eqa_panel_product_ols.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "\n",
    "import pandas as pd\n",
    "import statsmodels.api as sm\n",
    "from statsmodels.formula.api import ols\n",
    "\n",
    "warnings.filterwarnings(action='ignore', category=UserWarning)\n",
    "warnings.filterwarnings(action='ignore', category=RuntimeWarning)\n",
    "\n",
    "df = pd.read_table('./final2.tsv')  # same as the extended data 2.\n",
    "df['reference wesTMB'] = df['sample'].map({'A1': 103.0021, 'B1': 148.34,\n",
    "                                           'C1': 112.2377, 'D1': 4.97, 'E1': .0})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "type2: panel_id * product + sample + rTMB\n",
      "                               sum_sq      df         F    PR(>F)\n",
      "C(panel_id)                       NaN    49.0       NaN       NaN\n",
      "C(product)                        NaN    37.0       NaN       NaN\n",
      "C(sample)                  613.000600     4.0  0.178151  0.673427\n",
      "C(panel_id):C(product)  277842.521789  1813.0  0.178151  0.673427\n",
      "Q(\"reference wesTMB\")      153.250150     1.0  0.178151  0.673427\n",
      "Residual                169464.671345   197.0       NaN       NaN\n",
      "type3: panel_id * product + sample + rTMB\n",
      "                               sum_sq      df         F    PR(>F)\n",
      "Intercept                  153.250150     1.0  0.178151  0.673427\n",
      "C(panel_id)               7509.257346    49.0  0.178151  0.673427\n",
      "C(product)                5670.255547    37.0  0.178151  0.673427\n",
      "C(sample)                  613.000600     4.0  0.178151  0.673427\n",
      "C(panel_id):C(product)  277842.521789  1813.0  0.178151  0.673427\n",
      "Q(\"reference wesTMB\")      153.250150     1.0  0.178151  0.673427\n",
      "Residual                169464.671345   197.0       NaN       NaN\n",
      "type2: panel_id + product + sample + rTMB\n",
      "                              sum_sq     df           F        PR(>F)\n",
      "C(panel_id)            158547.559189   49.0    3.766008  2.274729e-11\n",
      "C(product)             156921.578492   37.0    4.936268  7.057307e-14\n",
      "C(sample)                8908.766497    4.0    2.592241  5.391362e-02\n",
      "Q(\"reference wesTMB\")  251487.136305    1.0  292.707311  8.139916e-41\n",
      "Residual               169257.698523  197.0         NaN           NaN\n",
      "type3: panel_id + product + sample + rTMB\n",
      "                              sum_sq     df           F        PR(>F)\n",
      "Intercept                1802.135105    1.0    2.097515  1.491284e-01\n",
      "C(panel_id)            158547.559189   49.0    3.766008  2.274729e-11\n",
      "C(product)             156921.578492   37.0    4.936268  7.057307e-14\n",
      "C(sample)                8908.766497    4.0    2.592241  5.391362e-02\n",
      "Q(\"reference wesTMB\")  251487.136305    1.0  292.707311  8.139916e-41\n",
      "Residual               169257.698523  197.0         NaN           NaN\n"
     ]
    }
   ],
   "source": [
    "tp2_model = ols('wesTMB ~ C(panel_id) * C(product) + C(sample) + Q(\"reference wesTMB\")', data=df).fit()\n",
    "tp2_ancova_table = sm.stats.anova_lm(tp2_model, typ=2)\n",
    "print('type2: panel_id * product + sample + rTMB')\n",
    "print(tp2_ancova_table)\n",
    "\n",
    "tp3_model = ols('wesTMB ~ C(panel_id) * C(product) + C(sample) + Q(\"reference wesTMB\")', data=df).fit()\n",
    "tp3_ancova_table = sm.stats.anova_lm(tp3_model, typ=3)\n",
    "print('type3: panel_id * product + sample + rTMB')\n",
    "print(tp3_ancova_table)\n",
    "\n",
    "tp2_model2 = ols('wesTMB ~ C(panel_id) + C(product) + C(sample) + Q(\"reference wesTMB\")', data=df).fit()\n",
    "tp2_ancova_table2 = sm.stats.anova_lm(tp2_model2, typ=2)\n",
    "print('type2: panel_id + product + sample + rTMB')\n",
    "print(tp2_ancova_table2)\n",
    "\n",
    "tp3_model2 = ols('wesTMB ~ C(panel_id) + C(product) + C(sample) + Q(\"reference wesTMB\")', data=df).fit()\n",
    "tp3_ancova_table2 = sm.stats.anova_lm(tp3_model2, typ=3)\n",
    "print('type3: panel_id + product + sample + rTMB')\n",
    "print(tp3_ancova_table2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "type2: panel_id * product + sample + rTMB\n",
      "                               sum_sq      df         F    PR(>F)    eta_sq\n",
      "C(panel_id)                       NaN    49.0       NaN       NaN       NaN\n",
      "C(product)                        NaN    37.0       NaN       NaN       NaN\n",
      "C(sample)                  613.000600     4.0  0.178151  0.673427  0.001368\n",
      "C(panel_id):C(product)  277842.521789  1813.0  0.178151  0.673427  0.620083\n",
      "Q(\"reference wesTMB\")      153.250150     1.0  0.178151  0.673427  0.000342\n",
      "Residual                169464.671345   197.0       NaN       NaN  0.378207\n",
      "type3: panel_id * product + sample + rTMB\n",
      "                               sum_sq      df         F    PR(>F)    eta_sq\n",
