paper-case-studies/case-1-pleiotropy.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "public-ethiopia",
   "metadata": {
    "papermill": {
     "duration": 0.024283,
     "end_time": "2021-03-22T10:46:22.261358",
     "exception": false,
     "start_time": "2021-03-22T10:46:22.237075",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Case study 1: Distinguishing vertical and horizontal pleiotropy for SNP-protein associations\n",
    "\n",
    "Mendelian randomization (MR), a technique to evaluate the causal role of modifiable exposures on a health outcome using a set of genetic instruments, tacitly assumes that a genetic variant (e.g. SNP) is only related to an outcome of interest through an exposure (i.e. the \"exclusion restriction criterion\"). \n",
    "Horizontal pleiotropy, where a SNP is associated with multiple phenotypes independently of the exposure of interest, potentially violates this assumption delivering misleading conclusions. \n",
    "In contrast, vertical pleiotropy, where a SNP is associated with multiple phenotypes on the same biological pathway, does not violate this assumption. \n",
    "\n",
    "Here, we use external evidence on biological pathways and protein-protein interactions to assess the pleiotropic profile for a genetic variant based on the genes (and proteins) with which that variant is associated. \n",
    "In a graph representation, we will show that:\n",
    "- We have evidence of potential **vertical** pleiotropy when the proteins associated with a variant belong to one connected group,\n",
    "- In contrast, if the proteins associated with a single variant are isolated from each other, this suggests evidence of **horizontal** pleiotropy.\n",
    "\n",
    "This case study goes as follows:\n",
    "\n",
    "- For a list of genes that we wish to assess their pleiotropic profiles (against a variant of interests), we retrieve from EpiGraphDB their mapped protein products.\n",
    "- Then respecively from the evidence on shared pathway network and protein-protein interaction network, we can extract the proteins groups formed in these networks as connected communities to assess the pleiotropic profiles of the genes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "smart-operations",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:22.300332Z",
     "iopub.status.busy": "2021-03-22T10:46:22.299885Z",
     "iopub.status.idle": "2021-03-22T10:46:22.847723Z",
     "shell.execute_reply": "2021-03-22T10:46:22.848245Z"
    },
    "papermill": {
     "duration": 0.57015,
     "end_time": "2021-03-22T10:46:22.848521",
     "exception": false,
     "start_time": "2021-03-22T10:46:22.278371",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "from itertools import permutations\n",
    "from typing import List\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import networkx as nx\n",
    "import requests\n",
    "import pandas as pd\n",
    "\n",
    "plt.rcParams[\"figure.figsize\"] = (12, 9)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "starting-springer",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:22.895992Z",
     "iopub.status.busy": "2021-03-22T10:46:22.895339Z",
     "iopub.status.idle": "2021-03-22T10:46:22.896744Z",
     "shell.execute_reply": "2021-03-22T10:46:22.897278Z"
    },
    "papermill": {
     "duration": 0.026683,
     "end_time": "2021-03-22T10:46:22.897433",
     "exception": false,
     "start_time": "2021-03-22T10:46:22.870750",
     "status": "completed"
    },
    "tags": [
     "parameters"
    ]
   },
   "outputs": [],
   "source": [
    "# Default parameters\n",
    "API_URL = \"https://api.epigraphdb.org\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "automated-archives",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:22.943286Z",
     "iopub.status.busy": "2021-03-22T10:46:22.942920Z",
     "iopub.status.idle": "2021-03-22T10:46:22.944602Z",
     "shell.execute_reply": "2021-03-22T10:46:22.944306Z"
    },
    "papermill": {
     "duration": 0.02295,
     "end_time": "2021-03-22T10:46:22.944685",
     "exception": false,
     "start_time": "2021-03-22T10:46:22.921735",
     "status": "completed"
    },
    "tags": [
     "injected-parameters"
    ]
   },
   "outputs": [],
   "source": [
    "# Parameters\n",
    "API_URL = \"https://api.epigraphdb.org\"\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "apparent-textbook",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:22.974935Z",
     "iopub.status.busy": "2021-03-22T10:46:22.974530Z",
     "iopub.status.idle": "2021-03-22T10:46:23.116053Z",
     "shell.execute_reply": "2021-03-22T10:46:23.116418Z"
    },
    "papermill": {
     "duration": 0.158829,
     "end_time": "2021-03-22T10:46:23.116541",
     "exception": false,
     "start_time": "2021-03-22T10:46:22.957712",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "https://api.epigraphdb.org\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(API_URL)\n",
    "requests.get(f\"{API_URL}/ping\").json()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "surprised-saskatchewan",
   "metadata": {
    "papermill": {
     "duration": 0.013267,
     "end_time": "2021-03-22T10:46:23.143607",
     "exception": false,
     "start_time": "2021-03-22T10:46:23.130340",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## Parameters\n",
    "\n",
    "In this example we will look at the variant **rs12720356**, a SNP located in chromosome 19 that has been associated with Crohn’s disease and psoriasis.\n",
    "\n",
    "We need two paramters for this case study:\n",
    "- First, the list of genes associated with our eQTL of interest (rs12720356) that can be retrieved from GTEx portal (https://www.gtexportal.org/home/snp/rs12720356), or using their API (see Appendix below).\n",
    "- Second, we need to specify the number of mediator proteins between a pair of proteins. Zero means a direct link, and higher values represent indirect links with N nodes in the middle.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "initial-weekly",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:23.173048Z",
     "iopub.status.busy": "2021-03-22T10:46:23.172704Z",
     "iopub.status.idle": "2021-03-22T10:46:23.174709Z",
     "shell.execute_reply": "2021-03-22T10:46:23.174400Z"
    },
    "papermill": {
     "duration": 0.017802,
     "end_time": "2021-03-22T10:46:23.174784",
     "exception": false,
     "start_time": "2021-03-22T10:46:23.156982",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "SNP = \"rs12720356\"\n",
    "\n",
    "GENE_LIST = [\n",
    "    \"ZGLP1\",\n",
    "    \"FDX1L\",\n",
    "    \"MRPL4\",\n",
    "    \"ICAM5\",\n",
    "    \"TYK2\",\n",
    "    \"GRAP2\",\n",
    "    \"KRI1\",\n",
    "    \"TMED1\",\n",
    "    \"ICAM1\",\n",
    "]\n",
    "\n",
