prep_main_df.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Import libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Find the number of sequences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"\\nimport pathlib\\nvir_path = pathlib.Path('./sequences/edwards/virus/')\\nhost_path = pathlib.Path('./sequences/edwards/host/')\\nvir_count = 0\\nhost_count = 0\\nfor vir_file in vir_path.iterdir():\\n    vir_count += 1\\nfor host_file in host_path.iterdir():\\n    host_count += 1\\n\\nprint(f'virus files: {vir_count}')\\nprint(f'host files: {host_count}')\\nprint(f'paired files: {vir_count * host_count}')\\n\""
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "'''\n",
    "import pathlib\n",
    "vir_path = pathlib.Path('./sequences/edwards/virus/')\n",
    "host_path = pathlib.Path('./sequences/edwards/host/')\n",
    "vir_count = 0\n",
    "host_count = 0\n",
    "for vir_file in vir_path.iterdir():\n",
    "    vir_count += 1\n",
    "for host_file in host_path.iterdir():\n",
    "    host_count += 1\n",
    "\n",
    "print(f'virus files: {vir_count}')\n",
    "print(f'host files: {host_count}')\n",
    "print(f'paired files: {vir_count * host_count}')\n",
    "'''"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Importing the dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Merge the datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2213180, 11)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pathlib\n",
    "from functools import reduce\n",
    "\n",
    "features_path = pathlib.Path('./features/')\n",
    "dfs = [pd.read_csv(feat_file, sep='\\t', header=0) \n",
    "                    for feat_file in features_path.iterdir()]\n",
    "final_df = reduce(lambda left,right: \n",
    "                pd.merge(left,right,\n",
    "                on=[\"#virus\", \"bacteria\"],how=\"outer\"), dfs)\n",
    "final_df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Replace the column names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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>virus</th>\n",
       "      <th>host</th>\n",
       "      <th>blastn</th>\n",
       "      <th>crisprdetect-2mismatch</th>\n",
       "      <th>gc_content</th>\n",
       "      <th>k25</th>\n",
       "      <th>k6-chebyshev</th>\n",
       "      <th>k6-kendalltau</th>\n",
       "      <th>k6-manhattan</th>\n",
       "      <th>piler-2mismatch</th>\n",
       "      <th>wish</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>NC_010363</td>\n",
       "      <td>NC_008527</td>\n",
       "      <td>60.8</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.021973</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.002122</td>\n",
       "      <td>0.398421</td>\n",
       "      <td>0.382144</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-1.33553</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>NC_010363</td>\n",
       "      <td>NC_002662</td>\n",
       "      <td>59.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.016709</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.001929</td>\n",
       "      <td>0.397773</td>\n",
       "      <td>0.377498</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-1.33035</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>NC_010363</td>\n",
       "      <td>NC_017949</td>\n",
       "      <td>59.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.020818</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.002088</td>\n",
       "      <td>0.396969</td>\n",
       "      <td>0.379686</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-1.33341</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>NC_010363</td>\n",
       "      <td>NC_017492</td>\n",
       "      <td>59.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.022209</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.002131</td>\n",
       "      <td>0.396148</td>\n",
       "      <td>0.380930</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-1.33767</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>NC_010363</td>\n",
       "      <td>NC_009004</td>\n",
       "      <td>59.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.020871</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.002088</td>\n",
       "      <td>0.397095</td>\n",
       "      <td>0.379834</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-1.33341</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       virus       host  blastn  crisprdetect-2mismatch  gc_content  k25  \\\n",
       "0  NC_010363  NC_008527    60.8                     NaN    0.021973  NaN   \n",
       "1  NC_010363  NC_002662    59.0                     NaN    0.016709  NaN   \n",
       "2  NC_010363  NC_017949    59.0                     NaN    0.020818  NaN   \n",
       "3  NC_010363  NC_017492    59.0                     NaN    0.022209  NaN   \n",
