integrate.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Integrate resources to create a drug repurposing hetnet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import pandas\n",
    "import seaborn\n",
    "\n",
    "import hetio.hetnet\n",
    "import hetio.readwrite\n",
    "import hetio.stats\n",
    "\n",
    "from utils import rawgit, obo_iri"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "## Define the metagraph and instantiate the graph"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "kind_to_abbev = {\n",
    "    \n",
    "    # metanodes\n",
    "    'Compound': 'C',\n",
    "    'Disease': 'D',\n",
    "    'Gene': 'G',\n",
    "    'Anatomy': 'A',\n",
    "    'Symptom': 'S',\n",
    "    'Side Effect': 'SE',\n",
    "    'Pathway': 'PW',\n",
    "    'Pharmacologic Class': 'PC',\n",
    "    'Biological Process': 'BP',\n",
    "    'Cellular Component': 'CC',\n",
    "    'Molecular Function': 'MF',\n",
    "    \n",
    "    # metaedges\n",
    "    'treats': 't',\n",
    "#     'palliates': 'p',\n",
    "    'binds': 'b',\n",
    "    'expresses': 'e',\n",
    "    'regulates': 'r',\n",
    "    'upregulates': 'u',\n",
    "    'downregulates': 'd',\n",
    "    'interacts': 'i',\n",
    "    'includes': 'i',\n",
    "    'covaries': 'c',\n",
    "    'regulates': 'r',\n",
    "    'participates': 'p',\n",
    "    'resembles': 'r',\n",
    "    'associates': 'a',\n",
    "    'localizes': 'l',\n",
    "    'presents': 'p',\n",
    "    'causes': 'c',\n",
    "}\n",
    "\n",
    "metaedge_tuples = [\n",
    "    ('Compound', 'Disease', 'treats', 'both'),\n",
    "#     ('Compound', 'Disease', 'palliates', 'both'),\n",
    "    ('Compound', 'Gene', 'binds', 'both'),\n",
    "    ('Compound', 'Gene', 'upregulates', 'both'),\n",
    "    ('Compound', 'Gene', 'downregulates', 'both'),\n",
    "    ('Compound', 'Compound', 'resembles', 'both'),\n",
    "    ('Compound', 'Side Effect', 'causes', 'both'),\n",
    "    ('Pharmacologic Class', 'Compound', 'includes', 'both'),\n",
    "    ('Anatomy', 'Gene', 'expresses', 'both'),\n",
    "    ('Anatomy', 'Gene', 'upregulates', 'both'),\n",
    "    ('Anatomy', 'Gene', 'downregulates', 'both'),\n",
    "    ('Gene', 'Gene', 'interacts', 'both'),\n",
    "    ('Gene', 'Gene', 'covaries', 'both'),\n",
    "    ('Gene', 'Gene', 'regulates', 'forward'),\n",
    "    ('Gene', 'Pathway', 'participates', 'both'),\n",
    "    ('Gene', 'Biological Process', 'participates', 'both'),\n",
    "    ('Gene', 'Cellular Component', 'participates', 'both'),\n",
    "    ('Gene', 'Molecular Function', 'participates', 'both'),\n",
    "    ('Disease', 'Disease', 'resembles', 'both'),\n",
    "    ('Disease', 'Gene', 'associates', 'both'),\n",
    "    ('Disease', 'Gene', 'upregulates', 'both'),\n",
    "    ('Disease', 'Gene', 'downregulates', 'both'),\n",
    "    ('Disease', 'Anatomy', 'localizes', 'both'),\n",
    "    ('Disease', 'Symptom', 'presents', 'both'),\n",
    "]\n",
    "metagraph = hetio.hetnet.MetaGraph.from_edge_tuples(metaedge_tuples, kind_to_abbev)\n",
    "graph = hetio.hetnet.Graph(metagraph)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Gene Nodes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>tax_id</th>\n",
       "      <th>GeneID</th>\n",
       "      <th>Symbol</th>\n",
       "      <th>chromosome</th>\n",
       "      <th>map_location</th>\n",
       "      <th>type_of_gene</th>\n",
       "      <th>description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>9606</td>\n",
       "      <td>1</td>\n",
       "      <td>A1BG</td>\n",
       "      <td>19</td>\n",
       "      <td>19q13.4</td>\n",
       "      <td>protein-coding</td>\n",
       "      <td>alpha-1-B glycoprotein</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>9606</td>\n",
       "      <td>2</td>\n",
       "      <td>A2M</td>\n",
       "      <td>12</td>\n",
       "      <td>12p13.31</td>\n",
       "      <td>protein-coding</td>\n",
       "      <td>alpha-2-macroglobulin</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   tax_id  GeneID Symbol chromosome map_location    type_of_gene  \\\n",
       "0    9606       1   A1BG         19      19q13.4  protein-coding   \n",
       "1    9606       2    A2M         12     12p13.31  protein-coding   \n",
       "\n",
       "              description  \n",
       "0  alpha-1-B glycoprotein  \n",
       "1   alpha-2-macroglobulin  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "commit = 'a7362748a34211e5df6f2d185bb3246279760546'\n",
    "url = rawgit('dhimmel', 'entrez-gene', commit, 'data/genes-human.tsv')\n",
    "gene_df = pandas.read_table(url)\n",
    "gene_df = gene_df[gene_df.type_of_gene == 'protein-coding']\n",
    "coding_genes = set(gene_df.GeneID)\n",
    "gene_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "for i, row in gene_df.iterrows():\n",
    "    if row.type_of_gene != 'protein-coding':\n",
    "        continue\n",
    "    data = {\n",
    "        'description': row['description'],\n",
    "        'source': 'Entrez Gene',\n",
    "        'url': 'http://identifiers.org/ncbigene/{}'.format(row.GeneID),\n",
    "        'license': 'CC0 1.0',\n",
    "    }\n",
    "    if pandas.notnull(row['chromosome']):\n",
    "        data['chromosome'] = row['chromosome']\n",
    "    graph.add_node(kind = 'Gene', identifier=row.GeneID, name=row.Symbol, data=data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Disease Nodes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>doid</th>\n",
       "      <th>name</th>\n",
       "      <th>source</th>\n",
       "      <th>pathophysiology</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DOID:2531</td>\n",
       "      <td>hematologic cancer</td>\n",
       "      <td>DOcancerslim</td>\n",
       "      <td>neoplastic</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>DOID:1319</td>\n",
       "      <td>brain cancer</td>\n",
       "      <td>DOcancerslim</td>\n",
       "      <td>neoplastic</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        doid                name        source pathophysiology\n",
       "0  DOID:2531  hematologic cancer  DOcancerslim      neoplastic\n",
       "1  DOID:1319        brain cancer  DOcancerslim      neoplastic"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "commit = '75050ea2d4f60e745d3f3578ae03560a2cc0e444'\n",
    "url = rawgit('dhimmel', 'disease-ontology', commit, 'data/slim-terms.tsv')\n",
    "disease_df = pandas.read_table(url)\n",
    "disease_df.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Read rare diseases"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "rare_dises = pandas.read_csv(\"data/rare_diseases.tsv\", sep='\\t')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>dise_name</th>\n",
       "      <th>orpha_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Endometrial stromal sarcoma</td>\n",
       "      <td>ORPHA:213711</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Small cell lung cancer</td>\n",
       "      <td>ORPHA:70573</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     dise_name      orpha_id\n",
       "0  Endometrial stromal sarcoma  ORPHA:213711\n",
       "1       Small cell lung cancer   ORPHA:70573"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rare_dises.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "disease_df = (disease_df\n",
    "    .append(\n",
    "        (rare_dises\n",
    "            .rename(columns={\"dise_name\": \"name\", \"orpha_id\": \"doid\"})\n",
    "            .assign(source=\"Orphanet\", pathophysiology=\"unknown\")\n",
    "        )\n",
    "    )\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>doid</th>\n",
       "      <th>name</th>\n",
       "      <th>pathophysiology</th>\n",
       "      <th>source</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4542</th>\n",
       "      <td>ORPHA:444002</td>\n",
       "      <td>11q22.2q22.3 microdeletion syndrome</td>\n",
       "      <td>unknown</td>\n",
       "      <td>Orphanet</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4543</th>\n",
       "      <td>ORPHA:444051</td>\n",
       "      <td>20q11.2 microdeletion syndrome</td>\n",
       "      <td>unknown</td>\n",
       "      <td>Orphanet</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4544</th>\n",
       "      <td>ORPHA:457059</td>\n",
       "      <td>Pseudohypoparathyroidism with Albright heredit...</td>\n",
       "      <td>unknown</td>\n",
       "      <td>Orphanet</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              doid                                               name  \\\n",
       "4542  ORPHA:444002                11q22.2q22.3 microdeletion syndrome   \n",
       "4543  ORPHA:444051                     20q11.2 microdeletion syndrome   \n",
       "4544  ORPHA:457059  Pseudohypoparathyroidism with Albright heredit...   \n",
       "\n",
       "     pathophysiology    source  \n",
       "4542         unknown  Orphanet  \n",
       "4543         unknown  Orphanet  \n",
       "4544         unknown  Orphanet  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "disease_df.tail(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4682, 4)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "disease_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "for i, row in disease_df.iterrows():\n",
    "    data = {\n",
    "        'source': 'Disease Ontology',\n",
    "        'url': obo_iri(row.doid),\n",
    "        'license': 'CC BY 3.0',\n",
    "    }\n",
    "    graph.add_node(kind='Disease', identifier=row.doid, name=row['name'], data=data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Compound Nodes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>drugbank_id</th>\n",
       "      <th>name</th>\n",
       "      <th>type</th>\n",
       "      <th>groups</th>\n",
       "      <th>atc_codes</th>\n",
       "      <th>categories</th>\n",
       "      <th>inchikey</th>\n",
       "      <th>inchi</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DB00014</td>\n",
       "      <td>Goserelin</td>\n",
       "      <td>small molecule</td>\n",
       "      <td>approved</td>\n",
       "      <td>L02AE03</td>\n",
       "      <td>NaN</td>\n",
       "      <td>InChIKey=BLCLNMBMMGCOAS-URPVMXJPSA-N</td>\n",
       "      <td>InChI=1S/C59H84N18O14/c1-31(2)22-40(49(82)68-3...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>DB00035</td>\n",
       "      <td>Desmopressin</td>\n",
       "      <td>small molecule</td>\n",
       "      <td>approved</td>\n",
       "      <td>H01BA02</td>\n",
       "      <td>Antidiuretic Agents|Hemostatics|Renal Agents</td>\n",
       "      <td>InChIKey=NFLWUMRGJYTJIN-NXBWRCJVSA-N</td>\n",
       "      <td>InChI=1S/C46H64N14O12S2/c47-35(62)15-14-29-40(...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  drugbank_id          name            type    groups atc_codes  \\\n",
       "0     DB00014     Goserelin  small molecule  approved   L02AE03   \n",
       "1     DB00035  Desmopressin  small molecule  approved   H01BA02   \n",
       "\n",
       "                                     categories  \\\n",
       "0                                           NaN   \n",
       "1  Antidiuretic Agents|Hemostatics|Renal Agents   \n",
       "\n",
       "                               inchikey  \\\n",
       "0  InChIKey=BLCLNMBMMGCOAS-URPVMXJPSA-N   \n",
