ...  \n",
       "763                                     0.000000  \n",
       "764                                     0.000000  \n",
       "765                                     0.000044  \n",
       "766                                     0.000156  \n",
       "767                                     0.000000  \n",
       "\n",
       "[768 rows x 849 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "42222761",
   "metadata": {},
   "source": [
    "# Pandas dataframe to R table"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "088723d8",
   "metadata": {},
   "outputs": [],
   "source": [
    "with localconverter(robjects.default_converter + pandas2ri.converter):\n",
    "    r_from_pd_df = robjects.conversion.py2rpy(X)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0244b3ab",
   "metadata": {},
   "source": [
    "# Zero replacement"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "2096b00c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No. adjusted imputations:  425572 \n"
     ]
    }
   ],
   "source": [
    "np.random.seed(333)\n",
    "R_X = zCompositions.cmultRepl(r_from_pd_df, label = 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5e3d73ca",
   "metadata": {},
   "source": [
    "# CLR"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "da946733",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(333)\n",
    "R_clr_X = robCompositions.cenLR(R_X)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1cc67ebe",
   "metadata": {},
   "source": [
    "# R table to Pandas dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "7fd8bdc0",
   "metadata": {},
   "outputs": [],
   "source": [
    "with localconverter(robjects.default_converter + pandas2ri.converter):\n",
    "    X = robjects.conversion.rpy2py(R_clr_X[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "185a3c64",
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>758</th>\n",
       "      <th>759</th>\n",
       "      <th>760</th>\n",
       "      <th>761</th>\n",
       "      <th>762</th>\n",
       "      <th>763</th>\n",
       "      <th>764</th>\n",
       "      <th>765</th>\n",
       "      <th>766</th>\n",
       "      <th>767</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Streptococcus_anginosus___ref_mOTU_v2_0004__</th>\n",
       "      <td>-0.639814</td>\n",
       "      <td>2.453491</td>\n",
       "      <td>2.598610</td>\n",
       "      <td>-0.596255</td>\n",
       "      <td>0.157178</td>\n",
       "      <td>0.510385</td>\n",
       "      <td>1.332754</td>\n",
       "      <td>-1.175181</td>\n",
       "      <td>-0.783341</td>\n",
       "      <td>-1.157931</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.189774</td>\n",
       "      <td>-1.373054</td>\n",
       "      <td>-1.359970</td>\n",
       "      <td>-1.188060</td>\n",
       "      <td>1.345367</td>\n",
       "      <td>-0.996471</td>\n",
       "      <td>-0.490840</td>\n",
       "      <td>-0.996933</td>\n",
       "      <td>-0.751492</td>\n",
       "      <td>-0.619330</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Enterobacteriaceae_sp.___ref_mOTU_v2_0036__</th>\n",
       "      <td>-1.123042</td>\n",
       "      <td>0.256614</td>\n",
       "      <td>0.033862</td>\n",
       "      <td>0.544960</td>\n",
       "      <td>5.573185</td>\n",
       "      <td>7.364506</td>\n",
       "      <td>4.062783</td>\n",
       "      <td>0.035616</td>\n",
       "      <td>0.380089</td>\n",
       "      <td>5.863864</td>\n",
       "      <td>...</td>\n",
       "      <td>2.013689</td>\n",
       "      <td>3.671123</td>\n",
       "      <td>1.055905</td>\n",
       "      <td>-1.671288</td>\n",
       "      <td>0.651934</td>\n",
       "      <td>0.650929</td>\n",
       "      <td>-0.490840</td>\n",
       "      <td>-1.480161</td>\n",
       "      <td>1.998640</td>\n",
       "      <td>-1.102559</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Citrobacter_sp.___ref_mOTU_v2_0076__</th>\n",
       "      <td>-0.638052</td>\n",
       "      <td>-0.833799</td>\n",
       "      <td>3.671247</td>\n",
       "      <td>-0.594493</td>\n",
       "      <td>2.722033</td>\n",
       "      <td>-0.814918</td>\n",
       "      <td>0.927289</td>\n",
       "      <td>-1.173420</td>\n",
       "      <td>-0.781579</td>\n",
       "      <td>-1.156169</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.188012</td>\n",
       "      <td>-1.371292</td>\n",
       "      <td>-1.358209</td>\n",
       "      <td>-1.186298</td>\n",
       "      <td>-0.635869</td>\n",
       "      <td>-0.994709</td>\n",
       "      <td>-1.000097</td>\n",
       "      <td>-0.995171</td>\n",
       "      <td>-0.749730</td>\n",
       "      <td>2.941765</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Klebsiella_michiganensis/oxytoca___ref_mOTU_v2_0079__</th>\n",
       "      <td>-0.627669</td>\n",
       "      <td>-0.823416</td>\n",
       "      <td>-1.061800</td>\n",
       "      <td>-0.584110</td>\n",
       "      <td>1.255696</td>\n",
       "      <td>-0.804534</td>\n",
       "      <td>-1.139914</td>\n",
       "      <td>-1.163036</td>\n",
       "      <td>-0.771196</td>\n",
       "      <td>-1.145786</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.177628</td>\n",
       "      <td>-1.360908</td>\n",
       "      <td>-1.347825</td>\n",
       "      <td>-1.175915</td>\n",
       "      <td>-0.625485</td>\n",
       "      <td>-0.984326</td>\n",
       "      <td>-0.989714</td>\n",
       "      <td>-0.984788</td>\n",
       "      <td>-0.739347</td>\n",
       "      <td>-0.607185</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Enterococcus_faecalis___ref_mOTU_v2_0116__</th>\n",
       "      <td>-0.766950</td>\n",
       "      <td>-0.962697</td>\n",
       "      <td>0.033862</td>\n",
       "      <td>-0.723391</td>\n",
       "      <td>0.157178</td>\n",
       "      <td>0.510385</td>\n",
       "      <td>-1.279196</td>\n",
       "      <td>0.035616</td>\n",
       "      <td>-0.910477</td>\n",
       "      <td>-1.285067</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.316910</td>\n",
       "      <td>-1.500190</td>\n",
       "      <td>-1.487106</td>\n",
       "      <td>-1.315196</td>\n",
       "      <td>-0.764766</td>\n",
       "      <td>-0.041564</td>\n",
       "      <td>-1.128995</td>\n",
       "      <td>-1.124069</td>\n",
       "      <td>1.593176</td>\n",
       "      <td>-0.746466</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Clostridium_sp._CAG__230___meta_mOTU_v2_7788__</th>\n",
       "      <td>-0.535670</td>\n",
       "      <td>-0.731417</td>\n",
       "      <td>2.598610</td>\n",
       "      <td>-0.492111</td>\n",
       "      <td>-1.088093</td>\n",
       "      <td>-0.712535</td>\n",
       "      <td>2.025901</td>\n",
       "      <td>-1.071037</td>\n",
       "      <td>-0.679197</td>\n",
       "      <td>2.180771</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.085629</td>\n",
       "      <td>-1.268909</td>\n",
       "      <td>0.362754</td>\n",
       "      <td>-1.083916</td>\n",
       "      <td>-0.533486</td>\n",
       "      <td>-0.892327</td>\n",
       "      <td>-0.897715</td>\n",
       "      <td>1.979854</td>\n",
       "      <td>-0.647348</td>\n",
       "      <td>-0.515186</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Clostridium_sp._CAG__1193___meta_mOTU_v2_7789__</th>\n",
       "      <td>-0.353747</td>\n",
       "      <td>-0.549494</td>\n",
       "      <td>-0.787878</td>\n",
       "      <td>-0.310188</td>\n",
       "      <td>-0.906170</td>\n",
       "      <td>-0.530612</td>\n",
       "      <td>-0.865992</td>\n",
       "      <td>0.729412</td>\n",
       "      <td>-0.497274</td>\n",
       "      <td>-0.871864</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.903706</td>\n",
       "      <td>-1.086986</td>\n",
       "      <td>-1.073903</td>\n",
       "      <td>-0.901993</td>\n",
       "      <td>-0.351563</td>\n",
       "      <td>-0.710404</td>\n",
       "      <td>-0.715792</td>\n",
       "      <td>-0.710866</td>\n",
       "      <td>-0.465425</td>\n",
       "      <td>-0.333263</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unknown_Erysipelotrichaceae___meta_mOTU_v2_7790__</th>\n",
       "      <td>-0.392983</td>\n",
       "      <td>-0.588730</td>\n",
       "      <td>-0.827114</td>\n",
       "      <td>-0.349424</td>\n",
       "      <td>-0.945406</td>\n",
       "      <td>-0.569848</td>\n",
       "      <td>0.234356</td>\n",
       "      <td>-0.928350</td>\n",
       "      <td>-0.536510</td>\n",
       "      <td>5.199317</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.942942</td>\n",
       "      <td>-1.126222</td>\n",
       "      <td>-1.113139</td>\n",
       "      <td>-0.941229</td>\n",
       "      <td>-0.390799</td>\n",
       "      <td>-0.749640</td>\n",
       "      <td>-0.755028</td>\n",
       "      <td>-0.750102</td>\n",
       "      <td>-0.504661</td>\n",
       "      <td>-0.372499</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unknown_Clostridiales___meta_mOTU_v2_7795__</th>\n",
       "      <td>0.657708</td>\n",
       "      <td>-0.970122</td>\n",
       "      <td>3.365865</td>\n",
       "      <td>-0.730816</td>\n",
       "      <td>-1.326798</td>\n",
       "      <td>-0.951240</td>\n",
       "      <td>-1.286620</td>\n",
       "      <td>3.127304</td>\n",
       "      <td>-0.917902</td>\n",
       "      <td>2.983117</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.324334</td>\n",
       "      <td>-1.507614</td>\n",
       "      <td>0.362754</td>\n",
       "      <td>-1.322621</td>\n",
       "      <td>-0.772191</td>\n",
       "      <td>-1.131032</td>\n",
       "      <td>-1.136420</td>\n",
       "      <td>-1.131494</td>\n",
       "      <td>-0.886053</td>\n",
       "      <td>-0.753891</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unknown_Clostridiales___meta_mOTU_v2_7800__</th>\n",
       "      <td>0.657708</td>\n",
       "      <td>-0.962697</td>\n",
       "      <td>-1.201081</td>\n",
       "      <td>-0.723391</td>\n",
       "      <td>-1.319373</td>\n",
       "      <td>-0.943815</td>\n",
       "      <td>0.234356</td>\n",
       "      <td>-1.302317</td>\n",
       "      <td>-0.910477</td>\n",
       "      <td>2.660344</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.625446</td>\n",
       "      <td>5.097799</td>\n",
       "      <td>-0.329932</td>\n",
       "      <td>-1.315196</td>\n",