       "4  NC_010363  NC_009004    59.0                     NaN    0.020871  NaN   \n",
       "\n",
       "   k6-chebyshev  k6-kendalltau  k6-manhattan  piler-2mismatch     wish  \n",
       "0      0.002122       0.398421      0.382144              NaN -1.33553  \n",
       "1      0.001929       0.397773      0.377498              NaN -1.33035  \n",
       "2      0.002088       0.396969      0.379686              NaN -1.33341  \n",
       "3      0.002131       0.396148      0.380930              NaN -1.33767  \n",
       "4      0.002088       0.397095      0.379834              NaN -1.33341  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "col_names = ['virus', 'host']\n",
    "col_names += [file.stem for file in features_path.iterdir()]\n",
    "final_df.columns = col_names\n",
    "final_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Handle missing values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Check for missing values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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>virus</th>\n",
       "      <th>host</th>\n",
       "      <th>blastn</th>\n",
       "      <th>crisprdetect-2mismatch</th>\n",
       "      <th>gc_content</th>\n",
       "      <th>k25</th>\n",
       "      <th>k6-chebyshev</th>\n",
       "      <th>k6-kendalltau</th>\n",
       "      <th>k6-manhattan</th>\n",
       "      <th>piler-2mismatch</th>\n",
       "      <th>wish</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3611</th>\n",
       "      <td>NC_011357</td>\n",
       "      <td>NC_012731</td>\n",
       "      <td>670.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.067620</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.002660</td>\n",
       "      <td>0.468791</td>\n",
       "      <td>0.410775</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-1.38249</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4182</th>\n",
       "      <td>NC_004585</td>\n",
       "      <td>NC_008022</td>\n",
       "      <td>22561.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.001751</td>\n",
       "      <td>14364.0</td>\n",
       "      <td>0.000713</td>\n",
       "      <td>0.285006</td>\n",
       "      <td>0.232619</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-1.33040</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4183</th>\n",
       "      <td>NC_004585</td>\n",
       "      <td>NC_008021</td>\n",
       "      <td>22561.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.002832</td>\n",
       "      <td>10850.0</td>\n",
       "      <td>0.000736</td>\n",
       "      <td>0.288697</td>\n",
       "      <td>0.235694</td>\n",
       "      <td>2.0</td>\n",
       "      <td>-1.33668</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4187</th>\n",
       "      <td>NC_004585</td>\n",
       "      <td>NC_002737</td>\n",
       "      <td>12311.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.002570</td>\n",
       "      <td>6765.0</td>\n",
       "      <td>0.000741</td>\n",
       "      <td>0.290344</td>\n",
       "      <td>0.237060</td>\n",
       "      <td>2.0</td>\n",
       "      <td>-1.34749</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4190</th>\n",
       "      <td>NC_004585</td>\n",
       "      <td>NC_020540</td>\n",
       "      <td>4621.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.002453</td>\n",
       "      <td>5271.0</td>\n",
       "      <td>0.000749</td>\n",
       "      <td>0.290872</td>\n",
       "      <td>0.237612</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-1.36364</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          virus       host   blastn  crisprdetect-2mismatch  gc_content  \\\n",
       "3611  NC_011357  NC_012731    670.0                     1.0    0.067620   \n",
       "4182  NC_004585  NC_008022  22561.0                     2.0    0.001751   \n",
       "4183  NC_004585  NC_008021  22561.0                     2.0    0.002832   \n",
       "4187  NC_004585  NC_002737  12311.0                     2.0    0.002570   \n",
       "4190  NC_004585  NC_020540   4621.0                     3.0    0.002453   \n",
       "\n",
       "          k25  k6-chebyshev  k6-kendalltau  k6-manhattan  piler-2mismatch  \\\n",
       "3611      2.0      0.002660       0.468791      0.410775              1.0   \n",
       "4182  14364.0      0.000713       0.285006      0.232619              1.0   \n",
       "4183  10850.0      0.000736       0.288697      0.235694              2.0   \n",
       "4187   6765.0      0.000741       0.290344      0.237060              2.0   \n",
       "4190   5271.0      0.000749       0.290872      0.237612              1.0   \n",
       "\n",
       "         wish  \n",
       "3611 -1.38249  \n",
       "4182 -1.33040  \n",
       "4183 -1.33668  \n",
       "4187 -1.34749  \n",
       "4190 -1.36364  "