       "1  InChIKey=NFLWUMRGJYTJIN-NXBWRCJVSA-N   \n",
       "\n",
       "                                               inchi  \n",
       "0  InChI=1S/C59H84N18O14/c1-31(2)22-40(49(82)68-3...  \n",
       "1  InChI=1S/C46H64N14O12S2/c47-35(62)15-14-29-40(...  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "commit = '3e87872db5fca5ac427ce27464ab945c0ceb4ec6'\n",
    "url = rawgit('dhimmel', 'drugbank', commit, 'data/drugbank-slim.tsv')\n",
    "compound_df = pandas.read_table(url)\n",
    "compound_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "for i, row in compound_df.iterrows():\n",
    "    url = 'http://www.drugbank.ca/drugs/' + row.drugbank_id\n",
    "    data = {\n",
    "        'source': 'DrugBank',\n",
    "        'inchikey': row.inchikey,\n",
    "        'inchi': row.inchi, 'url': url,\n",
    "        'license': 'CC BY-NC 4.0',\n",
    "    }\n",
    "    graph.add_node(kind='Compound', identifier=row.drugbank_id, name=row['name'], data=data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Anatomy nodes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>uberon_id</th>\n",
       "      <th>uberon_name</th>\n",
       "      <th>mesh_id</th>\n",
       "      <th>mesh_name</th>\n",
       "      <th>bto_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>UBERON:0000002</td>\n",
       "      <td>uterine cervix</td>\n",
       "      <td>D002584</td>\n",
       "      <td>Cervix Uteri</td>\n",
       "      <td>BTO:0001421</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>UBERON:0000004</td>\n",
       "      <td>nose</td>\n",
       "      <td>D009666</td>\n",
       "      <td>Nose</td>\n",
       "      <td>BTO:0000840</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        uberon_id     uberon_name  mesh_id     mesh_name       bto_id\n",
       "0  UBERON:0000002  uterine cervix  D002584  Cervix Uteri  BTO:0001421\n",
       "1  UBERON:0000004            nose  D009666          Nose  BTO:0000840"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "commit = '134f23479186abba03ba340fc6dc90e16c781920'\n",
    "url = rawgit('dhimmel', 'uberon', commit, 'data/hetio-slim.tsv')\n",
    "uberon_df = pandas.read_table(url)\n",
    "uberon_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for i, row in uberon_df.iterrows():\n",
    "    data = {\n",
    "        'source': 'Uberon',\n",
    "        'url': obo_iri(row['uberon_id']),\n",
    "        'license': 'CC BY 3.0',\n",
    "    }\n",
    "    for xref in 'mesh_id', 'bto_id':\n",
    "        if pandas.notnull(row[xref]):\n",
    "            data[xref] = row[xref]\n",
    "    graph.add_node(kind='Anatomy', identifier=row['uberon_id'], name=row['uberon_name'], data=data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Symptom Nodes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mesh_id</th>\n",
       "      <th>mesh_name</th>\n",
       "      <th>in_hsdn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>D000006</td>\n",
       "      <td>Abdomen, Acute</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>D000270</td>\n",
       "      <td>Adie Syndrome</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   mesh_id       mesh_name  in_hsdn\n",
       "0  D000006  Abdomen, Acute        1\n",
       "1  D000270   Adie Syndrome        0"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "commit = 'a7036a37302973b15ab949aab4056d9bc062910e'\n",
    "url = rawgit('dhimmel', 'mesh', commit, 'data/symptoms.tsv')\n",
    "symptom_df = pandas.read_table(url)\n",
    "symptom_df.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Rare disease phenotypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "pheno_links = (pandas\n",
    "    .read_csv(\"data/rare_dise_phenotype_links.tsv\", sep='\\t')\n",
    "    .assign(dise_orpha_id = lambda df: df[\"dise_id\"].map(lambda v: \"ORPHA:{}\".format(v)))\n",
    "    .drop(\"dise_id\", axis=1)\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Use the MeSH id if it exists in the symptoms dataframe, otherwise use the HPO id."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "pheno_links[\"fin_id\"] = pheno_links[[\"hpo_id\", \"mesh_id\"]].apply(\n",
    "    lambda r: r[\"mesh_id\"] if r[\"mesh_id\"] in set(symptom_df[\"mesh_id\"]) else r[\"hpo_id\"],\n",
    "    axis=1\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>dise_name</th>\n",
       "      <th>freq</th>\n",
       "      <th>hpo_id</th>\n",
       "      <th>hpo_name</th>\n",
       "      <th>mesh_id</th>\n",
       "      <th>dise_orpha_id</th>\n",
       "      <th>fin_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3-methylcrotonyl-CoA carboxylase deficiency</td>\n",
       "      <td>Very frequent (99-80%)</td>\n",
       "      <td>HP:0001252</td>\n",
       "      <td>Muscular hypotonia</td>\n",
       "      <td>D009123</td>\n",
       "      <td>ORPHA:6</td>\n",
       "      <td>D009123</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3-methylcrotonyl-CoA carboxylase deficiency</td>\n",
       "      <td>Occasional (29-5%)</td>\n",
       "      <td>HP:0001257</td>\n",
       "      <td>Spasticity</td>\n",
       "      <td>D009128</td>\n",
       "      <td>ORPHA:6</td>\n",
       "      <td>D009128</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3-methylcrotonyl-CoA carboxylase deficiency</td>\n",
       "      <td>Frequent (79-30%)</td>\n",
       "      <td>HP:0001531</td>\n",
       "      <td>Failure to thrive in infancy</td>\n",
       "      <td>NaN</td>\n",
       "      <td>ORPHA:6</td>\n",
       "      <td>HP:0001531</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3-methylcrotonyl-CoA carboxylase deficiency</td>\n",
       "      <td>Very frequent (99-80%)</td>\n",
       "      <td>HP:0001943</td>\n",
       "      <td>Hypoglycemia</td>\n",
       "      <td>D007003</td>\n",
       "      <td>ORPHA:6</td>\n",
       "      <td>HP:0001943</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3-methylcrotonyl-CoA carboxylase deficiency</td>\n",
       "      <td>Frequent (79-30%)</td>\n",
       "      <td>HP:0001987</td>\n",
       "      <td>Hyperammonemia</td>\n",
       "      <td>D022124</td>\n",
       "      <td>ORPHA:6</td>\n",
       "      <td>HP:0001987</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                     dise_name                    freq  \\\n",
       "0  3-methylcrotonyl-CoA carboxylase deficiency  Very frequent (99-80%)   \n",
       "1  3-methylcrotonyl-CoA carboxylase deficiency      Occasional (29-5%)   \n",
       "2  3-methylcrotonyl-CoA carboxylase deficiency       Frequent (79-30%)   \n",
       "3  3-methylcrotonyl-CoA carboxylase deficiency  Very frequent (99-80%)   \n",
       "4  3-methylcrotonyl-CoA carboxylase deficiency       Frequent (79-30%)   \n",
       "\n",
       "       hpo_id                      hpo_name  mesh_id dise_orpha_id      fin_id  \n",
       "0  HP:0001252            Muscular hypotonia  D009123       ORPHA:6     D009123  \n",
       "1  HP:0001257                    Spasticity  D009128       ORPHA:6     D009128  \n",
       "2  HP:0001531  Failure to thrive in infancy      NaN       ORPHA:6  HP:0001531  \n",
       "3  HP:0001943                  Hypoglycemia  D007003       ORPHA:6  HP:0001943  \n",
       "4  HP:0001987                Hyperammonemia  D022124       ORPHA:6  HP:0001987  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pheno_links.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Scrape out new HPO phenotype terms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "new_hpo_terms = (pheno_links\n",
    "    .assign(is_hp = lambda df: df[\"fin_id\"].map(lambda v: v.startswith(\"HP:\")))\n",
    "    .query(\"is_hp\")\n",
    "    [[\"hpo_id\", \"hpo_name\"]]\n",
    "    .drop_duplicates()\n",
    "    .rename(columns={\"hpo_id\": \"mesh_id\", \"hpo_name\": \"mesh_name\"})\n",
    "    .assign(in_hsdn = 0)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mesh_id</th>\n",
       "      <th>mesh_name</th>\n",
       "      <th>in_hsdn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HP:0001531</td>\n",
       "      <td>Failure to thrive in infancy</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HP:0001943</td>\n",
       "      <td>Hypoglycemia</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>HP:0001987</td>\n",
       "      <td>Hyperammonemia</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>HP:0001992</td>\n",
       "      <td>Organic aciduria</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>HP:0002093</td>\n",
       "      <td>Respiratory insufficiency</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      mesh_id                     mesh_name  in_hsdn\n",
       "2  HP:0001531  Failure to thrive in infancy        0\n",
       "3  HP:0001943                  Hypoglycemia        0\n",
       "4  HP:0001987                Hyperammonemia        0\n",
       "5  HP:0001992              Organic aciduria        0\n",
       "6  HP:0002093     Respiratory insufficiency        0"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_hpo_terms.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "symptom_df = symptom_df.append(new_hpo_terms)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4100, 3)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "symptom_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mesh_id</th>\n",
       "      <th>mesh_name</th>\n",
       "      <th>in_hsdn</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>D000006</td>\n",
       "      <td>Abdomen, Acute</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>D000270</td>\n",
       "      <td>Adie Syndrome</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>D000326</td>\n",
       "      <td>Adrenoleukodystrophy</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>D000334</td>\n",
       "      <td>Aerophagy</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>D000370</td>\n",
       "      <td>Ageusia</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   mesh_id             mesh_name  in_hsdn\n",
       "0  D000006        Abdomen, Acute        1\n",
       "1  D000270         Adie Syndrome        0\n",
       "2  D000326  Adrenoleukodystrophy        0\n",
       "3  D000334             Aerophagy        1\n",
       "4  D000370               Ageusia        1"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "symptom_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "for i, row in symptom_df.iterrows():\n",
    "    url = 'http://identifiers.org/mesh/{}'.format(row.mesh_id)\n",
    "    data = {\n",
    "        'source': 'MeSH',\n",
    "        'url': url,\n",
    "        'license': 'CC0 1.0',\n",
    "    }\n",
    "    graph.add_node(kind='Symptom', identifier=row.mesh_id, name=row.mesh_name, data=data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pathway Nodes and edges"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>identifier</th>\n",
       "      <th>name</th>\n",
       "      <th>url</th>\n",
       "      <th>n_genes</th>\n",
       "      <th>n_coding_genes</th>\n",
       "      <th>source</th>\n",
       "      <th>license</th>\n",
       "      <th>genes</th>\n",
       "      <th>coding_genes</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1860</th>\n",
       "      <td>WP80_r68938</td>\n",
       "      <td>Nucleotide GPCRs</td>\n",
