target_pred_21_demo.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Ligand-based Target Prediction example\n",
    "## George Papadatos, ChEMBL Group"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This is a short demo on getting target predictions based on the current multi-category Naive Bayesian ChEMBL_21 models for a single molecule input. The predictions are ranked by the output classifier probability.\n",
    "\n",
    "This is the outline of the demo:\n",
    "\n",
    "* Download the models\n",
    "* Input a molecule to get target predictions for it based on the models\n",
    "* Assess the predictions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1) Download the models (this is needed only once)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/Users/georgep/testmodels_21\n"
     ]
    }
   ],
   "source": [
    "cd testmodels_21/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--2016-03-29 15:16:53--  ftp://ftp.ebi.ac.uk/pub/databases/chembl/target_predictions/chembl_21_models.tar.gz\n",
      "           => 'chembl_21_models.tar.gz'\n",
      "Resolving ftp.ebi.ac.uk... 193.62.194.182\n",
      "Connecting to ftp.ebi.ac.uk|193.62.194.182|:21... connected.\n",
      "Logging in as anonymous ... Logged in!\n",
      "==> SYST ... done.    ==> PWD ... done.\n",
      "==> TYPE I ... done.  ==> CWD (1) /pub/databases/chembl/target_predictions ... done.\n",
      "==> SIZE chembl_21_models.tar.gz ... 64469262\n",
      "==> PASV ... done.    ==> RETR chembl_21_models.tar.gz ... done.\n",
      "Length: 64469262 (61M) (unauthoritative)\n",
      "\n",
      "100%[======================================>] 64,469,262  44.7MB/s   in 1.4s   \n",
      "\n",
      "2016-03-29 15:16:55 (44.7 MB/s) - 'chembl_21_models.tar.gz' saved [64469262]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "!wget ftp://ftp.ebi.ac.uk/pub/databases/chembl/target_predictions/chembl_21_models.tar.gz"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "!tar -zxf chembl_21_models.tar.gz"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The two models (1uM and 10uM, respectively) are now under ./models_21\n",
    "\n",
    "### 2) Read the models and get predictions\n",
    "Import the necessary modules:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from rdkit.Chem import AllChem as Chem\n",
    "from rdkit.Chem.Draw import IPythonConsole\n",
    "from rdkit.Chem import PandasTools\n",
    "from rdkit.Chem import Draw\n",
    "\n",
    "from rdkit import DataStructs\n",
    "\n",
    "import pandas as pd\n",
    "from pandas import concat\n",
    "\n",
    "from collections import OrderedDict\n",
    "import requests\n",
    "import numpy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.externals import joblib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2015.09.2\n"
     ]
    }
   ],
   "source": [
    "# Check the RDKit version\n",
    "from rdkit import rdBase\n",
    "print rdBase.rdkitVersion"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import matplotlib as mpl\n",
    "mpl.rcParams['figure.figsize'] = 10,10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "morgan_nb = joblib.load('./models_21/10uM/mNB_10uM_all.pkl')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "classes = list(morgan_nb.targets)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1429"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(classes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "That means that this model contains trained data for 1429 different targets (categories)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3) Read a compound and get predictions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's use a molecule (GS-5759) found in this [publication](http://www.ncbi.nlm.nih.gov/pubmed/24513870). It is a novel compound (not in ChEMBL) reported to have dual activity against PDE4 and beta-2 adrenergic receptor."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "smiles = 'O[C@@H](CNCCCC#CC1=CC=C(C=C1)NC(=O)C=1C=C(C=CC1)S(=O)(=O)C=1C=C2C(=C(C=NC2=C(C1)C)C(=O)N)NC1=CC(=CC=C1)OC)C1=C2C=CC(NC2=C(C=C1)O)=O'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "mol = Chem.MolFromSmiles(smiles)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAcIAAACWCAIAAADCEh9HAAAJe0lEQVR4nO3dUXakKhSFYenVM+r5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\nFABEiFEAECFGAUDkH2O2bLsEEwEtAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<rdkit.Chem.rdchem.Mol at 0x107d4a2f0>"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mol"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "info={}\n",