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "check_rows = final_df[~final_df.isnull().any(axis=1)]\n",
    "check_rows.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(469, 11)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "check_rows.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Check number of missing values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "virus                           0\n",
       "host                            0\n",
       "blastn                    1925573\n",
       "crisprdetect-2mismatch    2212430\n",
       "gc_content                      0\n",
       "k25                       2154290\n",
       "k6-chebyshev                    0\n",
       "k6-kendalltau                   0\n",
       "k6-manhattan                    0\n",
       "piler-2mismatch           2212592\n",
       "wish                            0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_df.isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "virus                     2213180\n",
       "host                      2213180\n",
       "blastn                     287607\n",
       "crisprdetect-2mismatch        750\n",
       "gc_content                2213180\n",
       "k25                         58890\n",
       "k6-chebyshev              2213180\n",
       "k6-kendalltau             2213180\n",
       "k6-manhattan              2213180\n",
       "piler-2mismatch               588\n",
       "wish                      2213180\n",
       "dtype: int64"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_df.notnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Fill missing values (according to promoter's suggestions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "final_df['blastn'].fillna(0, inplace=True)\n",
    "final_df['crisprdetect-2mismatch'].fillna(0, inplace=True)\n",
    "final_df['k25'].fillna(0, inplace=True)\n",
    "final_df['piler-2mismatch'].fillna(0, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "virus                     0\n",
       "host                      0\n",
       "blastn                    0\n",
       "crisprdetect-2mismatch    0\n",
       "gc_content                0\n",
       "k25                       0\n",
       "k6-chebyshev              0\n",
       "k6-kendalltau             0\n",
       "k6-manhattan              0\n",
       "piler-2mismatch           0\n",
       "wish                      0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_df.isna().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Find labels (which viruses infect which hosts in nature)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load in the taxonomy JSONs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['host', 'virus'])"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import json\n",
    "import pathlib\n",
    "\n",
    "orgs = {}\n",
    "for file in pathlib.Path('./taxonomy/').iterdir():\n",
    "    with open(file, 'r') as open_file:\n",
    "        orgs[file.stem] = json.load(open_file)\n",
    "        \n",
    "orgs.keys()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Find groups for LeaveOneGroupOut"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Bacteria-based"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"\\nfinal_df['group'] = final_df.apply(lambda row: \\n                    orgs['host'][row['host']]['lineage_names'][-3], axis=1)\\npd.unique(final_df['group'])\\n\""
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "'''\n",
    "final_df['group'] = final_df.apply(lambda row: \n",
    "                    orgs['host'][row['host']]['lineage_names'][-3], axis=1)\n",
    "pd.unique(final_df['group'])\n",
    "'''"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Virus-based"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "final_df['group'] = final_df.apply(lambda row: \n",
    "                    orgs['virus'][row['virus']]['host']['lineage_names'][-3], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Enterobacteriaceae                                   569489\n",
       "Pseudomonadaceae                                     251007\n",
       "Staphylococcaceae                                    199726\n",
       "Streptococcaceae                                     188930\n",
       "Bacillaceae                                          140348\n",
       "Burkholderiaceae                                     113358\n",
       "Vibrionaceae                                          94465\n",
       "Propionibacteriaceae                                  59378\n",
       "Mycobacteriaceae                                      59378\n",
       "Xanthomonadaceae                                      48582\n",
       "Listeriaceae                                          43184\n",
       "Lactobacillaceae                                      43184\n",
       "Clostridiaceae                                        35087\n",
