prediction/8-feature-example-viz.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Visualize subset of the feature matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import pandas\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "feature_df = pandas.read_table('features/transformed-features.tsv.bz2', low_memory=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>compound_id</th>\n",
       "      <th>compound_name</th>\n",
       "      <th>disease_id</th>\n",
       "      <th>disease_name</th>\n",
       "      <th>category</th>\n",
       "      <th>status</th>\n",
       "      <th>prior_prob</th>\n",
       "      <th>prior_logit</th>\n",
       "      <th>degree_CbG</th>\n",
       "      <th>degree_CcSE</th>\n",
       "      <th>...</th>\n",
       "      <th>dwpc_CuGaDuGaD</th>\n",
       "      <th>dwpc_CuGcG&lt;rGaD</th>\n",
       "      <th>dwpc_CuGcGiGaD</th>\n",
       "      <th>dwpc_CuGdDpSpD</th>\n",
       "      <th>dwpc_CuGr&gt;GbCtD</th>\n",
       "      <th>dwpc_CuGuCrCtD</th>\n",
       "      <th>dwpc_CuGuCtD</th>\n",
       "      <th>dwpc_CuGuCtDrD</th>\n",
       "      <th>dwpc_CuGuCuGaD</th>\n",
       "      <th>dwpc_CuGuDuGaD</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DB01048</td>\n",
       "      <td>Abacavir</td>\n",
       "      <td>DOID:10652</td>\n",
       "      <td>Alzheimer's disease</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0.004753</td>\n",
       "      <td>-5.344</td>\n",
       "      <td>-0.2955</td>\n",
       "      <td>0.8146</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.5028</td>\n",
       "      <td>-0.4211</td>\n",
       "      <td>-0.4581</td>\n",
       "      <td>-0.5399</td>\n",
       "      <td>-0.357</td>\n",
       "      <td>-0.3849</td>\n",
       "      <td>-0.375</td>\n",
       "      <td>-0.5642</td>\n",
       "      <td>-0.5555</td>\n",
       "      <td>-0.4952</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>DB05812</td>\n",
       "      <td>Abiraterone</td>\n",
       "      <td>DOID:10652</td>\n",
       "      <td>Alzheimer's disease</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0.004753</td>\n",
       "      <td>-5.344</td>\n",
       "      <td>0.5495</td>\n",
       "      <td>0.5554</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.5028</td>\n",
       "      <td>-0.4211</td>\n",
       "      <td>-0.4581</td>\n",
       "      <td>-0.5399</td>\n",
       "      <td>-0.357</td>\n",
       "      <td>-0.3849</td>\n",
       "      <td>-0.375</td>\n",
       "      <td>-0.5642</td>\n",
       "      <td>-0.5555</td>\n",
       "      <td>-0.4952</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 166 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  compound_id compound_name  disease_id         disease_name category  status  \\\n",
       "0     DB01048      Abacavir  DOID:10652  Alzheimer's disease      NaN       0   \n",
       "1     DB05812   Abiraterone  DOID:10652  Alzheimer's disease      NaN       0   \n",
       "\n",
       "   prior_prob  prior_logit  degree_CbG  degree_CcSE       ...        \\\n",
       "0    0.004753       -5.344     -0.2955       0.8146       ...         \n",
       "1    0.004753       -5.344      0.5495       0.5554       ...         \n",
       "\n",
       "   dwpc_CuGaDuGaD  dwpc_CuGcG<rGaD  dwpc_CuGcGiGaD  dwpc_CuGdDpSpD  \\\n",
       "0         -0.5028          -0.4211         -0.4581         -0.5399   \n",
       "1         -0.5028          -0.4211         -0.4581         -0.5399   \n",
       "\n",
       "   dwpc_CuGr>GbCtD  dwpc_CuGuCrCtD  dwpc_CuGuCtD  dwpc_CuGuCtDrD  \\\n",
       "0           -0.357         -0.3849        -0.375         -0.5642   \n",
       "1           -0.357         -0.3849        -0.375         -0.5642   \n",
       "\n",
       "   dwpc_CuGuCuGaD  dwpc_CuGuDuGaD  \n",
       "0         -0.5555         -0.4952  \n",
       "1         -0.5555         -0.4952  \n",
       "\n",