       "      <td>http://www.wikipathways.org/instance/WP80_r68938</td>\n",
       "      <td>11</td>\n",
       "      <td>11</td>\n",
       "      <td>WikiPathways</td>\n",
       "      <td>CC BY 3.0</td>\n",
       "      <td>10161|1241|134|135|136|140|2846|5028|5029|5030...</td>\n",
       "      <td>10161|1241|134|135|136|140|2846|5028|5029|5030...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1861</th>\n",
       "      <td>WP98_r72088</td>\n",
       "      <td>Prostaglandin Synthesis and Regulation</td>\n",
       "      <td>http://www.wikipathways.org/instance/WP98_r72088</td>\n",
       "      <td>30</td>\n",
       "      <td>30</td>\n",
       "      <td>WikiPathways</td>\n",
       "      <td>CC BY 3.0</td>\n",
       "      <td>1583|1906|1909|1910|301|302|306|307|308|309|32...</td>\n",
       "      <td>1583|1906|1909|1910|301|302|306|307|308|309|32...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       identifier                                    name  \\\n",
       "1860  WP80_r68938                        Nucleotide GPCRs   \n",
       "1861  WP98_r72088  Prostaglandin Synthesis and Regulation   \n",
       "\n",
       "                                                   url  n_genes  \\\n",
       "1860  http://www.wikipathways.org/instance/WP80_r68938       11   \n",
       "1861  http://www.wikipathways.org/instance/WP98_r72088       30   \n",
       "\n",
       "      n_coding_genes        source    license  \\\n",
       "1860              11  WikiPathways  CC BY 3.0   \n",
       "1861              30  WikiPathways  CC BY 3.0   \n",
       "\n",
       "                                                  genes  \\\n",
       "1860  10161|1241|134|135|136|140|2846|5028|5029|5030...   \n",
       "1861  1583|1906|1909|1910|301|302|306|307|308|309|32...   \n",
       "\n",
       "                                           coding_genes  \n",
       "1860  10161|1241|134|135|136|140|2846|5028|5029|5030...  \n",
       "1861  1583|1906|1909|1910|301|302|306|307|308|309|32...  "
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "commit = '1bd2c68853e38297d20f8f885419ff81fc0608a8'\n",
    "url = rawgit('dhimmel', 'pathways', commit, 'data/pathways.tsv')\n",
    "pathway_df = pandas.read_table(url)\n",
    "pathway_df = pathway_df.query(\"n_coding_genes > 1\")\n",
    "source_map = {\n",
    "    'wikipathways': 'WikiPathways',\n",
    "    'reactome': 'Reactome via Pathway Commons',\n",
    "    'pid': 'PID via Pathway Commons',\n",
    "}\n",
    "pathway_df.source = pathway_df.source.map(source_map)\n",
    "pathway_df.tail(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "for i, row in pathway_df.iterrows():\n",
    "    pathway_id = row.identifier\n",
    "    data = {'license': row.license, 'source': row.source}\n",
    "    if pandas.notnull(row.url):\n",
    "        data['url'] = row.url\n",
    "    graph.add_node(kind='Pathway', identifier=pathway_id, name=row['name'], data=data)\n",
    "    for gene in row.coding_genes.split('|'):\n",
    "        gene = int(gene)\n",
    "        source_id = 'Gene', gene\n",
    "        target_id = 'Pathway', pathway_id\n",
    "        edge_data = data.copy()\n",
    "        edge_data['unbiased'] = False\n",
    "        graph.add_edge(source_id, target_id, 'participates', 'both', edge_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pharmacologic Classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>class_id</th>\n",
       "      <th>class_name</th>\n",
       "      <th>class_source</th>\n",
       "      <th>class_type</th>\n",
       "      <th>url</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>240</th>\n",
       "      <td>N0000005760</td>\n",
       "      <td>Aminosalicylic Acids</td>\n",
       "      <td>FDA</td>\n",
       "      <td>Chemical/Ingredient</td>\n",
       "      <td>http://purl.bioontology.org/ontology/NDFRT/N00...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>241</th>\n",
       "      <td>N0000005934</td>\n",
       "      <td>Progesterone</td>\n",
       "      <td>FDA</td>\n",
       "      <td>Chemical/Ingredient</td>\n",
       "      <td>http://purl.bioontology.org/ontology/NDFRT/N00...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        class_id            class_name class_source           class_type  \\\n",
       "240  N0000005760  Aminosalicylic Acids          FDA  Chemical/Ingredient   \n",
       "241  N0000005934          Progesterone          FDA  Chemical/Ingredient   \n",
       "\n",
       "                                                   url  \n",
       "240  http://purl.bioontology.org/ontology/NDFRT/N00...  \n",
       "241  http://purl.bioontology.org/ontology/NDFRT/N00...  "
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "commit = '0e085ddb6aa42be367c85eba424edea716bebd94'\n",
    "url = rawgit('dhimmel', 'drugcentral', commit, 'rephetio/classes.tsv')\n",
    "class_df = pandas.read_table(url)\n",
    "class_types = {'Physiologic Effect', 'Mechanism of Action', 'Chemical/Ingredient'}\n",
    "class_df = class_df.query(\"class_type in @class_types\")\n",
    "class_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "for row in class_df.itertuples():\n",
    "    data = {\n",
    "        'source': '{} via DrugCentral'.format(row.class_source),\n",
    "        'class_type': row.class_type,\n",
    "        'license': 'CC BY 4.0',\n",
    "        'url': row.url,\n",
    "    }\n",
    "    graph.add_node(kind='Pharmacologic Class', identifier=row.class_id, name=row.class_name, data=data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>drugbank_id</th>\n",
       "      <th>drugbank_name</th>\n",
       "      <th>class_id</th>\n",
       "      <th>class_name</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>DB00014</td>\n",
       "      <td>Goserelin</td>\n",
       "      <td>N0000175654</td>\n",
       "      <td>Gonadotropin Releasing Hormone Receptor Agonists</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>DB00035</td>\n",
       "      <td>Desmopressin</td>\n",
       "      <td>N0000175914</td>\n",
       "      <td>Increased Coagulation Factor VIII Activity</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  drugbank_id drugbank_name     class_id  \\\n",
       "1     DB00014     Goserelin  N0000175654   \n",
       "5     DB00035  Desmopressin  N0000175914   \n",
       "\n",
       "                                         class_name  \n",
       "1  Gonadotropin Releasing Hormone Receptor Agonists  \n",
       "5        Increased Coagulation Factor VIII Activity  "
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "url = rawgit('dhimmel', 'drugcentral', commit, 'rephetio/drug-to-class.tsv')\n",
    "drug_class_df = pandas.read_table(url)\n",
    "drug_class_df = drug_class_df.query(\"class_id in @class_df.class_id\")\n",
    "drug_class_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for row in drug_class_df.itertuples():\n",
    "    data = {\n",
    "        'source': 'DrugCentral',\n",
    "        'license': 'CC BY 4.0',\n",
    "        'unbiased': False,\n",
    "    }\n",
    "    source_id = 'Pharmacologic Class', row.class_id\n",
    "    target_id = 'Compound', row.drugbank_id\n",
    "    graph.add_edge(source_id, target_id, 'includes', 'both', data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Gene ontology domains"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>go_id</th>\n",
       "      <th>go_name</th>\n",
       "      <th>go_domain</th>\n",
       "      <th>tax_id</th>\n",
       "      <th>annotation_type</th>\n",
       "      <th>size</th>\n",
       "      <th>gene_ids</th>\n",
       "      <th>gene_symbols</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>GO:0000002</td>\n",
       "      <td>mitochondrial genome maintenance</td>\n",
       "      <td>biological_process</td>\n",
       "      <td>9606</td>\n",
       "      <td>inferred</td>\n",
       "      <td>25</td>\n",
       "      <td>142|291|1763|1890|3980|4205|4358|4976|5428|708...</td>\n",
       "      <td>PARP1|SLC25A4|DNA2|TYMP|LIG3|MEF2A|MPV17|OPA1|...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>GO:0000003</td>\n",
       "      <td>reproduction</td>\n",
       "      <td>biological_process</td>\n",
       "      <td>9606</td>\n",
       "      <td>inferred</td>\n",
       "      <td>1333</td>\n",
       "      <td>18|49|51|90|91|92|100|109|113|116|117|133|150|...</td>\n",
       "      <td>ABAT|ACR|ACOX1|ACVR1|ACVR1B|ACVR2A|ADA|ADCY3|A...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        go_id                           go_name           go_domain  tax_id  \\\n",
       "0  GO:0000002  mitochondrial genome maintenance  biological_process    9606   \n",
       "1  GO:0000003                      reproduction  biological_process    9606   \n",
       "\n",
       "  annotation_type  size                                           gene_ids  \\\n",
       "0        inferred    25  142|291|1763|1890|3980|4205|4358|4976|5428|708...   \n",
       "1        inferred  1333  18|49|51|90|91|92|100|109|113|116|117|133|150|...   \n",
       "\n",
       "                                        gene_symbols  \n",
       "0  PARP1|SLC25A4|DNA2|TYMP|LIG3|MEF2A|MPV17|OPA1|...  \n",
       "1  ABAT|ACR|ACOX1|ACVR1|ACVR1B|ACVR2A|ADA|ADCY3|A...  "
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "commit = '87bab297f55db283e65a7a984607316b409415ae'\n",
    "url = rawgit('dhimmel', 'gene-ontology', commit, 'annotations/taxid_9606/GO_annotations-9606-inferred-allev.tsv')\n",
    "go_df = pandas.read_table(url)\n",
    "go_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for i, row in go_df.iterrows():\n",
    "    genes = coding_genes & set(map(int, row.gene_ids.split('|')))\n",
    "    if 2 > len(genes) or len(genes) > 1000:\n",
    "        continue\n",
    "    kind = row['go_domain'].replace('_', ' ').title()\n",
    "    data = {'source': 'Gene Ontology', 'url': obo_iri(row.go_id), 'license': 'CC BY 4.0'}\n",
    "    target = graph.add_node(kind=kind, identifier=row['go_id'], name=row['go_name'], data=data)\n",
    "    target_id = target.get_id()\n",
    "    for gene in genes:\n",
    "        source_id = 'Gene', gene\n",
    "        data = {'source': 'NCBI gene2go', 'unbiased': False, 'license': 'CC BY 4.0'}\n",
    "        graph.add_edge(source_id, target_id, 'participates', 'both', data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Disease-gene associations from compilation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>doid_id</th>\n",
       "      <th>doid_name</th>\n",
       "      <th>entrez_gene_id</th>\n",
       "      <th>gene_symbol</th>\n",
       "      <th>license</th>\n",
       "      <th>sources</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DOID:2531</td>\n",
       "      <td>Hematologic cancer</td>\n",
       "      <td>25</td>\n",
       "      <td>ABL1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>DISEASES|DisGeNET</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>DOID:2531</td>\n",
       "      <td>Hematologic cancer</td>\n",
       "      <td>27</td>\n",
       "      <td>ABL2</td>\n",
       "      <td>ODbL 1.0</td>\n",
       "      <td>DisGeNET</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     doid_id           doid_name  entrez_gene_id gene_symbol   license  \\\n",
       "0  DOID:2531  Hematologic cancer              25        ABL1       NaN   \n",
