Pathogen Prediction-LR.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predicting Pathogen from RNAseq data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from math import pow\n",
    "from sklearn.feature_selection import RFECV, RFE\n",
    "from sklearn.model_selection import train_test_split, cross_val_score, GridSearchCV\n",
    "from sklearn.svm import LinearSVC, LinearSVR\n",
    "\n",
    "random_state = 7"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Complete data set has 132 samples and 25342 features.\n",
      "Training set has 105 samples. Testing set has 27 samples.\n"
     ]
    }
   ],
   "source": [
    "patient_groups=[\"control\", \"viral\", \"bacterial\", \"fungal\"]\n",
    "group_id = lambda name: patient_groups.index(name)\n",
    "\n",
    "X = pd.DataFrame.from_csv(\"combineSV_WTcpmtable_v2.txt\", sep=\"\\s+\").T\n",
    "# + [group_id(\"fungal\")] * 10 \\\n",
    "y = [group_id(\"bacterial\")] * 29 \\\n",
    "    + [group_id(\"viral\")] * 42 \\\n",
    "    + [group_id(\"control\")] * 61\n",
    "    \n",
    "# Drop the fungal patients\n",
    "fv = range(29+42, 29+42+10)\n",
    "X = X.drop(X.index[fv])\n",
    "\n",
    "print \"Complete data set has %d samples and %d features.\" % (X.shape[0], X.shape[1])\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, \n",
    "                                                    random_state=random_state, \n",
    "                                                    stratify=y)\n",
    "print \"Training set has %d samples. Testing set has %d samples.\" % (len(X_train), len(X_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def print_gridsearch_results(clf):\n",
    "    print(\"Best: %f using %s\" % (clf.best_score_, clf.best_params_))\n",
    "    means = clf.cv_results_['mean_test_score']\n",
    "    stds = clf.cv_results_['std_test_score']\n",
    "    params = clf.cv_results_['params']\n",
    "    for mean, stdev, param in zip(means, stds, params):\n",
    "        print(\"%f (%f) with: %r\" % (mean, stdev, param))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RFECV(cv=4,\n",
       "   estimator=LogisticRegression(C=1.19209289551e-07, class_weight='balanced', dual=False,\n",
       "          fit_intercept=True, intercept_scaling=1, max_iter=100,\n",
       "          multi_class='ovr', n_jobs=1, penalty='l2', random_state=None,\n",
       "          solver='liblinear', tol=0.0001, verbose=0, warm_start=False),\n",
       "   n_jobs=6, scoring='neg_mean_squared_error', step=0.001, verbose=0)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.multiclass import OneVsRestClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "#parameters={'estimator__C': [pow(2, i) for i in xrange(-25, 4, 1)]}\n",
    "est = LogisticRegression(class_weight=\"balanced\", C=pow(2, -23))\n",
    "#est = LinearSVR(C=pow(2, -23))\n",
    "#clf = RFE(estimator=lr, step=0.01, n_features_to_select=1, verbose=1)\n",
    "clf = RFECV(cv=4, estimator=est, n_jobs=6, scoring='neg_mean_squared_error', step=.001, verbose=0)\n",
    "#clf = GridSearchCV(rfe, parameters, scoring='accuracy', n_jobs=8, cv=3, verbose=1)\n",
    "clf.fit(X_train, y_train)\n",
    "#print_gridsearch_results(clf)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "<h2>List of 167 genes found by RFE</h2>\n",
       "<p>Note: the NCBI link will open the target in a new window or tab.</p>\n",
       "<table>\n",
       "<tr></tr><tr>\n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=ABL2[Gene Name] AND Human[Organism]\">ABL2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=ACTB[Gene Name] AND Human[Organism]\">ACTB</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=ACTG1[Gene Name] AND Human[Organism]\">ACTG1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=ADAR[Gene Name] AND Human[Organism]\">ADAR</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=AGO1[Gene Name] AND Human[Organism]\">AGO1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=AKAP13[Gene Name] AND Human[Organism]\">AKAP13</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=ALDOA[Gene Name] AND Human[Organism]\">ALDOA</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=ARF6[Gene Name] AND Human[Organism]\">ARF6</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=ARHGAP26[Gene Name] AND Human[Organism]\">ARHGAP26</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=ATF6[Gene Name] AND Human[Organism]\">ATF6</a></td>\n",