    "PPI_N_INTERMEDIATE_PROTEINS = 1"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "beneficial-trunk",
   "metadata": {
    "papermill": {
     "duration": 0.013415,
     "end_time": "2021-03-22T10:46:23.201493",
     "exception": false,
     "start_time": "2021-03-22T10:46:23.188078",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## Mapping genes to proteins\n",
    "\n",
    "The first step of the analysis is to map each of these genes to their protein product, recovering one protein for each gene in our database.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "regional-bleeding",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:23.237241Z",
     "iopub.status.busy": "2021-03-22T10:46:23.236881Z",
     "iopub.status.idle": "2021-03-22T10:46:23.407161Z",
     "shell.execute_reply": "2021-03-22T10:46:23.407586Z"
    },
    "papermill": {
     "duration": 0.192948,
     "end_time": "2021-03-22T10:46:23.407735",
     "exception": false,
     "start_time": "2021-03-22T10:46:23.214787",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>protein_name</th>\n",
       "      <th>uniprot_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ZGLP1</td>\n",
       "      <td>P0C6A0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>FDX1L</td>\n",
       "      <td>Q6P4F2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>MRPL4</td>\n",
       "      <td>Q9BYD3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ICAM5</td>\n",
       "      <td>Q9UMF0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>TYK2</td>\n",
       "      <td>P29597</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>GRAP2</td>\n",
       "      <td>O75791</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>KRI1</td>\n",
       "      <td>Q8N9T8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>TMED1</td>\n",
       "      <td>Q13445</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>ICAM1</td>\n",
       "      <td>P05362</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  protein_name uniprot_id\n",
       "0        ZGLP1     P0C6A0\n",
       "1        FDX1L     Q6P4F2\n",
       "2        MRPL4     Q9BYD3\n",
       "3        ICAM5     Q9UMF0\n",
       "4         TYK2     P29597\n",
       "5        GRAP2     O75791\n",
       "6         KRI1     Q8N9T8\n",
       "7        TMED1     Q13445\n",
       "8        ICAM1     P05362"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def get_gene_protein(gene_list):\n",
    "\n",
    "    # get gene-protein\n",
    "    endpoint = \"/mappings/gene-to-protein\"\n",
    "    url = f\"{API_URL}{endpoint}\"\n",
    "    payload = {\"gene_name_list\": gene_list}\n",
    "    r = requests.post(url, json=payload)\n",
    "    r.raise_for_status()\n",
    "    protein_df = pd.json_normalize(r.json()[\"results\"])\n",
    "    if len(protein_df) > 0:\n",
    "        res_df = protein_df[[\"gene.name\", \"protein.uniprot_id\"]].rename(\n",
    "            columns={\n",
    "                \"gene.name\": \"protein_name\",\n",
    "                \"protein.uniprot_id\": \"uniprot_id\",\n",
    "            }\n",
    "        )\n",
    "    else:\n",
    "        return pd.DataFrame(columns=[\"protein_name\", \"uniprot_id\"])\n",
    "\n",
    "    return res_df\n",
    "\n",
    "\n",
    "gene_protein_df = get_gene_protein(gene_list=GENE_LIST)\n",
    "gene_protein_df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "czech-communications",
   "metadata": {
    "papermill": {
     "duration": 0.017854,
     "end_time": "2021-03-22T10:46:23.444037",
     "exception": false,
     "start_time": "2021-03-22T10:46:23.426183",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## Identifying the involvement in the same biological pathways\n",
    "\n",
    "Once we have the list of protein identifiers, for each protein we retrieve the pathways they are involved in, either as a single entity or as part of a complex.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "modified-symphony",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:23.494415Z",
     "iopub.status.busy": "2021-03-22T10:46:23.493973Z",
     "iopub.status.idle": "2021-03-22T10:46:23.654619Z",
     "shell.execute_reply": "2021-03-22T10:46:23.653755Z"
    },
    "papermill": {
     "duration": 0.188874,
     "end_time": "2021-03-22T10:46:23.654836",
     "exception": false,
     "start_time": "2021-03-22T10:46:23.465962",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>uniprot_id</th>\n",
       "      <th>pathway_count</th>\n",
       "      <th>pathway_reactome_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>P0C6A0</td>\n",
       "      <td>0</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Q6P4F2</td>\n",
       "      <td>15</td>\n",
       "      <td>[R-HSA-8957322, R-HSA-5668914, R-HSA-5579029, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Q9BYD3</td>\n",
       "      <td>6</td>\n",
       "      <td>[R-HSA-72766, R-HSA-5419276, R-HSA-5389840, R-...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Q9UMF0</td>\n",
       "      <td>5</td>\n",
       "      <td>[R-HSA-216083, R-HSA-198933, R-HSA-168256, R-H...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>P29597</td>\n",
       "      <td>28</td>\n",
       "      <td>[R-HSA-9679506, R-HSA-9679191, R-HSA-913531, R...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>O75791</td>\n",
       "      <td>15</td>\n",
       "      <td>[R-HSA-9006934, R-HSA-389356, R-HSA-388841, R-...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Q8N9T8</td>\n",
       "      <td>0</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Q13445</td>\n",
       "      <td>0</td>\n",
       "      <td>[]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>P05362</td>\n",
       "      <td>11</td>\n",
       "      <td>[R-HSA-913531, R-HSA-877300, R-HSA-6785807, R-...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  uniprot_id  pathway_count                                pathway_reactome_id\n",
       "0     P0C6A0              0                                                 []\n",
       "1     Q6P4F2             15  [R-HSA-8957322, R-HSA-5668914, R-HSA-5579029, ...\n",
       "2     Q9BYD3              6  [R-HSA-72766, R-HSA-5419276, R-HSA-5389840, R-...\n",
       "3     Q9UMF0              5  [R-HSA-216083, R-HSA-198933, R-HSA-168256, R-H...\n",
       "4     P29597             28  [R-HSA-9679506, R-HSA-9679191, R-HSA-913531, R...\n",
       "5     O75791             15  [R-HSA-9006934, R-HSA-389356, R-HSA-388841, R-...\n",
       "6     Q8N9T8              0                                                 []\n",
       "7     Q13445              0                                                 []\n",
       "8     P05362             11  [R-HSA-913531, R-HSA-877300, R-HSA-6785807, R-..."
