{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Basic querying of GO\n",
    "\n",
    "See https://github.com/geneontology/helpdesk/issues/22\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/cjm/repos/go-notebooks/venv/lib/python3.5/site-packages/cachier/mongo_core.py:24: UserWarning: Cachier warning: pymongo was not found. MongoDB cores will not work.\n",
      "  \"Cachier warning: pymongo was not found. MongoDB cores will not work.\")\n"
     ]
    }
   ],
   "source": [
    "## Create an ontology factory and use it to make an ontology object, using the handle 'go'\n",
    "## This will create an ontology object using remote services\n",
    "from ontobio.ontol_factory import OntologyFactory\n",
    "ofa = OntologyFactory()\n",
    "ont = ofa.create('go')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "## Create an association factory to get pombase GO annotations\n",
    "## (this uses an 'eager' method, where all annotations are fetched from services in advance)\n",
    "from ontobio.assoc_factory import AssociationSetFactory\n",
    "afactory = AssociationSetFactory()\n",
    "aset = afactory.create(ontology=ont, subject_category='gene', object_category='function', taxon='NCBITaxon:4896')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Annotation: GO:1903467 'negative regulation of mitotic DNA replication initiation'\n",
      " Annotation: GO:0004693 'cyclin-dependent protein serine/threonine kinase activity'\n",
      " Annotation: GO:0072686 'mitotic spindle'\n",
      " Annotation: GO:0005515 'protein binding'\n",
      " Annotation: GO:0031138 'negative regulation of conjugation with cellular fusion'\n",
      " Annotation: GO:1905785 'negative regulation of anaphase-promoting complex-dependent catabolic process'\n",
      " Annotation: GO:0010389 'regulation of G2/M transition of mitotic cell cycle'\n",
      " Annotation: GO:0005829 'cytosol'\n",
      " Annotation: GO:0031031 'positive regulation of septation initiation signaling'\n",
      " Annotation: GO:0000775 'chromosome, centromeric region'\n",
      " Annotation: GO:1900087 'positive regulation of G1/S transition of mitotic cell cycle'\n",
      " Annotation: GO:0072434 'signal transduction involved in mitotic G2 DNA damage checkpoint'\n",
      " Annotation: GO:0005737 'cytoplasm'\n",
      " Annotation: GO:1904537 'negative regulation of mitotic telomere tethering at nuclear periphery'\n",
      " Annotation: GO:0007089 'traversing start control point of mitotic cell cycle'\n",
      " Annotation: GO:0072435 'response to mitotic G2 DNA damage checkpoint signaling'\n",
      " Annotation: GO:1990820 'response to mitotic DNA integrity checkpoint signaling'\n",
      " Annotation: GO:0010971 'positive regulation of G2/M transition of mitotic cell cycle'\n",
      " Annotation: GO:0097472 'cyclin-dependent protein kinase activity'\n",
      " Annotation: GO:0004672 'protein kinase activity'\n",
      " Annotation: GO:1905168 'positive regulation of double-strand break repair via homologous recombination'\n",
      " Annotation: GO:1990023 'mitotic spindle midzone'\n",
      " Annotation: GO:1903465 'positive regulation of mitotic cell cycle DNA replication'\n",
      " Annotation: GO:0072429 'response to intra-S DNA damage checkpoint signaling'\n",
      " Annotation: GO:0004674 'protein serine/threonine kinase activity'\n",
      " Annotation: GO:0005634 'nucleus'\n",
      " Annotation: GO:1902845 'negative regulation of mitotic spindle elongation'\n",
      " Annotation: GO:0001100 'negative regulation of exit from mitosis'\n",
      " Annotation: GO:0005524 'ATP binding'\n",
      " Annotation: GO:1903380 'positive regulation of mitotic chromosome condensation'\n",
      " Annotation: GO:0044732 'mitotic spindle pole body'\n",
      " Annotation: GO:0045842 'positive regulation of mitotic metaphase/anaphase transition'\n",