      "Intercept                  153.250150     1.0  0.178151  0.673427       NaN\n",
      "C(panel_id)               7509.257346    49.0  0.178151  0.673427       NaN\n",
      "C(product)                5670.255547    37.0  0.178151  0.673427       NaN\n",
      "C(sample)                  613.000600     4.0  0.178151  0.673427  0.001329\n",
      "C(panel_id):C(product)  277842.521789  1813.0  0.178151  0.673427  0.602165\n",
      "Q(\"reference wesTMB\")      153.250150     1.0  0.178151  0.673427  0.000332\n",
      "Residual                169464.671345   197.0       NaN       NaN  0.367279\n",
      "type2: panel_id + product + sample + rTMB\n",
      "                              sum_sq     df           F        PR(>F)  \\\n",
      "C(panel_id)            158547.559189   49.0    3.766008  2.274729e-11   \n",
      "C(product)             156921.578492   37.0    4.936268  7.057307e-14   \n",
      "C(sample)                8908.766497    4.0    2.592241  5.391362e-02   \n",
      "Q(\"reference wesTMB\")  251487.136305    1.0  292.707311  8.139916e-41   \n",
      "Residual               169257.698523  197.0         NaN           NaN   \n",
      "\n",
      "                         eta_sq  \n",
      "C(panel_id)            0.212780  \n",
      "C(product)             0.210598  \n",
      "C(sample)              0.011956  \n",
      "Q(\"reference wesTMB\")  0.337511  \n",
      "Residual               0.227154  \n",
      "type3: panel_id + product + sample + rTMB\n",
      "                              sum_sq     df           F        PR(>F)  \\\n",
      "Intercept                1802.135105    1.0    2.097515  1.491284e-01   \n",
      "C(panel_id)            158547.559189   49.0    3.766008  2.274729e-11   \n",
      "C(product)             156921.578492   37.0    4.936268  7.057307e-14   \n",
      "C(sample)                8908.766497    4.0    2.592241  5.391362e-02   \n",
      "Q(\"reference wesTMB\")  251487.136305    1.0  292.707311  8.139916e-41   \n",
      "Residual               169257.698523  197.0         NaN           NaN   \n",
      "\n",
      "                         eta_sq  \n",
      "Intercept              0.002413  \n",
      "C(panel_id)            0.212267  \n",
      "C(product)             0.210090  \n",
      "C(sample)              0.011927  \n",
      "Q(\"reference wesTMB\")  0.336697  \n",
      "Residual               0.226606  \n"
     ]
    }
   ],
   "source": [
    "tp2_model = ols('wesTMB ~ C(panel_id) * C(product) + C(sample) + Q(\"reference wesTMB\")', data=df).fit()\n",
    "tp2_ancova_table = sm.stats.anova_lm(tp2_model, typ=2)\n",
    "tp2_ancova_table['eta_sq'] = tp2_ancova_table['sum_sq'] / tp2_ancova_table['sum_sq'].sum()\n",
    "print('type2: panel_id * product + sample + rTMB')\n",
    "print(tp2_ancova_table)\n",
    "\n",
    "tp3_model = ols('wesTMB ~ C(panel_id) * C(product) + C(sample) + Q(\"reference wesTMB\")', data=df).fit()\n",
    "tp3_ancova_table = sm.stats.anova_lm(tp3_model, typ=3)\n",
    "tp3_ancova_table['eta_sq'] = tp2_ancova_table['sum_sq'] / tp3_ancova_table['sum_sq'].sum()\n",
    "print('type3: panel_id * product + sample + rTMB')\n",
    "print(tp3_ancova_table)\n",
    "\n",
    "tp2_model2 = ols('wesTMB ~ C(panel_id) + C(product) + C(sample) + Q(\"reference wesTMB\")', data=df).fit()\n",
    "tp2_ancova_table2 = sm.stats.anova_lm(tp2_model2, typ=2)\n",
    "tp2_ancova_table2['eta_sq'] = tp2_ancova_table2['sum_sq'] / tp2_ancova_table2['sum_sq'].sum()\n",
    "print('type2: panel_id + product + sample + rTMB')\n",
    "print(tp2_ancova_table2)\n",
    "\n",
    "tp3_model2 = ols('wesTMB ~ C(panel_id) + C(product) + C(sample) + Q(\"reference wesTMB\")', data=df).fit()\n",
    "tp3_ancova_table2 = sm.stats.anova_lm(tp3_model2, typ=3)\n",
    "tp3_ancova_table2['eta_sq'] = tp3_ancova_table2['sum_sq'] / tp3_ancova_table2['sum_sq'].sum()\n",
    "print('type3: panel_id + product + sample + rTMB')\n",
    "print(tp3_ancova_table2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "type2: panel_id + product + sample * rTMB\n",
      "                                        sum_sq     df             F  \\\n",
      "C(panel_id)                      158547.559189   49.0  3.766008e+00   \n",
      "C(product)                       156921.578492   37.0  4.936268e+00   \n",
      "C(sample)                             0.001815    4.0  5.280675e-07   \n",
      "Q(\"reference wesTMB\")                      NaN    1.0           NaN   \n",
      "C(sample):Q(\"reference wesTMB\")    8908.766497    4.0  2.592241e+00   \n",
      "Residual                         169257.698523  197.0           NaN   \n",
      "\n",
      "                                       PR(>F)        eta_sq  \n",
      "C(panel_id)                      2.274729e-11  3.211834e-01  \n",
      "C(product)                       7.057307e-14  3.178895e-01  \n",
      "C(sample)                        1.000000e+00  3.676420e-09  \n",
      "Q(\"reference wesTMB\")                     NaN           NaN  \n",