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def get_protein_pathway(gene_protein_df):\n",
    "    endpoint = \"/protein/in-pathway\"\n",
    "    url = f\"{API_URL}{endpoint}\"\n",
    "    payload = {\n",
    "        \"uniprot_id_list\": gene_protein_df[\"uniprot_id\"].to_list(),\n",
    "    }\n",
    "    r = requests.post(url, json=payload)\n",
    "    r.raise_for_status()\n",
    "    df = pd.json_normalize(r.json()[\"results\"])\n",
    "\n",
    "    if len(df) > 0:\n",
    "        res_df = gene_protein_df[[\"uniprot_id\"]].merge(\n",
    "            df, left_on=\"uniprot_id\", right_on=\"uniprot_id\", how=\"left\"\n",
    "        )\n",
    "    else:\n",
    "        res_df = (\n",
    "            gene_protein_df[[\"uniprot_id\"]]\n",
    "            .assign(pathway_count=None)\n",
    "            .assign(pathway_reactome_id=None)\n",
    "        )\n",
    "    res_df = res_df.assign(\n",
    "        pathway_count=lambda df: df[\"pathway_count\"]\n",
    "        .apply(lambda x: 0 if pd.isna(x) else x)\n",
    "        .astype(int)\n",
    "    ).assign(\n",
    "        pathway_reactome_id=lambda df: df[\"pathway_reactome_id\"].apply(\n",
    "            lambda x: [] if not isinstance(x, list) else x\n",
    "        )\n",
    "    )\n",
    "\n",
    "    return res_df\n",
    "\n",
    "\n",
    "pathway_df = get_protein_pathway(gene_protein_df=gene_protein_df)\n",
    "pathway_df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "correct-dutch",
   "metadata": {
    "papermill": {
     "duration": 0.031709,
     "end_time": "2021-03-22T10:46:23.716134",
     "exception": false,
     "start_time": "2021-03-22T10:46:23.684425",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "Now, for each pair of proteins we match the pathways they have in common."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "minute-phase",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:23.828424Z",
     "iopub.status.busy": "2021-03-22T10:46:23.774063Z",
     "iopub.status.idle": "2021-03-22T10:46:23.845091Z",
     "shell.execute_reply": "2021-03-22T10:46:23.845677Z"
    },
    "papermill": {
     "duration": 0.108519,
     "end_time": "2021-03-22T10:46:23.845850",
     "exception": false,
     "start_time": "2021-03-22T10:46:23.737331",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Num. shared_pathway pairs:  14\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>protein</th>\n",
       "      <th>assoc_protein</th>\n",
       "      <th>combination</th>\n",
       "      <th>shared_pathway</th>\n",
       "      <th>count</th>\n",
       "      <th>connected</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>P29597</td>\n",
       "      <td>P05362</td>\n",
       "      <td>[P05362, P29597]</td>\n",
       "      <td>{R-HSA-6783783, R-HSA-168256, R-HSA-913531, R-...</td>\n",
       "      <td>6</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68</th>\n",
       "      <td>P05362</td>\n",
       "      <td>P29597</td>\n",
       "      <td>[P05362, P29597]</td>\n",
       "      <td>{R-HSA-6783783, R-HSA-168256, R-HSA-913531, R-...</td>\n",
       "      <td>6</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>Q9UMF0</td>\n",
       "      <td>P05362</td>\n",
       "      <td>[P05362, Q9UMF0]</td>\n",
       "      <td>{R-HSA-168256, R-HSA-198933, R-HSA-216083, R-H...</td>\n",
       "      <td>5</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67</th>\n",
       "      <td>P05362</td>\n",
       "      <td>Q9UMF0</td>\n",
       "      <td>[P05362, Q9UMF0]</td>\n",
       "      <td>{R-HSA-168256, R-HSA-198933, R-HSA-216083, R-H...</td>\n",
       "      <td>5</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>P29597</td>\n",
       "      <td>O75791</td>\n",
       "      <td>[O75791, P29597]</td>\n",
       "      <td>{R-HSA-168256, R-HSA-162582}</td>\n",
       "      <td>2</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>Q9UMF0</td>\n",
       "      <td>Q13445</td>\n",
       "      <td>[Q13445, Q9UMF0]</td>\n",
       "      <td>{}</td>\n",
       "      <td>0</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>P29597</td>\n",
       "      <td>P0C6A0</td>\n",
       "      <td>[P0C6A0, P29597]</td>\n",
       "      <td>{}</td>\n",
       "      <td>0</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>P29597</td>\n",
       "      <td>Q9BYD3</td>\n",
       "      <td>[P29597, Q9BYD3]</td>\n",
       "      <td>{}</td>\n",
       "      <td>0</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>P0C6A0</td>\n",
       "      <td>Q9BYD3</td>\n",
       "      <td>[P0C6A0, Q9BYD3]</td>\n",
       "      <td>{}</td>\n",
       "      <td>0</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>71</th>\n",
       "      <td>P05362</td>\n",
       "      <td>Q13445</td>\n",
       "      <td>[P05362, Q13445]</td>\n",
       "      <td>{}</td>\n",
       "      <td>0</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>72 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   protein assoc_protein       combination  \\\n",
       "39  P29597        P05362  [P05362, P29597]   \n",
       "68  P05362        P29597  [P05362, P29597]   \n",
       "31  Q9UMF0        P05362  [P05362, Q9UMF0]   \n",
       "67  P05362        Q9UMF0  [P05362, Q9UMF0]   \n",
       "36  P29597        O75791  [O75791, P29597]   \n",
       "..     ...           ...               ...   \n",
       "30  Q9UMF0        Q13445  [Q13445, Q9UMF0]   \n",
       "32  P29597        P0C6A0  [P0C6A0, P29597]   \n",
       "34  P29597        Q9BYD3  [P29597, Q9BYD3]   \n",
       "1   P0C6A0        Q9BYD3  [P0C6A0, Q9BYD3]   \n",
       "71  P05362        Q13445  [P05362, Q13445]   \n",
       "\n",
       "                                       shared_pathway  count  connected  \n",
       "39  {R-HSA-6783783, R-HSA-168256, R-HSA-913531, R-...      6       True  \n",
       "68  {R-HSA-6783783, R-HSA-168256, R-HSA-913531, R-...      6       True  \n",
       "31  {R-HSA-168256, R-HSA-198933, R-HSA-216083, R-H...      5       True  \n",
       "67  {R-HSA-168256, R-HSA-198933, R-HSA-216083, R-H...      5       True  \n",
       "36                       {R-HSA-168256, R-HSA-162582}      2       True  \n",
       "..                                                ...    ...        ...  \n",
       "30                                                 {}      0      False  \n",
       "32                                                 {}      0      False  \n",
       "34                                                 {}      0      False  \n",
       "1                                                  {}      0      False  \n",
       "71                                                 {}      0      False  \n",
       "\n",