       "      <td>-0.764766</td>\n",
       "      <td>-1.123607</td>\n",
       "      <td>-1.128995</td>\n",
       "      <td>-0.323276</td>\n",
       "      <td>1.305489</td>\n",
       "      <td>-0.746466</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>849 rows × 768 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                           0         1  \\\n",
       "Streptococcus_anginosus___ref_mOTU_v2_0004__       -0.639814  2.453491   \n",
       "Enterobacteriaceae_sp.___ref_mOTU_v2_0036__        -1.123042  0.256614   \n",
       "Citrobacter_sp.___ref_mOTU_v2_0076__               -0.638052 -0.833799   \n",
       "Klebsiella_michiganensis/oxytoca___ref_mOTU_v2_... -0.627669 -0.823416   \n",
       "Enterococcus_faecalis___ref_mOTU_v2_0116__         -0.766950 -0.962697   \n",
       "...                                                      ...       ...   \n",
       "Clostridium_sp._CAG__230___meta_mOTU_v2_7788__     -0.535670 -0.731417   \n",
       "Clostridium_sp._CAG__1193___meta_mOTU_v2_7789__    -0.353747 -0.549494   \n",
       "unknown_Erysipelotrichaceae___meta_mOTU_v2_7790__  -0.392983 -0.588730   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7795__         0.657708 -0.970122   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7800__         0.657708 -0.962697   \n",
       "\n",
       "                                                           2         3  \\\n",
       "Streptococcus_anginosus___ref_mOTU_v2_0004__        2.598610 -0.596255   \n",
       "Enterobacteriaceae_sp.___ref_mOTU_v2_0036__         0.033862  0.544960   \n",
       "Citrobacter_sp.___ref_mOTU_v2_0076__                3.671247 -0.594493   \n",
       "Klebsiella_michiganensis/oxytoca___ref_mOTU_v2_... -1.061800 -0.584110   \n",
       "Enterococcus_faecalis___ref_mOTU_v2_0116__          0.033862 -0.723391   \n",
       "...                                                      ...       ...   \n",
       "Clostridium_sp._CAG__230___meta_mOTU_v2_7788__      2.598610 -0.492111   \n",
       "Clostridium_sp._CAG__1193___meta_mOTU_v2_7789__    -0.787878 -0.310188   \n",
       "unknown_Erysipelotrichaceae___meta_mOTU_v2_7790__  -0.827114 -0.349424   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7795__         3.365865 -0.730816   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7800__        -1.201081 -0.723391   \n",
       "\n",
       "                                                           4         5  \\\n",
       "Streptococcus_anginosus___ref_mOTU_v2_0004__        0.157178  0.510385   \n",
       "Enterobacteriaceae_sp.___ref_mOTU_v2_0036__         5.573185  7.364506   \n",
       "Citrobacter_sp.___ref_mOTU_v2_0076__                2.722033 -0.814918   \n",
       "Klebsiella_michiganensis/oxytoca___ref_mOTU_v2_...  1.255696 -0.804534   \n",
       "Enterococcus_faecalis___ref_mOTU_v2_0116__          0.157178  0.510385   \n",
       "...                                                      ...       ...   \n",
       "Clostridium_sp._CAG__230___meta_mOTU_v2_7788__     -1.088093 -0.712535   \n",
       "Clostridium_sp._CAG__1193___meta_mOTU_v2_7789__    -0.906170 -0.530612   \n",
       "unknown_Erysipelotrichaceae___meta_mOTU_v2_7790__  -0.945406 -0.569848   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7795__        -1.326798 -0.951240   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7800__        -1.319373 -0.943815   \n",
       "\n",
       "                                                           6         7  \\\n",
       "Streptococcus_anginosus___ref_mOTU_v2_0004__        1.332754 -1.175181   \n",
       "Enterobacteriaceae_sp.___ref_mOTU_v2_0036__         4.062783  0.035616   \n",
       "Citrobacter_sp.___ref_mOTU_v2_0076__                0.927289 -1.173420   \n",
       "Klebsiella_michiganensis/oxytoca___ref_mOTU_v2_... -1.139914 -1.163036   \n",
       "Enterococcus_faecalis___ref_mOTU_v2_0116__         -1.279196  0.035616   \n",
       "...                                                      ...       ...   \n",
       "Clostridium_sp._CAG__230___meta_mOTU_v2_7788__      2.025901 -1.071037   \n",
       "Clostridium_sp._CAG__1193___meta_mOTU_v2_7789__    -0.865992  0.729412   \n",
       "unknown_Erysipelotrichaceae___meta_mOTU_v2_7790__   0.234356 -0.928350   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7795__        -1.286620  3.127304   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7800__         0.234356 -1.302317   \n",
       "\n",
       "                                                           8         9  ...  \\\n",
       "Streptococcus_anginosus___ref_mOTU_v2_0004__       -0.783341 -1.157931  ...   \n",
       "Enterobacteriaceae_sp.___ref_mOTU_v2_0036__         0.380089  5.863864  ...   \n",
       "Citrobacter_sp.___ref_mOTU_v2_0076__               -0.781579 -1.156169  ...   \n",
       "Klebsiella_michiganensis/oxytoca___ref_mOTU_v2_... -0.771196 -1.145786  ...   \n",
       "Enterococcus_faecalis___ref_mOTU_v2_0116__         -0.910477 -1.285067  ...   \n",
       "...                                                      ...       ...  ...   \n",
       "Clostridium_sp._CAG__230___meta_mOTU_v2_7788__     -0.679197  2.180771  ...   \n",
       "Clostridium_sp._CAG__1193___meta_mOTU_v2_7789__    -0.497274 -0.871864  ...   \n",
       "unknown_Erysipelotrichaceae___meta_mOTU_v2_7790__  -0.536510  5.199317  ...   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7795__        -0.917902  2.983117  ...   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7800__        -0.910477  2.660344  ...   \n",
       "\n",
       "                                                         758       759  \\\n",
       "Streptococcus_anginosus___ref_mOTU_v2_0004__       -1.189774 -1.373054   \n",
       "Enterobacteriaceae_sp.___ref_mOTU_v2_0036__         2.013689  3.671123   \n",
       "Citrobacter_sp.___ref_mOTU_v2_0076__               -1.188012 -1.371292   \n",
       "Klebsiella_michiganensis/oxytoca___ref_mOTU_v2_... -1.177628 -1.360908   \n",
       "Enterococcus_faecalis___ref_mOTU_v2_0116__         -1.316910 -1.500190   \n",
       "...                                                      ...       ...   \n",
       "Clostridium_sp._CAG__230___meta_mOTU_v2_7788__     -1.085629 -1.268909   \n",
       "Clostridium_sp._CAG__1193___meta_mOTU_v2_7789__    -0.903706 -1.086986   \n",
       "unknown_Erysipelotrichaceae___meta_mOTU_v2_7790__  -0.942942 -1.126222   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7795__        -1.324334 -1.507614   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7800__        -0.625446  5.097799   \n",
       "\n",
       "                                                         760       761  \\\n",
       "Streptococcus_anginosus___ref_mOTU_v2_0004__       -1.359970 -1.188060   \n",
       "Enterobacteriaceae_sp.___ref_mOTU_v2_0036__         1.055905 -1.671288   \n",
       "Citrobacter_sp.___ref_mOTU_v2_0076__               -1.358209 -1.186298   \n",
       "Klebsiella_michiganensis/oxytoca___ref_mOTU_v2_... -1.347825 -1.175915   \n",
       "Enterococcus_faecalis___ref_mOTU_v2_0116__         -1.487106 -1.315196   \n",
       "...                                                      ...       ...   \n",
       "Clostridium_sp._CAG__230___meta_mOTU_v2_7788__      0.362754 -1.083916   \n",
       "Clostridium_sp._CAG__1193___meta_mOTU_v2_7789__    -1.073903 -0.901993   \n",
       "unknown_Erysipelotrichaceae___meta_mOTU_v2_7790__  -1.113139 -0.941229   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7795__         0.362754 -1.322621   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7800__        -0.329932 -1.315196   \n",
       "\n",
       "                                                         762       763  \\\n",
       "Streptococcus_anginosus___ref_mOTU_v2_0004__        1.345367 -0.996471   \n",
       "Enterobacteriaceae_sp.___ref_mOTU_v2_0036__         0.651934  0.650929   \n",
       "Citrobacter_sp.___ref_mOTU_v2_0076__               -0.635869 -0.994709   \n",
       "Klebsiella_michiganensis/oxytoca___ref_mOTU_v2_... -0.625485 -0.984326   \n",
       "Enterococcus_faecalis___ref_mOTU_v2_0116__         -0.764766 -0.041564   \n",
       "...                                                      ...       ...   \n",
       "Clostridium_sp._CAG__230___meta_mOTU_v2_7788__     -0.533486 -0.892327   \n",
       "Clostridium_sp._CAG__1193___meta_mOTU_v2_7789__    -0.351563 -0.710404   \n",
       "unknown_Erysipelotrichaceae___meta_mOTU_v2_7790__  -0.390799 -0.749640   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7795__        -0.772191 -1.131032   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7800__        -0.764766 -1.123607   \n",
       "\n",
       "                                                         764       765  \\\n",
       "Streptococcus_anginosus___ref_mOTU_v2_0004__       -0.490840 -0.996933   \n",
       "Enterobacteriaceae_sp.___ref_mOTU_v2_0036__        -0.490840 -1.480161   \n",
       "Citrobacter_sp.___ref_mOTU_v2_0076__               -1.000097 -0.995171   \n",
       "Klebsiella_michiganensis/oxytoca___ref_mOTU_v2_... -0.989714 -0.984788   \n",
       "Enterococcus_faecalis___ref_mOTU_v2_0116__         -1.128995 -1.124069   \n",
       "...                                                      ...       ...   \n",
       "Clostridium_sp._CAG__230___meta_mOTU_v2_7788__     -0.897715  1.979854   \n",
       "Clostridium_sp._CAG__1193___meta_mOTU_v2_7789__    -0.715792 -0.710866   \n",
       "unknown_Erysipelotrichaceae___meta_mOTU_v2_7790__  -0.755028 -0.750102   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7795__        -1.136420 -1.131494   \n",
       "unknown_Clostridiales___meta_mOTU_v2_7800__        -1.128995 -0.323276   \n",
       "\n",
       "                                                         766       767  \n",
       "Streptococcus_anginosus___ref_mOTU_v2_0004__       -0.751492 -0.619330  \n",
       "Enterobacteriaceae_sp.___ref_mOTU_v2_0036__         1.998640 -1.102559  \n",
       "Citrobacter_sp.___ref_mOTU_v2_0076__               -0.749730  2.941765  \n",
       "Klebsiella_michiganensis/oxytoca___ref_mOTU_v2_... -0.739347 -0.607185  \n",
       "Enterococcus_faecalis___ref_mOTU_v2_0116__          1.593176 -0.746466  \n",
       "...                                                      ...       ...  \n",
       "Clostridium_sp._CAG__230___meta_mOTU_v2_7788__     -0.647348 -0.515186  \n",
       "Clostridium_sp._CAG__1193___meta_mOTU_v2_7789__    -0.465425 -0.333263  \n",