      " Annotation: GO:0000307 'cyclin-dependent protein kinase holoenzyme complex'\n",
      " Annotation: GO:0000790 'nuclear chromatin'\n",
      " Annotation: GO:1902424 'negative regulation of attachment of mitotic spindle microtubules to kinetochore'\n",
      " Annotation: GO:2001033 'negative regulation of double-strand break repair via nonhomologous end joining'\n",
      " Annotation: GO:0098783 'correction of merotelic kinetochore attachment, mitotic'\n",
      " Annotation: GO:0051445 'regulation of meiotic cell cycle'\n",
      " Annotation: GO:0035974 'meiotic spindle pole body'\n",
      " Annotation: GO:0008361 'regulation of cell size'\n"
     ]
    }
   ],
   "source": [
    "## Change this to your gene of interest\n",
    "CDC2 = 'PomBase:SPBC11B10.09'\n",
    "\n",
    "## Write id-labels for direct annotations to CDC2\n",
    "direct_anns = aset.annotations(CDC2)\n",
    "for t in direct_anns:\n",
    "    print(\" Annotation: {id} '{label}'\".format(id=t, label=ont.label(t)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "## TODO: Annotation extensions\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "## Get ancestors of all direct annotated terms, following is-a and part-of\n",
    "inferred_anns = ont.traverse_nodes(direct_anns, relations=['subClassOf', 'BFO:0000050'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Simple visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "## Render using graphviz, highlighting directly annotated terms\n",
    "from ontobio.io.ontol_renderers import GraphRenderer\n",
    "w = GraphRenderer.create('png')\n",
    "w.outfile = 'output/cdc2.png'\n",
    "w.write_subgraph(ont, inferred_anns, query_ids=direct_anns)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![img](output/cdc2.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Grouping annotations by slims or root nodes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['GO:0003674', 'GO:0008150', 'GO:0005575']"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "## Get 3 GO ontology roots\n",
    "[mf] = ont.search('molecular_function')\n",
    "[bp] = ont.search('biological_process')\n",
    "[cc] = ont.search('cellular_component')\n",
    "roots = [mf,bp,cc]\n",
    "roots"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "## Create a mapping of all nodes to the ontology root\n",
    "rootmap = ont.create_slim_mapping(subset_nodes=roots, relations='subClassOf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "GO:0003674 molecular_function\n",
      "  GO:0097472 cyclin-dependent protein kinase activity\n",
      "  GO:0005515 protein binding\n",
      "  GO:0005524 ATP binding\n",
      "  GO:0004674 protein serine/threonine kinase activity\n",
      "  GO:0004693 cyclin-dependent protein serine/threonine kinase activity\n",
      "  GO:0004672 protein kinase activity\n",
      "GO:0008150 biological_process\n",
      "  GO:1905168 positive regulation of double-strand break repair via homologous recombination\n",
      "  GO:1903465 positive regulation of mitotic cell cycle DNA replication\n",
      "  GO:0010971 positive regulation of G2/M transition of mitotic cell cycle\n",
      "  GO:1903467 negative regulation of mitotic DNA replication initiation\n",
      "  GO:0045842 positive regulation of mitotic metaphase/anaphase transition\n",
      "  GO:0031031 positive regulation of septation initiation signaling\n",
      "  GO:1902424 negative regulation of attachment of mitotic spindle microtubules to kinetochore\n",
      "  GO:1900087 positive regulation of G1/S transition of mitotic cell cycle\n",
      "  GO:1990820 response to mitotic DNA integrity checkpoint signaling\n",
      "  GO:0007089 traversing start control point of mitotic cell cycle\n",
      "  GO:0072435 response to mitotic G2 DNA damage checkpoint signaling\n",