       "    </tr><tr>\n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=B2M[Gene Name] AND Human[Organism]\">B2M</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=B4GALT5[Gene Name] AND Human[Organism]\">B4GALT5</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=BMS1[Gene Name] AND Human[Organism]\">BMS1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=BNIP3L[Gene Name] AND Human[Organism]\">BNIP3L</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=BRE[Gene Name] AND Human[Organism]\">BRE</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=C20orf194[Gene Name] AND Human[Organism]\">C20orf194</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CALM3[Gene Name] AND Human[Organism]\">CALM3</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CD74[Gene Name] AND Human[Organism]\">CD74</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CD96[Gene Name] AND Human[Organism]\">CD96</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CDR1[Gene Name] AND Human[Organism]\">CDR1</a></td>\n",
       "    </tr><tr>\n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CEP250[Gene Name] AND Human[Organism]\">CEP250</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CGN[Gene Name] AND Human[Organism]\">CGN</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CHP1[Gene Name] AND Human[Organism]\">CHP1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CLVS2[Gene Name] AND Human[Organism]\">CLVS2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CNBP[Gene Name] AND Human[Organism]\">CNBP</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CNTRL[Gene Name] AND Human[Organism]\">CNTRL</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CORO1A[Gene Name] AND Human[Organism]\">CORO1A</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CRKL[Gene Name] AND Human[Organism]\">CRKL</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CSF2RB[Gene Name] AND Human[Organism]\">CSF2RB</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CSPG4[Gene Name] AND Human[Organism]\">CSPG4</a></td>\n",
       "    </tr><tr>\n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CTSB[Gene Name] AND Human[Organism]\">CTSB</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CXCL10[Gene Name] AND Human[Organism]\">CXCL10</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CXCL8[Gene Name] AND Human[Organism]\">CXCL8</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=CXCR2[Gene Name] AND Human[Organism]\">CXCR2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=DIP2A[Gene Name] AND Human[Organism]\">DIP2A</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=DSCAM[Gene Name] AND Human[Organism]\">DSCAM</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=EEF1A1[Gene Name] AND Human[Organism]\">EEF1A1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=EEF2[Gene Name] AND Human[Organism]\">EEF2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=EIF2AK2[Gene Name] AND Human[Organism]\">EIF2AK2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=EPB41L2[Gene Name] AND Human[Organism]\">EPB41L2</a></td>\n",
       "    </tr><tr>\n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=ERICH3-AS1[Gene Name] AND Human[Organism]\">ERICH3-AS1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=ETS1[Gene Name] AND Human[Organism]\">ETS1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=EVI2B[Gene Name] AND Human[Organism]\">EVI2B</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=EXOC7[Gene Name] AND Human[Organism]\">EXOC7</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=FAM120A[Gene Name] AND Human[Organism]\">FAM120A</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=FAM192A[Gene Name] AND Human[Organism]\">FAM192A</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=FTH1[Gene Name] AND Human[Organism]\">FTH1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=FTL[Gene Name] AND Human[Organism]\">FTL</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=FXR1[Gene Name] AND Human[Organism]\">FXR1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=FYB[Gene Name] AND Human[Organism]\">FYB</a></td>\n",
       "    </tr><tr>\n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=G3BP2[Gene Name] AND Human[Organism]\">G3BP2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=GAS7[Gene Name] AND Human[Organism]\">GAS7</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=GBP5[Gene Name] AND Human[Organism]\">GBP5</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=GFAP[Gene Name] AND Human[Organism]\">GFAP</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=GLUL[Gene Name] AND Human[Organism]\">GLUL</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=GNAI2[Gene Name] AND Human[Organism]\">GNAI2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=GNB2L1[Gene Name] AND Human[Organism]\">GNB2L1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=GPX3[Gene Name] AND Human[Organism]\">GPX3</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=HDLBP[Gene Name] AND Human[Organism]\">HDLBP</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=HIVEP1[Gene Name] AND Human[Organism]\">HIVEP1</a></td>\n",