      "C(sample):Q(\"reference wesTMB\")  5.391362e-02  1.804725e-02  \n",
      "Residual                                  NaN  3.428798e-01  \n",
      "type3: panel_id + product + sample * rTMB\n",
      "                                        sum_sq     df          F  \\\n",
      "Intercept                           578.098251    1.0   0.672852   \n",
      "C(panel_id)                      158547.559189   49.0   3.766008   \n",
      "C(product)                       156921.578492   37.0   4.936268   \n",
      "C(sample)                          8908.766498    4.0   2.592241   \n",
      "Q(\"reference wesTMB\")             58439.748697    1.0  68.018357   \n",
      "C(sample):Q(\"reference wesTMB\")    8908.766497    4.0   2.592241   \n",
      "Residual                         169257.698523  197.0        NaN   \n",
      "\n",
      "                                       PR(>F)    eta_sq  \n",
      "Intercept                        4.130505e-01  0.001029  \n",
      "C(panel_id)                      2.274729e-11  0.282333  \n",
      "C(product)                       7.057307e-14  0.279438  \n",
      "C(sample)                        5.391362e-02  0.015864  \n",
      "Q(\"reference wesTMB\")            2.267849e-14  0.104066  \n",
      "C(sample):Q(\"reference wesTMB\")  5.391362e-02  0.015864  \n",
      "Residual                                  NaN  0.301405  \n",
      "type2: panel_id + product + sample + rTMB\n",
      "                              sum_sq     df           F        PR(>F)  \\\n",
      "C(panel_id)            158547.559189   49.0    3.766008  2.274729e-11   \n",
      "C(product)             156921.578492   37.0    4.936268  7.057307e-14   \n",
      "C(sample)                8908.766497    4.0    2.592241  5.391362e-02   \n",
      "Q(\"reference wesTMB\")  251487.136305    1.0  292.707311  8.139916e-41   \n",
      "Residual               169257.698523  197.0         NaN           NaN   \n",
      "\n",
      "                         eta_sq  \n",
      "C(panel_id)            0.212780  \n",
      "C(product)             0.210598  \n",
      "C(sample)              0.011956  \n",
      "Q(\"reference wesTMB\")  0.337511  \n",
      "Residual               0.227154  \n",
      "type3: panel_id + product + sample + rTMB\n",
      "                              sum_sq     df           F        PR(>F)  \\\n",
      "Intercept                1802.135105    1.0    2.097515  1.491284e-01   \n",
      "C(panel_id)            158547.559189   49.0    3.766008  2.274729e-11   \n",
      "C(product)             156921.578492   37.0    4.936268  7.057307e-14   \n",
      "C(sample)                8908.766497    4.0    2.592241  5.391362e-02   \n",
      "Q(\"reference wesTMB\")  251487.136305    1.0  292.707311  8.139916e-41   \n",
      "Residual               169257.698523  197.0         NaN           NaN   \n",
      "\n",
      "                         eta_sq  \n",
      "Intercept              0.002413  \n",
      "C(panel_id)            0.212267  \n",
      "C(product)             0.210090  \n",
      "C(sample)              0.011927  \n",
      "Q(\"reference wesTMB\")  0.336697  \n",
      "Residual               0.226606  \n"
     ]
    }
   ],
   "source": [
    "tp2_model3 = ols('wesTMB ~ C(panel_id) + C(product) + C(sample) * Q(\"reference wesTMB\")', data=df).fit()\n",
    "tp2_ancova_table3 = sm.stats.anova_lm(tp2_model3, typ=2)\n",
    "tp2_ancova_table3['eta_sq'] = tp2_ancova_table3['sum_sq'] / tp2_ancova_table3['sum_sq'].sum()\n",
    "print('type2: panel_id + product + sample * rTMB')\n",
    "print(tp2_ancova_table3)\n",
    "\n",
    "tp3_model3 = ols('wesTMB ~ C(panel_id) + C(product) + C(sample) * Q(\"reference wesTMB\")', data=df).fit()\n",
    "tp3_ancova_table3 = sm.stats.anova_lm(tp3_model3, typ=3)\n",
    "tp3_ancova_table3['eta_sq'] = tp3_ancova_table3['sum_sq'] / tp3_ancova_table3['sum_sq'].sum()\n",
    "print('type3: panel_id + product + sample * rTMB')\n",
    "print(tp3_ancova_table3)\n",
    "\n",
    "tp2_model4 = ols('wesTMB ~ C(panel_id) + C(product) + C(sample) + Q(\"reference wesTMB\")', data=df).fit()\n",
    "tp2_ancova_table4 = sm.stats.anova_lm(tp2_model4, typ=2)\n",
    "tp2_ancova_table4['eta_sq'] = tp2_ancova_table4['sum_sq'] / tp2_ancova_table4['sum_sq'].sum()\n",
    "print('type2: panel_id + product + sample + rTMB')\n",
    "print(tp2_ancova_table4)\n",
    "\n",
    "tp3_model4 = ols('wesTMB ~ C(panel_id) + C(product) + C(sample) + Q(\"reference wesTMB\")', data=df).fit()\n",
    "tp3_ancova_table4 = sm.stats.anova_lm(tp3_model4, typ=3)\n",
    "tp3_ancova_table4['eta_sq'] = tp3_ancova_table4['sum_sq'] / tp3_ancova_table4['sum_sq'].sum()\n",
    "print('type3: panel_id + product + sample + rTMB')\n",
    "print(tp3_ancova_table4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "type2: panel_id * product + sample + rTMB\n",
      "                               sum_sq    df           F        PR(>F)  \\\n",
      "C(panel_id)             491689.342882  17.0   27.005071  1.897209e-09   \n",