       "[72 rows x 6 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def get_shared_pathway(pathway_df):\n",
    "\n",
    "    # For the protein-pathway data\n",
    "    # Get protein-protein permutations where they share pathways\n",
    "    def per_permutation(pathway_df, permutation):\n",
    "        df = pathway_df[pathway_df[\"uniprot_id\"].isin(permutation)]\n",
    "        shared_pathway = set(df[\"pathway_reactome_id\"].iloc[0]).intersection(\n",
    "            set(df[\"pathway_reactome_id\"].iloc[1])\n",
    "        )\n",
    "        return shared_pathway\n",
    "\n",
    "    pairwise_permutations = permutations(pathway_df[\"uniprot_id\"], 2)\n",
    "\n",
    "    shared_pathway_df = (\n",
    "        pd.DataFrame(\n",
    "            [\n",
    "                {\n",
    "                    \"protein\": permutation[0],\n",
    "                    \"assoc_protein\": permutation[1],\n",
    "                    \"combination\": sorted(permutation),\n",
    "                    \"shared_pathway\": per_permutation(pathway_df, permutation),\n",
    "                }\n",
    "                for permutation in pairwise_permutations\n",
    "            ]\n",
    "        )\n",
    "        .assign(count=lambda df: df[\"shared_pathway\"].apply(lambda x: len(x)))\n",
    "        .assign(connected=lambda df: df[\"count\"].apply(lambda x: x > 0))\n",
    "    )\n",
    "    return shared_pathway_df\n",
    "\n",
    "\n",
    "shared_pathway_df = get_shared_pathway(pathway_df)\n",
    "print(\"Num. shared_pathway pairs: \", len(shared_pathway_df.query(\"count > 0\")))\n",
    "shared_pathway_df.sort_values(by=\"count\", ascending=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "primary-yahoo",
   "metadata": {
    "papermill": {
     "duration": 0.015045,
     "end_time": "2021-03-22T10:46:23.880621",
     "exception": false,
     "start_time": "2021-03-22T10:46:23.865576",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "We can further query EpiGraphDB regarding the pathway information using `GET /meta/nodes/Pathway/search`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "unknown-china",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:23.925109Z",
     "iopub.status.busy": "2021-03-22T10:46:23.924329Z",
     "iopub.status.idle": "2021-03-22T10:46:25.757055Z",
     "shell.execute_reply": "2021-03-22T10:46:25.756704Z"
    },
    "papermill": {
     "duration": 1.860924,
     "end_time": "2021-03-22T10:46:25.757155",
     "exception": false,
     "start_time": "2021-03-22T10:46:23.896231",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>node.name</th>\n",
       "      <th>node.id</th>\n",
       "      <th>node.url</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Interleukin-10 signaling</td>\n",
       "      <td>R-HSA-6783783</td>\n",
       "      <td>https://reactome.org/PathwayBrowser/#/R-HSA-67...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Immune System</td>\n",
       "      <td>R-HSA-168256</td>\n",
       "      <td>https://reactome.org/PathwayBrowser/#/R-HSA-16...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Signal Transduction</td>\n",
       "      <td>R-HSA-162582</td>\n",
       "      <td>https://reactome.org/PathwayBrowser/#/R-HSA-16...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Interferon Signaling</td>\n",
       "      <td>R-HSA-913531</td>\n",
       "      <td>https://reactome.org/PathwayBrowser/#/R-HSA-91...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Disease</td>\n",
       "      <td>R-HSA-1643685</td>\n",
       "      <td>https://reactome.org/PathwayBrowser/#/R-HSA-16...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>Interleukin-4 and Interleukin-13 signaling</td>\n",
       "      <td>R-HSA-6785807</td>\n",
       "      <td>https://reactome.org/PathwayBrowser/#/R-HSA-67...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Immunoregulatory interactions between a Lympho...</td>\n",
       "      <td>R-HSA-198933</td>\n",
       "      <td>https://reactome.org/PathwayBrowser/#/R-HSA-19...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Integrin cell surface interactions</td>\n",
       "      <td>R-HSA-216083</td>\n",
       "      <td>https://reactome.org/PathwayBrowser/#/R-HSA-21...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Extracellular matrix organization</td>\n",
       "      <td>R-HSA-1474244</td>\n",
       "      <td>https://reactome.org/PathwayBrowser/#/R-HSA-14...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Adaptive Immune System</td>\n",
       "      <td>R-HSA-1280218</td>\n",
       "      <td>https://reactome.org/PathwayBrowser/#/R-HSA-12...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>Signaling by Interleukins</td>\n",
       "      <td>R-HSA-449147</td>\n",
       "      <td>https://reactome.org/PathwayBrowser/#/R-HSA-44...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Cytokine Signaling in Immune system</td>\n",
       "      <td>R-HSA-1280215</td>\n",
       "      <td>https://reactome.org/PathwayBrowser/#/R-HSA-12...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                            node.name        node.id  \\\n",
       "0                            Interleukin-10 signaling  R-HSA-6783783   \n",
       "1                                       Immune System   R-HSA-168256   \n",
       "2                                 Signal Transduction   R-HSA-162582   \n",
       "3                                Interferon Signaling   R-HSA-913531   \n",
       "4                                             Disease  R-HSA-1643685   \n",
       "5          Interleukin-4 and Interleukin-13 signaling  R-HSA-6785807   \n",
       "6   Immunoregulatory interactions between a Lympho...   R-HSA-198933   \n",
       "7                  Integrin cell surface interactions   R-HSA-216083   \n",
       "8                   Extracellular matrix organization  R-HSA-1474244   \n",
       "9                              Adaptive Immune System  R-HSA-1280218   \n",
       "10                          Signaling by Interleukins   R-HSA-449147   \n",
       "11                Cytokine Signaling in Immune system  R-HSA-1280215   \n",
       "\n",
       "                                             node.url  \n",
       "0   https://reactome.org/PathwayBrowser/#/R-HSA-67...  \n",
       "1   https://reactome.org/PathwayBrowser/#/R-HSA-16...  \n",
       "2   https://reactome.org/PathwayBrowser/#/R-HSA-16...  \n",
       "3   https://reactome.org/PathwayBrowser/#/R-HSA-91...  \n",
       "4   https://reactome.org/PathwayBrowser/#/R-HSA-16...  \n",
       "5   https://reactome.org/PathwayBrowser/#/R-HSA-67...  \n",
       "6   https://reactome.org/PathwayBrowser/#/R-HSA-19...  \n",
       "7   https://reactome.org/PathwayBrowser/#/R-HSA-21...  \n",
