{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Univariate analyses\n",
    "\n",
    "Perform some univariate analyses to establish estimates of \"ground truth\" feature importance, after controlling for site effects."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "import pickle    \n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "sns.set_style('whitegrid')\n",
    "    \n",
    "from armed.models.lme import MixedLogisticGLM\n",
    "from scipy import stats"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Load pre-defined k-folds."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('./10x10_kfolds_top20sites.pkl', 'rb') as f:\n",
    "    kfolds = pickle.load(f)\n",
    "    \n",
    "dfData = kfolds.x\n",
    "arrLabels = kfolds.y.squeeze()\n",
    "dfSite = kfolds.z"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For the categorical features, compute relative risk stratified by site. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>Site</th>\n",
       "      <th>Variable</th>\n",
       "      <th>Relative Risk</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>ApoE4 - 1 allele</td>\n",
       "      <td>1.142857</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>21</td>\n",
       "      <td>ApoE4 - 1 allele</td>\n",
       "      <td>2.250000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>23</td>\n",
       "      <td>ApoE4 - 1 allele</td>\n",
       "      <td>0.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>27</td>\n",
       "      <td>ApoE4 - 1 allele</td>\n",
       "      <td>2.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>31</td>\n",
       "      <td>ApoE4 - 1 allele</td>\n",
       "      <td>1.178571</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>104</th>\n",
       "      <td>127</td>\n",
       "      <td>Marital Status - Widowed</td>\n",
       "      <td>4.666667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>105</th>\n",
       "      <td>128</td>\n",
       "      <td>Marital Status - Widowed</td>\n",
       "      <td>4.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>106</th>\n",
       "      <td>130</td>\n",
       "      <td>Marital Status - Widowed</td>\n",
       "      <td>1.333333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>107</th>\n",
       "      <td>137</td>\n",
       "      <td>Marital Status - Widowed</td>\n",
       "      <td>2.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>108</th>\n",
       "      <td>141</td>\n",
       "      <td>Marital Status - Widowed</td>\n",
       "      <td>0.916667</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>109 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    Site                  Variable  Relative Risk\n",
       "0      2          ApoE4 - 1 allele       1.142857\n",
       "1     21          ApoE4 - 1 allele       2.250000\n",
       "2     23          ApoE4 - 1 allele       0.666667\n",
       "3     27          ApoE4 - 1 allele       2.666667\n",
       "4     31          ApoE4 - 1 allele       1.178571\n",
       "..   ...                       ...            ...\n",
       "104  127  Marital Status - Widowed       4.666667\n",
       "105  128  Marital Status - Widowed       4.000000\n",
       "106  130  Marital Status - Widowed       1.333333\n",
       "107  137  Marital Status - Widowed       2.500000\n",
       "108  141  Marital Status - Widowed       0.916667\n",
       "\n",
       "[109 rows x 3 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lsVars = ['ApoE4 - 1 allele',\n",
    "         'ApoE4 - 2 alleles',\n",
    "         'Gender - Female',\n",
    "         'Ethnicity - Hispanic/Latino', \n",
    "         'Race - Am Indian/Alaskan',\n",
    "         'Race - Asian',\n",
    "         'Race - Black',\n",
    "         'Race - Multiple',\n",
    "         'Marital Status - Divorced',\n",
    "         'Marital Status - Never', \n",
    "         'Marital Status - Widowed']\n",
    "\n",
    "def contingency_table(data, labels, var):\n",
    "    nExposedConverters = np.sum((data[var] == 1) & (labels == 1))\n",
    "    nUnexposedConverters = np.sum((data[var] == 0) & (labels == 1))\n",
    "    nExposedNonconverters = np.sum((data[var] == 1) & (labels == 0))\n",