       "Enterococcaceae                                       35087\n",
       "Aeromonadaceae                                        32388\n",
       "Nocardiaceae                                          26990\n",
       "Moraxellaceae                                         21592\n",
       "Pasteurellaceae                                       21592\n",
       "Caulobacteraceae                                      18893\n",
       "Leuconostocaceae                                      18893\n",
       "Peptostreptococcaceae                                 18893\n",
       "Streptomycetaceae                                     16194\n",
       "Chlamydiaceae                                         16194\n",
       "Thermaceae                                            13495\n",
       "Rhodobacteraceae                                      13495\n",
       "Flavobacteriaceae                                     10796\n",
       "Helicobacteraceae                                     10796\n",
       "Alcaligenaceae                                         8097\n",
       "Campylobacteraceae                                     8097\n",
       "Mycoplasmataceae                                       8097\n",
       "Bdellovibrionaceae                                     8097\n",
       "Rhizobiaceae                                           5398\n",
       "Acholeplasmataceae                                     5398\n",
       "Brucellaceae                                           5398\n",
       "Bacteroidaceae                                         5398\n",
       "Microbacteriaceae                                      5398\n",
       "Myxococcaceae                                          2699\n",
       "Comamonadaceae                                         2699\n",
       "                                                       2699\n",
       "Nostocaceae                                            2699\n",
       "Thermoanaerobacterales Family III. Incertae Sedis      2699\n",
       "Tsukamurellaceae                                       2699\n",
       "Rhodospirillaceae                                      2699\n",
       "Corynebacteriaceae                                     2699\n",
       "Shewanellaceae                                         2699\n",
       "Rhodothermaceae                                        2699\n",
       "Prochlorococcaceae                                     2699\n",
       "Alteromonadaceae                                       2699\n",
       "Name: group, dtype: int64"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_df['group'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "48"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(pd.unique(final_df['group']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "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>virus</th>\n",
       "      <th>host</th>\n",
       "      <th>blastn</th>\n",
       "      <th>crisprdetect-2mismatch</th>\n",
       "      <th>gc_content</th>\n",
       "      <th>k25</th>\n",
       "      <th>k6-chebyshev</th>\n",
       "      <th>k6-kendalltau</th>\n",
       "      <th>k6-manhattan</th>\n",
       "      <th>piler-2mismatch</th>\n",
       "      <th>wish</th>\n",
       "      <th>group</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>NC_010363</td>\n",
       "      <td>NC_008527</td>\n",
       "      <td>60.8</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.021973</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.002122</td>\n",
       "      <td>0.398421</td>\n",
       "      <td>0.382144</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.33553</td>\n",
       "      <td>Streptococcaceae</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>NC_010363</td>\n",
       "      <td>NC_002662</td>\n",
       "      <td>59.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.016709</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.001929</td>\n",
       "      <td>0.397773</td>\n",
       "      <td>0.377498</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.33035</td>\n",
       "      <td>Streptococcaceae</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>NC_010363</td>\n",
       "      <td>NC_017949</td>\n",
       "      <td>59.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.020818</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.002088</td>\n",
       "      <td>0.396969</td>\n",
       "      <td>0.379686</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.33341</td>\n",
       "      <td>Streptococcaceae</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>NC_010363</td>\n",
       "      <td>NC_017492</td>\n",
       "      <td>59.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.022209</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.002131</td>\n",
       "      <td>0.396148</td>\n",
       "      <td>0.380930</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.33767</td>\n",