       "8   https://reactome.org/PathwayBrowser/#/R-HSA-14...  \n",
       "9   https://reactome.org/PathwayBrowser/#/R-HSA-12...  \n",
       "10  https://reactome.org/PathwayBrowser/#/R-HSA-44...  \n",
       "11  https://reactome.org/PathwayBrowser/#/R-HSA-12...  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def get_pathway_info(reactome_id):\n",
    "    \"\"\"Get additional info from Pathway name\"\"\"\n",
    "    endpoint = \"/meta/nodes/Pathway/search\"\n",
    "    url = f\"{API_URL}{endpoint}\"\n",
    "    params = {\"id\": reactome_id}\n",
    "    r = requests.get(url, params=params)\n",
    "    r.raise_for_status()\n",
    "    keep_columns = [\"node.name\", \"node.id\", \"node.url\"]\n",
    "    df = pd.json_normalize(r.json()[\"results\"])[keep_columns]\n",
    "    return df\n",
    "\n",
    "\n",
    "pathway = set()\n",
    "\n",
    "for elem in shared_pathway_df[shared_pathway_df[\"connected\"]][\n",
    "    \"shared_pathway\"\n",
    "].tolist():\n",
    "    pathway = pathway.union(elem)\n",
    "\n",
    "pathway_info = pd.concat(\n",
    "    [get_pathway_info(reactome_id) for reactome_id in pathway]\n",
    ").reset_index(drop=True)\n",
    "pathway_info"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "accomplished-tunnel",
   "metadata": {
    "papermill": {
     "duration": 0.022053,
     "end_time": "2021-03-22T10:46:25.797574",
     "exception": false,
     "start_time": "2021-03-22T10:46:25.775521",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "In order to extract protein groups from the shared pathways,\n",
    "the last step for this query is to convert the shared pathway data into a graph where\n",
    "\n",
    "- Proteins are represented as nodes.\n",
    "- And shared pathways are represented as edges.\n",
    "- Then the protein groups (via shared pathways) are the connected communities in the graph.\n",
    "\n",
    "We then count the number of nodes in each connected community and plot the graph."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "continuous-ladder",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:25.846874Z",
     "iopub.status.busy": "2021-03-22T10:46:25.846341Z",
     "iopub.status.idle": "2021-03-22T10:46:25.849077Z",
     "shell.execute_reply": "2021-03-22T10:46:25.849749Z"
    },
    "papermill": {
     "duration": 0.033814,
     "end_time": "2021-03-22T10:46:25.849926",
     "exception": false,
     "start_time": "2021-03-22T10:46:25.816112",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Convert protein data to a graph\n",
    "# then extract protein groups from the graph\n",
    "\n",
    "\n",
    "def protein_df_to_graph(df):\n",
    "    df_connected = (\n",
    "        df.query(\"`connected`\")\n",
    "        # Convert combination to a string singleton\n",
    "        .assign(\n",
    "            combination=lambda df: df[\"combination\"].apply(lambda x: str(x))\n",
    "        ).drop_duplicates(subset=[\"combination\"])\n",
    "    )\n",
    "\n",
    "    df_isolated = df.query(\"not `connected`\")[[\"protein\"]].drop_duplicates()\n",
    "\n",
    "    # Create the graph of the connected nodes, then\n",
    "    # add isolated nodes\n",
    "    graph = nx.from_pandas_edgelist(df_connected, \"protein\", \"assoc_protein\")\n",
    "    for node in df_isolated[\"protein\"].to_list():\n",
    "        nx.add_path(graph, [node])\n",
    "\n",
    "    return graph\n",
    "\n",
    "\n",
    "def graph_to_protein_groups(graph):\n",
    "    components = nx.connected_components(graph)\n",
    "    res = [{\"group_members\": item, \"group_size\": len(item)} for item in components]\n",
    "    return res\n",
    "\n",
    "\n",
    "def draw_graph(graph):\n",
    "    pos = nx.kamada_kawai_layout(graph)\n",
    "    nx.draw(\n",
    "        G=graph,\n",
    "        pos=pos,\n",
    "        node_size=2800,\n",
    "        alpha=0.8,\n",
    "        with_labels=True,\n",
    "        node_color=\"#e53935\",\n",
    "        font_color=\"k\",\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "applicable-profit",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:25.898316Z",
     "iopub.status.busy": "2021-03-22T10:46:25.895319Z",
     "iopub.status.idle": "2021-03-22T10:46:25.901860Z",
     "shell.execute_reply": "2021-03-22T10:46:25.902168Z"
    },
    "papermill": {
     "duration": 0.030337,
     "end_time": "2021-03-22T10:46:25.902279",
     "exception": false,
     "start_time": "2021-03-22T10:46:25.871942",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{'group_members': {'O75791', 'P05362', 'P29597', 'Q6P4F2', 'Q9UMF0'},\n",
       "  'group_size': 5},\n",
       " {'group_members': {'P0C6A0'}, 'group_size': 1},\n",
       " {'group_members': {'Q9BYD3'}, 'group_size': 1},\n",
       " {'group_members': {'Q8N9T8'}, 'group_size': 1},\n",
       " {'group_members': {'Q13445'}, 'group_size': 1}]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pathway_protein_graph = protein_df_to_graph(shared_pathway_df)\n",
    "pathway_protein_groups = graph_to_protein_groups(pathway_protein_graph)\n",
    "pathway_protein_groups"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "generic-reduction",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:25.943290Z",
     "iopub.status.busy": "2021-03-22T10:46:25.942581Z",
     "iopub.status.idle": "2021-03-22T10:46:26.258700Z",
     "shell.execute_reply": "2021-03-22T10:46:26.259243Z"
    },
    "papermill": {
     "duration": 0.33973,
     "end_time": "2021-03-22T10:46:26.259464",
     "exception": false,
     "start_time": "2021-03-22T10:46:25.919734",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x648 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "draw_graph(pathway_protein_graph)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "knowing-referral",
   "metadata": {
    "papermill": {
     "duration": 0.016887,
     "end_time": "2021-03-22T10:46:26.293772",
     "exception": false,
     "start_time": "2021-03-22T10:46:26.276885",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## Retrieving shared protein-protein interactions\n",
    "\n",
    "We can complement the information given by pathway ontologies with data of interactions between proteins. For each pair of proteins, we retrieve shared protein-protein interactions, either a direct interaction or an interaction where there's one mediator protein in the middle (see paramaters section)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "increasing-principal",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:26.337602Z",
     "iopub.status.busy": "2021-03-22T10:46:26.337223Z",
     "iopub.status.idle": "2021-03-22T10:46:27.486527Z",
     "shell.execute_reply": "2021-03-22T10:46:27.487129Z"
    },