       "    </tr><tr>\n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=HLA-DRA[Gene Name] AND Human[Organism]\">HLA-DRA</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=HSP90AA1[Gene Name] AND Human[Organism]\">HSP90AA1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=IDS[Gene Name] AND Human[Organism]\">IDS</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=IFI30[Gene Name] AND Human[Organism]\">IFI30</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=IFI44L[Gene Name] AND Human[Organism]\">IFI44L</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=IFITM1[Gene Name] AND Human[Organism]\">IFITM1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=IFITM2[Gene Name] AND Human[Organism]\">IFITM2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=IFITM3[Gene Name] AND Human[Organism]\">IFITM3</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=IL1B[Gene Name] AND Human[Organism]\">IL1B</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=KCNQ1OT1[Gene Name] AND Human[Organism]\">KCNQ1OT1</a></td>\n",
       "    </tr><tr>\n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=KIAA1109[Gene Name] AND Human[Organism]\">KIAA1109</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=KLF13[Gene Name] AND Human[Organism]\">KLF13</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=KMT2D[Gene Name] AND Human[Organism]\">KMT2D</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=KSR2[Gene Name] AND Human[Organism]\">KSR2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=LOC101929439[Gene Name] AND Human[Organism]\">LOC101929439</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=LYZ[Gene Name] AND Human[Organism]\">LYZ</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=MALAT1[Gene Name] AND Human[Organism]\">MALAT1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=MALT1[Gene Name] AND Human[Organism]\">MALT1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=MAP1B[Gene Name] AND Human[Organism]\">MAP1B</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=MARCH8[Gene Name] AND Human[Organism]\">MARCH8</a></td>\n",
       "    </tr><tr>\n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=MBP[Gene Name] AND Human[Organism]\">MBP</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=MCTP2[Gene Name] AND Human[Organism]\">MCTP2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=MED13L[Gene Name] AND Human[Organism]\">MED13L</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=MMP8[Gene Name] AND Human[Organism]\">MMP8</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=MOBP[Gene Name] AND Human[Organism]\">MOBP</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=MSN[Gene Name] AND Human[Organism]\">MSN</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=MTMR6[Gene Name] AND Human[Organism]\">MTMR6</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=MUC16[Gene Name] AND Human[Organism]\">MUC16</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=MUC3A[Gene Name] AND Human[Organism]\">MUC3A</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=MX1[Gene Name] AND Human[Organism]\">MX1</a></td>\n",
       "    </tr><tr>\n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=MYCBP2[Gene Name] AND Human[Organism]\">MYCBP2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=MYO5A[Gene Name] AND Human[Organism]\">MYO5A</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=OAS1[Gene Name] AND Human[Organism]\">OAS1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=OAS2[Gene Name] AND Human[Organism]\">OAS2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=OAS3[Gene Name] AND Human[Organism]\">OAS3</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=PABPC1[Gene Name] AND Human[Organism]\">PABPC1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=PDHA1[Gene Name] AND Human[Organism]\">PDHA1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=PDZD2[Gene Name] AND Human[Organism]\">PDZD2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=PHC3[Gene Name] AND Human[Organism]\">PHC3</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=PICALM[Gene Name] AND Human[Organism]\">PICALM</a></td>\n",
       "    </tr><tr>\n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=PIK3C2B[Gene Name] AND Human[Organism]\">PIK3C2B</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=PIK3R1[Gene Name] AND Human[Organism]\">PIK3R1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=PKM[Gene Name] AND Human[Organism]\">PKM</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=PLEK[Gene Name] AND Human[Organism]\">PLEK</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=PNRC1[Gene Name] AND Human[Organism]\">PNRC1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=PRRC2B[Gene Name] AND Human[Organism]\">PRRC2B</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=PSAP[Gene Name] AND Human[Organism]\">PSAP</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=PSMF1[Gene Name] AND Human[Organism]\">PSMF1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=PTPRC[Gene Name] AND Human[Organism]\">PTPRC</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=PUM1[Gene Name] AND Human[Organism]\">PUM1</a></td>\n",