       "[2 rows x 166 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "feature_df = feature_df.query(\"prior_prob > 0\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "metapaths = [\n",
    "    'CbGaD',\n",
    "    'CbGiGaD',\n",
    "    'CbGpPWpGaD',\n",
    "    'CbGeAlD',\n",
    "    'CcSEcCtD',\n",
    "    'CuGuCtD',\n",
    "    'CdGdCtD',\n",
    "]\n",
    "\n",
    "features = ['dwpc_' + metapath for metapath in metapaths]\n",
    "renamer = {'dwpc_' + metapath: metapath for metapath in metapaths}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>status</th>\n",
       "      <th>prior_logit</th>\n",
       "      <th>CbGaD</th>\n",
       "      <th>CbGiGaD</th>\n",
       "      <th>CbGpPWpGaD</th>\n",
       "      <th>CbGeAlD</th>\n",
       "      <th>CcSEcCtD</th>\n",
       "      <th>CuGuCtD</th>\n",
       "      <th>CdGdCtD</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Calcitriol–dilated cardiomyopathy</th>\n",
       "      <td>0</td>\n",
       "      <td>-4.911</td>\n",
       "      <td>-0.358</td>\n",
       "      <td>1.5360</td>\n",
       "      <td>-0.2313</td>\n",
       "      <td>-0.71350</td>\n",
       "      <td>0.75460</td>\n",
       "      <td>-0.375</td>\n",
       "      <td>1.1030</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Nilutamide–metabolic syndrome X</th>\n",
       "      <td>0</td>\n",
       "      <td>-6.757</td>\n",
       "      <td>-0.358</td>\n",
       "      <td>0.6863</td>\n",
       "      <td>0.7708</td>\n",
       "      <td>0.02763</td>\n",
       "      <td>0.01044</td>\n",
       "      <td>-0.375</td>\n",
       "      <td>-0.4084</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                   status  prior_logit  CbGaD  CbGiGaD  \\\n",
       "name                                                                     \n",
       "Calcitriol–dilated cardiomyopathy       0       -4.911 -0.358   1.5360   \n",
       "Nilutamide–metabolic syndrome X         0       -6.757 -0.358   0.6863   \n",
       "\n",
       "                                   CbGpPWpGaD  CbGeAlD  CcSEcCtD  CuGuCtD  \\\n",
       "name                                                                        \n",
       "Calcitriol–dilated cardiomyopathy     -0.2313 -0.71350   0.75460   -0.375   \n",
       "Nilutamide–metabolic syndrome X        0.7708  0.02763   0.01044   -0.375   \n",
       "\n",
       "                                   CdGdCtD  \n",
       "name                                        \n",
       "Calcitriol–dilated cardiomyopathy   1.1030  \n",
       "Nilutamide–metabolic syndrome X    -0.4084  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "subset_df = (feature_df\n",
    "    .assign(name = feature_df.compound_name + '–' + feature_df.disease_name)\n",
    "    .groupby('status')\n",
    "    .apply(lambda df: df.sample(n=6, random_state=8))\n",
    "    .set_index(['name'])\n",
    "    [['status', 'prior_logit'] + features]\n",
    "    .rename(columns=renamer)\n",
    ")\n",
    "subset_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZgAAAFwCAYAAACFLG26AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f605cd137f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "status_colors = subset_df.status.map({0: '#ECF0F1', 1: '#66A58B'})\n",
    "heatmap_df = subset_df.drop(['status', 'prior_logit'], axis=1)\n",
    "grid = seaborn.clustermap(heatmap_df, row_cluster=False,\n",
    "    col_cluster=False, row_colors=status_colors, figsize=(4, 5),\n",
    "    linewidths=1)\n",
    "grid.cax.set_visible(False)\n",
    "plt.setp(grid.ax_heatmap.get_yticklabels(), rotation=0);\n",
    "grid.ax_heatmap.set_ylabel(\"\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "grid.savefig('figure/example-feature-matrix.svg', bbox_inches='tight', transparent=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Convert to PNG and trim whitespace\n",
    "! convert -density 300 figure/example-feature-matrix.svg -trim +repage figure/example-feature-matrix.png"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.5.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}