       "unknown_Erysipelotrichaceae___meta_mOTU_v2_7790__  -0.504661 -0.372499  \n",
       "unknown_Clostridiales___meta_mOTU_v2_7795__        -0.886053 -0.753891  \n",
       "unknown_Clostridiales___meta_mOTU_v2_7800__         1.305489 -0.746466  \n",
       "\n",
       "[849 rows x 768 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "XT = X.T\n",
    "XT"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "88aaa089",
   "metadata": {},
   "source": [
    "# Clustering the features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "dbc1a83c",
   "metadata": {},
   "outputs": [],
   "source": [
    "d_sample = 'euclidean' #See possible values: https://docs.scipy.org/doc/scipy/reference/generated/scipy.spatial.distance.pdist.html#scipy.spatial.distance.pdist\n",
    "d_group = 'ward' #See possible values (method): https://docs.scipy.org/doc/scipy/reference/generated/scipy.cluster.hierarchy.linkage.html\n",
    "N_leafs = 20 #Number of leaf nodes. n=400"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "d020cc51",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "np.random.seed(333)\n",
    "hierarchy.set_link_color_palette(['b', 'r', 'g'])\n",
    "Z = hierarchy.linkage(XT , method=d_group, metric=d_sample) #method denotes cluster distance, metric denotes sample distance\n",
    "plt.figure()\n",
    "# plt.title('Hierarchical Clustering Dendrogram')\n",
    "plt.ylabel('features in cluster')\n",
    "plt.xlabel('distance')\n",
    "den = hierarchy.dendrogram(\n",
    "    Z,\n",
    "#     leaf_rotation=90.,\n",
    "    leaf_font_size=10.,\n",
    "    truncate_mode='lastp',\n",
    "    p = N_leafs,\n",
    "    count_sort = \"ascending\",\n",
    "    orientation='right'\n",
    ")\n",
    "plt.savefig(\"plots/LOSO_CLR_featadd_4.jpg\",dpi=300)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "4c969857",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The associated labels with cluster 1 is:\n",
      "[  2   3   4  10  12  15  19  21  22  33  36  37  39  40  41  48  51  52\n",
      "  58  63  70  87  89  95  97 104 117 119 124 125 135 136 145 146 147 148\n",
      " 149 151 154 155 157 161 168 174 188 190 203 211 214 216 222 224 233 234\n",
      " 241 244 250 274 279 281 282 292 303 306 311 318 332 336 338 339 340 342\n",
      " 345 353 354 359 361 370 373 374 380 387 389 391 392 394 404 406 410 412\n",
      " 424 428 440 446 447 457 463 468 471 476 478 481 483 484 491 495 497 500\n",
      " 503 510 511 514 517 518 520 523 534 536 544 546 549 550 554 555 560 563\n",
      " 566 569 575 578 582 587 592 599 603 606 619 624 625 628 629 640 641 654\n",
      " 666 668 674 676 684 690 710 719 730 744 748 751 752 753 754 768 771 772\n",
      " 787 790 792 794 795 805 814 818 822 830 831 832 833 837 838 839 845]\n",
      "\n",
      "\n",
      "The associated labels with cluster 2 is:\n",
      "[  6  32  73  75  88 112 113 120 121 123 131 143 153 156 173 185 194 221\n",
      " 256 260 288 320 398 417 426 436 459 465 537 548 551 609 648 678 717 758\n",
      " 780 781 804 817 825]\n",
      "\n",
      "\n",
      "The associated labels with cluster 3 is:\n",
      "[ 20  45 103 105 132 138 139 160 165 171 181 195 198 207 232 238 239 254\n",
      " 255 258 263 265 269 270 271 275 280 284 290 291 293 294 295 296 297 299\n",
      " 305 307 309 310 312 313 314 315 317 319 322 323 325 326 328 333 335 337\n",
      " 343 344 346 348 351 352 356 357 367 369 371 376 377 379 381 382 383 388\n",
      " 390 397 408 414 416 418 421 422 423 425 427 429 430 431 433 435 441 442\n",
      " 444 448 449 450 452 453 456 461 466 467 470 477 479 482 485 486 487 489\n",
      " 492 494 499 501 502 504 507 508 512 515 516 522 526 528 529 530 535 539\n",
      " 545 564 565 570 571 574 577 581 584 586 588 590 594 605 608 611 621 626\n",
      " 630 635 636 638 645 646 651 652 655 657 658 659 660 661 663 667 669 670\n",
      " 673 675 679 681 682 683 685 687 691 693 696 700 706 708 709 711 712 713\n",
      " 714 716 718 720 722 725 729 732 733 734 737 740 746 747 750 755 757 759\n",
      " 760 764 766 769 773 778 784 786 789 793 796 797 798 799 800 806 807 809\n",
      " 810 811 816 821 823 824 826 829 834 841 843 844 846]\n",
      "\n",
      "\n",
      "The associated labels with cluster 4 is:\n",
      "[141 180 199 246 253 264 277 278 285 286 302 304 308 330 349 360 378 385\n",
      " 399 400 407 413 419 437 439 443 455 460 473 475 488 490 498 506 519 524\n",
      " 532 543 568 572 580 585 596 602 615 634 639 649 662 672 692 704 707 715\n",
      " 723 728 749 756 761 765 782 783 785 801 820 847 848]\n",
      "\n",
      "\n",
      "The associated labels with cluster 5 is:\n",
      "[ 31  76  79  83  96 126 140 142 152 163 167 193 201 205 208 229 231 240\n",
      " 261 267 268 300 324 327 334 347 402 411 415 438 464 472 496 533 540 556\n",
      " 589 613 617 642 643 644 647 695 724 726 735 739 770 791 802 827 836]\n",
      "\n",
      "\n",
      "The associated labels with cluster 6 is:\n",
      "[ 13  14  28  38  46  54  55  85 102 166 187 189 191 245 248 251 287 329\n",
      " 355 364 384 386 393 405 445 462 469 525 531 541 542 558 576 579 593 604\n",
      " 614 623 631 633 664 665 689 698 699 721 727 736 738 741 767 776 777 812\n",
      " 813 815 828]\n",
      "\n",
      "\n",
      "The associated labels with cluster 7 is:\n",
      "[  0   5   7  24  25  30  34  35  47  49  50  53  59  60  62  64  67  68\n",
      "  69  90  91  98  99 122 137 206 209 213 217 242 243 350 365 375 480 618\n",
      " 803]\n",
      "\n",
      "\n",
      "The associated labels with cluster 8 is:\n",
      "[ 11  44  72  86 101 107 115 175 176 179 192 218 220 283 547 686]\n",
      "\n",
      "\n",
      "The associated labels with cluster 9 is:\n",
      "[111 129 150 182 200 215 228 236 259 331 403 538 552 627 637]\n",
      "\n",
      "\n",
      "The associated labels with cluster 10 is:\n",
      "[81 82]\n",
      "\n",
      "\n",
      "The associated labels with cluster 11 is:\n",
      "[  1  66  71  80  92  93  94 100 108 109 223 509]\n",
      "\n",
      "\n",
      "The associated labels with cluster 12 is:\n",
      "[116 273 298 372 395 409 420 451 454 458 474 493 513 559 583 597 598 600\n",
      " 607 694 697 701 703 705 731 742 745 774 779 819 835 840 842]\n",
      "\n",
      "\n",
      "The associated labels with cluster 13 is:\n",
      "[183 257 262 266 316 358 362 366 401 505 527 601 616 620 632 653 671 688\n",
      " 743 762 763 775]\n",
      "\n",
      "\n",
      "The associated labels with cluster 14 is:\n",
      "[ 43  65  84 127 128 178 196 202 235 237 788]\n",
      "\n",
      "\n",
      "The associated labels with cluster 15 is:\n",
      "[ 26  29  57  74  77  78 133 159 197 561]\n",
      "\n",
      "\n",
      "The associated labels with cluster 16 is:\n",
      "[  8   9  16  17  18 321]\n",
      "\n",
      "\n",
      "The associated labels with cluster 17 is:\n",
      "[ 56 106 134 164 184 212 249 252 567 680]\n",
      "\n",
      "\n",
      "The associated labels with cluster 18 is:\n",
      "[ 27 130 158 162 169 172 177 219 227 276 289 341 368 432 521 553 557 573\n",
      " 591 595 610 612 650 656 677 808]\n",
      "\n",
      "\n",
      "The associated labels with cluster 19 is:\n",
      "[110 114 118 144 170 186 210 225 226 230 272 363 562 622 702]\n",
      "\n",
      "\n",
      "The associated labels with cluster 20 is:\n",
      "[ 23  42  61 204 247 301 396 434]\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "np.random.seed(333)\n",
    "k = 20\n",
    "cluster_designation = fcluster(Z, k, criterion = \"maxclust\")\n",
    "\n",
    "for i in range(1,k+1):\n",
    "    datapoints = np.where(cluster_designation == i)\n",
    "    print(\"The associated labels with cluster {} is:\".format(i))\n",
    "    print(datapoints[0])\n",
    "    print(\"\\n\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "8222d3c4",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_new = pd.DataFrame()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f593734c",
   "metadata": {},
   "source": [
    "# Summation of features in the same cluster"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "6036ccce",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "k = 20\n",
    "cluster_designation = fcluster(Z, k, criterion = \"maxclust\")\n",
    "for i in range(1,k+1):\n",
    "    datapoints = np.where(cluster_designation == i)\n",
    "    X_new[f\"{i}\"] = (X.iloc[:,datapoints[0]]).sum(axis=1)\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "787a3f56",
   "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>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>11</th>\n",
       "      <th>12</th>\n",
       "      <th>13</th>\n",
       "      <th>14</th>\n",
       "      <th>15</th>\n",
       "      <th>16</th>\n",
       "      <th>17</th>\n",
       "      <th>18</th>\n",
       "      <th>19</th>\n",
       "      <th>20</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-105.642589</td>\n",
       "      <td>3.523409</td>\n",
       "      <td>-92.246457</td>\n",
       "      <td>-10.842514</td>\n",
       "      <td>-14.564065</td>\n",
       "      <td>-9.463218</td>\n",
       "      <td>-13.855369</td>\n",
       "      <td>76.950658</td>\n",
       "      <td>49.846658</td>\n",
       "      <td>7.948933</td>\n",
       "      <td>11.997859</td>\n",
       "      <td>-17.600960</td>\n",
       "      <td>15.747128</td>\n",
       "      <td>-5.359820</td>\n",
       "      <td>7.656326</td>\n",
       "      <td>14.626570</td>\n",
       "      <td>17.383610</td>\n",
       "      <td>53.427690</td>\n",
       "      <td>8.527089</td>\n",
       "      <td>1.939061</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-132.559431</td>\n",
       "      <td>0.686573</td>\n",
       "      <td>-117.879118</td>\n",
       "      <td>-33.839001</td>\n",
       "      <td>0.368988</td>\n",
       "      <td>-5.450244</td>\n",
       "      <td>13.086981</td>\n",
       "      <td>72.105231</td>\n",
       "      <td>29.220280</td>\n",
       "      <td>6.478679</td>\n",
       "      <td>16.046478</td>\n",
       "      <td>0.346984</td>\n",
       "      <td>20.771332</td>\n",
       "      <td>-1.210382</td>\n",
       "      <td>6.475880</td>\n",
       "      <td>24.276589</td>\n",
       "      <td>31.065191</td>\n",
       "      <td>43.517274</td>\n",
       "      <td>31.887459</td>\n",