       "      <td>Streptococcaceae</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>NC_010363</td>\n",
       "      <td>NC_009004</td>\n",
       "      <td>59.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.020871</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.002088</td>\n",
       "      <td>0.397095</td>\n",
       "      <td>0.379834</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.33341</td>\n",
       "      <td>Streptococcaceae</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       virus       host  blastn  crisprdetect-2mismatch  gc_content  k25  \\\n",
       "0  NC_010363  NC_008527    60.8                     0.0    0.021973  0.0   \n",
       "1  NC_010363  NC_002662    59.0                     0.0    0.016709  0.0   \n",
       "2  NC_010363  NC_017949    59.0                     0.0    0.020818  0.0   \n",
       "3  NC_010363  NC_017492    59.0                     0.0    0.022209  0.0   \n",
       "4  NC_010363  NC_009004    59.0                     0.0    0.020871  0.0   \n",
       "\n",
       "   k6-chebyshev  k6-kendalltau  k6-manhattan  piler-2mismatch     wish  \\\n",
       "0      0.002122       0.398421      0.382144              0.0 -1.33553   \n",
       "1      0.001929       0.397773      0.377498              0.0 -1.33035   \n",
       "2      0.002088       0.396969      0.379686              0.0 -1.33341   \n",
       "3      0.002131       0.396148      0.380930              0.0 -1.33767   \n",
       "4      0.002088       0.397095      0.379834              0.0 -1.33341   \n",
       "\n",
       "              group  \n",
       "0  Streptococcaceae  \n",
       "1  Streptococcaceae  \n",
       "2  Streptococcaceae  \n",
       "3  Streptococcaceae  \n",
       "4  Streptococcaceae  "
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:ylabel='group'>"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "final_df['group'].value_counts().plot(kind=\"pie\", autopct=\"%1.1f%%\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Make 'group' parameter categorical – optimalization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dtype('O')"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_df['group'].dtype"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 2213180 entries, 0 to 2213179\n",
      "Data columns (total 12 columns):\n",
      " #   Column                  Dtype  \n",
      "---  ------                  -----  \n",
      " 0   virus                   object \n",
      " 1   host                    object \n",
      " 2   blastn                  float64\n",
      " 3   crisprdetect-2mismatch  float64\n",
      " 4   gc_content              float64\n",
      " 5   k25                     float64\n",
      " 6   k6-chebyshev            float64\n",
      " 7   k6-kendalltau           float64\n",
      " 8   k6-manhattan            float64\n",
      " 9   piler-2mismatch         float64\n",
      " 10  wish                    float64\n",
      " 11  group                   object \n",
      "dtypes: float64(9), object(3)\n",
      "memory usage: 219.5+ MB\n"
     ]
    }
   ],
   "source": [
    "final_df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "CategoricalDtype(categories=['', 'Acholeplasmataceae', 'Aeromonadaceae', 'Alcaligenaceae',\n",
       "                  'Alteromonadaceae', 'Bacillaceae', 'Bacteroidaceae',\n",
       "                  'Bdellovibrionaceae', 'Brucellaceae', 'Burkholderiaceae',\n",
       "                  'Campylobacteraceae', 'Caulobacteraceae', 'Chlamydiaceae',\n",
       "                  'Clostridiaceae', 'Comamonadaceae', 'Corynebacteriaceae',\n",
       "                  'Enterobacteriaceae', 'Enterococcaceae', 'Flavobacteriaceae',\n",
       "                  'Helicobacteraceae', 'Lactobacillaceae', 'Leuconostocaceae',\n",
       "                  'Listeriaceae', 'Microbacteriaceae', 'Moraxellaceae',\n",
       "                  'Mycobacteriaceae', 'Mycoplasmataceae', 'Myxococcaceae',\n",
       "                  'Nocardiaceae', 'Nostocaceae', 'Pasteurellaceae',\n",
       "                  'Peptostreptococcaceae', 'Prochlorococcaceae',\n",
       "                  'Propionibacteriaceae', 'Pseudomonadaceae', 'Rhizobiaceae',\n",
       "                  'Rhodobacteraceae', 'Rhodospirillaceae', 'Rhodothermaceae',\n",
       "                  'Shewanellaceae', 'Staphylococcaceae', 'Streptococcaceae',\n",
       "                  'Streptomycetaceae', 'Thermaceae',\n",
       "                  'Thermoanaerobacterales Family III. Incertae Sedis',\n",
       "                  'Tsukamurellaceae', 'Vibrionaceae', 'Xanthomonadaceae'],\n",
       ", ordered=False)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_df['group'] = pd.Categorical(final_df['group'])\n",