       "    </tr><tr>\n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=QKI[Gene Name] AND Human[Organism]\">QKI</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=RN7SL1[Gene Name] AND Human[Organism]\">RN7SL1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=RN7SL2[Gene Name] AND Human[Organism]\">RN7SL2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=RNF213[Gene Name] AND Human[Organism]\">RNF213</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=RPL11[Gene Name] AND Human[Organism]\">RPL11</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=RPL28[Gene Name] AND Human[Organism]\">RPL28</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=RPL5[Gene Name] AND Human[Organism]\">RPL5</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=RPL6[Gene Name] AND Human[Organism]\">RPL6</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=RPS11[Gene Name] AND Human[Organism]\">RPS11</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=RPS3[Gene Name] AND Human[Organism]\">RPS3</a></td>\n",
       "    </tr><tr>\n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=RPS6[Gene Name] AND Human[Organism]\">RPS6</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=RSAD2[Gene Name] AND Human[Organism]\">RSAD2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=S100PBP[Gene Name] AND Human[Organism]\">S100PBP</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=SAMHD1[Gene Name] AND Human[Organism]\">SAMHD1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=SAT1[Gene Name] AND Human[Organism]\">SAT1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=SEC31A[Gene Name] AND Human[Organism]\">SEC31A</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=SERPINA1[Gene Name] AND Human[Organism]\">SERPINA1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=SETX[Gene Name] AND Human[Organism]\">SETX</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=SLC13A4[Gene Name] AND Human[Organism]\">SLC13A4</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=SLC25A6[Gene Name] AND Human[Organism]\">SLC25A6</a></td>\n",
       "    </tr><tr>\n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=SMAP2[Gene Name] AND Human[Organism]\">SMAP2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=SON[Gene Name] AND Human[Organism]\">SON</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=SPTAN1[Gene Name] AND Human[Organism]\">SPTAN1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=SPTBN1[Gene Name] AND Human[Organism]\">SPTBN1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=SYNE1[Gene Name] AND Human[Organism]\">SYNE1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=SZT2[Gene Name] AND Human[Organism]\">SZT2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=THAP6[Gene Name] AND Human[Organism]\">THAP6</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=TLN1[Gene Name] AND Human[Organism]\">TLN1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=TMPRSS3[Gene Name] AND Human[Organism]\">TMPRSS3</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=TNFAIP3[Gene Name] AND Human[Organism]\">TNFAIP3</a></td>\n",
       "    </tr><tr>\n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=TPM3[Gene Name] AND Human[Organism]\">TPM3</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=TRAK1[Gene Name] AND Human[Organism]\">TRAK1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=TRIB1[Gene Name] AND Human[Organism]\">TRIB1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=TRPM3[Gene Name] AND Human[Organism]\">TRPM3</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=TTN[Gene Name] AND Human[Organism]\">TTN</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=TTR[Gene Name] AND Human[Organism]\">TTR</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=TXNIP[Gene Name] AND Human[Organism]\">TXNIP</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=UBA1[Gene Name] AND Human[Organism]\">UBA1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=UBR4[Gene Name] AND Human[Organism]\">UBR4</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=URI1[Gene Name] AND Human[Organism]\">URI1</a></td>\n",