      "C(product)              144614.512612   5.0   27.005071  1.897209e-09   \n",
      "C(panel_id):C(product)  397586.188188  85.0    4.367328  1.544940e-05   \n",
      "Q(\"reference wesTMB\")   604056.197198   1.0  564.002187  1.663160e-35   \n",
      "Residual                 76042.240626  71.0         NaN           NaN   \n",
      "\n",
      "                          eta_sq  \n",
      "C(panel_id)             0.286869  \n",
      "C(product)              0.084373  \n",
      "C(panel_id):C(product)  0.231965  \n",
      "Q(\"reference wesTMB\")   0.352427  \n",
      "Residual                0.044366  \n",
      "type3: panel_id * product + sample + rTMB\n",
      "                               sum_sq    df           F        PR(>F)  \\\n",
      "Intercept                   47.867366   1.0    0.044693  8.331745e-01   \n",
      "C(panel_id)              65834.326544  17.0    3.615821  6.896689e-05   \n",
      "C(product)               63869.381434   5.0   11.926861  2.158167e-08   \n",
      "C(panel_id):C(product)  397586.188188  85.0    4.367328  1.544940e-05   \n",
      "Q(\"reference wesTMB\")   604056.197198   1.0  564.002187  1.663160e-35   \n",
      "Residual                 76042.240626  71.0         NaN           NaN   \n",
      "\n",
      "                          eta_sq  \n",
      "Intercept               0.000040  \n",
      "C(panel_id)             0.054524  \n",
      "C(product)              0.052897  \n",
      "C(panel_id):C(product)  0.329281  \n",
      "Q(\"reference wesTMB\")   0.500280  \n",
      "Residual                0.062978  \n",
      "type2: panel_id + product + sample + rTMB\n",
      "                              sum_sq    df           F        PR(>F)    eta_sq\n",
      "C(panel_id)             65834.326544  17.0    3.615821  6.896689e-05  0.081297\n",
      "C(product)              63869.381434   5.0   11.926861  2.158167e-08  0.078870\n",
      "Q(\"reference wesTMB\")  604056.197198   1.0  564.002187  1.663160e-35  0.745931\n",
      "Residual                76042.240626  71.0         NaN           NaN  0.093902\n",
      "type3: panel_id + product + sample + rTMB\n",
      "                              sum_sq    df           F        PR(>F)    eta_sq\n",
      "Intercept                  47.867366   1.0    0.044693  8.331745e-01  0.000059\n",
      "C(panel_id)             65834.326544  17.0    3.615821  6.896689e-05  0.081292\n",
      "C(product)              63869.381434   5.0   11.926861  2.158167e-08  0.078866\n",
      "Q(\"reference wesTMB\")  604056.197198   1.0  564.002187  1.663160e-35  0.745887\n",
      "Residual                76042.240626  71.0         NaN           NaN  0.093897\n"
     ]
    }
   ],
   "source": [
    "only_abcde = df.query('group != \"other\"').copy(deep=True)\n",
    "only_abcde_tp2_model = ols('wesTMB ~ C(panel_id) * C(product) + Q(\"reference wesTMB\")', data=only_abcde).fit()\n",
    "only_abcde_tp2_ancova_table = sm.stats.anova_lm(only_abcde_tp2_model, typ=2)\n",
    "only_abcde_tp2_ancova_table['eta_sq'] = only_abcde_tp2_ancova_table['sum_sq'] / only_abcde_tp2_ancova_table['sum_sq'].sum()\n",
    "print('type2: panel_id * product + sample + rTMB')\n",
    "print(only_abcde_tp2_ancova_table)\n",
    "\n",
    "only_abcde_tp3_model = ols('wesTMB ~ C(panel_id) * C(product) + Q(\"reference wesTMB\")', data=only_abcde).fit()\n",
    "only_abcde_tp3_ancova_table = sm.stats.anova_lm(only_abcde_tp3_model, typ=3)\n",
    "only_abcde_tp3_ancova_table['eta_sq'] = only_abcde_tp3_ancova_table['sum_sq'] / only_abcde_tp3_ancova_table['sum_sq'].sum()\n",
    "print('type3: panel_id * product + sample + rTMB')\n",
    "print(only_abcde_tp3_ancova_table)\n",
    "\n",
    "only_abcde_tp2_model2 = ols('wesTMB ~ C(panel_id) + C(product) + Q(\"reference wesTMB\")', data=only_abcde).fit()\n",
    "only_abcde_tp2_ancova_table2 = sm.stats.anova_lm(only_abcde_tp2_model2, typ=2)\n",
    "only_abcde_tp2_ancova_table2['eta_sq'] = only_abcde_tp2_ancova_table2['sum_sq'] / only_abcde_tp2_ancova_table2['sum_sq'].sum()\n",
    "print('type2: panel_id + product + sample + rTMB')\n",
    "print(only_abcde_tp2_ancova_table2)\n",
    "\n",
    "only_abcde_tp3_model2 = ols('wesTMB ~ C(panel_id) + C(product) + Q(\"reference wesTMB\")', data=only_abcde).fit()\n",
    "only_abcde_tp3_ancova_table2 = sm.stats.anova_lm(only_abcde_tp3_model2, typ=3)\n",
    "only_abcde_tp3_ancova_table2['eta_sq'] = only_abcde_tp3_ancova_table2['sum_sq'] / only_abcde_tp3_ancova_table2['sum_sq'].sum()\n",
    "print('type3: panel_id + product + sample + rTMB')\n",
    "print(only_abcde_tp3_ancova_table2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "type2: panel_id * product + sample + rTMB\n",
      "                          df            F        PR(>F)    eta_sq\n",
      "C(panel_id)             49.0    37.596391  1.410286e-14  0.363417\n",
      "C(group)                 6.0    32.038802  8.660349e-13  0.037922\n",
      "C(panel_id):C(group)   294.0     4.285354  1.094826e-07  0.248541\n",
      "Q(\"reference wesTMB\")    1.0  1575.814965  1.669817e-96  0.310863\n",