    "papermill": {
     "duration": 1.176418,
     "end_time": "2021-03-22T10:46:27.487331",
     "exception": false,
     "start_time": "2021-03-22T10:46:26.310913",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>path_size</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>protein</th>\n",
       "      <th>assoc_protein</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">O75791</th>\n",
       "      <th>O75791</th>\n",
       "      <td>152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>P05362</th>\n",
       "      <td>64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>P29597</th>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"4\" valign=\"top\">P05362</th>\n",
       "      <th>O75791</th>\n",
       "      <td>64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>P05362</th>\n",
       "      <td>314</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>P29597</th>\n",
       "      <td>48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Q9UMF0</th>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>P0C6A0</th>\n",
       "      <th>P0C6A0</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">P29597</th>\n",
       "      <th>O75791</th>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>P05362</th>\n",
       "      <td>48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>P29597</th>\n",
       "      <td>218</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Q13445</th>\n",
       "      <th>Q13445</th>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Q6P4F2</th>\n",
       "      <th>Q6P4F2</th>\n",
       "      <td>36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">Q8N9T8</th>\n",
       "      <th>Q8N9T8</th>\n",
       "      <td>268</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Q9BYD3</th>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">Q9BYD3</th>\n",
       "      <th>Q8N9T8</th>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Q9BYD3</th>\n",
       "      <td>370</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">Q9UMF0</th>\n",
       "      <th>P05362</th>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Q9UMF0</th>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       path_size\n",
       "protein assoc_protein           \n",
       "O75791  O75791               152\n",
       "        P05362                64\n",
       "        P29597                20\n",
       "P05362  O75791                64\n",
       "        P05362               314\n",
       "        P29597                48\n",
       "        Q9UMF0                 8\n",
       "P0C6A0  P0C6A0                 2\n",
       "P29597  O75791                20\n",
       "        P05362                48\n",
       "        P29597               218\n",
       "Q13445  Q13445                 8\n",
       "Q6P4F2  Q6P4F2                36\n",
       "Q8N9T8  Q8N9T8               268\n",
       "        Q9BYD3                16\n",
       "Q9BYD3  Q8N9T8                16\n",
       "        Q9BYD3               370\n",
       "Q9UMF0  P05362                 8\n",
       "        Q9UMF0                 6"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def get_ppi(gene_protein_df, n_intermediate_proteins: int = 0):\n",
    "    endpoint = \"/protein/ppi/pairwise\"\n",
    "    url = f\"{API_URL}{endpoint}\"\n",
    "    payload = {\n",
    "        \"uniprot_id_list\": gene_protein_df[\"uniprot_id\"].to_list(),\n",
    "        \"n_intermediate_proteins\": n_intermediate_proteins,\n",
    "    }\n",
    "    r = requests.post(url, json=payload)\n",
    "    r.raise_for_status()\n",
    "    df = pd.json_normalize(r.json()[\"results\"])\n",
    "\n",
    "    if len(df) > 0:\n",
    "        res_df = (\n",
    "            gene_protein_df[[\"uniprot_id\"]]\n",
    "            .rename(columns={\"uniprot_id\": \"protein\"})\n",
    "            .merge(df, left_on=\"protein\", right_on=\"protein\", how=\"left\")\n",
    "        )\n",
    "    else:\n",
    "        res_df = (\n",
    "            gene_protein_df[[\"uniprot_id\"]]\n",
    "            .rename(columns={\"uniprot_id\": \"protein\"})\n",
    "            .assign(assoc_protein=None, path_size=None)\n",
    "        )\n",
    "    return res_df\n",
    "\n",
    "\n",
    "ppi_df = get_ppi(gene_protein_df, n_intermediate_proteins=PPI_N_INTERMEDIATE_PROTEINS)\n",
    "ppi_count = ppi_df.groupby([\"protein\", \"assoc_protein\"]).count()\n",
    "ppi_count.rename(columns={\"path_size\": \"count\"})\n",
    "ppi_count"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "parallel-storage",
   "metadata": {
    "papermill": {
     "duration": 0.035584,
     "end_time": "2021-03-22T10:46:27.558808",
     "exception": false,
     "start_time": "2021-03-22T10:46:27.523224",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "Like in the query for pathway data, we convert and analyse these results as a graph."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "approximate-debut",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:27.610107Z",
     "iopub.status.busy": "2021-03-22T10:46:27.606669Z",
     "iopub.status.idle": "2021-03-22T10:46:27.802310Z",
     "shell.execute_reply": "2021-03-22T10:46:27.802624Z"
    },
    "papermill": {
     "duration": 0.222666,
     "end_time": "2021-03-22T10:46:27.802737",
     "exception": false,
     "start_time": "2021-03-22T10:46:27.580071",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Num. shared_ppi pairs:  10\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>protein</th>\n",
       "      <th>assoc_protein</th>\n",
       "      <th>combination</th>\n",
       "      <th>shared_ppi</th>\n",
       "      <th>connected</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>P0C6A0</td>\n",
       "      <td>Q6P4F2</td>\n",
       "      <td>[P0C6A0, Q6P4F2]</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>P0C6A0</td>\n",
       "      <td>Q9BYD3</td>\n",
       "      <td>[P0C6A0, Q9BYD3]</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>P0C6A0</td>\n",
       "      <td>Q9UMF0</td>\n",
       "      <td>[P0C6A0, Q9UMF0]</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>P0C6A0</td>\n",
       "      <td>P29597</td>\n",
       "      <td>[P0C6A0, P29597]</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>P0C6A0</td>\n",
       "      <td>O75791</td>\n",
       "      <td>[O75791, P0C6A0]</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67</th>\n",
       "      <td>P05362</td>\n",
       "      <td>Q9UMF0</td>\n",
       "      <td>[P05362, Q9UMF0]</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68</th>\n",
       "      <td>P05362</td>\n",
       "      <td>P29597</td>\n",
       "      <td>[P05362, P29597]</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>69</th>\n",