       "    </tr><tr>\n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=USP24[Gene Name] AND Human[Organism]\">USP24</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=USP34[Gene Name] AND Human[Organism]\">USP34</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=VCAN[Gene Name] AND Human[Organism]\">VCAN</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=VIM[Gene Name] AND Human[Organism]\">VIM</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=WARS[Gene Name] AND Human[Organism]\">WARS</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=WASF2[Gene Name] AND Human[Organism]\">WASF2</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=WDFY3[Gene Name] AND Human[Organism]\">WDFY3</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=WDFY4[Gene Name] AND Human[Organism]\">WDFY4</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=WIPF1[Gene Name] AND Human[Organism]\">WIPF1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=XAF1[Gene Name] AND Human[Organism]\">XAF1</a></td>\n",
       "    </tr><tr>\n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=XIST[Gene Name] AND Human[Organism]\">XIST</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=ZBTB20[Gene Name] AND Human[Organism]\">ZBTB20</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=ZC3HAV1[Gene Name] AND Human[Organism]\">ZC3HAV1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=ZEB1[Gene Name] AND Human[Organism]\">ZEB1</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=ZFX[Gene Name] AND Human[Organism]\">ZFX</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=ZNF264[Gene Name] AND Human[Organism]\">ZNF264</a></td>\n",
       "    \n",
       "    <td><a target=\"_blank\" href=\"https://www.ncbi.nlm.nih.gov/gene/?term=ZNF37BP[Gene Name] AND Human[Organism]\">ZNF37BP</a></td>\n",
       "    </tr></table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from IPython.core.display import display, HTML\n",
    "\n",
    "rfe_features = []\n",
    "for (rank, name) in zip(clf.ranking_, X.columns):\n",
    "    if rank == 1:\n",
    "        rfe_features.append(name)\n",
    "\n",
    "s=\"\"\"\n",
    "<h2>List of %d genes found by RFE</h2>\n",
    "<p>Note: the NCBI link will open the target in a new window or tab.</p>\n",
    "<table>\n",
    "\"\"\" % (clf.n_features_)\n",
    "\n",
    "ncbi_url = lambda gene: \"https://www.ncbi.nlm.nih.gov/gene/?term=%s[Gene Name] AND Human[Organism]\" % (gene)\n",
    "\n",
    "s += \"<tr>\"\n",
    "for (i, gene) in enumerate(rfe_features):\n",
    "    if not i % 10:\n",
    "        s += \"</tr><tr>\"\n",
    "    s += \"\"\"\n",
    "    <td><a target=\\\"_blank\\\" href=\\\"%s\\\">%s</a></td>\n",
    "    \"\"\" %(ncbi_url(gene), gene)\n",
    "s += \"</tr></table>\"\n",
    "\n",
    "display(HTML(s))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimal number of features : 167\n",
      "Recursive Feature Elimination (RFE) eliminated 25175 features\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjsAAAFyCAYAAAATaRtiAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x110e2f8d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "best_estimator = clf\n",
    "\n",
    "print(\"Optimal number of features : %d\" % best_estimator.n_features_)\n",
    "print(\"Recursive Feature Elimination (RFE) eliminated %d features\" % (X.shape[1] - best_estimator.n_features_))\n",
    "\n",
    "# Plot number of features VS. cross-validation scores\n",
    "plt.figure()\n",
    "plt.xlabel(\"Number of feature subsets selected\")\n",
    "plt.ylabel(\"Cross validation score\")\n",
    "plt.plot(range(1, len(best_estimator.grid_scores_) + 1), best_estimator.grid_scores_)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "None\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiIAAAGHCAYAAACNjTnqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x110e63110>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiIAAAGHCAYAAACNjTnqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nFAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x106820ad0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "from learning_curves import plot_learning_curve\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import StratifiedShuffleSplit\n",
    "\n",
    "def create_learning_curve(title, model):\n",
    "    cv = StratifiedShuffleSplit(n_splits=2, test_size=0.3, random_state=random_state)                                     \n",
    "    plot_learning_curve(model, title, X, y, (0.2, 1.01), cv=cv, n_jobs=1)\n",
    "    \n",
    "create_learning_curve(\"Learning Curves Full Feature Set\", est)\n",
    "create_learning_curve(\"Learning Curves RFE Feature Set\", clf.estimator_)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Make predictions based on the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Full Metrics\n",
      "Accuracy was 66.67%\n",
      "\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "    control       0.70      0.58      0.64        12\n",
      "      viral       0.71      0.56      0.63         9\n",