       "      <td>-5.395743</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-125.820963</td>\n",
       "      <td>-29.318014</td>\n",
       "      <td>-122.626954</td>\n",
       "      <td>40.305080</td>\n",
       "      <td>-26.607432</td>\n",
       "      <td>9.570022</td>\n",
       "      <td>19.152528</td>\n",
       "      <td>74.205878</td>\n",
       "      <td>36.133930</td>\n",
       "      <td>4.056305</td>\n",
       "      <td>2.475786</td>\n",
       "      <td>37.753835</td>\n",
       "      <td>-4.098960</td>\n",
       "      <td>1.987441</td>\n",
       "      <td>-8.440176</td>\n",
       "      <td>27.583215</td>\n",
       "      <td>11.461091</td>\n",
       "      <td>42.995740</td>\n",
       "      <td>16.015795</td>\n",
       "      <td>-6.784145</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-109.827708</td>\n",
       "      <td>-2.130972</td>\n",
       "      <td>-73.175728</td>\n",
       "      <td>-33.102889</td>\n",
       "      <td>-4.895685</td>\n",
       "      <td>6.343304</td>\n",
       "      <td>-14.379102</td>\n",
       "      <td>66.353085</td>\n",
       "      <td>42.533320</td>\n",
       "      <td>1.782594</td>\n",
       "      <td>14.423091</td>\n",
       "      <td>-14.711778</td>\n",
       "      <td>10.534064</td>\n",
       "      <td>-4.101555</td>\n",
       "      <td>10.320962</td>\n",
       "      <td>4.699534</td>\n",
       "      <td>22.557302</td>\n",
       "      <td>71.811843</td>\n",
       "      <td>6.837782</td>\n",
       "      <td>-1.871465</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-157.076006</td>\n",
       "      <td>-2.560414</td>\n",
       "      <td>-154.402908</td>\n",
       "      <td>-34.053480</td>\n",
       "      <td>32.223315</td>\n",
       "      <td>5.883290</td>\n",
       "      <td>-26.036329</td>\n",
       "      <td>41.949521</td>\n",
       "      <td>53.124888</td>\n",
       "      <td>10.032831</td>\n",
       "      <td>37.293703</td>\n",
       "      <td>51.091919</td>\n",
       "      <td>19.164965</td>\n",
       "      <td>-1.440777</td>\n",
       "      <td>8.477097</td>\n",
       "      <td>13.370941</td>\n",
       "      <td>24.086364</td>\n",
       "      <td>35.228973</td>\n",
       "      <td>41.661200</td>\n",
       "      <td>1.980905</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>763</th>\n",
       "      <td>-135.482146</td>\n",
       "      <td>4.662305</td>\n",
       "      <td>-152.819028</td>\n",
       "      <td>-50.540514</td>\n",
       "      <td>25.331998</td>\n",
       "      <td>-2.516600</td>\n",
       "      <td>-31.505189</td>\n",
       "      <td>52.733280</td>\n",
       "      <td>62.210393</td>\n",
       "      <td>14.143995</td>\n",
       "      <td>37.520076</td>\n",
       "      <td>27.989724</td>\n",
       "      <td>16.184726</td>\n",
       "      <td>-10.239194</td>\n",
       "      <td>6.294204</td>\n",
       "      <td>0.992845</td>\n",
       "      <td>28.950867</td>\n",
       "      <td>61.940471</td>\n",
       "      <td>30.870036</td>\n",
       "      <td>13.277754</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>764</th>\n",
       "      <td>-119.300415</td>\n",
       "      <td>-4.276729</td>\n",
       "      <td>-140.446478</td>\n",
       "      <td>-43.307099</td>\n",
       "      <td>42.011823</td>\n",
       "      <td>-15.461389</td>\n",
       "      <td>-7.324946</td>\n",
       "      <td>44.725894</td>\n",
       "      <td>57.869686</td>\n",
       "      <td>8.734915</td>\n",
       "      <td>38.412462</td>\n",
       "      <td>26.337352</td>\n",
       "      <td>-10.345786</td>\n",
       "      <td>-4.509702</td>\n",
       "      <td>4.708844</td>\n",
       "      <td>2.614399</td>\n",
       "      <td>30.080757</td>\n",
       "      <td>32.968018</td>\n",
       "      <td>52.712497</td>\n",
       "      <td>3.795897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>765</th>\n",
       "      <td>-172.142869</td>\n",
       "      <td>-18.304606</td>\n",
       "      <td>-136.150350</td>\n",
       "      <td>-27.622497</td>\n",
       "      <td>37.332445</td>\n",
       "      <td>-7.721931</td>\n",
       "      <td>-39.142331</td>\n",
       "      <td>41.211884</td>\n",
       "      <td>64.103819</td>\n",
       "      <td>11.788660</td>\n",
       "      <td>37.141524</td>\n",
       "      <td>21.923082</td>\n",
       "      <td>36.069846</td>\n",
       "      <td>-1.041638</td>\n",
       "      <td>2.717797</td>\n",
       "      <td>-4.894031</td>\n",
       "      <td>25.678374</td>\n",
       "      <td>50.417046</td>\n",
       "      <td>60.466569</td>\n",
       "      <td>18.169207</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>766</th>\n",
       "      <td>-88.603652</td>\n",
       "      <td>19.872399</td>\n",
       "      <td>-91.702457</td>\n",
       "      <td>2.176852</td>\n",
       "      <td>24.157245</td>\n",
       "      <td>-3.398766</td>\n",
       "      <td>-23.370564</td>\n",
       "      <td>18.979715</td>\n",
       "      <td>20.380324</td>\n",
       "      <td>0.412639</td>\n",
       "      <td>14.867101</td>\n",
       "      <td>14.679876</td>\n",
       "      <td>-2.292881</td>\n",
       "      <td>4.328534</td>\n",
       "      <td>29.531122</td>\n",
       "      <td>-2.852520</td>\n",
       "      <td>11.329032</td>\n",
       "      <td>30.099205</td>\n",
       "      <td>11.842104</td>\n",
       "      <td>9.564691</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>767</th>\n",
       "      <td>-96.757244</td>\n",
       "      <td>60.310860</td>\n",
       "      <td>-83.014671</td>\n",
       "      <td>-38.856395</td>\n",
       "      <td>0.045580</td>\n",
       "      <td>-2.270765</td>\n",
       "      <td>-14.747818</td>\n",
       "      <td>32.516680</td>\n",
       "      <td>45.835674</td>\n",
       "      <td>0.410615</td>\n",
       "      <td>23.006240</td>\n",
       "      <td>-8.351489</td>\n",
       "      <td>-9.630631</td>\n",
       "      <td>-5.761053</td>\n",
       "      <td>19.915662</td>\n",
       "      <td>-2.296542</td>\n",
       "      <td>12.848735</td>\n",
       "      <td>40.759094</td>\n",
       "      <td>31.686205</td>\n",
       "      <td>-5.648737</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>768 rows × 20 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              1          2           3          4          5          6  \\\n",
       "0   -105.642589   3.523409  -92.246457 -10.842514 -14.564065  -9.463218   \n",
       "1   -132.559431   0.686573 -117.879118 -33.839001   0.368988  -5.450244   \n",
       "2   -125.820963 -29.318014 -122.626954  40.305080 -26.607432   9.570022   \n",
       "3   -109.827708  -2.130972  -73.175728 -33.102889  -4.895685   6.343304   \n",
       "4   -157.076006  -2.560414 -154.402908 -34.053480  32.223315   5.883290   \n",
       "..          ...        ...         ...        ...        ...        ...   \n",
       "763 -135.482146   4.662305 -152.819028 -50.540514  25.331998  -2.516600   \n",
       "764 -119.300415  -4.276729 -140.446478 -43.307099  42.011823 -15.461389   \n",
       "765 -172.142869 -18.304606 -136.150350 -27.622497  37.332445  -7.721931   \n",
       "766  -88.603652  19.872399  -91.702457   2.176852  24.157245  -3.398766   \n",
       "767  -96.757244  60.310860  -83.014671 -38.856395   0.045580  -2.270765   \n",
       "\n",
       "             7          8          9         10         11         12  \\\n",
       "0   -13.855369  76.950658  49.846658   7.948933  11.997859 -17.600960   \n",
       "1    13.086981  72.105231  29.220280   6.478679  16.046478   0.346984   \n",
       "2    19.152528  74.205878  36.133930   4.056305   2.475786  37.753835   \n",
       "3   -14.379102  66.353085  42.533320   1.782594  14.423091 -14.711778   \n",
       "4   -26.036329  41.949521  53.124888  10.032831  37.293703  51.091919   \n",
       "..         ...        ...        ...        ...        ...        ...   \n",
       "763 -31.505189  52.733280  62.210393  14.143995  37.520076  27.989724   \n",
       "764  -7.324946  44.725894  57.869686   8.734915  38.412462  26.337352   \n",
       "765 -39.142331  41.211884  64.103819  11.788660  37.141524  21.923082   \n",
       "766 -23.370564  18.979715  20.380324   0.412639  14.867101  14.679876   \n",
       "767 -14.747818  32.516680  45.835674   0.410615  23.006240  -8.351489   \n",
       "\n",
       "            13         14         15         16         17         18  \\\n",
       "0    15.747128  -5.359820   7.656326  14.626570  17.383610  53.427690   \n",
       "1    20.771332  -1.210382   6.475880  24.276589  31.065191  43.517274   \n",
       "2    -4.098960   1.987441  -8.440176  27.583215  11.461091  42.995740   \n",
       "3    10.534064  -4.101555  10.320962   4.699534  22.557302  71.811843   \n",
       "4    19.164965  -1.440777   8.477097  13.370941  24.086364  35.228973   \n",
       "..         ...        ...        ...        ...        ...        ...   \n",
       "763  16.184726 -10.239194   6.294204   0.992845  28.950867  61.940471   \n",
       "764 -10.345786  -4.509702   4.708844   2.614399  30.080757  32.968018   \n",
       "765  36.069846  -1.041638   2.717797  -4.894031  25.678374  50.417046   \n",
       "766  -2.292881   4.328534  29.531122  -2.852520  11.329032  30.099205   \n",
       "767  -9.630631  -5.761053  19.915662  -2.296542  12.848735  40.759094   \n",
       "\n",
       "            19         20  \n",
       "0     8.527089   1.939061  \n",
       "1    31.887459  -5.395743  \n",
       "2    16.015795  -6.784145  \n",
       "3     6.837782  -1.871465  \n",
       "4    41.661200   1.980905  \n",
       "..         ...        ...  \n",
       "763  30.870036  13.277754  \n",
       "764  52.712497   3.795897  \n",
       "765  60.466569  18.169207  \n",
       "766  11.842104   9.564691  \n",
       "767  31.686205  -5.648737  \n",
       "\n",
       "[768 rows x 20 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_new"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "62656e2e",
   "metadata": {},
   "source": [
    "# Feature selection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "fe930237",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_colnames = X_new.columns.values.tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "5dbc36e6",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(333)\n",
    "importances = mutual_info_classif(X_new,y)\n",