    "    nUnexposedNonconverters = np.sum((data[var] == 0) & (labels == 0))\n",
    "\n",
    "    return np.array([[nExposedConverters, nExposedNonconverters],\n",
    "                     [nUnexposedConverters, nUnexposedNonconverters]])\n",
    "\n",
    "dfRR = pd.DataFrame(columns=['Site', 'Variable', 'Relative Risk'])\n",
    "iRow = 0\n",
    "\n",
    "# for strExposure, strUnexposed in lsComparisons:\n",
    "for strVar in lsVars:\n",
    "\n",
    "    lsSiteContingency = []\n",
    "    for iSite in range(dfSite.shape[1]):\n",
    "        dfDataSite = dfData.loc[dfSite.iloc[:, iSite] == 1]\n",
    "        arrLabelsSite = arrLabels[dfSite.iloc[:, iSite] == 1]\n",
    "        arrSiteContingency = contingency_table(dfDataSite, arrLabelsSite, strVar)\n",
    "        # Omit sites with zero exposed, unexposed, or unexposed converters\n",
    "        if (arrSiteContingency[0, :].sum() > 0) & (arrSiteContingency[1, :].sum() > 0) & (arrSiteContingency[1, 0] > 0):\n",
    "            rr = (arrSiteContingency[0, 0] / arrSiteContingency[0, :].sum()) \\\n",
    "                / (arrSiteContingency[1, 0] / arrSiteContingency[1, :].sum())\n",
    "                \n",
    "            dfRR.loc[iRow] = [dfSite.columns[iSite], strVar, rr]\n",
    "            iRow += 1\n",
    "                    \n",
    "dfRR"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plot the inter-site variance of each feature's relative risk."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'Relative Risk s.d.')"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 360x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dfRRVar = dfRR.groupby('Variable').std()\n",
    "fig, ax = plt.subplots(figsize=(5, 10))\n",
    "ax.barh(y=dfRRVar.index, width=dfRRVar['Relative Risk'])\n",
    "ax.set_ylabel('Variable')\n",
    "ax.set_xlabel('Relative Risk s.d.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "dfRRVar.to_csv('./categorical_relativerisk_sd.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For the continuous features, fit univariate logistic mixed effects models with site-specific random slopes and intercepts. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>Coef</th>\n",
       "      <th>S.D.</th>\n",
       "      <th>Random slope S.D.</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Years of education</th>\n",
       "      <td>-0.0526124</td>\n",
       "      <td>0.116909</td>\n",
       "      <td>0.071996</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Age</th>\n",
       "      <td>0.304627</td>\n",
       "      <td>0.120916</td>\n",
       "      <td>0.114411</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CDR-SB</th>\n",
       "      <td>1.08176</td>\n",
       "      <td>0.134881</td>\n",
       "      <td>0.229683</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ADAS11</th>\n",
       "      <td>1.41601</td>\n",
       "      <td>0.153318</td>\n",
       "      <td>0.241421</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ADAS13</th>\n",
       "      <td>1.59391</td>\n",
       "      <td>0.159107</td>\n",
       "      <td>0.279189</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ADAS Word Recognition</th>\n",
       "      <td>1.30647</td>\n",
       "      <td>0.140108</td>\n",
       "      <td>0.270639</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MMSE</th>\n",
       "      <td>-0.769754</td>\n",
       "      <td>0.122371</td>\n",
       "      <td>0.216995</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RAVLT immediate recall</th>\n",
       "      <td>-1.55386</td>\n",
       "      <td>0.160542</td>\n",
       "      <td>0.0561487</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RAVLT learning score</th>\n",
       "      <td>-0.813063</td>\n",
       "      <td>0.134996</td>\n",
       "      <td>0.170272</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RAVLT forgetting score</th>\n",
       "      <td>0.270179</td>\n",
       "      <td>0.122214</td>\n",
       "      <td>0.119079</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RAVLT % forgetting</th>\n",
       "      <td>1.00354</td>\n",
       "      <td>0.13377</td>\n",
       "      <td>0.214415</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Trails B score</th>\n",