    "final_df['group'].dtype"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "NOTE: there are hosts without designated families"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "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>virus</th>\n",
       "      <th>host</th>\n",
       "      <th>blastn</th>\n",
       "      <th>crisprdetect-2mismatch</th>\n",
       "      <th>gc_content</th>\n",
       "      <th>k25</th>\n",
       "      <th>k6-chebyshev</th>\n",
       "      <th>k6-kendalltau</th>\n",
       "      <th>k6-manhattan</th>\n",
       "      <th>piler-2mismatch</th>\n",
       "      <th>wish</th>\n",
       "      <th>group</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>189108</th>\n",
       "      <td>NC_008562</td>\n",
       "      <td>NC_011729</td>\n",
       "      <td>1936.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.073441</td>\n",
       "      <td>50.0</td>\n",
       "      <td>0.003370</td>\n",
       "      <td>0.851879</td>\n",
       "      <td>0.639265</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.43287</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>189109</th>\n",
       "      <td>NC_008562</td>\n",
       "      <td>NC_010296</td>\n",
       "      <td>1930.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.036224</td>\n",
       "      <td>1431.0</td>\n",
       "      <td>0.002731</td>\n",
       "      <td>0.891840</td>\n",
       "      <td>0.601082</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.44261</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>189110</th>\n",
       "      <td>NC_008562</td>\n",
       "      <td>NC_016640</td>\n",
       "      <td>1159.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.023027</td>\n",
       "      <td>50.0</td>\n",
       "      <td>0.002908</td>\n",
       "      <td>0.864091</td>\n",
       "      <td>0.557508</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.44114</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>189111</th>\n",
       "      <td>NC_008562</td>\n",
       "      <td>NC_019753</td>\n",
       "      <td>801.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.057910</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.002481</td>\n",
       "      <td>0.797970</td>\n",
       "      <td>0.577121</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.42135</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>189112</th>\n",
       "      <td>NC_008562</td>\n",
       "      <td>NC_014501</td>\n",
       "      <td>720.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.057305</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.003333</td>\n",
       "      <td>0.830270</td>\n",
       "      <td>0.592474</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.42505</td>\n",
       "      <td></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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>905267</th>\n",
       "      <td>NC_008562</td>\n",
       "      <td>NC_018417</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.314027</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.048148</td>\n",
       "      <td>0.767260</td>\n",
       "      <td>1.275939</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.39383</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>905268</th>\n",
       "      <td>NC_008562</td>\n",
       "      <td>NC_018415</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.317681</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.053895</td>\n",
       "      <td>0.757731</td>\n",
       "      <td>1.293647</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.39432</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>905269</th>\n",
       "      <td>NC_008562</td>\n",
       "      <td>NC_018416</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.317754</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.050627</td>\n",
       "      <td>0.765630</td>\n",
       "      <td>1.289442</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.39417</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>905270</th>\n",
       "      <td>NC_008562</td>\n",
       "      <td>NC_018414</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.319773</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.050478</td>\n",
       "      <td>0.763087</td>\n",
       "      <td>1.298034</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.39409</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>905271</th>\n",
       "      <td>NC_008562</td>\n",
       "      <td>NC_014497</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.324143</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.040648</td>\n",
       "      <td>0.746222</td>\n",
       "      <td>1.300492</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.39656</td>\n",
       "      <td></td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2699 rows × 12 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            virus       host  blastn  crisprdetect-2mismatch  gc_content  \\\n",