    "f_i = list(zip(X_colnames,importances))\n",
    "\n",
    "f_i.sort(key = lambda x : x[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "022d7e3e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('6', 0.0),\n",
       " ('8', 0.0),\n",
       " ('10', 0.0),\n",
       " ('14', 0.0),\n",
       " ('17', 0.0),\n",
       " ('18', 0.0),\n",
       " ('19', 0.0),\n",
       " ('15', 0.0032482742345829596),\n",
       " ('16', 0.01116219314105682),\n",
       " ('4', 0.01650498971353631),\n",
       " ('3', 0.017918094837578558),\n",
       " ('12', 0.02026531372528706),\n",
       " ('1', 0.02552714174102122),\n",
       " ('11', 0.0280507471677085),\n",
       " ('7', 0.028238099651494375),\n",
       " ('2', 0.0287300873375127),\n",
       " ('5', 0.03359536844224209),\n",
       " ('20', 0.04322037757881869),\n",
       " ('9', 0.048967133438106325),\n",
       " ('13', 0.050182129795516284)]"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f_i"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "195155c7",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.bar([x[0] for x in f_i[-20:]], [x[1] for x in f_i[-20:]], color = \"tab:blue\")\n",
    "plt.xlabel(\"cluster number\") \n",
    "plt.ylabel(\"importance\") \n",
    "plt.savefig(\"plots/LOSO_CLR_featadd_3.jpg\",dpi=300)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "3457c1a9",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_reduced = pd.DataFrame()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "01f1fd0a",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "for x in f_i:\n",
    "    if x[1]>0:\n",
    "        X_reduced[x[0]] = X_new.loc[:,x[0]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "c51152bb",
   "metadata": {
    "scrolled": true
   },
   "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>15</th>\n",
       "      <th>16</th>\n",
       "      <th>4</th>\n",
       "      <th>3</th>\n",
       "      <th>12</th>\n",
       "      <th>1</th>\n",
       "      <th>11</th>\n",
       "      <th>7</th>\n",
       "      <th>2</th>\n",
       "      <th>5</th>\n",
       "      <th>20</th>\n",
       "      <th>9</th>\n",
       "      <th>13</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>7.656326</td>\n",
       "      <td>14.626570</td>\n",
       "      <td>-10.842514</td>\n",
       "      <td>-92.246457</td>\n",
       "      <td>-17.600960</td>\n",
       "      <td>-105.642589</td>\n",
       "      <td>11.997859</td>\n",
       "      <td>-13.855369</td>\n",
       "      <td>3.523409</td>\n",
       "      <td>-14.564065</td>\n",
       "      <td>1.939061</td>\n",
       "      <td>49.846658</td>\n",
       "      <td>15.747128</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>6.475880</td>\n",
       "      <td>24.276589</td>\n",
       "      <td>-33.839001</td>\n",
       "      <td>-117.879118</td>\n",
       "      <td>0.346984</td>\n",
       "      <td>-132.559431</td>\n",
       "      <td>16.046478</td>\n",
       "      <td>13.086981</td>\n",
       "      <td>0.686573</td>\n",
       "      <td>0.368988</td>\n",
       "      <td>-5.395743</td>\n",
       "      <td>29.220280</td>\n",
       "      <td>20.771332</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-8.440176</td>\n",
       "      <td>27.583215</td>\n",
       "      <td>40.305080</td>\n",
       "      <td>-122.626954</td>\n",
       "      <td>37.753835</td>\n",
       "      <td>-125.820963</td>\n",
       "      <td>2.475786</td>\n",
       "      <td>19.152528</td>\n",
       "      <td>-29.318014</td>\n",
       "      <td>-26.607432</td>\n",
       "      <td>-6.784145</td>\n",
       "      <td>36.133930</td>\n",
       "      <td>-4.098960</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10.320962</td>\n",
       "      <td>4.699534</td>\n",
       "      <td>-33.102889</td>\n",
       "      <td>-73.175728</td>\n",
       "      <td>-14.711778</td>\n",
       "      <td>-109.827708</td>\n",
       "      <td>14.423091</td>\n",
       "      <td>-14.379102</td>\n",
       "      <td>-2.130972</td>\n",
       "      <td>-4.895685</td>\n",
       "      <td>-1.871465</td>\n",
       "      <td>42.533320</td>\n",
       "      <td>10.534064</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>8.477097</td>\n",
       "      <td>13.370941</td>\n",
       "      <td>-34.053480</td>\n",
       "      <td>-154.402908</td>\n",
       "      <td>51.091919</td>\n",
       "      <td>-157.076006</td>\n",
       "      <td>37.293703</td>\n",
       "      <td>-26.036329</td>\n",
       "      <td>-2.560414</td>\n",
       "      <td>32.223315</td>\n",
       "      <td>1.980905</td>\n",
       "      <td>53.124888</td>\n",
       "      <td>19.164965</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>763</th>\n",
       "      <td>6.294204</td>\n",
       "      <td>0.992845</td>\n",
       "      <td>-50.540514</td>\n",
       "      <td>-152.819028</td>\n",
       "      <td>27.989724</td>\n",
       "      <td>-135.482146</td>\n",
       "      <td>37.520076</td>\n",
       "      <td>-31.505189</td>\n",
       "      <td>4.662305</td>\n",
       "      <td>25.331998</td>\n",
       "      <td>13.277754</td>\n",
       "      <td>62.210393</td>\n",
       "      <td>16.184726</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>764</th>\n",
       "      <td>4.708844</td>\n",
       "      <td>2.614399</td>\n",
       "      <td>-43.307099</td>\n",
       "      <td>-140.446478</td>\n",
       "      <td>26.337352</td>\n",
       "      <td>-119.300415</td>\n",
       "      <td>38.412462</td>\n",
       "      <td>-7.324946</td>\n",
       "      <td>-4.276729</td>\n",
       "      <td>42.011823</td>\n",
       "      <td>3.795897</td>\n",
       "      <td>57.869686</td>\n",
       "      <td>-10.345786</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>765</th>\n",
       "      <td>2.717797</td>\n",
       "      <td>-4.894031</td>\n",
       "      <td>-27.622497</td>\n",
       "      <td>-136.150350</td>\n",
       "      <td>21.923082</td>\n",
       "      <td>-172.142869</td>\n",
       "      <td>37.141524</td>\n",
       "      <td>-39.142331</td>\n",
       "      <td>-18.304606</td>\n",
       "      <td>37.332445</td>\n",
       "      <td>18.169207</td>\n",
       "      <td>64.103819</td>\n",
       "      <td>36.069846</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>766</th>\n",
       "      <td>29.531122</td>\n",
       "      <td>-2.852520</td>\n",
       "      <td>2.176852</td>\n",
       "      <td>-91.702457</td>\n",
       "      <td>14.679876</td>\n",
       "      <td>-88.603652</td>\n",
       "      <td>14.867101</td>\n",
       "      <td>-23.370564</td>\n",
       "      <td>19.872399</td>\n",
       "      <td>24.157245</td>\n",
       "      <td>9.564691</td>\n",
       "      <td>20.380324</td>\n",
       "      <td>-2.292881</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>767</th>\n",
       "      <td>19.915662</td>\n",
       "      <td>-2.296542</td>\n",
       "      <td>-38.856395</td>\n",
       "      <td>-83.014671</td>\n",
       "      <td>-8.351489</td>\n",
       "      <td>-96.757244</td>\n",
       "      <td>23.006240</td>\n",
       "      <td>-14.747818</td>\n",
       "      <td>60.310860</td>\n",
       "      <td>0.045580</td>\n",
       "      <td>-5.648737</td>\n",
       "      <td>45.835674</td>\n",
       "      <td>-9.630631</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>768 rows × 13 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            15         16          4           3         12           1  \\\n",
       "0     7.656326  14.626570 -10.842514  -92.246457 -17.600960 -105.642589   \n",
       "1     6.475880  24.276589 -33.839001 -117.879118   0.346984 -132.559431   \n",
       "2    -8.440176  27.583215  40.305080 -122.626954  37.753835 -125.820963   \n",
       "3    10.320962   4.699534 -33.102889  -73.175728 -14.711778 -109.827708   \n",
       "4     8.477097  13.370941 -34.053480 -154.402908  51.091919 -157.076006   \n",
       "..         ...        ...        ...         ...        ...         ...   \n",
       "763   6.294204   0.992845 -50.540514 -152.819028  27.989724 -135.482146   \n",
       "764   4.708844   2.614399 -43.307099 -140.446478  26.337352 -119.300415   \n",
       "765   2.717797  -4.894031 -27.622497 -136.150350  21.923082 -172.142869   \n",
       "766  29.531122  -2.852520   2.176852  -91.702457  14.679876  -88.603652   \n",
       "767  19.915662  -2.296542 -38.856395  -83.014671  -8.351489  -96.757244   \n",
       "\n",
       "            11          7          2          5         20          9  \\\n",
       "0    11.997859 -13.855369   3.523409 -14.564065   1.939061  49.846658   \n",
       "1    16.046478  13.086981   0.686573   0.368988  -5.395743  29.220280   \n",
       "2     2.475786  19.152528 -29.318014 -26.607432  -6.784145  36.133930   \n",
       "3    14.423091 -14.379102  -2.130972  -4.895685  -1.871465  42.533320   \n",
       "4    37.293703 -26.036329  -2.560414  32.223315   1.980905  53.124888   \n",
       "..         ...        ...        ...        ...        ...        ...   \n",
       "763  37.520076 -31.505189   4.662305  25.331998  13.277754  62.210393   \n",
       "764  38.412462  -7.324946  -4.276729  42.011823   3.795897  57.869686   \n",
       "765  37.141524 -39.142331 -18.304606  37.332445  18.169207  64.103819   \n",
       "766  14.867101 -23.370564  19.872399  24.157245   9.564691  20.380324   \n",
       "767  23.006240 -14.747818  60.310860   0.045580  -5.648737  45.835674   \n",
       "\n",
       "            13  \n",
       "0    15.747128  \n",
       "1    20.771332  \n",
       "2    -4.098960  \n",
       "3    10.534064  \n",
       "4    19.164965  \n",
       "..         ...  \n",
       "763  16.184726  \n",
       "764 -10.345786  \n",
       "765  36.069846  \n",
       "766  -2.292881  \n",
       "767  -9.630631  \n",
       "\n",
       "[768 rows x 13 columns]"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_reduced"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "874330dc",
   "metadata": {},
   "source": [
    "## Dataset names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "3d31b47b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['AT-CRC_species',\n",
       " 'CN-CRC_species',\n",
       " 'DE-CRC_species',\n",
       " 'FR-CRC_species',\n",
       " 'IT-CRC-2_species',\n",
       " 'IT-CRC_species',\n",