      "Residual               199.0          NaN           NaN  0.039257\n",
      "type3: panel_id * product + sample + rTMB\n",
      "                          df            F        PR(>F)    eta_sq\n",
      "Intercept                1.0     0.956487  3.292610e-01  0.000291\n",
      "C(panel_id)             49.0     3.677393  4.872544e-11  0.054751\n",
      "C(group)                 6.0    12.548069  5.517090e-12  0.022876\n",
      "C(panel_id):C(group)   294.0     4.285354  1.094826e-07  0.382813\n",
      "Q(\"reference wesTMB\")    1.0  1575.814965  1.669817e-96  0.478804\n",
      "Residual               199.0          NaN           NaN  0.060465\n",
      "type2: panel_id + product + sample + rTMB\n",
      "                         df           F        PR(>F)    eta_sq\n",
      "C(panel_id)            17.0    3.615821  6.896689e-05  0.081297\n",
      "C(product)              5.0   11.926861  2.158167e-08  0.078870\n",
      "Q(\"reference wesTMB\")   1.0  564.002187  1.663160e-35  0.745931\n",
      "Residual               71.0         NaN           NaN  0.093902\n",
      "type3: panel_id + product + sample + rTMB\n",
      "                         df           F        PR(>F)    eta_sq\n",
      "Intercept               1.0    0.044693  8.331745e-01  0.000059\n",
      "C(panel_id)            17.0    3.615821  6.896689e-05  0.081292\n",
      "C(product)              5.0   11.926861  2.158167e-08  0.078866\n",
      "Q(\"reference wesTMB\")   1.0  564.002187  1.663160e-35  0.745887\n",
      "Residual               71.0         NaN           NaN  0.093897\n"
     ]
    }
   ],
   "source": [
    "group_tp2_model = ols('wesTMB ~ C(panel_id) * C(group) + Q(\"reference wesTMB\")', data=df).fit()\n",
    "group_tp2_ancova_table = sm.stats.anova_lm(group_tp2_model, typ=2)\n",
    "group_tp2_ancova_table['eta_sq'] = group_tp2_ancova_table['sum_sq'] / group_tp2_ancova_table['sum_sq'].sum()\n",
    "print('type2: panel_id * product + sample + rTMB')\n",
    "print(group_tp2_ancova_table[['df', 'F', 'PR(>F)', 'eta_sq']])\n",
    "\n",
    "group_tp3_model = ols('wesTMB ~ C(panel_id) * C(group) + Q(\"reference wesTMB\")', data=df).fit()\n",
    "group_tp3_ancova_table = sm.stats.anova_lm(group_tp3_model, typ=3)\n",
    "group_tp3_ancova_table['eta_sq'] = group_tp3_ancova_table['sum_sq'] / group_tp3_ancova_table['sum_sq'].sum()\n",
    "print('type3: panel_id * product + sample + rTMB')\n",
    "print(group_tp3_ancova_table[['df', 'F', 'PR(>F)', 'eta_sq']])\n",
    "\n",
    "group_tp2_model2 = ols('wesTMB ~ C(panel_id) + C(group) + Q(\"reference wesTMB\")', data=df).fit()\n",
    "group_tp2_ancova_table2 = sm.stats.anova_lm(group_tp2_model2, typ=2)\n",
    "group_tp2_ancova_table2['eta_sq'] = group_tp2_ancova_table2['sum_sq'] / group_tp2_ancova_table2['sum_sq'].sum()\n",
    "print('type2: panel_id + product + sample + rTMB')\n",
    "print(only_abcde_tp2_ancova_table2[['df', 'F', 'PR(>F)', 'eta_sq']])\n",
    "\n",
    "group_tp3_model2 = ols('wesTMB ~ C(panel_id) + C(group) + Q(\"reference wesTMB\")', data=df).fit()\n",
    "group_tp3_ancova_table2 = sm.stats.anova_lm(group_tp3_model2, typ=3)\n",
    "group_tp3_ancova_table2['eta_sq'] = group_tp3_ancova_table2['sum_sq'] / group_tp3_ancova_table2['sum_sq'].sum()\n",
    "print('type3: panel_id + product + sample + rTMB')\n",
    "print(only_abcde_tp3_ancova_table2[['df', 'F', 'PR(>F)', 'eta_sq']])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>index</th>\n",
       "      <th>df</th>\n",
       "      <th>F</th>\n",
       "      <th>PR(&gt;F)</th>\n",
       "      <th>eta_sq</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Intercept</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.956487</td>\n",
       "      <td>3.292610e-01</td>\n",
       "      <td>0.000291</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>C(panel_id)</td>\n",
       "      <td>49.0</td>\n",
       "      <td>3.677393</td>\n",
       "      <td>4.872544e-11</td>\n",
       "      <td>0.054751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>C(group)</td>\n",
       "      <td>6.0</td>\n",
       "      <td>12.548069</td>\n",
       "      <td>5.517090e-12</td>\n",
       "      <td>0.022876</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>C(panel_id):C(group)</td>\n",
       "      <td>294.0</td>\n",
       "      <td>4.285354</td>\n",
       "      <td>1.094826e-07</td>\n",
       "      <td>0.382813</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Q(\"reference wesTMB\")</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1575.814965</td>\n",
       "      <td>1.669817e-96</td>\n",
       "      <td>0.478804</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Residual</td>\n",
       "      <td>199.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.060465</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   index     df            F        PR(>F)    eta_sq\n",
       "0              Intercept    1.0     0.956487  3.292610e-01  0.000291\n",