    "fp = Chem.GetMorganFingerprintAsBitVect(mol,2,nBits=2048, bitInfo=info)\n",
    "res = numpy.zeros(len(fp),numpy.int32)\n",
    "DataStructs.ConvertToNumpyArray(fp,res)  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "probas = list(morgan_nb.predict_proba(res.reshape(1,-1))[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "predictions = pd.DataFrame(zip(classes, probas), columns=['id','proba'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "collapsed": false,
    "scrolled": 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>id</th>\n",
       "      <th>proba</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CHEMBL1075051</td>\n",
       "      <td>7.053404e-33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CHEMBL1075104</td>\n",
       "      <td>4.911063e-17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CHEMBL1075108</td>\n",
       "      <td>4.848016e-38</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CHEMBL1075111</td>\n",
       "      <td>2.056597e-40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>CHEMBL1075115</td>\n",
       "      <td>7.705190e-10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>CHEMBL1075138</td>\n",
       "      <td>6.682079e-07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>CHEMBL1075140</td>\n",
       "      <td>5.666627e-41</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>CHEMBL1075145</td>\n",
       "      <td>2.389783e-36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>CHEMBL1075152</td>\n",
       "      <td>4.561929e-42</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>CHEMBL1075167</td>\n",
       "      <td>8.753342e-09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>CHEMBL1075189</td>\n",
       "      <td>4.621086e-09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>CHEMBL1075228</td>\n",
       "      <td>2.019287e-31</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>CHEMBL1075269</td>\n",
       "      <td>2.000263e-33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>CHEMBL1075280</td>\n",
       "      <td>8.740582e-25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>CHEMBL1075282</td>\n",
       "      <td>1.461456e-29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>CHEMBL1075284</td>\n",
       "      <td>1.198526e-23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>CHEMBL1075294</td>\n",
       "      <td>8.631092e-25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>CHEMBL1075319</td>\n",
       "      <td>2.336758e-29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>CHEMBL1075322</td>\n",
       "      <td>1.376716e-13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>CHEMBL1075323</td>\n",
       "      <td>2.354106e-28</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               id         proba\n",
       "0   CHEMBL1075051  7.053404e-33\n",
       "1   CHEMBL1075104  4.911063e-17\n",
       "2   CHEMBL1075108  4.848016e-38\n",
       "3   CHEMBL1075111  2.056597e-40\n",
       "4   CHEMBL1075115  7.705190e-10\n",
       "5   CHEMBL1075138  6.682079e-07\n",
       "6   CHEMBL1075140  5.666627e-41\n",
       "7   CHEMBL1075145  2.389783e-36\n",
       "8   CHEMBL1075152  4.561929e-42\n",
       "9   CHEMBL1075167  8.753342e-09\n",
       "10  CHEMBL1075189  4.621086e-09\n",
       "11  CHEMBL1075228  2.019287e-31\n",
       "12  CHEMBL1075269  2.000263e-33\n",
       "13  CHEMBL1075280  8.740582e-25\n",
       "14  CHEMBL1075282  1.461456e-29\n",
       "15  CHEMBL1075284  1.198526e-23\n",
       "16  CHEMBL1075294  8.631092e-25\n",
       "17  CHEMBL1075319  2.336758e-29\n",
       "18  CHEMBL1075322  1.376716e-13\n",
       "19  CHEMBL1075323  2.354106e-28"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions.head(20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The above is a pandas data frame with all the predictions across the 1429 targets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x10c77b0d0>"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmEAAAJPCAYAAAA0UwMNAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAQCsQAkDAFiBEgYAsAIlDABgBf8/SsGTxmgf7xwAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10d790910>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "predictions['proba'].hist()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As expected, most of the targets will have very low probability for this particular compound. But what about the targets with the highest probability scores?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "top_preds = predictions.sort_values(by='proba', ascending=False).head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "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>id</th>\n",
