interaction.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Integrating PPI data to identify OPC-endothelial cell interactors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import urllib\n",
    "import os\n",
    "import json\n",
    "\n",
    "import pandas\n",
    "import numpy\n",
    "import networkx\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Download STRING if missing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "url_to_path = {\n",
    "    # STRING PPI\n",
    "    'http://string.embl.de/newstring_download/protein.aliases.v10/10090.protein.aliases.v10.txt.gz':\n",
    "        os.path.join('download', 'string', '10090.protein.aliases.v10.txt.gz'),\n",
    "    'http://string.embl.de/newstring_download/protein.links.v10/10090.protein.links.v10.txt.gz':\n",
    "        os.path.join('download', 'string', '10090.protein.links.v10.txt.gz'),\n",
    "}\n",
    "\n",
    "for url, path in url_to_path.items():\n",
    "    if not os.path.exists(path):\n",
    "        urllib.request.urlretrieve(url, path)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Read and prepare STRING"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "path = os.path.join('download', 'string', '10090.protein.aliases.v10.txt.gz')\n",
    "alias_df = pandas.read_table(path, skiprows=1, names = ['string_protein_id', 'alias', 'source'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "alias_df = alias_df.query(\"source == 'BLAST_KEGG_NAME BLAST_UniProt_GN Ensembl_EntrezGene Ensembl_IKMCs_ES_cells_available Ensembl_MGI Ensembl_UniProt_GN Ensembl_WikiGene'\")\n",
    "alias_df = alias_df[['string_protein_id', 'alias']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>opc_candidate</th>\n",
       "      <th>endo_candidate</th>\n",
       "      <th>combined_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3478100</th>\n",
       "      <td>0610007P14Rik</td>\n",
       "      <td>2610028H24Rik</td>\n",
       "      <td>648</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3849791</th>\n",
       "      <td>0610007P14Rik</td>\n",
       "      <td>Aacs</td>\n",
       "      <td>153</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         opc_candidate endo_candidate  combined_score\n",
       "3478100  0610007P14Rik  2610028H24Rik             648\n",
       "3849791  0610007P14Rik           Aacs             153"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "path = os.path.join('download', 'string', '10090.protein.links.v10.txt.gz')\n",
    "string_df = pandas.read_table(path, sep=' ')\n",
    "\n",
    "string_df = string_df.merge(\n",
    "    alias_df.rename(columns={'string_protein_id': 'protein1', 'alias': 'opc_candidate'})).merge(\n",
    "    alias_df.rename(columns={'string_protein_id': 'protein2', 'alias': 'endo_candidate'}))\n",
    "\n",
    "string_df = string_df[['opc_candidate', 'endo_candidate', 'combined_score']].drop_duplicates()\n",
    "string_df = string_df.sort_values(['opc_candidate', 'endo_candidate'])\n",
    "string_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4556514"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(string_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZoAAAEACAYAAACK+7BGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ng/EAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fc21e5dd860>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "string_df.combined_score.hist(bins=30);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Integrate expression data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Gene symbol</th>\n",
       "      <th>opc_L2FC</th>\n",
       "      <th>opc_microglia_ratio</th>\n",
       "      <th>endothelial_L2FC</th>\n",
       "      <th>DA_Cat_L2FC</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0610005C13Rik</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.11760</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0610007C21Rik</td>\n",
       "      <td>0.39919</td>\n",
       "      <td>0.798934</td>\n",
       "      <td>0.11792</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Gene symbol  opc_L2FC  opc_microglia_ratio  endothelial_L2FC  DA_Cat_L2FC\n",
       "0  0610005C13Rik   0.00000             1.000000          -0.11760          NaN\n",
       "1  0610007C21Rik   0.39919             0.798934           0.11792          NaN"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fc_df = pandas.read_table('data/enrichment.tsv')\n",
    "fc_df['opc_microglia_ratio'] = fc_df['Oligodendrocyte Precursor Cell'] / fc_df['Microglia']\n",
    "fc_df = fc_df[['Gene symbol', 'opc_L2FC', 'opc_microglia_ratio', 'endothelial_L2FC', 'DA_Cat_L2FC']]\n",
    "fc_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4430882"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_renamer = lambda x: {\n",