       "      <td>P05362</td>\n",
       "      <td>O75791</td>\n",
       "      <td>[O75791, P05362]</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70</th>\n",
       "      <td>P05362</td>\n",
       "      <td>Q8N9T8</td>\n",
       "      <td>[P05362, Q8N9T8]</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>71</th>\n",
       "      <td>P05362</td>\n",
       "      <td>Q13445</td>\n",
       "      <td>[P05362, Q13445]</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>72 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   protein assoc_protein       combination  shared_ppi  connected\n",
       "0   P0C6A0        Q6P4F2  [P0C6A0, Q6P4F2]       False      False\n",
       "1   P0C6A0        Q9BYD3  [P0C6A0, Q9BYD3]       False      False\n",
       "2   P0C6A0        Q9UMF0  [P0C6A0, Q9UMF0]       False      False\n",
       "3   P0C6A0        P29597  [P0C6A0, P29597]       False      False\n",
       "4   P0C6A0        O75791  [O75791, P0C6A0]       False      False\n",
       "..     ...           ...               ...         ...        ...\n",
       "67  P05362        Q9UMF0  [P05362, Q9UMF0]        True       True\n",
       "68  P05362        P29597  [P05362, P29597]        True       True\n",
       "69  P05362        O75791  [O75791, P05362]        True       True\n",
       "70  P05362        Q8N9T8  [P05362, Q8N9T8]       False      False\n",
       "71  P05362        Q13445  [P05362, Q13445]       False      False\n",
       "\n",
       "[72 rows x 5 columns]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def get_shared_ppi(ppi_df):\n",
    "    def per_row(ppi_df, protein, assoc_protein):\n",
    "        df = ppi_df.query(\n",
    "            f\"`protein` == '{protein}' & `assoc_protein` == '{assoc_protein}'\"\n",
    "        )\n",
    "        return True if len(df) > 0 else False\n",
    "\n",
    "    pairwise_permutations = permutations(ppi_df[\"protein\"].drop_duplicates(), 2)\n",
    "\n",
    "    shared_ppi_df = pd.DataFrame(\n",
    "        [\n",
    "            {\n",
    "                \"protein\": permutation[0],\n",
    "                \"assoc_protein\": permutation[1],\n",
    "                \"combination\": sorted(permutation),\n",
    "                \"shared_ppi\": per_row(ppi_df, *permutation),\n",
    "            }\n",
    "            for permutation in pairwise_permutations\n",
    "        ]\n",
    "    ).assign(connected=lambda df: df[\"shared_ppi\"])\n",
    "    return shared_ppi_df\n",
    "\n",
    "\n",
    "shared_ppi_df = get_shared_ppi(ppi_df)\n",
    "print(\"Num. shared_ppi pairs: \", len(shared_ppi_df.query(\"shared_ppi\")))\n",
    "shared_ppi_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "south-general",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:27.865456Z",
     "iopub.status.busy": "2021-03-22T10:46:27.864920Z",
     "iopub.status.idle": "2021-03-22T10:46:27.869434Z",
     "shell.execute_reply": "2021-03-22T10:46:27.869749Z"
    },
    "papermill": {
     "duration": 0.046351,
     "end_time": "2021-03-22T10:46:27.869864",
     "exception": false,
     "start_time": "2021-03-22T10:46:27.823513",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{'group_members': {'Q8N9T8', 'Q9BYD3'}, 'group_size': 2},\n",
       " {'group_members': {'O75791', 'P05362', 'P29597', 'Q9UMF0'}, 'group_size': 4},\n",
       " {'group_members': {'P0C6A0'}, 'group_size': 1},\n",
       " {'group_members': {'Q6P4F2'}, 'group_size': 1},\n",
       " {'group_members': {'Q13445'}, 'group_size': 1}]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ppi_protein_graph = protein_df_to_graph(shared_ppi_df)\n",
    "ppi_protein_groups = graph_to_protein_groups(ppi_protein_graph)\n",
    "ppi_protein_groups"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "smart-concert",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:27.916649Z",
     "iopub.status.busy": "2021-03-22T10:46:27.916195Z",
     "iopub.status.idle": "2021-03-22T10:46:28.176406Z",
     "shell.execute_reply": "2021-03-22T10:46:28.176955Z"
    },
    "papermill": {
     "duration": 0.286707,
     "end_time": "2021-03-22T10:46:28.177145",
     "exception": false,
     "start_time": "2021-03-22T10:46:27.890438",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x648 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "draw_graph(ppi_protein_graph)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "colonial-visiting",
   "metadata": {
    "papermill": {
     "duration": 0.039727,
     "end_time": "2021-03-22T10:46:28.257966",
     "exception": false,
     "start_time": "2021-03-22T10:46:28.218239",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "\n",
    "## Appendix: Querying GTEx for single tissue eQTLs\n",
    "\n",
    "Below is the query to retrieve and parse the single tissue eQTL data for the SNP rs12720356."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "violent-friendship",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:28.306246Z",
     "iopub.status.busy": "2021-03-22T10:46:28.305516Z",
     "iopub.status.idle": "2021-03-22T10:46:28.860327Z",
     "shell.execute_reply": "2021-03-22T10:46:28.860932Z"
    },
    "papermill": {
     "duration": 0.583568,
     "end_time": "2021-03-22T10:46:28.861112",
     "exception": false,
     "start_time": "2021-03-22T10:46:28.277544",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>pval</th>\n",
       "      <th>tissue</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ensembl_id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ENSG00000090339</th>\n",
       "      <td>1.035110e-04</td>\n",
       "      <td>[Whole_Blood]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ENSG00000099203</th>\n",
       "      <td>4.804360e-05</td>\n",
       "      <td>[Whole_Blood]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ENSG00000105364</th>\n",
       "      <td>6.797370e-09</td>\n",
       "      <td>[Skin_Not_Sun_Exposed_Suprapubic, Artery_Tibia...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ENSG00000105376</th>\n",
       "      <td>7.561840e-07</td>\n",
       "      <td>[Whole_Blood]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ENSG00000105397</th>\n",
       "      <td>3.017820e-06</td>\n",
       "      <td>[Whole_Blood, Adrenal_Gland]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ENSG00000129347</th>\n",
       "      <td>9.941250e-06</td>\n",
       "      <td>[Skin_Sun_Exposed_Lower_leg]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ENSG00000161847</th>\n",
       "      <td>5.448880e-06</td>\n",
       "      <td>[Whole_Blood]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ENSG00000220201</th>\n",
       "      <td>1.128130e-15</td>\n",