       "1            C(panel_id)   49.0     3.677393  4.872544e-11  0.054751\n",
       "2               C(group)    6.0    12.548069  5.517090e-12  0.022876\n",
       "3   C(panel_id):C(group)  294.0     4.285354  1.094826e-07  0.382813\n",
       "4  Q(\"reference wesTMB\")    1.0  1575.814965  1.669817e-96  0.478804\n",
       "5               Residual  199.0          NaN           NaN  0.060465"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "group_tp3_ancova_table[['df', 'F', 'PR(>F)', 'eta_sq']].copy(deep=True).reset_index(drop=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>factor</th>\n",
       "      <th>F</th>\n",
       "      <th>PR(&gt;F)</th>\n",
       "      <th>eta_sq</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Intercept</td>\n",
       "      <td>0.956487</td>\n",
       "      <td>3.292610e-01</td>\n",
       "      <td>0.000291</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>C(panel_id)</td>\n",
       "      <td>3.677393</td>\n",
       "      <td>4.872544e-11</td>\n",
       "      <td>0.054751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>C(group)</td>\n",
       "      <td>12.548069</td>\n",
       "      <td>5.517090e-12</td>\n",
       "      <td>0.022876</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>C(panel_id):C(group)</td>\n",
       "      <td>4.285354</td>\n",
       "      <td>1.094826e-07</td>\n",
       "      <td>0.382813</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Q(\"reference wesTMB\")</td>\n",
       "      <td>1575.814965</td>\n",
       "      <td>1.669817e-96</td>\n",
       "      <td>0.478804</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Residual</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.060465</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  factor            F        PR(>F)    eta_sq\n",
       "0              Intercept     0.956487  3.292610e-01  0.000291\n",
       "1            C(panel_id)     3.677393  4.872544e-11  0.054751\n",
       "2               C(group)    12.548069  5.517090e-12  0.022876\n",
       "3   C(panel_id):C(group)     4.285354  1.094826e-07  0.382813\n",
       "4  Q(\"reference wesTMB\")  1575.814965  1.669817e-96  0.478804\n",
       "5               Residual          NaN           NaN  0.060465"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "group_tp3_ancova_table[['F', 'PR(>F)', 'eta_sq']].copy(deep=True).reset_index(drop=False).rename(columns={'index': 'factor'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:                 wesTMB   R-squared:                       0.898\n",
      "Model:                            OLS   Adj. R-squared:                  0.872\n",
      "Method:                 Least Squares   F-statistic:                     34.75\n",
      "Date:                Wed, 22 Nov 2023   Prob (F-statistic):           4.36e-28\n",
      "Time:                        21:06:43   Log-Likelihood:                -430.97\n",
      "No. Observations:                  90   AIC:                             899.9\n",
      "Df Residuals:                      71   BIC:                             947.4\n",
      "Df Model:                          18                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "=========================================================================================================\n",
      "                                            coef    std err          t      P>|t|      [0.025      0.975]\n",
      "---------------------------------------------------------------------------------------------------------\n",
      "Intercept                                 3.2293     15.275      0.211      0.833     -27.229      33.687\n",
      "C(panel_id)[T.p07b]                   -1.663e-12     20.698  -8.03e-14      1.000     -41.271      41.271\n",
      "C(panel_id)[T.p07c]                      -1.7803     20.698     -0.086      0.932     -43.051      39.490\n",
      "C(panel_id)[T.p07d]                      -0.5342     20.698     -0.026      0.979     -41.805      40.736\n",
      "C(panel_id)[T.p10a]                      -2.2655      6.934     -0.327      0.745     -16.092      11.561\n",
      "C(panel_id)[T.p10b]                      -1.4088      6.934     -0.203      0.840     -15.236      12.418\n",
      "C(panel_id)[T.p10c]                      -1.9447      6.934     -0.280      0.780     -15.772      11.882\n",
      "C(panel_id)[T.p10d]                      -1.5202      6.934     -0.219      0.827     -15.347      12.307\n",
      "C(panel_id)[T.p10e]                      -1.4088      6.934     -0.203      0.840     -15.236      12.418\n",
      "C(panel_id)[T.p23a]                      -8.7784      6.865     -1.279      0.205     -22.466       4.910\n",
      "C(panel_id)[T.p23b]                       1.8221      6.865      0.265      0.791     -11.866      15.510\n",
      "C(panel_id)[T.p23c]                      -0.7194      6.865     -0.105      0.917     -14.407      12.969\n",
      "C(panel_id)[T.p32a]                      -9.8858      6.845     -1.444      0.153     -23.535       3.763\n",