       " 'JP-CRC_species',\n",
       " 'US-CRC_species']"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "unique_groups = list(np.unique(groups))\n",
    "unique_groups"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a331e388",
   "metadata": {},
   "source": [
    "# Creating dataframes for results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "be80a857",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "AUC_scores = pd.DataFrame(index=unique_groups, columns=unique_groups)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "21e98d1d",
   "metadata": {},
   "outputs": [],
   "source": [
    "ROC_curve = pd.DataFrame(index=unique_groups, columns=[\"fpr\",\"tpr\",\"auc\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b6ec21a3",
   "metadata": {},
   "source": [
    "# LOSO validation with random forests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "82e26546",
   "metadata": {},
   "outputs": [],
   "source": [
    "SEED = 123\n",
    "show_fold_stats = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "44d0a12c",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------------------------------------------------------------------------\n",
      "Fold : 1 CN-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7843\n",
      "F1 : 0.6733, Recall : 0.6296 , Precision : 0.7234\n",
      "[[61 13]\n",
      " [20 34]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 2 DE-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6940\n",
      "F1 : 0.7338, Recall : 0.8500 , Precision : 0.6456\n",
      "[[32 28]\n",
      " [ 9 51]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 3 FR-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7932\n",
      "F1 : 0.6847, Recall : 0.6230 , Precision : 0.7600\n",
      "[[41 12]\n",
      " [23 38]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 4 IT-CRC-2_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6825\n",
      "F1 : 0.6765, Recall : 0.8214 , Precision : 0.5750\n",
      "[[15 17]\n",
      " [ 5 23]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 5 IT-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6853\n",
      "F1 : 0.4500, Recall : 0.3750 , Precision : 0.5625\n",
      "[[22  7]\n",
      " [15  9]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 6 JP-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6419\n",
      "F1 : 0.4262, Recall : 0.3250 , Precision : 0.6190\n",
      "[[32  8]\n",
      " [27 13]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 7 US-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6585\n",
      "F1 : 0.5684, Recall : 0.5192 , Precision : 0.6279\n",
      "[[36 16]\n",
      " [25 27]]\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "Group : 1 AT-CRC_species\n",
      "ROC AUC score for RandomForest model, LOGO test set: 0.7876\n",
      "F1 : 0.7846, Recall : 0.8095 , Precision : 0.7612\n",
      "[[30 16]\n",
      " [12 51]]\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 1 AT-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7778\n",
      "F1 : 0.7846, Recall : 0.8095 , Precision : 0.7612\n",
      "[[30 16]\n",
      " [12 51]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 2 DE-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6757\n",
      "F1 : 0.6617, Recall : 0.7333 , Precision : 0.6027\n",
      "[[31 29]\n",
      " [16 44]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 3 FR-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.8017\n",
      "F1 : 0.7207, Recall : 0.6557 , Precision : 0.8000\n",
      "[[43 10]\n",
      " [21 40]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 4 IT-CRC-2_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6713\n",
      "F1 : 0.6667, Recall : 0.8214 , Precision : 0.5610\n",
      "[[14 18]\n",
      " [ 5 23]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 5 IT-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6307\n",
      "F1 : 0.3784, Recall : 0.2917 , Precision : 0.5385\n",
      "[[23  6]\n",
      " [17  7]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 6 JP-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6863\n",
      "F1 : 0.5588, Recall : 0.4750 , Precision : 0.6786\n",
      "[[31  9]\n",
      " [21 19]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 7 US-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6553\n",
      "F1 : 0.5918, Recall : 0.5577 , Precision : 0.6304\n",
      "[[35 17]\n",
      " [23 29]]\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "Group : 2 CN-CRC_species\n",
      "ROC AUC score for RandomForest model, LOGO test set: 0.7760\n",
      "F1 : 0.6346, Recall : 0.6111 , Precision : 0.6600\n",
      "[[57 17]\n",
      " [21 33]]\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 1 AT-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.8171\n",
      "F1 : 0.7820, Recall : 0.8254 , Precision : 0.7429\n",
      "[[28 18]\n",
      " [11 52]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 2 CN-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7895\n",
      "F1 : 0.6667, Recall : 0.7037 , Precision : 0.6333\n",
      "[[52 22]\n",
      " [16 38]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 3 FR-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7744\n",
      "F1 : 0.7458, Recall : 0.7213 , Precision : 0.7719\n",
      "[[40 13]\n",
      " [17 44]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 4 IT-CRC-2_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7081\n",
      "F1 : 0.6866, Recall : 0.8214 , Precision : 0.5897\n",
      "[[16 16]\n",
      " [ 5 23]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 5 IT-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6925\n",
      "F1 : 0.5714, Recall : 0.5000 , Precision : 0.6667\n",
      "[[23  6]\n",
      " [12 12]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 6 JP-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6878\n",
      "F1 : 0.5161, Recall : 0.4000 , Precision : 0.7273\n",
      "[[34  6]\n",
      " [24 16]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 7 US-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6842\n",
      "F1 : 0.6667, Recall : 0.6346 , Precision : 0.7021\n",
      "[[38 14]\n",
      " [19 33]]\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "Group : 3 DE-CRC_species\n",
      "ROC AUC score for RandomForest model, LOGO test set: 0.6992\n",
      "F1 : 0.7092, Recall : 0.8333 , Precision : 0.6173\n",
      "[[29 31]\n",
      " [10 50]]\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 1 AT-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7640\n",
      "F1 : 0.7059, Recall : 0.6667 , Precision : 0.7500\n",
      "[[32 14]\n",
      " [21 42]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 2 CN-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7671\n",
      "F1 : 0.6296, Recall : 0.6296 , Precision : 0.6296\n",
      "[[54 20]\n",
      " [20 34]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 3 DE-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6637\n",
      "F1 : 0.7111, Recall : 0.8000 , Precision : 0.6400\n",
      "[[33 27]\n",
      " [12 48]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 4 IT-CRC-2_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6663\n",
      "F1 : 0.6567, Recall : 0.7857 , Precision : 0.5641\n",
      "[[15 17]\n",
      " [ 6 22]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 5 IT-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6386\n",
      "F1 : 0.4103, Recall : 0.3333 , Precision : 0.5333\n",
      "[[22  7]\n",
      " [16  8]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 6 JP-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6872\n",
      "F1 : 0.4918, Recall : 0.3750 , Precision : 0.7143\n",
      "[[34  6]\n",
      " [25 15]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 7 US-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6304\n",
      "F1 : 0.5217, Recall : 0.4615 , Precision : 0.6000\n",
      "[[36 16]\n",
      " [28 24]]\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "Group : 4 FR-CRC_species\n",
      "ROC AUC score for RandomForest model, LOGO test set: 0.7974\n",
      "F1 : 0.7193, Recall : 0.6721 , Precision : 0.7736\n",
      "[[41 12]\n",
      " [20 41]]\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------------------------------------------------------------------------\n",
      "Fold : 1 AT-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7771\n",
      "F1 : 0.7761, Recall : 0.8254 , Precision : 0.7324\n",
      "[[27 19]\n",
      " [11 52]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 2 CN-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7829\n",
      "F1 : 0.6346, Recall : 0.6111 , Precision : 0.6600\n",
      "[[57 17]\n",
      " [21 33]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 3 DE-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7154\n",
      "F1 : 0.7194, Recall : 0.8333 , Precision : 0.6329\n",
      "[[31 29]\n",
      " [10 50]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 4 FR-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.8037\n",
      "F1 : 0.7731, Recall : 0.7541 , Precision : 0.7931\n",
      "[[41 12]\n",
      " [15 46]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 5 IT-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6832\n",
      "F1 : 0.4737, Recall : 0.3750 , Precision : 0.6429\n",
      "[[24  5]\n",
      " [15  9]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 6 JP-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6650\n",
      "F1 : 0.5000, Recall : 0.3750 , Precision : 0.7500\n",
      "[[35  5]\n",
      " [25 15]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 7 US-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6514\n",
      "F1 : 0.6061, Recall : 0.5769 , Precision : 0.6383\n",
      "[[35 17]\n",
      " [22 30]]\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "Group : 5 IT-CRC-2_species\n",
      "ROC AUC score for RandomForest model, LOGO test set: 0.6696\n",
      "F1 : 0.6471, Recall : 0.7857 , Precision : 0.5500\n",
      "[[14 18]\n",
      " [ 6 22]]\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 1 AT-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7942\n",
      "F1 : 0.7939, Recall : 0.8254 , Precision : 0.7647\n",
      "[[30 16]\n",
      " [11 52]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 2 CN-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7884\n",
      "F1 : 0.6355, Recall : 0.6296 , Precision : 0.6415\n",
      "[[55 19]\n",
