rcsb_affinity_graphql.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a4226b9e-f975-4a5b-b716-7cc0bd2843a9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['10GS',\n",
       " '13GS',\n",
       " '16PK',\n",
       " '181L',\n",
       " '182L',\n",
       " '183L',\n",
       " '184L',\n",
       " '185L',\n",
       " '186L',\n",
       " '187L']"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import json\n",
    "import requests \n",
    "import urllib\n",
    "from urllib.request import urlopen\n",
    "\n",
    "\n",
    "def get_all_entries():\n",
    "    base_url = 'https://search.rcsb.org/rcsbsearch/v1/query'\n",
    "    json_query_string = \"\"\"\n",
    "        {\n",
    "      \"query\": {\n",
    "        \"type\": \"terminal\",\n",
    "        \"service\": \"text\",\n",
    "        \"parameters\": {\n",
    "          \"attribute\": \"rcsb_binding_affinity.value\",\n",
    "          \"operator\": \"exists\"\n",
    "        }\n",
    "      },\n",
    "      \"request_options\": {\n",
    "        \"results_verbosity\": \"compact\",\n",
    "        \"return_all_hits\": true\n",
    "      },\n",
    "      \"return_type\": \"entry\"\n",
    "    }  \n",
    "        \"\"\"\n",
    "    #formulate the query:\n",
    "    query = urllib.parse.quote(json_query_string)\n",
    "    req_url = base_url+'?json={request}'\n",
    "    url_query = req_url.format(request=query)\n",
    "    response = urlopen(url_query)\n",
    "    #read the output into json/dict:\n",
    "    output = json.loads(response.read())\n",
    "    return output\n",
    "    \n",
    "all_entries = get_all_entries()\n",
    "all_entries['result_set'][:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e48f6708-05c3-458f-b0c2-c5bcf2c00d6e",
   "metadata": {},
   "outputs": [],
   "source": [
    "def query_for_component_id(ids):\n",
    "    \n",
    "    \"\"\"\n",
    "    Input a list of PDB codes and return all the measured affinity \n",
    "    data.\n",
    "    \"\"\"\n",
    "    #base graphql query:\n",
    "    base_url = 'https://data.rcsb.org/graphql'\n",
    "    \n",
    "    #all graphql queries require format: [\"id1\", \"id2\", so on...]\n",
    "    query_fmt = '[' + ', '.join(['\"' + i + '\"' for i in ids]) + ']'\n",
    "    \n",
    "    #put the query together\n",
    "    graphql_query = \"\"\"\n",
    "    {\n",
    "      entries(entry_ids:\"\"\" + query_fmt +\"\"\") {\n",
    "      rcsb_binding_affinity {\n",
    "                comp_id \n",
    "          value\n",
    "          type\n",
    "          unit\n",
    "          symbol\n",
    "          link\n",
    "          provenance_code\n",
    "          \n",
    "        }\n",
    "        rcsb_primary_citation {\n",
    "          pdbx_database_id_PubMed\n",
    "          pdbx_database_id_DOI\n",
    "        }\n",
    "      }\n",
    "    }\n",
    "    \"\"\"\n",
    "\n",
    "    #encode as an html request and get the json data\n",
    "    r = requests.post(base_url, json={'query': graphql_query})\n",
    "    output = r.json()\n",
    "    return output\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a3f1a2d4-4998-47a4-ba5d-11f780ae2ce1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'data': {'entries': [{'rcsb_binding_affinity': [{'comp_id': 'VWW',\n",
       "      'value': 400.0,\n",
       "      'type': 'Ki',\n",
       "      'unit': 'nM',\n",
       "      'symbol': None,\n",
       "      'link': 'http://www.pdbbind-cn.org/quickpdb.php?quickpdb=10gs',\n",
       "      'provenance_code': 'PDBBind'},\n",
       "     {'comp_id': 'VWW',\n",
       "      'value': 400.0,\n",
       "      'type': 'Ki',\n",
       "      'unit': 'nM',\n",
       "      'symbol': None,\n",
       "      'link': 'http://www.bindingmoad.org/pdbrecords/index/10gs',\n",
       "      'provenance_code': 'Binding MOAD'}],\n",
       "    'rcsb_primary_citation': {'pdbx_database_id_PubMed': 9398518,\n",
       "     'pdbx_database_id_DOI': '10.1006/jmbi.1997.1364'}},\n",
       "   {'rcsb_binding_affinity': [{'comp_id': 'SAS',\n",
       "      'value': 24000.0,\n",
       "      'type': 'Ki',\n",
       "      'unit': 'nM',\n",
       "      'symbol': None,\n",
       "      'link': 'http://www.bindingmoad.org/pdbrecords/index/13gs',\n",
       "      'provenance_code': 'Binding MOAD'},\n",
       "     {'comp_id': 'SAS',\n",
       "      'value': 24000.0,\n",
       "      'type': 'Ki',\n",
       "      'unit': 'nM',\n",
       "      'symbol': None,\n",
       "      'link': 'http://www.pdbbind-cn.org/quickpdb.php?quickpdb=13gs',\n",
       "      'provenance_code': 'PDBBind'}],\n",
       "    'rcsb_primary_citation': {'pdbx_database_id_PubMed': 10452896,\n",
       "     'pdbx_database_id_DOI': '10.1006/jmbi.1999.3029'}}]}}"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "query_for_component_id(all_entries['result_set'][0:2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "310cf748-5bff-40b8-b969-aa0201fb0c94",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}