    "    'Gene symbol': x,\n",
    "    'opc_L2FC': '{}_opc_l2fc'.format(x),\n",
    "    'opc_microglia_ratio': '{}_opc_microglia_ratio'.format(x),\n",
    "    'endothelial_L2FC': '{}_endo_l2fc'.format(x),\n",
    "    'DA_Cat_L2FC': '{}_dac_l2fc'.format(x),\n",
    "}\n",
    "\n",
    "ixn_df = string_df.merge(\n",
    "    fc_df.rename(columns=get_renamer('opc_candidate'))).merge(\n",
    "    fc_df.rename(columns=get_renamer('endo_candidate')))\n",
    "\n",
    "ixn_df = ixn_df.sort_values(['opc_candidate', 'endo_candidate'])\n",
    "\n",
    "len(ixn_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>opc_candidate</th>\n",
       "      <th>endo_candidate</th>\n",
       "      <th>combined_score</th>\n",
       "      <th>opc_candidate_opc_l2fc</th>\n",
       "      <th>opc_candidate_opc_microglia_ratio</th>\n",
       "      <th>opc_candidate_endo_l2fc</th>\n",
       "      <th>opc_candidate_dac_l2fc</th>\n",
       "      <th>endo_candidate_opc_l2fc</th>\n",
       "      <th>endo_candidate_opc_microglia_ratio</th>\n",
       "      <th>endo_candidate_endo_l2fc</th>\n",
       "      <th>endo_candidate_dac_l2fc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0610007P14Rik</td>\n",
       "      <td>2610028H24Rik</td>\n",
       "      <td>648</td>\n",
       "      <td>-0.23887</td>\n",
       "      <td>5.232446</td>\n",
       "      <td>-2.0939</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>3.203400e-16</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>0610007P14Rik</td>\n",
       "      <td>Aacs</td>\n",
       "      <td>153</td>\n",
       "      <td>-0.23887</td>\n",
       "      <td>5.232446</td>\n",
       "      <td>-2.0939</td>\n",
       "      <td>NaN</td>\n",
       "      <td>-2.4177</td>\n",
       "      <td>1.798442</td>\n",
       "      <td>-3.773100e+00</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    opc_candidate endo_candidate  combined_score  opc_candidate_opc_l2fc  \\\n",
       "0   0610007P14Rik  2610028H24Rik             648                -0.23887   \n",
       "47  0610007P14Rik           Aacs             153                -0.23887   \n",
       "\n",
       "    opc_candidate_opc_microglia_ratio  opc_candidate_endo_l2fc  \\\n",
       "0                            5.232446                  -2.0939   \n",
       "47                           5.232446                  -2.0939   \n",
       "\n",
       "    opc_candidate_dac_l2fc  endo_candidate_opc_l2fc  \\\n",
       "0                      NaN                   0.0000   \n",
       "47                     NaN                  -2.4177   \n",
       "\n",
       "    endo_candidate_opc_microglia_ratio  endo_candidate_endo_l2fc  \\\n",
       "0                             1.000000              3.203400e-16   \n",
       "47                            1.798442             -3.773100e+00   \n",
       "\n",
       "    endo_candidate_dac_l2fc  \n",
       "0                       NaN  \n",
       "47                      NaN  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ixn_df.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Read GO Annotations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['cell-cell adhesion', 'cell surface']"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "with open('data/go-to-genes.json') as fp:\n",
    "    go_to_genes = json.load(fp)\n",
    "\n",
    "# GO Terms to choose from\n",
    "list(go_to_genes.keys())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Filter for candidate interactions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Parameters\n",
    "l2fc_cutoff = 1.5\n",
    "string_score = 700\n",
    "opc_microglia_min_ratio = 0.05"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(140, 140)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "opc_go_terms = [\n",
    "    'cell surface',\n",
    "    'cell-cell adhesion'\n",
    "]\n",
    "\n",
    "endo_go_terms = [\n",
    "    'cell surface',\n",
    "    'cell-cell adhesion'\n",
    "]\n",
    "\n",
    "opc_go_subset = set.intersection(*[set(go_to_genes[term]) for term in opc_go_terms])\n",
    "endo_go_subset = set.intersection(*[set(go_to_genes[term]) for term in endo_go_terms])\n",
    "\n",
    "len(opc_go_subset), len(endo_go_subset)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Save parameters\n",
    "param_df = pandas.DataFrame({\n",
    "    'parameter': ['log2_fold_change_cutoff', 'opc_microglia_min_ratio', 'string_score_cutoff', 'opc_GO_terms', 'endo_GO_terms'],\n",
    "    'value': [l2fc_cutoff, opc_microglia_min_ratio, string_score, opc_go_terms, endo_go_terms]})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "29"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filtered_df = ixn_df.copy()\n",
    "\n",
    "# Filter interactions based on STRING score\n",