      " [20 34]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 3 DE-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7163\n",
      "F1 : 0.7299, Recall : 0.8333 , Precision : 0.6494\n",
      "[[33 27]\n",
      " [10 50]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 4 FR-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7935\n",
      "F1 : 0.6786, Recall : 0.6230 , Precision : 0.7451\n",
      "[[40 13]\n",
      " [23 38]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 5 IT-CRC-2_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6512\n",
      "F1 : 0.6667, Recall : 0.7857 , Precision : 0.5789\n",
      "[[16 16]\n",
      " [ 6 22]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 6 JP-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6650\n",
      "F1 : 0.4839, Recall : 0.3750 , Precision : 0.6818\n",
      "[[33  7]\n",
      " [25 15]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 7 US-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6633\n",
      "F1 : 0.5859, Recall : 0.5577 , Precision : 0.6170\n",
      "[[34 18]\n",
      " [23 29]]\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "Group : 6 IT-CRC_species\n",
      "ROC AUC score for RandomForest model, LOGO test set: 0.6760\n",
      "F1 : 0.4615, Recall : 0.3750 , Precision : 0.6000\n",
      "[[23  6]\n",
      " [15  9]]\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 1 AT-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7462\n",
      "F1 : 0.7143, Recall : 0.7143 , Precision : 0.7143\n",
      "[[28 18]\n",
      " [18 45]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 2 CN-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7779\n",
      "F1 : 0.6226, Recall : 0.6111 , Precision : 0.6346\n",
      "[[55 19]\n",
      " [21 33]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 3 DE-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7138\n",
      "F1 : 0.7246, Recall : 0.8333 , Precision : 0.6410\n",
      "[[32 28]\n",
      " [10 50]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 4 FR-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7891\n",
      "F1 : 0.7414, Recall : 0.7049 , Precision : 0.7818\n",
      "[[41 12]\n",
      " [18 43]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 5 IT-CRC-2_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6797\n",
      "F1 : 0.6667, Recall : 0.8214 , Precision : 0.5610\n",
      "[[14 18]\n",
      " [ 5 23]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 6 IT-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6616\n",
      "F1 : 0.4324, Recall : 0.3333 , Precision : 0.6154\n",
      "[[24  5]\n",
      " [16  8]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 7 US-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6688\n",
      "F1 : 0.5714, Recall : 0.5385 , Precision : 0.6087\n",
      "[[34 18]\n",
      " [24 28]]\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "Group : 7 JP-CRC_species\n",
      "ROC AUC score for RandomForest model, LOGO test set: 0.6738\n",
      "F1 : 0.5397, Recall : 0.4250 , Precision : 0.7391\n",
      "[[34  6]\n",
      " [23 17]]\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 1 AT-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.8042\n",
      "F1 : 0.7939, Recall : 0.8254 , Precision : 0.7647\n",
      "[[30 16]\n",
      " [11 52]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 2 CN-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7643\n",
      "F1 : 0.5941, Recall : 0.5556 , Precision : 0.6383\n",
      "[[57 17]\n",
      " [24 30]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 3 DE-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6958\n",
      "F1 : 0.6861, Recall : 0.7833 , Precision : 0.6104\n",
      "[[30 30]\n",
      " [13 47]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 4 FR-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.7929\n",
      "F1 : 0.6792, Recall : 0.5902 , Precision : 0.8000\n",
      "[[44  9]\n",
      " [25 36]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 5 IT-CRC-2_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6735\n",
      "F1 : 0.6567, Recall : 0.7857 , Precision : 0.5641\n",
      "[[15 17]\n",
      " [ 6 22]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 6 IT-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6487\n",
      "F1 : 0.4324, Recall : 0.3333 , Precision : 0.6154\n",
      "[[24  5]\n",
      " [16  8]]\n",
      "--------------------------------------------------------------------------------\n",
      "Fold : 7 JP-CRC_species\n",
      "ROC AUC score for RandomForest model, validation set: 0.6941\n",
      "F1 : 0.5882, Recall : 0.5000 , Precision : 0.7143\n",
      "[[32  8]\n",
      " [20 20]]\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n",
      "Group : 8 US-CRC_species\n",
      "ROC AUC score for RandomForest model, LOGO test set: 0.6526\n",
      "F1 : 0.6275, Recall : 0.6154 , Precision : 0.6400\n",
      "[[34 18]\n",
      " [20 32]]\n",
      "--------------------------------------------------------------------------------\n",
      "--------------------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "grp = 0\n",
    "for i in range(len(unique_groups)):\n",
    "    outer_test_index = list(np.where(df[\"Country\"]==unique_groups[i])[0])\n",
    "    \n",
    "    inner_group_list = []\n",
    "    inner_group_list.append(unique_groups[:i] + unique_groups[i+1:])\n",
    "    \n",
    "    outer_train_index = list(np.where(df[\"Country\"]!=unique_groups[i])[0])\n",
    "    \n",
    "    #Training dataset\n",
    "    outer_train_X = X_reduced.iloc[outer_train_index]\n",
    "    outer_train_y = y[outer_train_index]\n",
    "    \n",
    "    #Test dataset\n",
    "    outer_test_X =  X_reduced.iloc[outer_test_index]\n",
    "    outer_test_y =  y[outer_test_index]\n",
    "    \n",
    "    for k,name in enumerate(inner_group_list[0]):\n",
    "        inner_val_index = list(np.where(df[\"Country\"]==name)[0])\n",
    "        \n",
    "        inner_train_names = []\n",
    "        inner_train_names.append(inner_group_list[:k] + inner_group_list[k+1:])\n",
    "        \n",
    "        inner_train_index = list(np.where((df[\"Country\"]!=name) & (df[\"Country\"]!=unique_groups[i]))[0])\n",
    "\n",
    "        #Inner training dataset\n",
    "        inner_train_X = X_reduced.iloc[inner_train_index]\n",
    "        inner_train_y = y[inner_train_index]\n",
    "        \n",
    "        #Inner validation dataset\n",
    "        inner_val_X =  X_reduced.iloc[inner_val_index]\n",
    "        inner_val_y =  y[inner_val_index]\n",
    "        \n",
    "        \n",
    "        \n",
    "        # Random Forest\n",
    "        clf = RandomForestClassifier(n_estimators = 500, random_state = SEED, verbose = 0, bootstrap=True)\n",
    "        clf.fit(inner_train_X, inner_train_y)\n",
    "\n",
    "        RF_pred_class = clf.predict(inner_val_X)\n",
    "        RF_preds = clf.predict_proba(inner_val_X)\n",
    "\n",
    "        RF_AUC_test_score = roc_auc_score(inner_val_y, RF_preds[:,1])\n",
    "        RF_f1_test = f1_score(inner_val_y, RF_pred_class)\n",
    "        RF_recall_test = recall_score(inner_val_y, RF_pred_class)\n",
    "        RF_precision_test = precision_score(inner_val_y, RF_pred_class)\n",
    "        \n",
    "        #Specificity and sensitivity\n",
    "        tn, fp, fn, tp = confusion_matrix(inner_val_y, RF_pred_class).ravel()\n",
    "        specificity = tn / (tn + fp)\n",
    "        sensitivity = tp / (tp + fn)\n",
    "        \n",
    "        #Adding auc to dataframe\n",
    "        AUC_scores.at[unique_groups[i], name] = RF_AUC_test_score\n",
    "\n",
    "        if show_fold_stats:\n",
    "            print('-' * 80)\n",
    "            print('Fold : %s'%(k+1),name)\n",
    "            print('ROC AUC score for RandomForest model, validation set: %.4f'%RF_AUC_test_score)\n",
    "            print('F1 : %.4f, Recall : %.4f , Precision : %.4f'%(RF_f1_test, RF_recall_test, RF_precision_test))\n",
    "            print(confusion_matrix(inner_val_y, RF_pred_class))\n",
    "                \n",
    "    # Random Forest on logo_test data\n",
    "    clf = RandomForestClassifier(n_estimators = 500, random_state = SEED, verbose = 0, bootstrap=True)\n",
    "    clf.fit(outer_train_X, outer_train_y)\n",
    "        \n",
    "    RF_pred_class = clf.predict(outer_test_X)\n",
    "    RF_preds = clf.predict_proba(outer_test_X)\n",
    "\n",
    "    RF_AUC_test_score = roc_auc_score(outer_test_y, RF_preds[:,1])\n",
    "    RF_f1_test = f1_score(outer_test_y, RF_pred_class)\n",
    "    RF_recall_test = recall_score(outer_test_y, RF_pred_class)\n",
    "    RF_precision_test = precision_score(outer_test_y, RF_pred_class)\n",
    "    \n",
    "    #Specificity and sensitivity\n",
    "    tn, fp, fn, tp = confusion_matrix(outer_test_y, RF_pred_class).ravel()\n",
    "    specificity = tn / (tn + fp)\n",
    "    sensitivity = tp / (tp + fn)\n",
    "    \n",
    "    #Adding auc to dataframe\n",
    "    AUC_scores.at[unique_groups[i], unique_groups[i]] = RF_AUC_test_score\n",
    "    \n",
    "    #Getting AUCROC data\n",
    "    fpr, tpr, _ = roc_curve(outer_test_y,  RF_preds[:,1])\n",
    "\n",
    "    #Adding roc curve values to dataframe\n",
    "    ROC_curve.at[unique_groups[i], \"fpr\"] = \",\".join(str(x) for x in fpr)\n",
    "    ROC_curve.at[unique_groups[i], \"tpr\"] = \",\".join(str(x) for x in tpr)\n",
    "    ROC_curve.at[unique_groups[i], \"auc\"] = RF_AUC_test_score\n",
    "    \n",
    "\n",
    "    if show_fold_stats:\n",
    "        print('-' * 80)\n",
    "        print('-' * 80)\n",
    "        print('Group : %s'%(grp+1),unique_groups[i])\n",
    "        print('ROC AUC score for RandomForest model, LOGO test set: %.4f'%RF_AUC_test_score)\n",
    "        print('F1 : %.4f, Recall : %.4f , Precision : %.4f'%(RF_f1_test, RF_recall_test, RF_precision_test))\n",
    "        print(confusion_matrix(outer_test_y, RF_pred_class))\n",
    "        print('-' * 80)\n",
    "        print('-' * 80)\n",
    "    grp += 1\n",
    "        "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "48727aee",
   "metadata": {},
   "source": [
    "# Results and plots"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "401268e7",
   "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>AT-CRC_species</th>\n",
       "      <th>CN-CRC_species</th>\n",
       "      <th>DE-CRC_species</th>\n",
       "      <th>FR-CRC_species</th>\n",