       "189108  NC_008562  NC_011729  1936.0                     0.0    0.073441   \n",
       "189109  NC_008562  NC_010296  1930.0                     0.0    0.036224   \n",
       "189110  NC_008562  NC_016640  1159.0                     0.0    0.023027   \n",
       "189111  NC_008562  NC_019753   801.0                     0.0    0.057910   \n",
       "189112  NC_008562  NC_014501   720.0                     0.0    0.057305   \n",
       "...           ...        ...     ...                     ...         ...   \n",
       "905267  NC_008562  NC_018417     0.0                     0.0    0.314027   \n",
       "905268  NC_008562  NC_018415     0.0                     0.0    0.317681   \n",
       "905269  NC_008562  NC_018416     0.0                     0.0    0.317754   \n",
       "905270  NC_008562  NC_018414     0.0                     0.0    0.319773   \n",
       "905271  NC_008562  NC_014497     0.0                     0.0    0.324143   \n",
       "\n",
       "           k25  k6-chebyshev  k6-kendalltau  k6-manhattan  piler-2mismatch  \\\n",
       "189108    50.0      0.003370       0.851879      0.639265              0.0   \n",
       "189109  1431.0      0.002731       0.891840      0.601082              0.0   \n",
       "189110    50.0      0.002908       0.864091      0.557508              0.0   \n",
       "189111     2.0      0.002481       0.797970      0.577121              0.0   \n",
       "189112     6.0      0.003333       0.830270      0.592474              0.0   \n",
       "...        ...           ...            ...           ...              ...   \n",
       "905267     0.0      0.048148       0.767260      1.275939              0.0   \n",
       "905268     0.0      0.053895       0.757731      1.293647              0.0   \n",
       "905269     0.0      0.050627       0.765630      1.289442              0.0   \n",
       "905270     0.0      0.050478       0.763087      1.298034              0.0   \n",
       "905271     0.0      0.040648       0.746222      1.300492              0.0   \n",
       "\n",
       "           wish group  \n",
       "189108 -1.43287        \n",
       "189109 -1.44261        \n",
       "189110 -1.44114        \n",
       "189111 -1.42135        \n",
       "189112 -1.42505        \n",
       "...         ...   ...  \n",
       "905267 -1.39383        \n",
       "905268 -1.39432        \n",
       "905269 -1.39417        \n",
       "905270 -1.39409        \n",
       "905271 -1.39656        \n",
       "\n",
       "[2699 rows x 12 columns]"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_df[final_df['group'] == '']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "16    569489\n",
       "34    251007\n",
       "40    199726\n",
       "41    188930\n",
       "5     140348\n",
       "9     113358\n",
       "46     94465\n",
       "25     59378\n",
       "33     59378\n",
       "47     48582\n",
       "22     43184\n",
       "20     43184\n",
       "17     35087\n",
       "13     35087\n",
       "2      32388\n",
       "28     26990\n",
       "30     21592\n",
       "24     21592\n",
       "21     18893\n",
       "11     18893\n",
       "31     18893\n",
       "42     16194\n",
       "12     16194\n",
       "43     13495\n",
       "36     13495\n",
       "18     10796\n",
       "19     10796\n",
       "3       8097\n",
       "26      8097\n",
       "10      8097\n",
       "7       8097\n",
       "6       5398\n",
       "35      5398\n",
       "8       5398\n",
       "1       5398\n",
       "23      5398\n",
       "32      2699\n",
       "29      2699\n",
       "27      2699\n",
       "37      2699\n",
       "38      2699\n",
       "39      2699\n",
       "4       2699\n",
       "14      2699\n",
       "44      2699\n",
       "45      2699\n",
       "15      2699\n",
       "0       2699\n",
       "Name: group_code, dtype: int64"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_df['group_code'] = final_df['group'].cat.codes\n",
    "final_df['group_code'] = pd.Categorical(final_df['group_code'])\n",
    "final_df['group_code'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Species-based labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "final_df['y'] = final_df.apply(lambda row: \n",
    "                    1 if orgs['host'][row['host']]['lineage_names'][-1] == \n",
    "                    orgs['virus'][row['virus']]['host']['lineage_names'][-1] \n",
    "                    else 0, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    2196423\n",
       "1      16757\n",
       "Name: y, dtype: int64"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_df['y'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Enterobacteriaceae                                   8419\n",
       "Staphylococcaceae                                    3342\n",