      "C(panel_id)[T.p32b]                      -6.5248      6.845     -0.953      0.344     -20.174       7.124\n",
      "C(panel_id)[T.p35a]                      -4.3858      6.845     -0.641      0.524     -18.035       9.263\n",
      "C(panel_id)[T.p35b]                      -9.5397      6.845     -1.394      0.168     -23.189       4.109\n",
      "C(panel_id)[T.p37a]                      16.2022      6.845      2.367      0.021       2.553      29.851\n",
      "C(panel_id)[T.p37b]                      16.2022      6.845      2.367      0.021       2.553      29.851\n",
      "C(product)[T.p10]                        -8.5479     11.452     -0.746      0.458     -31.382      14.286\n",
      "C(product)[T.p23]                        -7.6756     10.140     -0.757      0.452     -27.894      12.543\n",
      "C(product)[T.p32]                       -16.4107      8.962     -1.831      0.071     -34.281       1.460\n",
      "C(product)[T.p35]                       -13.9255      8.962     -1.554      0.125     -31.796       3.945\n",
      "C(product)[T.p37]                        32.4043      8.962      3.616      0.001      14.534      50.275\n",
      "C(panel_id)[T.p07b]:C(product)[T.p10]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p07c]:C(product)[T.p10]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p07d]:C(product)[T.p10]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10a]:C(product)[T.p10]    -2.2655      6.934     -0.327      0.745     -16.092      11.561\n",
      "C(panel_id)[T.p10b]:C(product)[T.p10]    -1.4088      6.934     -0.203      0.840     -15.236      12.418\n",
      "C(panel_id)[T.p10c]:C(product)[T.p10]    -1.9447      6.934     -0.280      0.780     -15.772      11.882\n",
      "C(panel_id)[T.p10d]:C(product)[T.p10]    -1.5202      6.934     -0.219      0.827     -15.347      12.307\n",
      "C(panel_id)[T.p10e]:C(product)[T.p10]    -1.4088      6.934     -0.203      0.840     -15.236      12.418\n",
      "C(panel_id)[T.p23a]:C(product)[T.p10]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p23b]:C(product)[T.p10]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p23c]:C(product)[T.p10]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p32a]:C(product)[T.p10]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p32b]:C(product)[T.p10]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p35a]:C(product)[T.p10]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p35b]:C(product)[T.p10]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p37a]:C(product)[T.p10]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p37b]:C(product)[T.p10]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p07b]:C(product)[T.p23]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p07c]:C(product)[T.p23]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p07d]:C(product)[T.p23]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10a]:C(product)[T.p23]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10b]:C(product)[T.p23]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10c]:C(product)[T.p23]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10d]:C(product)[T.p23]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10e]:C(product)[T.p23]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p23a]:C(product)[T.p23]    -8.7784      6.865     -1.279      0.205     -22.466       4.910\n",
      "C(panel_id)[T.p23b]:C(product)[T.p23]     1.8221      6.865      0.265      0.791     -11.866      15.510\n",
      "C(panel_id)[T.p23c]:C(product)[T.p23]    -0.7194      6.865     -0.105      0.917     -14.407      12.969\n",
      "C(panel_id)[T.p32a]:C(product)[T.p23]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p32b]:C(product)[T.p23]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p35a]:C(product)[T.p23]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p35b]:C(product)[T.p23]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p37a]:C(product)[T.p23]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p37b]:C(product)[T.p23]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p07b]:C(product)[T.p32]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p07c]:C(product)[T.p32]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p07d]:C(product)[T.p32]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10a]:C(product)[T.p32]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10b]:C(product)[T.p32]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10c]:C(product)[T.p32]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10d]:C(product)[T.p32]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10e]:C(product)[T.p32]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p23a]:C(product)[T.p32]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p23b]:C(product)[T.p32]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p23c]:C(product)[T.p32]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p32a]:C(product)[T.p32]    -9.8858      6.845     -1.444      0.153     -23.535       3.763\n",