       "      <th>IT-CRC-2_species</th>\n",
       "      <th>IT-CRC_species</th>\n",
       "      <th>JP-CRC_species</th>\n",
       "      <th>US-CRC_species</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>AT-CRC_species</th>\n",
       "      <td>0.787612</td>\n",
       "      <td>0.784284</td>\n",
       "      <td>0.694028</td>\n",
       "      <td>0.793226</td>\n",
       "      <td>0.682478</td>\n",
       "      <td>0.685345</td>\n",
       "      <td>0.641875</td>\n",
       "      <td>0.658469</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CN-CRC_species</th>\n",
       "      <td>0.777778</td>\n",
       "      <td>0.776026</td>\n",
       "      <td>0.675694</td>\n",
       "      <td>0.801732</td>\n",
       "      <td>0.671317</td>\n",
       "      <td>0.630747</td>\n",
       "      <td>0.68625</td>\n",
       "      <td>0.655325</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DE-CRC_species</th>\n",
       "      <td>0.817115</td>\n",
       "      <td>0.78954</td>\n",
       "      <td>0.699167</td>\n",
       "      <td>0.774358</td>\n",
       "      <td>0.708147</td>\n",
       "      <td>0.692529</td>\n",
       "      <td>0.687813</td>\n",
       "      <td>0.684172</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FR-CRC_species</th>\n",
       "      <td>0.763975</td>\n",
       "      <td>0.767142</td>\n",
       "      <td>0.66375</td>\n",
       "      <td>0.797402</td>\n",
       "      <td>0.666295</td>\n",
       "      <td>0.638649</td>\n",
       "      <td>0.687188</td>\n",
       "      <td>0.630362</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IT-CRC-2_species</th>\n",
       "      <td>0.777088</td>\n",
       "      <td>0.782908</td>\n",
       "      <td>0.715417</td>\n",
       "      <td>0.803743</td>\n",
       "      <td>0.669643</td>\n",
       "      <td>0.68319</td>\n",
       "      <td>0.665</td>\n",
       "      <td>0.651442</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IT-CRC_species</th>\n",
       "      <td>0.794168</td>\n",
       "      <td>0.788413</td>\n",
       "      <td>0.71625</td>\n",
       "      <td>0.793535</td>\n",
       "      <td>0.651228</td>\n",
       "      <td>0.676006</td>\n",
       "      <td>0.665</td>\n",
       "      <td>0.663277</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>JP-CRC_species</th>\n",
       "      <td>0.746204</td>\n",
       "      <td>0.777903</td>\n",
       "      <td>0.71375</td>\n",
       "      <td>0.78905</td>\n",
       "      <td>0.679688</td>\n",
       "      <td>0.661638</td>\n",
       "      <td>0.67375</td>\n",
       "      <td>0.668824</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>US-CRC_species</th>\n",
       "      <td>0.804175</td>\n",
       "      <td>0.764264</td>\n",
       "      <td>0.695833</td>\n",
       "      <td>0.792917</td>\n",
       "      <td>0.673549</td>\n",
       "      <td>0.648707</td>\n",
       "      <td>0.694063</td>\n",
       "      <td>0.652552</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 AT-CRC_species CN-CRC_species DE-CRC_species FR-CRC_species  \\\n",
       "AT-CRC_species         0.787612       0.784284       0.694028       0.793226   \n",
       "CN-CRC_species         0.777778       0.776026       0.675694       0.801732   \n",
       "DE-CRC_species         0.817115        0.78954       0.699167       0.774358   \n",
       "FR-CRC_species         0.763975       0.767142        0.66375       0.797402   \n",
       "IT-CRC-2_species       0.777088       0.782908       0.715417       0.803743   \n",
       "IT-CRC_species         0.794168       0.788413        0.71625       0.793535   \n",
       "JP-CRC_species         0.746204       0.777903        0.71375        0.78905   \n",
       "US-CRC_species         0.804175       0.764264       0.695833       0.792917   \n",
       "\n",
       "                 IT-CRC-2_species IT-CRC_species JP-CRC_species US-CRC_species  \n",
       "AT-CRC_species           0.682478       0.685345       0.641875       0.658469  \n",
       "CN-CRC_species           0.671317       0.630747        0.68625       0.655325  \n",
       "DE-CRC_species           0.708147       0.692529       0.687813       0.684172  \n",
       "FR-CRC_species           0.666295       0.638649       0.687188       0.630362  \n",
       "IT-CRC-2_species         0.669643        0.68319          0.665       0.651442  \n",
       "IT-CRC_species           0.651228       0.676006          0.665       0.663277  \n",
       "JP-CRC_species           0.679688       0.661638        0.67375       0.668824  \n",
       "US-CRC_species           0.673549       0.648707       0.694063       0.652552  "
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "AUC_scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "b6feff72",
   "metadata": {},
   "outputs": [],
   "source": [
    "AUC_scores = AUC_scores.apply(pd.to_numeric, errors='coerce')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "d04d8142",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,10))\n",
    "sns.set_palette(\"husl\", 8)\n",
    "ax = sns.heatmap(AUC_scores, annot = True, cmap=\"crest\", vmin=0.6, vmax=1.0, annot_kws={\"fontsize\":14})\n",
    "ax.set(xlabel=\"Validation Datasets\", ylabel=\"Test Datasets\")\n",
    "ax.tick_params(axis='x', rotation=45)\n",
    "plt.savefig(\"plots/LOSO_CLR_featadd_1.jpg\",dpi=300)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "30cc75f1",
   "metadata": {},
   "source": [
    "# ROC curves"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "e9ecc1a3",
   "metadata": {
    "scrolled": true
   },
   "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>fpr</th>\n",
       "      <th>tpr</th>\n",
       "      <th>auc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>AT-CRC_species</th>\n",
       "      <td>0.0,0.0,0.0,0.0,0.0,0.021739130434782608,0.021...</td>\n",
       "      <td>0.0,0.015873015873015872,0.031746031746031744,...</td>\n",
       "      <td>0.787612</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CN-CRC_species</th>\n",
       "      <td>0.0,0.0,0.0,0.02702702702702703,0.027027027027...</td>\n",
       "      <td>0.0,0.018518518518518517,0.24074074074074073,0...</td>\n",
       "      <td>0.776026</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DE-CRC_species</th>\n",
       "      <td>0.0,0.016666666666666666,0.03333333333333333,0...</td>\n",
       "      <td>0.0,0.0,0.0,0.016666666666666666,0.01666666666...</td>\n",
       "      <td>0.699167</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FR-CRC_species</th>\n",
       "      <td>0.0,0.018867924528301886,0.018867924528301886,...</td>\n",
       "      <td>0.0,0.01639344262295082,0.11475409836065574,0....</td>\n",
       "      <td>0.797402</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IT-CRC-2_species</th>\n",
       "      <td>0.0,0.0,0.0625,0.125,0.125,0.15625,0.15625,0.1...</td>\n",
       "      <td>0.0,0.03571428571428571,0.03571428571428571,0....</td>\n",
       "      <td>0.669643</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>IT-CRC_species</th>\n",
       "      <td>0.0,0.034482758620689655,0.034482758620689655,...</td>\n",
       "      <td>0.0,0.0,0.041666666666666664,0.041666666666666...</td>\n",
       "      <td>0.676006</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>JP-CRC_species</th>\n",
       "      <td>0.0,0.025,0.025,0.05,0.1,0.1,0.125,0.125,0.125...</td>\n",
       "      <td>0.0,0.0,0.025,0.025,0.025,0.05,0.05,0.225,0.27...</td>\n",
       "      <td>0.67375</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>US-CRC_species</th>\n",
       "      <td>0.0,0.0,0.0,0.019230769230769232,0.01923076923...</td>\n",
       "      <td>0.0,0.019230769230769232,0.038461538461538464,...</td>\n",
       "      <td>0.652552</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                                fpr  \\\n",
       "AT-CRC_species    0.0,0.0,0.0,0.0,0.0,0.021739130434782608,0.021...   \n",
       "CN-CRC_species    0.0,0.0,0.0,0.02702702702702703,0.027027027027...   \n",
       "DE-CRC_species    0.0,0.016666666666666666,0.03333333333333333,0...   \n",
       "FR-CRC_species    0.0,0.018867924528301886,0.018867924528301886,...   \n",
       "IT-CRC-2_species  0.0,0.0,0.0625,0.125,0.125,0.15625,0.15625,0.1...   \n",
       "IT-CRC_species    0.0,0.034482758620689655,0.034482758620689655,...   \n",
       "JP-CRC_species    0.0,0.025,0.025,0.05,0.1,0.1,0.125,0.125,0.125...   \n",
       "US-CRC_species    0.0,0.0,0.0,0.019230769230769232,0.01923076923...   \n",
       "\n",
       "                                                                tpr       auc  \n",
       "AT-CRC_species    0.0,0.015873015873015872,0.031746031746031744,...  0.787612  \n",
       "CN-CRC_species    0.0,0.018518518518518517,0.24074074074074073,0...  0.776026  \n",
       "DE-CRC_species    0.0,0.0,0.0,0.016666666666666666,0.01666666666...  0.699167  \n",
       "FR-CRC_species    0.0,0.01639344262295082,0.11475409836065574,0....  0.797402  \n",
       "IT-CRC-2_species  0.0,0.03571428571428571,0.03571428571428571,0....  0.669643  \n",
       "IT-CRC_species    0.0,0.0,0.041666666666666664,0.041666666666666...  0.676006  \n",
       "JP-CRC_species    0.0,0.0,0.025,0.025,0.025,0.05,0.05,0.225,0.27...   0.67375  \n",
       "US-CRC_species    0.0,0.019230769230769232,0.038461538461538464,...  0.652552  "
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ROC_curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "6e4207df",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(10,10))\n",
    "\n",
    "for i in ROC_curve.index:\n",
    "    plt.plot([float(s) for s in ROC_curve.loc[i]['fpr'].split(\",\")], \n",
    "             [float(s) for s in ROC_curve.loc[i]['tpr'].split(\",\")], \n",
    "             label=\"{}, AUC={:.2f}\".format(i, ROC_curve.loc[i]['auc']))\n",
    "\n",
    "plt.plot([0,1], [0,1], color='orange', linestyle='--')\n",
    "\n",
    "plt.xticks(np.arange(0.0, 1.1, step=0.1))\n",
    "plt.xlabel(\"Flase Positive Rate\")#, fontsize=15)\n",
    "\n",
    "plt.yticks(np.arange(0.0, 1.1, step=0.1))\n",
    "plt.ylabel(\"True Positive Rate\")#, fontsize=15)\n",
    "\n",
    "plt.legend(prop={'size':10}, loc='lower right')\n",
    "\n",
    "plt.savefig(\"plots/LOSO_CLR_featadd_2.jpg\",dpi=300)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9c3c6d71",
   "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": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}