       "1  DOID:2531  Hematologic cancer              27        ABL2  ODbL 1.0   \n",
       "\n",
       "             sources  \n",
       "0  DISEASES|DisGeNET  \n",
       "1           DisGeNET  "
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "association_df = pandas.read_table('compile/DaG-association.tsv')\n",
    "association_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for i, row in association_df.iterrows():\n",
    "    source_id = 'Disease', row.doid_id\n",
    "    target_id = 'Gene', row.entrez_gene_id\n",
    "    sources = sorted(row.sources.split('|'))\n",
    "    data = {'sources': sources, 'unbiased': 'GWAS Catalog' in sources}\n",
    "    if pandas.notnull(row['license']):\n",
    "        data['license'] = row['license']\n",
    "    graph.add_edge(source_id, target_id, 'associates', 'both', data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Disease-gene differential expression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>slim_id</th>\n",
       "      <th>slim_name</th>\n",
       "      <th>entrez_gene_id</th>\n",
       "      <th>gene_symbol</th>\n",
       "      <th>direction</th>\n",
       "      <th>log2_fold_change</th>\n",
       "      <th>p_adjusted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DOID:0050156</td>\n",
       "      <td>idiopathic pulmonary fibrosis</td>\n",
       "      <td>55768</td>\n",
       "      <td>NGLY1</td>\n",
       "      <td>down</td>\n",
       "      <td>-0.004826</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>DOID:0050156</td>\n",
       "      <td>idiopathic pulmonary fibrosis</td>\n",
       "      <td>4927</td>\n",
       "      <td>NUP88</td>\n",
       "      <td>down</td>\n",
       "      <td>-0.005913</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        slim_id                      slim_name  entrez_gene_id gene_symbol  \\\n",
       "0  DOID:0050156  idiopathic pulmonary fibrosis           55768       NGLY1   \n",
       "1  DOID:0050156  idiopathic pulmonary fibrosis            4927       NUP88   \n",
       "\n",
       "  direction  log2_fold_change  p_adjusted  \n",
       "0      down         -0.004826         0.0  \n",
       "1      down         -0.005913         0.0  "
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "commit = '1a11633b5e0095454453335be82012a9f0f482e4'\n",
    "url = rawgit('dhimmel', 'stargeo', commit, 'data/diffex.tsv')\n",
    "stargeo_df = pandas.read_table(url)\n",
    "# Filter to at most 250 up and 250 down-regulated genes per disease\n",
    "stargeo_df = stargeo_df.groupby(['slim_id', 'direction']).apply(\n",
    "    lambda df: df.nsmallest(250, 'p_adjusted')).reset_index(drop=True)\n",
    "stargeo_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "for row in stargeo_df.itertuples():\n",
    "    source_id = 'Disease', row.slim_id\n",
    "    target_id = 'Gene', row.entrez_gene_id\n",
    "    kind = row.direction + 'regulates'\n",
    "    data = {\n",
    "        'source': 'STARGEO',\n",
    "        'log2_fold_change': round(row.log2_fold_change, 5),\n",
    "        'unbiased': True,\n",
    "        'license': 'CC0 1.0'\n",
    "    }\n",
    "    \n",
    "    try:\n",
    "        graph.add_edge(source_id, target_id, kind, 'both', data)\n",
    "    except AssertionError:\n",
    "        # edge already exists\n",
    "        continue"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Chemical Similarity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>compound0</th>\n",
       "      <th>compound1</th>\n",
       "      <th>similarity</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1557</th>\n",
       "      <td>DB01489</td>\n",
       "      <td>DB00683</td>\n",
       "      <td>0.5362</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1571</th>\n",
       "      <td>DB01489</td>\n",
       "      <td>DB01567</td>\n",
       "      <td>0.6260</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     compound0 compound1  similarity\n",
       "1557   DB01489   DB00683      0.5362\n",
       "1571   DB01489   DB01567      0.6260"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "commit = '3e87872db5fca5ac427ce27464ab945c0ceb4ec6'\n",
    "url = rawgit('dhimmel', 'drugbank', commit, 'data/similarity-slim.tsv.gz')\n",
    "chemical_df = pandas.read_table(url, compression='gzip')\n",
    "chemical_df = chemical_df[chemical_df.similarity >= 0.5]\n",
    "chemical_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "for i, row in chemical_df.iterrows():\n",
    "    source_id = 'Compound', row.compound0\n",
    "    target_id = 'Compound', row.compound1\n",
    "    data = {\n",
    "        'source': 'Dice similarity of ECFPs',\n",
    "        'similarity': round(row.similarity, 4),\n",
    "        'unbiased': True,\n",
    "        'license': 'CC0 1.0',\n",
    "    }\n",
    "    graph.add_edge(source_id, target_id, 'resembles', 'both', data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## MEDLINE cooccurrence data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "medline_data = {\n",
    "    'source': 'MEDLINE cooccurrence',\n",
    "    'unbiased': False,\n",
    "    'license': 'CC0 1.0',\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Symptom edges"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>doid_code</th>\n",
       "      <th>doid_name</th>\n",
       "      <th>mesh_id</th>\n",
       "      <th>mesh_name</th>\n",
       "      <th>cooccurrence</th>\n",
       "      <th>expected</th>\n",
       "      <th>enrichment</th>\n",
       "      <th>odds_ratio</th>\n",
       "      <th>p_fisher</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DOID:10652</td>\n",
       "      <td>Alzheimer's disease</td>\n",
       "      <td>D004314</td>\n",
       "      <td>Down Syndrome</td>\n",
       "      <td>800</td>\n",
       "      <td>35.619601</td>\n",
       "      <td>22.459544</td>\n",
       "      <td>39.918352</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>DOID:10652</td>\n",
       "      <td>Alzheimer's disease</td>\n",
       "      <td>D008569</td>\n",
       "      <td>Memory Disorders</td>\n",
       "      <td>1593</td>\n",
       "      <td>76.580532</td>\n",
       "      <td>20.801631</td>\n",
       "      <td>41.885877</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    doid_code            doid_name  mesh_id         mesh_name  cooccurrence  \\\n",
       "0  DOID:10652  Alzheimer's disease  D004314     Down Syndrome           800   \n",
       "1  DOID:10652  Alzheimer's disease  D008569  Memory Disorders          1593   \n",
       "\n",
       "    expected  enrichment  odds_ratio  p_fisher  \n",
       "0  35.619601   22.459544   39.918352       0.0  \n",
       "1  76.580532   20.801631   41.885877       0.0  "
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "commit = '60d611892bf387b5b23c5f2e2e3bc472cfce85f3'\n",
    "url = rawgit('dhimmel', 'medline', commit, 'data/disease-symptom-cooccurrence.tsv')\n",
    "disease_symptom_df = pandas.read_table(url)\n",
    "disease_symptom_df = disease_symptom_df[disease_symptom_df.p_fisher < 0.005]\n",
    "disease_symptom_df.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Rare disease phenotypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "rare_pheno_links = (pheno_links\n",
    "    [[\"dise_orpha_id\", \"dise_name\", \"fin_id\"]]\n",
    "    .drop_duplicates()\n",
    "    .merge(\n",
    "        symptom_df.drop(\"in_hsdn\", axis=1),\n",
    "        how=\"inner\", left_on=\"fin_id\", right_on=\"mesh_id\"\n",
    "    )\n",
    "    .drop(\"fin_id\", axis=1)\n",
    "    .rename(columns={\n",
    "        \"dise_orpha_id\": \"doid_code\",\n",
    "        \"dise_name\": \"doid_name\"\n",
    "    })\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>doid_code</th>\n",
       "      <th>doid_name</th>\n",
       "      <th>mesh_id</th>\n",
       "      <th>mesh_name</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ORPHA:6</td>\n",
       "      <td>3-methylcrotonyl-CoA carboxylase deficiency</td>\n",
       "      <td>D009123</td>\n",
       "      <td>Muscle Hypotonia</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  doid_code                                    doid_name  mesh_id  \\\n",
       "0   ORPHA:6  3-methylcrotonyl-CoA carboxylase deficiency  D009123   \n",
       "\n",
       "          mesh_name  \n",
       "0  Muscle Hypotonia  "
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rare_pheno_links.head(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(34289, 4)"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rare_pheno_links.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "disease_symptom_df = disease_symptom_df.append(rare_pheno_links)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for i, row in disease_symptom_df.iterrows():\n",
    "    source_id = 'Disease', row.doid_code\n",
    "    target_id = 'Symptom', row.mesh_id\n",
    "    data = medline_data.copy()\n",
    "    graph.add_edge(source_id, target_id, 'presents', 'both', data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Disease-localization edges"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>doid_code</th>\n",
       "      <th>doid_name</th>\n",
       "      <th>uberon_id</th>\n",
       "      <th>uberon_name</th>\n",
       "      <th>cooccurrence</th>\n",
       "      <th>expected</th>\n",
       "      <th>enrichment</th>\n",
       "      <th>odds_ratio</th>\n",
       "      <th>p_fisher</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DOID:10652</td>\n",
       "      <td>Alzheimer's disease</td>\n",
       "      <td>UBERON:0000955</td>\n",
       "      <td>brain</td>\n",
       "      <td>11209</td>\n",
       "      <td>1182.634069</td>\n",
       "      <td>9.477995</td>\n",
       "      <td>74.210761</td>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>DOID:10652</td>\n",
       "      <td>Alzheimer's disease</td>\n",
       "      <td>UBERON:0001890</td>\n",
       "      <td>forebrain</td>\n",
       "      <td>114</td>\n",
       "      <td>7.326350</td>\n",
       "      <td>15.560272</td>\n",
       "      <td>21.733764</td>\n",
       "      <td>5.971023e-99</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    doid_code            doid_name       uberon_id uberon_name  cooccurrence  \\\n",
       "0  DOID:10652  Alzheimer's disease  UBERON:0000955       brain         11209   \n",
       "1  DOID:10652  Alzheimer's disease  UBERON:0001890   forebrain           114   \n",
       "\n",
       "      expected  enrichment  odds_ratio      p_fisher  \n",
       "0  1182.634069    9.477995   74.210761  0.000000e+00  \n",
       "1     7.326350   15.560272   21.733764  5.971023e-99  "
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "commit = '60d611892bf387b5b23c5f2e2e3bc472cfce85f3'\n",
    "url = rawgit('dhimmel', 'medline', commit, 'data/disease-uberon-cooccurrence.tsv')\n",
    "disease_anatomy_df = pandas.read_table(url)\n",
    "disease_anatomy_df = disease_anatomy_df[disease_anatomy_df.p_fisher < 0.005]\n",
    "disease_anatomy_df = disease_anatomy_df[disease_anatomy_df.uberon_id.isin(uberon_df['uberon_id'])]\n",
    "disease_anatomy_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "for i, row in disease_anatomy_df.iterrows():\n",
    "    source_id = 'Disease', row.doid_code\n",