    "filtered_df = filtered_df.query(\"combined_score >= @string_score\")\n",
    "\n",
    "# Filter for OPC candidates that are enriched in expression in OPCs\n",
    "filtered_df = filtered_df.query(\"opc_candidate_opc_l2fc >= @l2fc_cutoff and opc_candidate_opc_microglia_ratio > @opc_microglia_min_ratio\")\n",
    "\n",
    "# Filter for endothelial candidates that are enriched in expression in endothelial cells\n",
    "filtered_df = filtered_df.query(\"endo_candidate_endo_l2fc >= @l2fc_cutoff\")\n",
    "\n",
    "# Filter for OPC candidates by GO term membership\n",
    "filtered_df = filtered_df.query(\"opc_candidate in @opc_go_subset\")\n",
    "\n",
    "# Filter for endothelial candidates by GO term membership\n",
    "filtered_df = filtered_df.query(\"endo_candidate in @endo_go_subset\")\n",
    "\n",
    "len(filtered_df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>opc_candidate</th>\n",
       "      <th>endo_candidate</th>\n",
       "      <th>combined_score</th>\n",
       "      <th>opc_candidate_opc_l2fc</th>\n",
       "      <th>opc_candidate_opc_microglia_ratio</th>\n",
       "      <th>opc_candidate_endo_l2fc</th>\n",
       "      <th>opc_candidate_dac_l2fc</th>\n",
       "      <th>endo_candidate_opc_l2fc</th>\n",
       "      <th>endo_candidate_opc_microglia_ratio</th>\n",
       "      <th>endo_candidate_endo_l2fc</th>\n",
       "      <th>endo_candidate_dac_l2fc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2812820</th>\n",
       "      <td>Alcam</td>\n",
       "      <td>Itgb1</td>\n",
       "      <td>858</td>\n",
       "      <td>5.5683</td>\n",
       "      <td>174.060237</td>\n",
       "      <td>-3.5519</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0750</td>\n",
       "      <td>1.570693</td>\n",
       "      <td>2.3607</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4342010</th>\n",
       "      <td>Anxa1</td>\n",
       "      <td>Cxcr4</td>\n",
       "      <td>903</td>\n",
       "      <td>2.6875</td>\n",
       "      <td>0.617111</td>\n",
       "      <td>5.4665</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.1729</td>\n",
       "      <td>0.468453</td>\n",
       "      <td>2.7818</td>\n",
       "      <td>1.1388</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1180048</th>\n",
       "      <td>Anxa1</td>\n",
       "      <td>Il6</td>\n",
       "      <td>705</td>\n",
       "      <td>2.6875</td>\n",
       "      <td>0.617111</td>\n",
       "      <td>5.4665</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.879275</td>\n",
       "      <td>2.4959</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1102123</th>\n",
       "      <td>Cd1d1</td>\n",
       "      <td>B2m</td>\n",
       "      <td>993</td>\n",
       "      <td>2.0542</td>\n",
       "      <td>0.385555</td>\n",
       "      <td>1.4210</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.7299</td>\n",
       "      <td>0.132059</td>\n",
       "      <td>1.6164</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3443688</th>\n",
       "      <td>Cxcl12</td>\n",
       "      <td>Anxa1</td>\n",
       "      <td>901</td>\n",
       "      <td>1.5265</td>\n",
       "      <td>7.270368</td>\n",
       "      <td>1.4406</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.6875</td>\n",
       "      <td>0.617111</td>\n",
       "      <td>5.4665</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        opc_candidate endo_candidate  combined_score  opc_candidate_opc_l2fc  \\\n",
       "2812820         Alcam          Itgb1             858                  5.5683   \n",
       "4342010         Anxa1          Cxcr4             903                  2.6875   \n",
       "1180048         Anxa1            Il6             705                  2.6875   \n",
       "1102123         Cd1d1            B2m             993                  2.0542   \n",
       "3443688        Cxcl12          Anxa1             901                  1.5265   \n",
       "\n",
       "         opc_candidate_opc_microglia_ratio  opc_candidate_endo_l2fc  \\\n",
       "2812820                         174.060237                  -3.5519   \n",
       "4342010                           0.617111                   5.4665   \n",
       "1180048                           0.617111                   5.4665   \n",
       "1102123                           0.385555                   1.4210   \n",
       "3443688                           7.270368                   1.4406   \n",
       "\n",
       "         opc_candidate_dac_l2fc  endo_candidate_opc_l2fc  \\\n",
       "2812820                     NaN                   1.0750   \n",
       "4342010                     NaN                   1.1729   \n",
       "1180048                     NaN                   0.0000   \n",