      "  bacterial       0.60      1.00      0.75         6\n",
      "\n",
      "avg / total       0.68      0.67      0.66        27\n",
      "\n",
      "Confusion Matrix: cols = predictions, rows = actual\n",
      "\n",
      "                       control          viral      bacterial         fungal\n",
      "        control              7              2              3              0\n",
      "          viral              3              5              1              0\n",
      "      bacterial              0              0              6              0\n",
      "         fungal              0              0              0              0\n",
      "--------------------------------------------------------------------------------\n",
      "\n",
      "RFE Metrics\n",
      "Accuracy was 74.07%\n",
      "\n",
      "             precision    recall  f1-score   support\n",
      "\n",
      "    control       0.88      0.58      0.70        12\n",
      "      viral       0.78      0.78      0.78         9\n",
      "  bacterial       0.60      1.00      0.75         6\n",
      "\n",
      "avg / total       0.78      0.74      0.74        27\n",
      "\n",
      "Confusion Matrix: cols = predictions, rows = actual\n",
      "\n",
      "                       control          viral      bacterial         fungal\n",
      "        control              7              2              3              0\n",
      "          viral              1              7              1              0\n",
      "      bacterial              0              0              6              0\n",
      "         fungal              0              0              0              0\n"
     ]
    }
   ],
   "source": [
    "from classification_metrics import classification_metrics\n",
    "\n",
    "est.fit(X_train, y_train)\n",
    "est_predicted = est.predict(X_test)\n",
    "print \"Full Metrics\"\n",
    "classification_metrics(y_test, est_predicted, patient_groups)\n",
    "rfe_predicted = clf.predict(X_test)\n",
    "print \"-\" * 80\n",
    "print \"\\nRFE Metrics\"\n",
    "classification_metrics(y_test, rfe_predicted, patient_groups)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Review model predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Actual</th>\n",
       "      <th>Predicted</th>\n",
       "      <th>Prob. bacteria</th>\n",
       "      <th>Prob. control</th>\n",
       "      <th>Prob. viral</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>MN_223</th>\n",
       "      <td>bacterial</td>\n",
       "      <td>bacterial</td>\n",
       "      <td>65.10</td>\n",
       "      <td>0.00</td>\n",
       "      <td>34.90</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MN_140</th>\n",
       "      <td>bacterial</td>\n",
       "      <td>bacterial</td>\n",
       "      <td>96.85</td>\n",
       "      <td>0.11</td>\n",
       "      <td>3.04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MN_304</th>\n",
       "      <td>bacterial</td>\n",
       "      <td>bacterial</td>\n",
       "      <td>75.90</td>\n",
       "      <td>10.18</td>\n",
       "      <td>13.91</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MN_224</th>\n",
       "      <td>bacterial</td>\n",
       "      <td>bacterial</td>\n",
       "      <td>46.74</td>\n",
       "      <td>21.19</td>\n",
       "      <td>32.07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MN_324</th>\n",
       "      <td>bacterial</td>\n",
       "      <td>bacterial</td>\n",
       "      <td>90.39</td>\n",
       "      <td>9.61</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.473</th>\n",
       "      <td>bacterial</td>\n",
       "      <td>bacterial</td>\n",
       "      <td>65.08</td>\n",
       "      <td>25.68</td>\n",
       "      <td>9.23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.571</th>\n",
       "      <td>control</td>\n",
       "      <td>control</td>\n",
       "      <td>2.84</td>\n",
       "      <td>76.51</td>\n",
       "      <td>20.65</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MN_368</th>\n",
       "      <td>control</td>\n",
       "      <td>bacterial</td>\n",
       "      <td>95.01</td>\n",
       "      <td>4.73</td>\n",
       "      <td>0.26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MN_366</th>\n",
       "      <td>control</td>\n",
       "      <td>control</td>\n",
       "      <td>2.68</td>\n",
       "      <td>81.32</td>\n",
       "      <td>16.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.213</th>\n",
       "      <td>control</td>\n",
       "      <td>control</td>\n",
       "      <td>25.97</td>\n",
       "      <td>42.77</td>\n",
       "      <td>31.26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.294</th>\n",