    "    target_id = 'Anatomy', row.uberon_id\n",
    "    data = medline_data.copy()\n",
    "    graph.add_edge(source_id, target_id, 'localizes', 'both', data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Disease-disease similarity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>doid_code_0</th>\n",
       "      <th>doid_name_0</th>\n",
       "      <th>doid_code_1</th>\n",
       "      <th>doid_name_1</th>\n",
       "      <th>cooccurrence</th>\n",
       "      <th>expected</th>\n",
       "      <th>enrichment</th>\n",
       "      <th>odds_ratio</th>\n",
       "      <th>p_fisher</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DOID:10652</td>\n",
       "      <td>Alzheimer's disease</td>\n",
       "      <td>DOID:14330</td>\n",
       "      <td>Parkinson's disease</td>\n",
       "      <td>2760</td>\n",
       "      <td>827.098152</td>\n",
       "      <td>3.336968</td>\n",
       "      <td>3.577398</td>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>DOID:10652</td>\n",
       "      <td>Alzheimer's disease</td>\n",
       "      <td>DOID:11949</td>\n",
       "      <td>Creutzfeldt-Jakob disease</td>\n",
       "      <td>332</td>\n",
       "      <td>96.723002</td>\n",
       "      <td>3.432482</td>\n",
       "      <td>3.593306</td>\n",
       "      <td>3.377672e-80</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  doid_code_0          doid_name_0 doid_code_1                doid_name_1  \\\n",
       "0  DOID:10652  Alzheimer's disease  DOID:14330        Parkinson's disease   \n",
       "1  DOID:10652  Alzheimer's disease  DOID:11949  Creutzfeldt-Jakob disease   \n",
       "\n",
       "   cooccurrence    expected  enrichment  odds_ratio      p_fisher  \n",
       "0          2760  827.098152    3.336968    3.577398  0.000000e+00  \n",
       "1           332   96.723002    3.432482    3.593306  3.377672e-80  "
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "commit = '60d611892bf387b5b23c5f2e2e3bc472cfce85f3'\n",
    "url = rawgit('dhimmel', 'medline', commit, 'data/disease-disease-cooccurrence.tsv')\n",
    "disease_similarity_df = pandas.read_table(url)\n",
    "disease_similarity_df = disease_similarity_df[-disease_similarity_df[['doid_code_0', 'doid_code_1']].apply(frozenset, 1).duplicated()]\n",
    "disease_similarity_df = disease_similarity_df[disease_similarity_df.p_fisher < 0.005]\n",
    "disease_similarity_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "for i, row in disease_similarity_df.iterrows():\n",
    "    source_id = 'Disease', row.doid_code_0\n",
    "    target_id = 'Disease', row.doid_code_1\n",
    "    data = medline_data.copy()\n",
    "    graph.add_edge(source_id, target_id, 'resembles', 'both', data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Anatomy-gene expression presence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>uberon_id</th>\n",
       "      <th>entrez_gene_id</th>\n",
       "      <th>sources</th>\n",
       "      <th>unbiased</th>\n",
       "      <th>license</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>41138</th>\n",
       "      <td>UBERON:0000002</td>\n",
       "      <td>2</td>\n",
       "      <td>Bgee</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41139</th>\n",
       "      <td>UBERON:0000002</td>\n",
       "      <td>9</td>\n",
       "      <td>Bgee</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            uberon_id  entrez_gene_id sources  unbiased license\n",
       "41138  UBERON:0000002               2    Bgee         1     NaN\n",
       "41139  UBERON:0000002               9    Bgee         1     NaN"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "expr_df = pandas.read_table('compile/AeG-expression.tsv.gz', low_memory=False)\n",
    "expr_df = expr_df[expr_df.uberon_id.isin(uberon_df.uberon_id) & expr_df.entrez_gene_id.isin(coding_genes)]\n",
    "expr_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for i, row in expr_df.iterrows():\n",
    "    source_id = 'Gene', row['entrez_gene_id']\n",
    "    target_id = 'Anatomy', row['uberon_id']\n",
    "    data = {'unbiased': bool(row['unbiased'])}\n",
    "    if pandas.notnull(row['license']):\n",
    "        data['license'] = row['license']\n",
    "    data['sources'] = row['sources'].split('|')\n",
    "    graph.add_edge(source_id, target_id, 'expresses', 'both', data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Anatomy-gene differential expression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>GeneID</th>\n",
       "      <th>uberon_id</th>\n",
       "      <th>direction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>37241</th>\n",
       "      <td>2</td>\n",
       "      <td>UBERON:0000002</td>\n",
       "      <td>upregulates</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37249</th>\n",
       "      <td>18</td>\n",
       "      <td>UBERON:0000002</td>\n",
       "      <td>downregulates</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       GeneID       uberon_id      direction\n",
       "37241       2  UBERON:0000002    upregulates\n",
       "37249      18  UBERON:0000002  downregulates"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "commit = '08ba54e83ee8e28dec22b4351d29e23f1d034d30'\n",
    "url = rawgit('dhimmel', 'bgee', commit, 'data/diffex.tsv.gz')\n",
    "diffex_df = pandas.read_table(url, compression='gzip')\n",
    "diffex_df = pandas.melt(diffex_df, id_vars='GeneID', var_name='uberon_id', value_name='direction')\n",
    "diffex_df = diffex_df.query(\"direction != 0\")\n",
    "diffex_df = diffex_df[diffex_df.uberon_id.isin(uberon_df.uberon_id) & diffex_df.GeneID.isin(coding_genes)]\n",
    "diffex_df = diffex_df.replace({'direction': {-1: 'downregulates', 1: 'upregulates'}})\n",
    "diffex_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for i, row in diffex_df.iterrows():\n",
    "    source_id = 'Gene', row['GeneID']\n",
    "    target_id = 'Anatomy', row['uberon_id']\n",
    "    data = {'source': 'Bgee', 'unbiased': True}\n",
    "    graph.add_edge(source_id, target_id, row['direction'], 'both', data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Compound bindings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>drugbank_id</th>\n",
       "      <th>entrez_gene_id</th>\n",
       "      <th>sources</th>\n",
       "      <th>pubmed_ids</th>\n",
       "      <th>actions</th>\n",
       "      <th>affinity_nM</th>\n",
       "      <th>license</th>\n",
       "      <th>urls</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DB00014</td>\n",
       "      <td>2798</td>\n",
       "      <td>DrugBank (target)|DrugCentral (ChEMBL)</td>\n",
       "      <td>20053189</td>\n",
       "      <td>agonist</td>\n",
       "      <td>NaN</td>\n",
       "      <td>CC BY-NC 4.0</td>\n",
       "      <td>https://www.ebi.ac.uk/chembl/compound/inspect/...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>DB00014</td>\n",
       "      <td>3973</td>\n",
       "      <td>DrugBank (target)</td>\n",
       "      <td>17016423|17139284|20053189</td>\n",
       "      <td>agonist</td>\n",
       "      <td>NaN</td>\n",
       "      <td>CC BY-NC 4.0</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  drugbank_id  entrez_gene_id                                 sources  \\\n",
       "0     DB00014            2798  DrugBank (target)|DrugCentral (ChEMBL)   \n",
       "1     DB00014            3973                       DrugBank (target)   \n",
       "\n",
       "                   pubmed_ids  actions  affinity_nM       license  \\\n",
       "0                    20053189  agonist          NaN  CC BY-NC 4.0   \n",
       "1  17016423|17139284|20053189  agonist          NaN  CC BY-NC 4.0   \n",
       "\n",
       "                                                urls  \n",
       "0  https://www.ebi.ac.uk/chembl/compound/inspect/...  \n",
       "1                                                NaN  "
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "binding_df = pandas.read_table('compile/CbG-binding.tsv')\n",
    "binding_df = binding_df.merge(compound_df[['drugbank_id']])\n",
    "binding_df = binding_df[binding_df.entrez_gene_id.isin(coding_genes)]\n",
    "binding_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "for i, row in binding_df.iterrows():\n",
    "    source_id = 'Compound', row.drugbank_id\n",
    "    target_id = 'Gene', row.entrez_gene_id\n",
    "    data = {'unbiased': False}\n",
    "    # singular fields\n",
    "    for key in 'affinity_nM', 'license':\n",
    "        value = row[key]\n",
    "        if pandas.notnull(value):\n",
    "            data[key] = value\n",
    "    # compound fields\n",
    "    for key in 'sources', 'pubmed_ids', 'actions', 'urls':\n",
    "        value = row[key]\n",
    "        if pandas.notnull(value):\n",
    "            data[key] = value.split('|')\n",
    "    graph.add_edge(source_id, target_id, 'binds', 'both', data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Protein interactions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>gene_0</th>\n",
       "      <th>gene_1</th>\n",
       "      <th>sources</th>\n",
       "      <th>unbiased</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>310</td>\n",
       "      <td>hetio-dag</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1026</td>\n",
       "      <td>hetio-dag</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   gene_0  gene_1    sources  unbiased\n",
       "0       1     310  hetio-dag         0\n",
       "1       1    1026  hetio-dag         0"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "commit = 'f6a7edbc8de6ba2d7fe1ef3fee4d89e5b8d0b900'\n",
    "url = rawgit('dhimmel', 'ppi', commit, 'data/ppi-hetio-ind.tsv')\n",
    "ppi_df = pandas.read_table(url)\n",
    "ppi_df = ppi_df[ppi_df.gene_0.isin(coding_genes) & ppi_df.gene_1.isin(coding_genes)]\n",
    "# Consider allowing self-interactions in future versions of Hetionet by removing following \n",
    "# line of code. See https://github.com/dhimmel/hetio/issues/10\n",
    "ppi_df = ppi_df.query(\"gene_0 != gene_1\")\n",
    "ppi_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for i, row in ppi_df.iterrows():\n",
    "    source_id = 'Gene', row.gene_0\n",
    "    target_id = 'Gene', row.gene_1\n",
    "    data = {\n",
    "        'sources': row.sources.split('|'),\n",
    "        'unbiased': bool(row.unbiased),\n",
    "    }\n",
    "    graph.add_edge(source_id, target_id, 'interacts', 'both', data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Evolutionary rate covariation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>source_entrez</th>\n",
       "      <th>target_entrez</th>\n",
       "      <th>correlation</th>\n",
       "      <th>n_ucsc_genes</th>\n",
       "      <th>gene_pair</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>9</td>\n",
       "      <td>3295</td>\n",
       "      <td>0.77623</td>\n",
       "      <td>1</td>\n",
       "      <td>(9, 3295)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>9</td>\n",
       "      <td>23313</td>\n",
       "      <td>0.79472</td>\n",
       "      <td>1</td>\n",
       "      <td>(9, 23313)</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    source_entrez  target_entrez  correlation  n_ucsc_genes   gene_pair\n",
       "29              9           3295      0.77623             1   (9, 3295)\n",