       "      <td>[Thyroid, Skin_Not_Sun_Exposed_Suprapubic, Pro...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ENSG00000267673</th>\n",
       "      <td>2.929330e-11</td>\n",
       "      <td>[Thyroid, Artery_Tibial, Nerve_Tibial, Brain_C...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         pval  \\\n",
       "ensembl_id                      \n",
       "ENSG00000090339  1.035110e-04   \n",
       "ENSG00000099203  4.804360e-05   \n",
       "ENSG00000105364  6.797370e-09   \n",
       "ENSG00000105376  7.561840e-07   \n",
       "ENSG00000105397  3.017820e-06   \n",
       "ENSG00000129347  9.941250e-06   \n",
       "ENSG00000161847  5.448880e-06   \n",
       "ENSG00000220201  1.128130e-15   \n",
       "ENSG00000267673  2.929330e-11   \n",
       "\n",
       "                                                            tissue  \n",
       "ensembl_id                                                          \n",
       "ENSG00000090339                                      [Whole_Blood]  \n",
       "ENSG00000099203                                      [Whole_Blood]  \n",
       "ENSG00000105364  [Skin_Not_Sun_Exposed_Suprapubic, Artery_Tibia...  \n",
       "ENSG00000105376                                      [Whole_Blood]  \n",
       "ENSG00000105397                       [Whole_Blood, Adrenal_Gland]  \n",
       "ENSG00000129347                       [Skin_Sun_Exposed_Lower_leg]  \n",
       "ENSG00000161847                                      [Whole_Blood]  \n",
       "ENSG00000220201  [Thyroid, Skin_Not_Sun_Exposed_Suprapubic, Pro...  \n",
       "ENSG00000267673  [Thyroid, Artery_Tibial, Nerve_Tibial, Brain_C...  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def get_gtex_eqtl(snp):\n",
    "\n",
    "    # get single tissue eQTL\n",
    "    url = \"https://gtexportal.org/rest/v1/association/singleTissueEqtl\"\n",
    "    payload = {\"format\": \"json\", \"snpId\": snp, \"datasetId\": \"gtex_v8\"}\n",
    "    r = requests.get(url, params=payload)\n",
    "    r.raise_for_status()\n",
    "    eqtl_df = pd.json_normalize(r.json()[\"singleTissueEqtl\"])\n",
    "    if len(eqtl_df) > 0:\n",
    "        res_df = (\n",
    "            eqtl_df[[\"gencodeId\", \"pValue\", \"tissueSiteDetailId\"]]\n",
    "            .rename(\n",
    "                columns={\n",
    "                    \"gencodeId\": \"ensembl_id\",\n",
    "                    \"pValue\": \"pval\",\n",
    "                    \"tissueSiteDetailId\": \"tissue\",\n",
    "                }\n",
    "            )\n",
    "            .assign(\n",
    "                ensembl_id=lambda df: df[\"ensembl_id\"].apply(lambda x: x.split(\".\")[0])\n",
    "            )\n",
    "        )\n",
    "\n",
    "    else:\n",
    "        return pd.DataFrame(columns=[\"ensembl_id\", \"pval\", \"tissue\"])\n",
    "\n",
    "    return res_df\n",
    "\n",
    "\n",
    "eqtl_df = (\n",
    "    get_gtex_eqtl(snp=SNP)\n",
    "    # For each gene, select the lowest pval, then collect all tissues\n",
    "    .groupby(\"ensembl_id\")\n",
    "    .apply(\n",
    "        lambda group_df: pd.DataFrame(\n",
    "            {\n",
    "                \"pval\": group_df[\"pval\"].min(),\n",
    "                \"tissue\": [group_df[\"tissue\"].tolist()],\n",
    "            }\n",
    "        ).rename_axis(\"level\")\n",
    "    )\n",
    "    .reset_index(level=\"level\", drop=True)\n",
    ")\n",
    "\n",
    "eqtl_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "atmospheric-witch",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-03-22T10:46:28.943415Z",
     "iopub.status.busy": "2021-03-22T10:46:28.943018Z",
     "iopub.status.idle": "2021-03-22T10:46:30.480412Z",
     "shell.execute_reply": "2021-03-22T10:46:30.480790Z"
    },
    "papermill": {
     "duration": 1.583226,
     "end_time": "2021-03-22T10:46:30.480905",
     "exception": false,
     "start_time": "2021-03-22T10:46:28.897679",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ensembl_id</th>\n",
       "      <th>name</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ENSG00000090339</td>\n",
       "      <td>ICAM1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ENSG00000099203</td>\n",
       "      <td>TMED1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ENSG00000105364</td>\n",
       "      <td>MRPL4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ENSG00000105376</td>\n",
       "      <td>ICAM5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ENSG00000105397</td>\n",
       "      <td>TYK2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>ENSG00000129347</td>\n",
       "      <td>KRI1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>ENSG00000161847</td>\n",
       "      <td>RAVER1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>ENSG00000220201</td>\n",
       "      <td>ZGLP1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>ENSG00000267673</td>\n",
       "      <td>FDX1L</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        ensembl_id    name\n",
       "0  ENSG00000090339   ICAM1\n",
       "1  ENSG00000099203   TMED1\n",
       "2  ENSG00000105364   MRPL4\n",
       "3  ENSG00000105376   ICAM5\n",
       "4  ENSG00000105397    TYK2\n",
       "5  ENSG00000129347    KRI1\n",
       "6  ENSG00000161847  RAVER1\n",
       "7  ENSG00000220201   ZGLP1\n",
       "8  ENSG00000267673   FDX1L"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def get_epigraphdb_gene_name(ensembl_id):\n",
    "\n",
    "    endpoint = \"/meta/nodes/Gene/search\"\n",
    "    url = f\"{API_URL}{endpoint}\"\n",
    "    params = {\"id\": ensembl_id, \"full_data\": False}\n",
    "    r = requests.get(url, params=params)\n",
    "    r.raise_for_status()\n",
    "    res = r.json()\n",
    "    if not res[\"metadata\"][\"empty_results\"]:\n",
    "        return res[\"results\"][0][\"name\"]\n",
    "\n",
    "\n",
    "gene_name_df = pd.DataFrame.from_records(\n",
    "    [\n",
    "        {\n",
    "            \"ensembl_id\": ensembl_id,\n",
    "            \"name\": get_epigraphdb_gene_name(ensembl_id),\n",
    "        }\n",
    "        for ensembl_id in list(eqtl_df.index)\n",
    "    ]\n",
    ")\n",
    "gene_name_df"
   ]
  }
 ],
 "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": "3.7.7"
  },
  "papermill": {
   "default_parameters": {},
   "duration": 10.200629,
   "end_time": "2021-03-22T10:46:31.661545",
   "environment_variables": {},
   "exception": null,
   "input_path": "paper-case-studies/case-1-pleiotropy.ipynb",
   "output_path": "paper-case-studies/case-1-pleiotropy.ipynb",
   "parameters": {
    "API_URL": "https://api.epigraphdb.org"
   },
   "start_time": "2021-03-22T10:46:21.460916",
   "version": "2.3.3"
  },
  "toc-autonumbering": false
 },
 "nbformat": 4,
 "nbformat_minor": 5
}