       "      <th>proba</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1283</th>\n",
       "      <td>CHEMBL5414</td>\n",
       "      <td>0.999912</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>243</th>\n",
       "      <td>CHEMBL210</td>\n",
       "      <td>0.994003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>652</th>\n",
       "      <td>CHEMBL3252</td>\n",
       "      <td>0.836839</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>249</th>\n",
       "      <td>CHEMBL213</td>\n",
       "      <td>0.747240</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>477</th>\n",
       "      <td>CHEMBL275</td>\n",
       "      <td>0.660680</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1123</th>\n",
       "      <td>CHEMBL4697</td>\n",
       "      <td>0.000231</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>851</th>\n",
       "      <td>CHEMBL3864</td>\n",
       "      <td>0.000114</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>707</th>\n",
       "      <td>CHEMBL3397</td>\n",
       "      <td>0.000102</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>401</th>\n",
       "      <td>CHEMBL2535</td>\n",
       "      <td>0.000058</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>710</th>\n",
       "      <td>CHEMBL340</td>\n",
       "      <td>0.000031</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              id     proba\n",
       "1283  CHEMBL5414  0.999912\n",
       "243    CHEMBL210  0.994003\n",
       "652   CHEMBL3252  0.836839\n",
       "249    CHEMBL213  0.747240\n",
       "477    CHEMBL275  0.660680\n",
       "1123  CHEMBL4697  0.000231\n",
       "851   CHEMBL3864  0.000114\n",
       "707   CHEMBL3397  0.000102\n",
       "401   CHEMBL2535  0.000058\n",
       "710    CHEMBL340  0.000031"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "top_preds"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Finally, let's use the ChEMBL WS to get some more info on the most likely predicted targets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def fetch_WS(trgt):\n",
    "    re = requests.get('https://www.ebi.ac.uk/chembl/api/data/target/{0}.json'.format(trgt))\n",
    "    return (trgt, re.json()['pref_name'], re.json()['organism'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "plist = []\n",
    "for i,e in enumerate(top_preds['id']):\n",
    "    plist.append(fetch_WS(e))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "target_info = pd.DataFrame(plist, columns =['id','name', 'organism'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "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>id</th>\n",
       "      <th>name</th>\n",
       "      <th>organism</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CHEMBL5414</td>\n",
       "      <td>Beta-2 adrenergic receptor</td>\n",
       "      <td>Cavia porcellus</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CHEMBL210</td>\n",
       "      <td>Beta-2 adrenergic receptor</td>\n",
       "      <td>Homo sapiens</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CHEMBL3252</td>\n",
       "      <td>Beta-1 adrenergic receptor</td>\n",
       "      <td>Rattus norvegicus</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CHEMBL213</td>\n",
       "      <td>Beta-1 adrenergic receptor</td>\n",
       "      <td>Homo sapiens</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>CHEMBL275</td>\n",
       "      <td>Phosphodiesterase 4B</td>\n",
       "      <td>Homo sapiens</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>CHEMBL4697</td>\n",
       "      <td>Hexose transporter 1</td>\n",
       "      <td>Plasmodium falciparum</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>CHEMBL3864</td>\n",
       "      <td>Protein-tyrosine phosphatase 2C</td>\n",
       "      <td>Homo sapiens</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>CHEMBL3397</td>\n",
       "      <td>Cytochrome P450 2C9</td>\n",
       "      <td>Homo sapiens</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>CHEMBL2535</td>\n",
       "      <td>Glucose transporter</td>\n",