       "46              9          23313      0.79472             1  (9, 23313)"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "commit = '757733f77a89499439c887acb88456e011c5322e'\n",
    "url = rawgit('dhimmel', 'erc', commit, 'data/erc_mam33-entrez-gt-0.6.tsv.gz')\n",
    "erc_df = pandas.read_table(url, compression='gzip')\n",
    "erc_df = erc_df[erc_df.correlation >= 0.75]\n",
    "erc_df = erc_df[erc_df.source_entrez.isin(coding_genes) & erc_df.target_entrez.isin(coding_genes)]\n",
    "# Drop self-loops (see https://github.com/dhimmel/erc/issues/2)\n",
    "erc_df = erc_df.query(\"source_entrez != target_entrez\")\n",
    "# Drop duplicate gene pair(s) (see https://github.com/dhimmel/erc/issues/1)\n",
    "erc_df['gene_pair'] = list(map(frozenset, zip(erc_df.source_entrez, erc_df.target_entrez)))\n",
    "erc_df = erc_df.drop_duplicates('gene_pair', keep='last')\n",
    "erc_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for i, row in erc_df.iterrows():\n",
    "    source_id = 'Gene', row.source_entrez\n",
    "    target_id = 'Gene', row.target_entrez\n",
    "    data = {\n",
    "        'source': 'ERC',\n",
    "        'unbiased': True,\n",
    "    }\n",
    "    graph.add_edge(source_id, target_id, 'covaries', 'both', data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Indications from the PharmacotherapyDB"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "with open(\"../crossval_idx.txt\", \"r\") as fin:\n",
    "    piece_idx = int(fin.read().strip())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "piece_idx"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# choose one subset of the data to use\n",
    "floc = \"../../crossval/data/training/training_piece{}.tsv\".format(piece_idx)\n",
    "indication_df = pandas.read_csv(floc, sep = '\\t')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(138639, 7)"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "indication_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>category</th>\n",
       "      <th>chemical_id</th>\n",
       "      <th>chemical_name</th>\n",
       "      <th>dise_type</th>\n",
       "      <th>disease_id</th>\n",
       "      <th>disease_name</th>\n",
       "      <th>piece</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>NaN</td>\n",
       "      <td>DB00035</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>ORPHA:100054</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>NaN</td>\n",
       "      <td>DB00035</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>ORPHA:100976</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>NaN</td>\n",
       "      <td>DB00035</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>ORPHA:1010</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>NaN</td>\n",
       "      <td>DB00035</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>ORPHA:101041</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>NaN</td>\n",
       "      <td>DB00035</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>ORPHA:103908</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  category chemical_id chemical_name dise_type    disease_id disease_name  \\\n",
       "0      NaN     DB00035           NaN       NaN  ORPHA:100054          NaN   \n",
       "1      NaN     DB00035           NaN       NaN  ORPHA:100976          NaN   \n",
       "2      NaN     DB00035           NaN       NaN    ORPHA:1010          NaN   \n",
       "3      NaN     DB00035           NaN       NaN  ORPHA:101041          NaN   \n",
       "4      NaN     DB00035           NaN       NaN  ORPHA:103908          NaN   \n",
       "\n",
       "   piece  \n",
       "0    NaN  \n",
       "1    NaN  \n",
       "2    NaN  \n",
       "3    NaN  \n",
       "4    NaN  "
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "indication_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "indication_df[\"category\"].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "categories = {'DM'}\n",
    "\n",
    "indication_df = indication_df.query(\"category in @categories\").copy()\n",
    "\n",
    "indication_df['kind'] = indication_df.category.map({'DM': 'treats', 'SYM': 'palliates'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "indication_df = indication_df[indication_df[\"disease_name\"].notnull()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1631, 7)"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "indication_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "indication_df = indication_df.assign(kind = \"treats\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>category</th>\n",
       "      <th>chemical_id</th>\n",
       "      <th>chemical_name</th>\n",
       "      <th>dise_type</th>\n",
       "      <th>disease_id</th>\n",
       "      <th>disease_name</th>\n",
       "      <th>piece</th>\n",
       "      <th>kind</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>DM</td>\n",
       "      <td>DB00035</td>\n",
       "      <td>Desmopressin</td>\n",
       "      <td>rare_genetic</td>\n",
       "      <td>ORPHA:166084</td>\n",
       "      <td>Von Willebrand disease type 2A</td>\n",
       "      <td>4.0</td>\n",
       "      <td>treats</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>DM</td>\n",
       "      <td>DB00035</td>\n",
       "      <td>Desmopressin</td>\n",
       "      <td>rare_genetic</td>\n",
       "      <td>ORPHA:166087</td>\n",
       "      <td>Von Willebrand disease type 2B</td>\n",
       "      <td>2.0</td>\n",
       "      <td>treats</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>DM</td>\n",
       "      <td>DB00035</td>\n",
       "      <td>Desmopressin</td>\n",
       "      <td>rare_genetic</td>\n",
       "      <td>ORPHA:166090</td>\n",
       "      <td>Von Willebrand disease type 2M</td>\n",
       "      <td>2.0</td>\n",
       "      <td>treats</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>DM</td>\n",
       "      <td>DB00035</td>\n",
       "      <td>Desmopressin</td>\n",
       "      <td>rare_genetic</td>\n",
       "      <td>ORPHA:166093</td>\n",
       "      <td>Von Willebrand disease type 2N</td>\n",
       "      <td>2.0</td>\n",
       "      <td>treats</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>DM</td>\n",
       "      <td>DB00035</td>\n",
       "      <td>Desmopressin</td>\n",
       "      <td>rare_genetic</td>\n",
       "      <td>ORPHA:166096</td>\n",
       "      <td>Von Willebrand disease type 3</td>\n",
       "      <td>1.0</td>\n",
       "      <td>treats</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   category chemical_id chemical_name     dise_type    disease_id  \\\n",
       "28       DM     DB00035  Desmopressin  rare_genetic  ORPHA:166084   \n",
       "29       DM     DB00035  Desmopressin  rare_genetic  ORPHA:166087   \n",
       "30       DM     DB00035  Desmopressin  rare_genetic  ORPHA:166090   \n",
       "31       DM     DB00035  Desmopressin  rare_genetic  ORPHA:166093   \n",
       "32       DM     DB00035  Desmopressin  rare_genetic  ORPHA:166096   \n",
       "\n",
       "                      disease_name  piece    kind  \n",
       "28  Von Willebrand disease type 2A    4.0  treats  \n",
       "29  Von Willebrand disease type 2B    2.0  treats  \n",
       "30  Von Willebrand disease type 2M    2.0  treats  \n",
       "31  Von Willebrand disease type 2N    2.0  treats  \n",
       "32   Von Willebrand disease type 3    1.0  treats  "
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "indication_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "551"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "indication_df[\"disease_id\"].nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "297"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "indication_df[\"chemical_id\"].nunique()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Add to graph"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for i, row in indication_df.iterrows():\n",
    "    source_id = ('Disease', row.disease_id)\n",
    "    target_id = ('Compound', row.chemical_id)\n",
    "    data = {'source': 'PharmacotherapyDB', 'unbiased': False, 'license': 'CC0 1.0'}\n",
    "    graph.add_edge(source_id, target_id, row['kind'], 'both', data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LINCS L1000 relationships"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "commit = 'abcb12f942f93e3ee839e5e3593f930df2c56845'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def filter_l1000_df(df, n):\n",
    "    \"\"\"\n",
    "    Filter LINCS L1000 differentially expression genes to at most `n` genes\n",
    "    per perturbagen-direction-status combination.\n",
    "    \"\"\"\n",
    "    df = df.groupby(['perturbagen', 'direction', 'status']).apply(\n",
    "        lambda x: x.nlargest(n, 'nlog10_bonferroni_pval')).reset_index(drop=True)\n",
    "    return df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LINCS compound-gene dysregulation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>perturbagen</th>\n",
       "      <th>entrez_gene_id</th>\n",
       "      <th>z_score</th>\n",
       "      <th>symbol</th>\n",
       "      <th>status</th>\n",
       "      <th>direction</th>\n",
       "      <th>nlog10_bonferroni_pval</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>39856</th>\n",
       "      <td>DB09020</td>\n",
       "      <td>3895</td>\n",
       "      <td>4.077</td>\n",
       "      <td>KTN1</td>\n",
       "      <td>measured</td>\n",
       "      <td>up</td>\n",
       "      <td>1.351</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39857</th>\n",
       "      <td>DB09020</td>\n",
       "      <td>23368</td>\n",
       "      <td>4.062</td>\n",
       "      <td>PPP1R13B</td>\n",
       "      <td>measured</td>\n",
       "      <td>up</td>\n",
       "      <td>1.323</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      perturbagen  entrez_gene_id  z_score    symbol    status direction  \\\n",
       "39856     DB09020            3895    4.077      KTN1  measured        up   \n",
       "39857     DB09020           23368    4.062  PPP1R13B  measured        up   \n",
       "\n",
       "       nlog10_bonferroni_pval  \n",
       "39856                   1.351  \n",
       "39857                   1.323  "
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "url = rawgit('dhimmel', 'lincs', commit, 'data/consensi/signif/dysreg-drugbank.tsv')\n",
    "l1000_df = pandas.read_table(url)\n",
    "l1000_df = l1000_df.query(\"perturbagen in @compound_df.drugbank_id and entrez_gene_id in @coding_genes\")\n",
    "l1000_df = filter_l1000_df(l1000_df, n=125)\n",
    "l1000_df.tail(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "mapper = {'up': 'upregulates', 'down': 'downregulates'}\n",
    "for row in l1000_df.itertuples():\n",
    "    source_id = 'Compound', row.perturbagen\n",
    "    target_id = 'Gene', row.entrez_gene_id\n",
    "    data = {\n",
    "        'source': 'LINCS L1000',\n",
    "        'z_score': round(row.z_score, 3),\n",
    "        'method': row.status,\n",
    "        'unbiased': True,\n",
    "    }\n",
    "    kind = mapper[row.direction]\n",
    "    \n",
    "    # duplicate edges\n",
    "    try:\n",
    "        graph.add_edge(source_id, target_id, kind, 'both', data)\n",
    "    except AssertionError:\n",