      "C(panel_id)[T.p32b]:C(product)[T.p32]    -6.5248      6.845     -0.953      0.344     -20.174       7.124\n",
      "C(panel_id)[T.p35a]:C(product)[T.p32]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p35b]:C(product)[T.p32]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p37a]:C(product)[T.p32]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p37b]:C(product)[T.p32]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p07b]:C(product)[T.p35]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p07c]:C(product)[T.p35]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p07d]:C(product)[T.p35]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10a]:C(product)[T.p35]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10b]:C(product)[T.p35]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10c]:C(product)[T.p35]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10d]:C(product)[T.p35]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10e]:C(product)[T.p35]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p23a]:C(product)[T.p35]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p23b]:C(product)[T.p35]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p23c]:C(product)[T.p35]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p32a]:C(product)[T.p35]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p32b]:C(product)[T.p35]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p35a]:C(product)[T.p35]    -4.3858      6.845     -0.641      0.524     -18.035       9.263\n",
      "C(panel_id)[T.p35b]:C(product)[T.p35]    -9.5397      6.845     -1.394      0.168     -23.189       4.109\n",
      "C(panel_id)[T.p37a]:C(product)[T.p35]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p37b]:C(product)[T.p35]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p07b]:C(product)[T.p37]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p07c]:C(product)[T.p37]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p07d]:C(product)[T.p37]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10a]:C(product)[T.p37]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10b]:C(product)[T.p37]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10c]:C(product)[T.p37]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10d]:C(product)[T.p37]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p10e]:C(product)[T.p37]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p23a]:C(product)[T.p37]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p23b]:C(product)[T.p37]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p23c]:C(product)[T.p37]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p32a]:C(product)[T.p37]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p32b]:C(product)[T.p37]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p35a]:C(product)[T.p37]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p35b]:C(product)[T.p37]          0          0        nan        nan           0           0\n",
      "C(panel_id)[T.p37a]:C(product)[T.p37]    16.2022      6.845      2.367      0.021       2.553      29.851\n",
      "C(panel_id)[T.p37b]:C(product)[T.p37]    16.2022      6.845      2.367      0.021       2.553      29.851\n",
      "Q(\"reference wesTMB\")                     1.3905      0.059     23.749      0.000       1.274       1.507\n",
      "==============================================================================\n",
      "Omnibus:                       36.877   Durbin-Watson:                   1.677\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):              157.521\n",
      "Skew:                           1.178   Prob(JB):                     6.23e-35\n",
      "Kurtosis:                       9.038   Cond. No.                     4.63e+16\n",
      "==============================================================================\n",
      "\n",
      "Notes:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n",
      "[2] The input rank is higher than the number of observations.\n",
      "[3] The smallest eigenvalue is 3.8e-28. This might indicate that there are\n",
      "strong multicollinearity problems or that the design matrix is singular.\n"
     ]
    }
   ],
   "source": [
    "print(only_abcde_tp3_model.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "(only_abcde_tp3_ancova_table\n",
    " [['F', 'PR(>F)', 'eta_sq']]\n",
    " .copy(deep=True)\n",
    " .reset_index(drop=False)\n",
    " .rename(columns={'index': 'factor'})\n",
    " .to_csv('./only_abcde_anova.tsv',\n",
    "         sep='\\t', lineterminator='\\n', encoding='utf-8', index=False))\n",
    "\n",
    "(group_tp3_ancova_table\n",
    " [['F', 'PR(>F)', 'eta_sq']]\n",
    " .copy(deep=True)\n",
    " .reset_index(drop=False)\n",
    " .rename(columns={'index': 'factor'})\n",
    " .to_csv('./group_anova.tsv',\n",
    "         sep='\\t', lineterminator='\\n', encoding='utf-8', index=False))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "data",
   "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": "3.11.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}