       "      <td>control</td>\n",
       "      <td>control</td>\n",
       "      <td>0.01</td>\n",
       "      <td>99.98</td>\n",
       "      <td>0.01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.533</th>\n",
       "      <td>control</td>\n",
       "      <td>control</td>\n",
       "      <td>1.60</td>\n",
       "      <td>90.78</td>\n",
       "      <td>7.61</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.631</th>\n",
       "      <td>control</td>\n",
       "      <td>bacterial</td>\n",
       "      <td>99.55</td>\n",
       "      <td>0.40</td>\n",
       "      <td>0.05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.675</th>\n",
       "      <td>control</td>\n",
       "      <td>control</td>\n",
       "      <td>0.39</td>\n",
       "      <td>85.34</td>\n",
       "      <td>14.27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.116</th>\n",
       "      <td>control</td>\n",
       "      <td>viral</td>\n",
       "      <td>25.64</td>\n",
       "      <td>3.90</td>\n",
       "      <td>70.46</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.054</th>\n",
       "      <td>control</td>\n",
       "      <td>viral</td>\n",
       "      <td>20.20</td>\n",
       "      <td>8.06</td>\n",
       "      <td>71.74</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.151</th>\n",
       "      <td>control</td>\n",
       "      <td>control</td>\n",
       "      <td>0.43</td>\n",
       "      <td>91.56</td>\n",
       "      <td>8.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.291</th>\n",
       "      <td>control</td>\n",
       "      <td>bacterial</td>\n",
       "      <td>70.70</td>\n",
       "      <td>5.79</td>\n",
       "      <td>23.51</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.234</th>\n",
       "      <td>viral</td>\n",
       "      <td>viral</td>\n",
       "      <td>34.94</td>\n",
       "      <td>4.32</td>\n",
       "      <td>60.74</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.033</th>\n",
       "      <td>viral</td>\n",
       "      <td>viral</td>\n",
       "      <td>18.34</td>\n",
       "      <td>33.67</td>\n",
       "      <td>48.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MN_171</th>\n",
       "      <td>viral</td>\n",
       "      <td>viral</td>\n",
       "      <td>4.29</td>\n",
       "      <td>42.05</td>\n",
       "      <td>53.66</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MN_282</th>\n",
       "      <td>viral</td>\n",
       "      <td>viral</td>\n",
       "      <td>31.02</td>\n",
       "      <td>15.04</td>\n",
       "      <td>53.94</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.215</th>\n",
       "      <td>viral</td>\n",
       "      <td>control</td>\n",
       "      <td>0.59</td>\n",
       "      <td>71.32</td>\n",
       "      <td>28.09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.098</th>\n",
       "      <td>viral</td>\n",
       "      <td>viral</td>\n",
       "      <td>5.28</td>\n",
       "      <td>24.66</td>\n",
       "      <td>70.06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.176</th>\n",
       "      <td>viral</td>\n",
       "      <td>viral</td>\n",
       "      <td>0.98</td>\n",
       "      <td>11.19</td>\n",
       "      <td>87.83</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.331</th>\n",
       "      <td>viral</td>\n",
       "      <td>viral</td>\n",
       "      <td>4.97</td>\n",
       "      <td>46.34</td>\n",
       "      <td>48.69</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.015</th>\n",
       "      <td>viral</td>\n",
       "      <td>bacterial</td>\n",
       "      <td>57.11</td>\n",
       "      <td>26.78</td>\n",
       "      <td>16.12</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Actual  Predicted Prob. bacteria Prob. control Prob. viral\n",
       "MN_223   bacterial  bacterial          65.10          0.00       34.90\n",
       "MN_140   bacterial  bacterial          96.85          0.11        3.04\n",
       "MN_304   bacterial  bacterial          75.90         10.18       13.91\n",
       "MN_224   bacterial  bacterial          46.74         21.19       32.07\n",
       "MN_324   bacterial  bacterial          90.39          9.61        0.00\n",
       "MNC.473  bacterial  bacterial          65.08         25.68        9.23\n",
       "MNC.571    control    control           2.84         76.51       20.65\n",
       "MN_368     control  bacterial          95.01          4.73        0.26\n",
       "MN_366     control    control           2.68         81.32       16.00\n",
       "MNC.213    control    control          25.97         42.77       31.26\n",
       "MNC.294    control    control           0.01         99.98        0.01\n",