    "        continue"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LINCS genetic perturbations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "url = rawgit('dhimmel', 'lincs', commit, 'data/consensi/signif/dysreg-knockdown.tsv')\n",
    "l1000_kd_df = filter_l1000_df(pandas.read_table(url), n=50)\n",
    "\n",
    "url = rawgit('dhimmel', 'lincs', commit, 'data/consensi/signif/dysreg-overexpression.tsv')\n",
    "l1000_oe_df = filter_l1000_df(pandas.read_table(url), n=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>perturbagen</th>\n",
       "      <th>entrez_gene_id</th>\n",
       "      <th>z_score</th>\n",
       "      <th>symbol</th>\n",
       "      <th>status</th>\n",
       "      <th>direction</th>\n",
       "      <th>nlog10_bonferroni_pval</th>\n",
       "      <th>kind</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>7422</td>\n",
       "      <td>-7.823</td>\n",
       "      <td>VEGFA</td>\n",
       "      <td>imputed</td>\n",
       "      <td>down</td>\n",
       "      <td>10.475</td>\n",
       "      <td>knockdown downregulates</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>6382</td>\n",
       "      <td>-6.402</td>\n",
       "      <td>SDC1</td>\n",
       "      <td>imputed</td>\n",
       "      <td>down</td>\n",
       "      <td>6.002</td>\n",
       "      <td>knockdown downregulates</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   perturbagen  entrez_gene_id  z_score symbol   status direction  \\\n",
       "0            2            7422   -7.823  VEGFA  imputed      down   \n",
       "1            2            6382   -6.402   SDC1  imputed      down   \n",
       "\n",
       "   nlog10_bonferroni_pval                     kind  \n",
       "0                  10.475  knockdown downregulates  \n",
       "1                   6.002  knockdown downregulates  "
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mapper = {'up': 'knockdown upregulates', 'down': 'knockdown downregulates'}\n",
    "l1000_kd_df['kind'] = l1000_kd_df.direction.map(lambda x: mapper[x])\n",
    "\n",
    "mapper = {'up': 'overexpression upregulates', 'down': 'overexpression downregulates'}\n",
    "l1000_oe_df['kind'] = l1000_oe_df.direction.map(lambda x: mapper[x])\n",
    "\n",
    "l1000_genetic_df = pandas.concat([l1000_kd_df, l1000_oe_df])\n",
    "l1000_genetic_df = l1000_genetic_df.query('perturbagen in @coding_genes and entrez_gene_id in @coding_genes')\n",
    "l1000_genetic_df = l1000_genetic_df.query('perturbagen != entrez_gene_id')\n",
    "\n",
    "l1000_genetic_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for (pert, gene), df in l1000_genetic_df.groupby(['perturbagen', 'entrez_gene_id'], sort=False):\n",
    "    source_id = 'Gene', pert\n",
    "    target_id = 'Gene', gene\n",
    "    method, = df.status.unique()\n",
    "    data = {'source': 'LINCS L1000', 'subtypes': list(df.kind), 'method': method, 'unbiased': True}\n",
    "    graph.add_edge(source_id, target_id, 'regulates', 'forward', data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Side effects - SIDER"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>umls_cui_from_meddra</th>\n",
       "      <th>side_effect_name</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>C0863129</td>\n",
       "      <td>AGEP</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>C0001849</td>\n",
       "      <td>AIDS dementia complex</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  umls_cui_from_meddra       side_effect_name\n",
       "0             C0863129                   AGEP\n",
       "1             C0001849  AIDS dementia complex"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "commit = 'be3adebc0d845baaddb907a880890cb5e85f5801'\n",
    "url = rawgit('dhimmel', 'SIDER4', commit, 'data/side-effect-terms.tsv')\n",
    "side_effect_df = pandas.read_table(url)\n",
    "side_effect_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for i, row in side_effect_df.iterrows():\n",
    "    umls_id = row['umls_cui_from_meddra']\n",
    "    data = {\n",
    "        'source': 'UMLS via SIDER 4.1',\n",
    "        'url': 'http://identifiers.org/umls/{}'.format(umls_id),\n",
    "        'license': 'CC BY-NC-SA 4.0',\n",
    "    }\n",
    "    graph.add_node(kind='Side Effect', identifier=umls_id, name=row['side_effect_name'], data=data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>drugbank_id</th>\n",
       "      <th>drugbank_name</th>\n",
       "      <th>umls_cui_from_meddra</th>\n",
       "      <th>side_effect_name</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6398</th>\n",
       "      <td>DB01048</td>\n",
       "      <td>Abacavir</td>\n",
       "      <td>C0001849</td>\n",
       "      <td>AIDS dementia complex</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6399</th>\n",
       "      <td>DB01048</td>\n",
       "      <td>Abacavir</td>\n",
       "      <td>C0232487</td>\n",
       "      <td>Abdominal discomfort</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     drugbank_id drugbank_name umls_cui_from_meddra       side_effect_name\n",
       "6398     DB01048      Abacavir             C0001849  AIDS dementia complex\n",
       "6399     DB01048      Abacavir             C0232487   Abdominal discomfort"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "url = rawgit('dhimmel', 'SIDER4', commit, 'data/side-effects.tsv')\n",
    "sider_df = pandas.read_table(url)\n",
    "sider_df = sider_df[sider_df.drugbank_id.isin(compound_df.drugbank_id)]\n",
    "sider_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "for i, row in sider_df.iterrows():\n",
    "    umls_id = row.umls_cui_from_meddra\n",
    "    source_id = 'Compound', row.drugbank_id\n",
    "    target_id = 'Side Effect', umls_id\n",
    "    data = {\n",
    "        'source': 'SIDER 4.1',\n",
    "        'url': 'http://sideeffects.embl.de/se/{}/'.format(umls_id),\n",
    "        'unbiased': False,\n",
    "        'license': 'CC BY-NC-SA 4.0',\n",
    "    }\n",
    "    graph.add_edge(source_id, target_id, 'causes', 'both', data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Network visualizations and stats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Export node degree tables\n",
    "hetio.stats.degrees_to_excel(graph, 'data/summary/degrees.xlsx')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# Create and save degree distribution vizualizations\n",
    "# hetio.stats.plot_degrees(graph, 'viz/degrees.pdf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>metanode</th>\n",
       "      <th>abbreviation</th>\n",
       "      <th>metaedges</th>\n",
       "      <th>nodes</th>\n",
       "      <th>unconnected_nodes</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Anatomy</td>\n",
       "      <td>A</td>\n",
       "      <td>4</td>\n",
       "      <td>402</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Biological Process</td>\n",
       "      <td>BP</td>\n",
       "      <td>1</td>\n",
       "      <td>11381</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Cellular Component</td>\n",
       "      <td>CC</td>\n",
       "      <td>1</td>\n",
       "      <td>1391</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Compound</td>\n",
       "      <td>C</td>\n",
       "      <td>7</td>\n",
       "      <td>1552</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Disease</td>\n",
       "      <td>D</td>\n",
       "      <td>7</td>\n",
       "      <td>4682</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Gene</td>\n",
       "      <td>G</td>\n",
       "      <td>16</td>\n",
       "      <td>20945</td>\n",
       "      <td>1799</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Molecular Function</td>\n",
       "      <td>MF</td>\n",
       "      <td>1</td>\n",
       "      <td>2884</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Pathway</td>\n",
       "      <td>PW</td>\n",
       "      <td>1</td>\n",
       "      <td>1822</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Pharmacologic Class</td>\n",
       "      <td>PC</td>\n",
       "      <td>1</td>\n",
       "      <td>345</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Side Effect</td>\n",
       "      <td>SE</td>\n",
       "      <td>1</td>\n",
       "      <td>5734</td>\n",
       "      <td>33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Symptom</td>\n",
       "      <td>S</td>\n",
       "      <td>1</td>\n",
       "      <td>4100</td>\n",
       "      <td>22</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               metanode abbreviation  metaedges  nodes  unconnected_nodes\n",
       "6               Anatomy            A          4    402                  2\n",
       "4    Biological Process           BP          1  11381                  0\n",
       "2    Cellular Component           CC          1   1391                  0\n",
       "5              Compound            C          7   1552                 14\n",
       "8               Disease            D          7   4682                  1\n",
       "10                 Gene            G         16  20945               1799\n",
       "9    Molecular Function           MF          1   2884                  0\n",
       "7               Pathway           PW          1   1822                  0\n",
       "3   Pharmacologic Class           PC          1    345                  0\n",
       "1           Side Effect           SE          1   5734                 33\n",
       "0               Symptom            S          1   4100                 22"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Summary of metanodes and cooresponding nodes\n",
    "metanode_df = hetio.stats.get_metanode_df(graph)\n",
    "metanode_df.to_csv('data/summary/metanodes.tsv', sep='\\t', index=False)\n",
    "metanode_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "55238"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Total number of nodes\n",
    "metanode_df.nodes.sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>metaedge</th>\n",
       "      <th>abbreviation</th>\n",
       "      <th>edges</th>\n",
       "      <th>source_nodes</th>\n",
       "      <th>target_nodes</th>\n",
       "      <th>unbiased</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Anatomy - downregulates - Gene</td>\n",
       "      <td>AdG</td>\n",
       "      <td>102240</td>\n",
       "      <td>36</td>\n",
       "      <td>15097</td>\n",
       "      <td>102240</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Anatomy - expresses - Gene</td>\n",
       "      <td>AeG</td>\n",
       "      <td>526407</td>\n",
       "      <td>241</td>\n",
       "      <td>18094</td>\n",
       "      <td>453477</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Anatomy - upregulates - Gene</td>\n",
       "      <td>AuG</td>\n",
       "      <td>97848</td>\n",
       "      <td>36</td>\n",
       "      <td>15929</td>\n",
       "      <td>97848</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Compound - binds - Gene</td>\n",
       "      <td>CbG</td>\n",
       "      <td>11571</td>\n",
       "      <td>1389</td>\n",
       "      <td>1689</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Compound - causes - Side Effect</td>\n",
       "      <td>CcSE</td>\n",
       "      <td>138944</td>\n",
       "      <td>1071</td>\n",
       "      <td>5701</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Compound - downregulates - Gene</td>\n",