       "      <td>Homo sapiens</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>CHEMBL340</td>\n",
       "      <td>Cytochrome P450 3A4</td>\n",
       "      <td>Homo sapiens</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           id                             name               organism\n",
       "0  CHEMBL5414       Beta-2 adrenergic receptor        Cavia porcellus\n",
       "1   CHEMBL210       Beta-2 adrenergic receptor           Homo sapiens\n",
       "2  CHEMBL3252       Beta-1 adrenergic receptor      Rattus norvegicus\n",
       "3   CHEMBL213       Beta-1 adrenergic receptor           Homo sapiens\n",
       "4   CHEMBL275             Phosphodiesterase 4B           Homo sapiens\n",
       "5  CHEMBL4697             Hexose transporter 1  Plasmodium falciparum\n",
       "6  CHEMBL3864  Protein-tyrosine phosphatase 2C           Homo sapiens\n",
       "7  CHEMBL3397              Cytochrome P450 2C9           Homo sapiens\n",
       "8  CHEMBL2535              Glucose transporter           Homo sapiens\n",
       "9   CHEMBL340              Cytochrome P450 3A4           Homo sapiens"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "target_info"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "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>id</th>\n",
       "      <th>proba</th>\n",
       "      <th>name</th>\n",
       "      <th>organism</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CHEMBL5414</td>\n",
       "      <td>0.999912</td>\n",
       "      <td>Beta-2 adrenergic receptor</td>\n",
       "      <td>Cavia porcellus</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CHEMBL210</td>\n",
       "      <td>0.994003</td>\n",
       "      <td>Beta-2 adrenergic receptor</td>\n",
       "      <td>Homo sapiens</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CHEMBL3252</td>\n",
       "      <td>0.836839</td>\n",
       "      <td>Beta-1 adrenergic receptor</td>\n",
       "      <td>Rattus norvegicus</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CHEMBL213</td>\n",
       "      <td>0.747240</td>\n",
       "      <td>Beta-1 adrenergic receptor</td>\n",
       "      <td>Homo sapiens</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>CHEMBL275</td>\n",
       "      <td>0.660680</td>\n",
       "      <td>Phosphodiesterase 4B</td>\n",
       "      <td>Homo sapiens</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>CHEMBL4697</td>\n",
       "      <td>0.000231</td>\n",
       "      <td>Hexose transporter 1</td>\n",
       "      <td>Plasmodium falciparum</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>CHEMBL3864</td>\n",
       "      <td>0.000114</td>\n",
       "      <td>Protein-tyrosine phosphatase 2C</td>\n",
       "      <td>Homo sapiens</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>CHEMBL3397</td>\n",
       "      <td>0.000102</td>\n",
       "      <td>Cytochrome P450 2C9</td>\n",
       "      <td>Homo sapiens</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>CHEMBL2535</td>\n",
       "      <td>0.000058</td>\n",
       "      <td>Glucose transporter</td>\n",
       "      <td>Homo sapiens</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>CHEMBL340</td>\n",
       "      <td>0.000031</td>\n",
       "      <td>Cytochrome P450 3A4</td>\n",
       "      <td>Homo sapiens</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           id     proba                             name               organism\n",
       "0  CHEMBL5414  0.999912       Beta-2 adrenergic receptor        Cavia porcellus\n",
       "1   CHEMBL210  0.994003       Beta-2 adrenergic receptor           Homo sapiens\n",
       "2  CHEMBL3252  0.836839       Beta-1 adrenergic receptor      Rattus norvegicus\n",
       "3   CHEMBL213  0.747240       Beta-1 adrenergic receptor           Homo sapiens\n",
       "4   CHEMBL275  0.660680             Phosphodiesterase 4B           Homo sapiens\n",
       "5  CHEMBL4697  0.000231             Hexose transporter 1  Plasmodium falciparum\n",
       "6  CHEMBL3864  0.000114  Protein-tyrosine phosphatase 2C           Homo sapiens\n",
       "7  CHEMBL3397  0.000102              Cytochrome P450 2C9           Homo sapiens\n",
       "8  CHEMBL2535  0.000058              Glucose transporter           Homo sapiens\n",
       "9   CHEMBL340  0.000031              Cytochrome P450 3A4           Homo sapiens"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.merge(top_preds, target_info)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}