       "      <td>0.908189</td>\n",
       "      <td>0.132521</td>\n",
       "      <td>0.18401</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Functional Activities Questionnaire</th>\n",
       "      <td>1.20108</td>\n",
       "      <td>0.141569</td>\n",
       "      <td>0.235654</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mPACC with Digit symbol substitution</th>\n",
       "      <td>-1.61594</td>\n",
       "      <td>0.153289</td>\n",
       "      <td>0.407114</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mPACC with Trails B</th>\n",
       "      <td>-1.60095</td>\n",
       "      <td>0.152856</td>\n",
       "      <td>0.356037</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Ventricular volume</th>\n",
       "      <td>0.134191</td>\n",
       "      <td>0.118779</td>\n",
       "      <td>0.0563372</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Hippocampal volume</th>\n",
       "      <td>-0.948827</td>\n",
       "      <td>0.144906</td>\n",
       "      <td>0.233789</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Whole brain volume</th>\n",
       "      <td>-0.457185</td>\n",
       "      <td>0.125555</td>\n",
       "      <td>0.102163</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Entorhinal volume</th>\n",
       "      <td>-0.861474</td>\n",
       "      <td>0.139893</td>\n",
       "      <td>0.100432</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fusiform volume</th>\n",
       "      <td>-0.851055</td>\n",
       "      <td>0.149274</td>\n",
       "      <td>0.0566998</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Mid temporal volume</th>\n",
       "      <td>-0.993274</td>\n",
       "      <td>0.150967</td>\n",
       "      <td>0.183983</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Intracranial volume</th>\n",
       "      <td>-0.0200011</td>\n",
       "      <td>0.118089</td>\n",
       "      <td>0.0869825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PET FDG</th>\n",
       "      <td>-1.46815</td>\n",
       "      <td>0.176709</td>\n",
       "      <td>0.359235</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CSF phosphorylated tau</th>\n",
       "      <td>1.18846</td>\n",
       "      <td>0.168199</td>\n",
       "      <td>0.197653</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CSF tau</th>\n",
       "      <td>1.19336</td>\n",
       "      <td>0.169094</td>\n",
       "      <td>0.173539</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CSF beta-amyloid</th>\n",
       "      <td>-0.965536</td>\n",
       "      <td>0.165458</td>\n",
       "      <td>0.182874</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                           Coef      S.D. Random slope S.D.\n",
       "Years of education                   -0.0526124  0.116909          0.071996\n",
       "Age                                    0.304627  0.120916          0.114411\n",
       "CDR-SB                                  1.08176  0.134881          0.229683\n",
       "ADAS11                                  1.41601  0.153318          0.241421\n",
       "ADAS13                                  1.59391  0.159107          0.279189\n",
       "ADAS Word Recognition                   1.30647  0.140108          0.270639\n",
       "MMSE                                  -0.769754  0.122371          0.216995\n",
       "RAVLT immediate recall                 -1.55386  0.160542         0.0561487\n",
       "RAVLT learning score                  -0.813063  0.134996          0.170272\n",
       "RAVLT forgetting score                 0.270179  0.122214          0.119079\n",
       "RAVLT % forgetting                      1.00354   0.13377          0.214415\n",
       "Trails B score                         0.908189  0.132521           0.18401\n",
       "Functional Activities Questionnaire     1.20108  0.141569          0.235654\n",
       "mPACC with Digit symbol substitution   -1.61594  0.153289          0.407114\n",
       "mPACC with Trails B                    -1.60095  0.152856          0.356037\n",
       "Ventricular volume                     0.134191  0.118779         0.0563372\n",
       "Hippocampal volume                    -0.948827  0.144906          0.233789\n",