       "      <td>CdG</td>\n",
       "      <td>21102</td>\n",
       "      <td>734</td>\n",
       "      <td>2880</td>\n",
       "      <td>21102</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Compound - resembles - Compound</td>\n",
       "      <td>CrC</td>\n",
       "      <td>6486</td>\n",
       "      <td>1042</td>\n",
       "      <td>1054</td>\n",
       "      <td>6486</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Compound - treats - Disease</td>\n",
       "      <td>CtD</td>\n",
       "      <td>1631</td>\n",
       "      <td>297</td>\n",
       "      <td>551</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Compound - upregulates - Gene</td>\n",
       "      <td>CuG</td>\n",
       "      <td>18756</td>\n",
       "      <td>703</td>\n",
       "      <td>3247</td>\n",
       "      <td>18756</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Disease - associates - Gene</td>\n",
       "      <td>DaG</td>\n",
       "      <td>19452</td>\n",
       "      <td>3700</td>\n",
       "      <td>7088</td>\n",
       "      <td>1284</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Disease - downregulates - Gene</td>\n",
       "      <td>DdG</td>\n",
       "      <td>7623</td>\n",
       "      <td>44</td>\n",
       "      <td>5745</td>\n",
       "      <td>7623</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Disease - localizes - Anatomy</td>\n",
       "      <td>DlA</td>\n",
       "      <td>3602</td>\n",
       "      <td>133</td>\n",
       "      <td>398</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Disease - presents - Symptom</td>\n",
       "      <td>DpS</td>\n",
       "      <td>37646</td>\n",
       "      <td>2436</td>\n",
       "      <td>4078</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>Disease - resembles - Disease</td>\n",
       "      <td>DrD</td>\n",
       "      <td>543</td>\n",
       "      <td>112</td>\n",
       "      <td>106</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>Disease - upregulates - Gene</td>\n",
       "      <td>DuG</td>\n",
       "      <td>7731</td>\n",
       "      <td>44</td>\n",
       "      <td>5630</td>\n",
       "      <td>7731</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Gene - covaries - Gene</td>\n",
       "      <td>GcG</td>\n",
       "      <td>61690</td>\n",
       "      <td>9043</td>\n",
       "      <td>9532</td>\n",
       "      <td>61690</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>Gene - interacts - Gene</td>\n",
       "      <td>GiG</td>\n",
       "      <td>147164</td>\n",
       "      <td>9526</td>\n",
       "      <td>14084</td>\n",
       "      <td>15517</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Gene - participates - Biological Process</td>\n",
       "      <td>GpBP</td>\n",
       "      <td>559504</td>\n",
       "      <td>14772</td>\n",
       "      <td>11381</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>Gene - participates - Cellular Component</td>\n",
       "      <td>GpCC</td>\n",
       "      <td>73566</td>\n",
       "      <td>10580</td>\n",
       "      <td>1391</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>Gene - participates - Molecular Function</td>\n",
       "      <td>GpMF</td>\n",
       "      <td>97222</td>\n",
       "      <td>13063</td>\n",
       "      <td>2884</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>Gene - participates - Pathway</td>\n",
       "      <td>GpPW</td>\n",
       "      <td>84372</td>\n",
       "      <td>8979</td>\n",
       "      <td>1822</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>Gene &gt; regulates &gt; Gene</td>\n",
       "      <td>Gr&gt;G</td>\n",
       "      <td>265672</td>\n",
       "      <td>4634</td>\n",
       "      <td>7048</td>\n",
       "      <td>265672</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>Pharmacologic Class - includes - Compound</td>\n",
       "      <td>PCiC</td>\n",
       "      <td>1029</td>\n",
       "      <td>345</td>\n",
       "      <td>724</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                     metaedge abbreviation   edges  \\\n",
       "0              Anatomy - downregulates - Gene          AdG  102240   \n",
       "1                  Anatomy - expresses - Gene          AeG  526407   \n",
       "2                Anatomy - upregulates - Gene          AuG   97848   \n",
       "3                     Compound - binds - Gene          CbG   11571   \n",
       "4             Compound - causes - Side Effect         CcSE  138944   \n",
       "5             Compound - downregulates - Gene          CdG   21102   \n",
       "6             Compound - resembles - Compound          CrC    6486   \n",
       "7                 Compound - treats - Disease          CtD    1631   \n",
       "8               Compound - upregulates - Gene          CuG   18756   \n",
       "9                 Disease - associates - Gene          DaG   19452   \n",
       "10             Disease - downregulates - Gene          DdG    7623   \n",
       "11              Disease - localizes - Anatomy          DlA    3602   \n",
       "12               Disease - presents - Symptom          DpS   37646   \n",
       "13              Disease - resembles - Disease          DrD     543   \n",
       "14               Disease - upregulates - Gene          DuG    7731   \n",
       "15                     Gene - covaries - Gene          GcG   61690   \n",
       "16                    Gene - interacts - Gene          GiG  147164   \n",
       "17   Gene - participates - Biological Process         GpBP  559504   \n",
       "18   Gene - participates - Cellular Component         GpCC   73566   \n",
       "19   Gene - participates - Molecular Function         GpMF   97222   \n",
       "20              Gene - participates - Pathway         GpPW   84372   \n",
       "21                    Gene > regulates > Gene         Gr>G  265672   \n",
       "22  Pharmacologic Class - includes - Compound         PCiC    1029   \n",
       "\n",
       "    source_nodes  target_nodes  unbiased  \n",
       "0             36         15097    102240  \n",
       "1            241         18094    453477  \n",
       "2             36         15929     97848  \n",
       "3           1389          1689         0  \n",
       "4           1071          5701         0  \n",
       "5            734          2880     21102  \n",
       "6           1042          1054      6486  \n",
       "7            297           551         0  \n",
       "8            703          3247     18756  \n",
       "9           3700          7088      1284  \n",
       "10            44          5745      7623  \n",
       "11           133           398         0  \n",
       "12          2436          4078         0  \n",
       "13           112           106         0  \n",
       "14            44          5630      7731  \n",
       "15          9043          9532     61690  \n",
       "16          9526         14084     15517  \n",
       "17         14772         11381         0  \n",
       "18         10580          1391         0  \n",
       "19         13063          2884         0  \n",
       "20          8979          1822         0  \n",
       "21          4634          7048    265672  \n",
       "22           345           724         0  "
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Summary of metaedges and cooresponding edges\n",
    "metaedge_df = hetio.stats.get_metaedge_df(graph)\n",
    "\n",
    "# Calculate number of unbiased edges\n",
    "rows = list()\n",
    "for metaedge, edges in graph.get_metaedge_to_edges(exclude_inverts=True).items():\n",
    "    unbiased = sum(edge.data['unbiased'] for edge in edges)\n",
    "    rows.append({'metaedge': str(metaedge), 'unbiased': unbiased})\n",
    "\n",
    "metaedge_df = metaedge_df.merge(pandas.DataFrame(rows))\n",
    "metaedge_df.to_csv('data/summary/metaedges.tsv', sep='\\t', index=False)\n",
    "metaedge_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Summary of different styles for representing each metaedge\n",
    "metaedge_style_df = hetio.stats.get_metaedge_style_df(metagraph)\n",
    "metaedge_style_df.to_csv('data/summary/metaedge-styles.tsv', sep='\\t', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2291801"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Number of edges in the network\n",
    "metaedge_df.edges.sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Save graph"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Write nodes to a table\n",
    "path = 'data/nodes.tsv'\n",
    "hetio.readwrite.write_nodetable(graph, path)\n",
    "\n",
    "# Write edges to a table\n",
    "path = 'data/edges.sif.gz'\n",
    "hetio.readwrite.write_sif(graph, path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Write a subset of edges to a table\n",
    "path = 'data/edges-10.sif'\n",
    "hetio.readwrite.write_sif(graph, path, max_edges=10)\n",
    "\n",
    "path = 'data/edges-5k.sif.gz'\n",
    "hetio.readwrite.write_sif(graph, path, max_edges=5000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Write metagraph as json\n",
    "path = 'data/metagraph.json'\n",
    "hetio.readwrite.write_metagraph(metagraph, path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Write graph as json\n",
    "path = 'data/hetnet.json.bz2'\n",
    "hetio.readwrite.write_graph(graph, path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3956af526e68ef09b9fd1feb273a8151bf7ebd726a77ad7ad45d5257920dc2a7  data/hetnet.json.bz2\r\n"
     ]
    }
   ],
   "source": [
    "! sha256sum data/hetnet.json.bz2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Barplots of metaedge and metanode counts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/tongli/anaconda3/envs/integrate/lib/python3.5/site-packages/matplotlib/font_manager.py:1297: UserWarning: findfont: Font family ['sans-serif'] not found. Falling back to DejaVu Sans\n",
      "  (prop.get_family(), self.defaultFamily[fontext]))\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZIAAAFVCAYAAADFb2n2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n4kBiZmZtcSAxM7O2OJCYmVlb/j9H+/AZ88lxsAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f256c15cdd8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = seaborn.barplot(x='metanode', y='nodes', data=metanode_df.sort_values('nodes'))\n",
    "for tick in ax.get_xticklabels():\n",
    "    tick.set_rotation(90)\n",
    "ax.set_xlabel(''); ax.set_ylabel('nodes');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/tongli/anaconda3/envs/integrate/lib/python3.5/site-packages/matplotlib/font_manager.py:1297: UserWarning: findfont: Font family ['sans-serif'] not found. Falling back to DejaVu Sans\n",
      "  (prop.get_family(), self.defaultFamily[fontext]))\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZgAAAHJCAYAAAC46S8HAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nqVQqlVnh/wPZztVIXMwLCgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f256c0be898>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = seaborn.barplot(x='metaedge', y='edges', data=metaedge_df.sort_values('edges'))\n",
    "for tick in ax.get_xticklabels():\n",
    "    tick.set_rotation(90)\n",
    "ax.set_xlabel(''); ax.set_ylabel('edges');"
   ]
  }
 ],
 "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}