       "1102123                     NaN                   2.7299   \n",
       "3443688                     NaN                   2.6875   \n",
       "\n",
       "         endo_candidate_opc_microglia_ratio  endo_candidate_endo_l2fc  \\\n",
       "2812820                            1.570693                    2.3607   \n",
       "4342010                            0.468453                    2.7818   \n",
       "1180048                            0.879275                    2.4959   \n",
       "1102123                            0.132059                    1.6164   \n",
       "3443688                            0.617111                    5.4665   \n",
       "\n",
       "         endo_candidate_dac_l2fc  \n",
       "2812820                      NaN  \n",
       "4342010                   1.1388  \n",
       "1180048                      NaN  \n",
       "1102123                      NaN  \n",
       "3443688                      NaN  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "filtered_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>gene</th>\n",
       "      <th>interactions</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Ptprc</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Cxcl12</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Vcam1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Anxa1</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     gene  interactions\n",
       "0   Ptprc             6\n",
       "1  Cxcl12             4\n",
       "2   Vcam1             3\n",
       "3   Anxa1             2"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Genes with the most OPC candidate interactions\n",
    "opc_count_df = filtered_df.opc_candidate.value_counts().reset_index()\n",
    "opc_count_df.columns = ['gene', 'interactions']\n",
    "opc_count_df.head(4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>gene</th>\n",
       "      <th>interactions</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Itgb1</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Il6</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Itga4</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>B2m</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    gene  interactions\n",
       "0  Itgb1             9\n",
       "1    Il6             5\n",
       "2  Itga4             4\n",
       "3    B2m             3"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Genes with the most endothelial candidate interactions\n",
    "endo_count_df = filtered_df.endo_candidate.value_counts().reset_index()\n",
    "endo_count_df.columns = ['gene', 'interactions']\n",
    "endo_count_df.head(4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "23"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Genes with interactions\n",
    "genes = set(filtered_df.opc_candidate) | set(filtered_df.endo_candidate)\n",
    "len(genes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Save to excel\n",
    "path = os.path.join('data', 'filtered.xlsx')\n",
    "writer = pandas.ExcelWriter(path, engine='xlsxwriter')\n",
    "filtered_df.to_excel(writer, index=False, sheet_name='interactions')\n",
    "opc_count_df.to_excel(writer, index=False, sheet_name='opc_candidate_gene_counts')\n",
    "endo_count_df.to_excel(writer, index=False, sheet_name='endo_candidate_gene_counts')\n",
    "param_df.to_excel(writer, index=False, sheet_name='parameters')\n",
    "for sheet in writer.sheets.values():\n",
    "    sheet.freeze_panes(1, 0)\n",
    "writer.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "graph = networkx.DiGraph()\n",
    "for i, row in fc_df.iterrows():\n",
    "    gene = row['Gene symbol']\n",
    "    if not gene in genes:\n",
    "        continue\n",
    "    graph.add_node(gene,\n",
    "                   opc_l2fc = row['opc_L2FC'],\n",
    "                   endo_l2fc = row['endothelial_L2FC'],\n",
    "                   dac_l2fc = row['DA_Cat_L2FC'])\n",
    "for i, row in filtered_df.iterrows():\n",
    "    graph.add_edge(row.opc_candidate, row.endo_candidate, weight=row.combined_score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAEACAYAAABVtcpZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n0ZWAjo6OTglGVwI6Ojo6JZj/A8iSZCNa3nfyAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fc218352da0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pos = networkx.spring_layout(graph, weight='weight', iterations=500, k=2)\n",
    "pc = networkx.draw_networkx_nodes(graph, pos)\n",
    "pc = networkx.draw_networkx_edges(graph, pos)\n",
    "labels = networkx.draw_networkx_labels(graph, pos)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Save network\n",
    "path = os.path.join('data', 'network.graphml')\n",
    "networkx.readwrite.write_graphml(graph, path)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}