       "MNC.533    control    control           1.60         90.78        7.61\n",
       "MNC.631    control  bacterial          99.55          0.40        0.05\n",
       "MNC.675    control    control           0.39         85.34       14.27\n",
       "MNC.116    control      viral          25.64          3.90       70.46\n",
       "MNC.054    control      viral          20.20          8.06       71.74\n",
       "MNC.151    control    control           0.43         91.56        8.00\n",
       "MNC.291    control  bacterial          70.70          5.79       23.51\n",
       "MNC.234      viral      viral          34.94          4.32       60.74\n",
       "MNC.033      viral      viral          18.34         33.67       48.00\n",
       "MN_171       viral      viral           4.29         42.05       53.66\n",
       "MN_282       viral      viral          31.02         15.04       53.94\n",
       "MNC.215      viral    control           0.59         71.32       28.09\n",
       "MNC.098      viral      viral           5.28         24.66       70.06\n",
       "MNC.176      viral      viral           0.98         11.19       87.83\n",
       "MNC.331      viral      viral           4.97         46.34       48.69\n",
       "MNC.015      viral  bacterial          57.11         26.78       16.12"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "probs = pd.DataFrame(clf.predict_proba(X_test))\n",
    "probstrs = lambda vals: [\"%.2f\" % (p*100) for p in vals]\n",
    "\n",
    "d = {\"Predicted\": [patient_groups[i] for i in rfe_predicted],\n",
    "     \"Actual\": [patient_groups[i] for i in y_test],\n",
    "     \"Prob. control\": probstrs(probs[0]),\n",
    "     \"Prob. viral\": probstrs(probs[1]),\n",
    "     \"Prob. bacteria\": probstrs(probs[2])}\n",
    "\n",
    "patient_df = pd.DataFrame(d, index=X_test.index)\n",
    "patient_df.sort_values(by=\"Actual\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Review patients the model misclassified"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Actual</th>\n",
       "      <th>Predicted</th>\n",
       "      <th>Prob. bacteria</th>\n",
       "      <th>Prob. control</th>\n",
       "      <th>Prob. viral</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>MNC.054</th>\n",
       "      <td>control</td>\n",
       "      <td>viral</td>\n",
       "      <td>20.20</td>\n",
       "      <td>8.06</td>\n",
       "      <td>71.74</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.116</th>\n",
       "      <td>control</td>\n",
       "      <td>viral</td>\n",
       "      <td>25.64</td>\n",
       "      <td>3.90</td>\n",
       "      <td>70.46</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.631</th>\n",
       "      <td>control</td>\n",
       "      <td>bacterial</td>\n",
       "      <td>99.55</td>\n",
       "      <td>0.40</td>\n",
       "      <td>0.05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.291</th>\n",
       "      <td>control</td>\n",
       "      <td>bacterial</td>\n",
       "      <td>70.70</td>\n",
       "      <td>5.79</td>\n",
       "      <td>23.51</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.015</th>\n",
       "      <td>viral</td>\n",
       "      <td>bacterial</td>\n",
       "      <td>57.11</td>\n",
       "      <td>26.78</td>\n",
       "      <td>16.12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MNC.215</th>\n",
       "      <td>viral</td>\n",
       "      <td>control</td>\n",
       "      <td>0.59</td>\n",
       "      <td>71.32</td>\n",
       "      <td>28.09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MN_368</th>\n",
       "      <td>control</td>\n",
       "      <td>bacterial</td>\n",
       "      <td>95.01</td>\n",
       "      <td>4.73</td>\n",
       "      <td>0.26</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          Actual  Predicted Prob. bacteria Prob. control Prob. viral\n",
       "MNC.054  control      viral          20.20          8.06       71.74\n",
       "MNC.116  control      viral          25.64          3.90       70.46\n",
       "MNC.631  control  bacterial          99.55          0.40        0.05\n",
       "MNC.291  control  bacterial          70.70          5.79       23.51\n",
       "MNC.015    viral  bacterial          57.11         26.78       16.12\n",
       "MNC.215    viral    control           0.59         71.32       28.09\n",
       "MN_368   control  bacterial          95.01          4.73        0.26"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "patient_df[patient_df[\"Predicted\"] != patient_df[\"Actual\"]]"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [conda root]",
   "language": "python",
   "name": "conda-root-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}