       "Pseudomonadaceae                                     1012\n",
       "Streptococcaceae                                      782\n",
       "Listeriaceae                                          624\n",
       "Bacillaceae                                           454\n",
       "Vibrionaceae                                          436\n",
       "Burkholderiaceae                                      342\n",
       "Helicobacteraceae                                     212\n",
       "Propionibacteriaceae                                  211\n",
       "Mycobacteriaceae                                      156\n",
       "Moraxellaceae                                         113\n",
       "Lactobacillaceae                                       74\n",
       "Xanthomonadaceae                                       71\n",
       "Enterococcaceae                                        64\n",
       "Chlamydiaceae                                          64\n",
       "Peptostreptococcaceae                                  49\n",
       "Clostridiaceae                                         49\n",
       "Pasteurellaceae                                        46\n",
       "Campylobacteraceae                                     39\n",
       "Thermaceae                                             20\n",
       "Brucellaceae                                           18\n",
       "Caulobacteraceae                                       14\n",
       "Aeromonadaceae                                         14\n",
       "Alteromonadaceae                                       13\n",
       "Prochlorococcaceae                                     12\n",
       "Leuconostocaceae                                       12\n",
       "Rhizobiaceae                                           12\n",
       "Nocardiaceae                                           11\n",
       "Alcaligenaceae                                          9\n",
       "Flavobacteriaceae                                       8\n",
       "Corynebacteriaceae                                      7\n",
       "Microbacteriaceae                                       6\n",
       "Bdellovibrionaceae                                      6\n",
       "Bacteroidaceae                                          6\n",
       "Streptomycetaceae                                       6\n",
       "Rhodobacteraceae                                        5\n",
       "Mycoplasmataceae                                        5\n",
       "Acholeplasmataceae                                      2\n",
       "Rhodothermaceae                                         2\n",
       "Nostocaceae                                             2\n",
       "Shewanellaceae                                          2\n",
       "Rhodospirillaceae                                       1\n",
       "Thermoanaerobacterales Family III. Incertae Sedis       1\n",
       "Comamonadaceae                                          1\n",
       "Tsukamurellaceae                                        1\n",
       "Myxococcaceae                                           1\n",
       "                                                        1\n",
       "Name: group, dtype: int64"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_df[final_df['y'] == 1]['group'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 2213180 entries, 0 to 2213179\n",
      "Data columns (total 14 columns):\n",
      " #   Column                  Dtype   \n",
      "---  ------                  -----   \n",
      " 0   virus                   object  \n",
      " 1   host                    object  \n",
      " 2   blastn                  float64 \n",
      " 3   crisprdetect-2mismatch  float64 \n",
      " 4   gc_content              float64 \n",
      " 5   k25                     float64 \n",
      " 6   k6-chebyshev            float64 \n",
      " 7   k6-kendalltau           float64 \n",
      " 8   k6-manhattan            float64 \n",
      " 9   piler-2mismatch         float64 \n",
      " 10  wish                    float64 \n",
      " 11  group                   category\n",
      " 12  group_code              category\n",
      " 13  y                       int64   \n",
      "dtypes: category(2), float64(9), int64(1), object(2)\n",
      "memory usage: 223.7+ MB\n"
     ]
    }
   ],
   "source": [
    "final_df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Save the dataframe to files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "final_df.to_csv('main_df.csv',index=False)\n",
    "final_df.to_pickle('./main_df.pkl')"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "b3ba2566441a7c06988d0923437866b63cedc61552a5af99d1f4fb67d367b25f"
  },
  "kernelspec": {
   "display_name": "Python 3.8.8 64-bit (conda)",
   "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": "3.10.4"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}