       "Whole brain volume                    -0.457185  0.125555          0.102163\n",
       "Entorhinal volume                     -0.861474  0.139893          0.100432\n",
       "Fusiform volume                       -0.851055  0.149274         0.0566998\n",
       "Mid temporal volume                   -0.993274  0.150967          0.183983\n",
       "Intracranial volume                  -0.0200011  0.118089         0.0869825\n",
       "PET FDG                                -1.46815  0.176709          0.359235\n",
       "CSF phosphorylated tau                  1.18846  0.168199          0.197653\n",
       "CSF tau                                 1.19336  0.169094          0.173539\n",
       "CSF beta-amyloid                      -0.965536  0.165458          0.182874"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lsVars = [\n",
    "       'Years of education', 'Age', 'CDR-SB', 'ADAS11', 'ADAS13',\n",
    "       'ADAS Word Recognition', 'MMSE', 'RAVLT immediate recall',\n",
    "       'RAVLT learning score', 'RAVLT forgetting score', 'RAVLT % forgetting',\n",
    "       'Trails B score', 'Functional Activities Questionnaire',\n",
    "       'mPACC with Digit symbol substitution', 'mPACC with Trails B',\n",
    "       'Ventricular volume', 'Hippocampal volume', 'Whole brain volume',\n",
    "       'Entorhinal volume', 'Fusiform volume', 'Mid temporal volume',\n",
    "       'Intracranial volume', 'PET FDG', 'CSF phosphorylated tau', 'CSF tau',\n",
    "       'CSF beta-amyloid']\n",
    "\n",
    "dfCoefs = pd.DataFrame(columns=['Coef', 'S.D.', 'Random slope S.D.'], index=lsVars)\n",
    "\n",
    "for iRow, strVariable in enumerate(lsVars):\n",
    "\n",
    "    strFormula = 'Conversion ~ Var'\n",
    "    dfLogRegData = dfData[[strVariable]].copy()\n",
    "    dfLogRegData.columns = ['Var']\n",
    "    dfLogRegData['Conversion'] = arrLabels\n",
    "    dfLogRegData['Site'] = dfSite.idxmax(axis=1)\n",
    "    \n",
    "    # Drop missing values\n",
    "    dfLogRegData.dropna(inplace=True)\n",
    "    # Drop sites with 100% or 0% conversion\n",
    "    dfSiteConversionRate = dfLogRegData.groupby('Site')['Conversion'].mean()\n",
    "    dfBadSites = dfSiteConversionRate.loc[(dfSiteConversionRate == 0) | (dfSiteConversionRate == 1)]\n",
    "    dfLogRegData = dfLogRegData.loc[~dfLogRegData['Site'].isin(dfBadSites.index)]\n",
    "\n",
    "    dfLogRegData['Var'] -= dfLogRegData['Var'].mean()\n",
    "    dfLogRegData['Var'] /= dfLogRegData['Var'].std()\n",
    "\n",
    "    # Logistic mixed effects model with site-specific random slope and intercept\n",
    "    mixedeffects = MixedLogisticGLM('Conversion ~ Var', \n",
    "                                    {'Site_slope': '0 + C(Site):Var',\n",
    "                                        'Site_intercept': 'C(Site)'},\n",
    "                                    'Site')\n",
    "                \n",
    "    mixedeffects.fit(dfLogRegData)\n",
    "    coef = mixedeffects.result.fe_mean[1]\n",
    "    sd = mixedeffects.result.fe_sd[1]\n",
    "        \n",
    "    # Compute random slope s.d.\n",
    "    dfRE = mixedeffects.result.random_effects()\n",
    "    dfRESlopes = dfRE.loc[['Var' in x for x in dfRE.index]]\n",
    "    re_sd = dfRESlopes['Mean'].std()\n",
    "\n",
    "    dfCoefs.loc[strVariable] = [coef, sd, re_sd]\n",
    "    \n",
    "dfCoefs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plot the inter-site variance of each feature's random slopes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'Random slope s.d.')"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 360x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(5, 10))\n",
    "ax.barh(y=dfCoefs.index, width=dfCoefs['Random slope S.D.'])\n",
    "ax.set_ylabel('Variable')\n",
    "ax.set_xlabel('Random slope s.d.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "dfCoefs.to_csv('./continuous_univariate_lme.csv')"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "a5dc4aa0b7969a772c348262b9338538f97e30fb3762c91cf138c2ab2be38e85"
  },
  "kernelspec": {
   "display_name": "Python 3.8.5 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": "3.8.5"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}