      "  GO:0051445 regulation of meiotic cell cycle\n",
      "  GO:2001033 negative regulation of double-strand break repair via nonhomologous end joining\n",
      "  GO:0008361 regulation of cell size\n",
      "  GO:1902845 negative regulation of mitotic spindle elongation\n",
      "  GO:0010389 regulation of G2/M transition of mitotic cell cycle\n",
      "  GO:0001100 negative regulation of exit from mitosis\n",
      "  GO:1903380 positive regulation of mitotic chromosome condensation\n",
      "  GO:0072429 response to intra-S DNA damage checkpoint signaling\n",
      "  GO:0072434 signal transduction involved in mitotic G2 DNA damage checkpoint\n",
      "  GO:0098783 correction of merotelic kinetochore attachment, mitotic\n",
      "  GO:0031138 negative regulation of conjugation with cellular fusion\n",
      "  GO:1904537 negative regulation of mitotic telomere tethering at nuclear periphery\n",
      "  GO:1905785 negative regulation of anaphase-promoting complex-dependent catabolic process\n",
      "GO:0005575 cellular_component\n",
      "  GO:0000775 chromosome, centromeric region\n",
      "  GO:0044732 mitotic spindle pole body\n",
      "  GO:0035974 meiotic spindle pole body\n",
      "  GO:0005829 cytosol\n",
      "  GO:0005737 cytoplasm\n",
      "  GO:0005634 nucleus\n",
      "  GO:0000790 nuclear chromatin\n",
      "  GO:0000307 cyclin-dependent protein kinase holoenzyme complex\n",
      "  GO:0072686 mitotic spindle\n",
      "  GO:1990023 mitotic spindle midzone\n"
     ]
    }
   ],
   "source": [
    "## Define a simple formatter that breaks annotations into groups/slims\n",
    "def print_by_group(group_nodes, anns, nodemap):\n",
    "    for r in group_nodes:\n",
    "        filtered_anns = [x for x in anns if r in nodemap[x]]\n",
    "        if len(filtered_anns) > 0:\n",
    "            print('{} {}'.format(r,ont.label(r)))\n",
    "            for a in filtered_anns:\n",
    "                print('  {} {}'.format(a,ont.label(a)))\n",
    "            \n",
    "## Print all annotations to CDC2, grouped by ontology\n",
    "print_by_group(roots, direct_anns, rootmap)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Removing redundant links\n",
    "\n",
    "Note in the report above we see some redundant annotations - nucleus and nuclear chromatin for example\n",
    "\n",
    "Next we'll filter these out. We use a simple model of redundancy, using minimal relations, and ignoring\n",
    "evidence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reduced 40 annotations -> 32 nonredundant annotations\n"
     ]
    }
   ],
   "source": [
    "## Create a go-basic ontology, restricted to is-a and part-of\n",
    "basic_ont = ont.subontology(relations=['subClassOf', 'BFO:0000050'])\n",
    "nr_anns = basic_ont.filter_redundant(direct_anns)\n",
    "print(\"Reduced {} annotations -> {} nonredundant annotations\".format(len(direct_anns), len(nr_anns)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "GO:0003674 molecular_function\n",
      "  GO:0005515 protein binding\n",
      "  GO:0005524 ATP binding\n",
      "  GO:0004693 cyclin-dependent protein serine/threonine kinase activity\n",
      "GO:0008150 biological_process\n",
      "  GO:1905168 positive regulation of double-strand break repair via homologous recombination\n",
      "  GO:1903465 positive regulation of mitotic cell cycle DNA replication\n",
      "  GO:1903467 negative regulation of mitotic DNA replication initiation\n",
      "  GO:0010971 positive regulation of G2/M transition of mitotic cell cycle\n",
      "  GO:0045842 positive regulation of mitotic metaphase/anaphase transition\n",
      "  GO:0031031 positive regulation of septation initiation signaling\n",
      "  GO:1902424 negative regulation of attachment of mitotic spindle microtubules to kinetochore\n",
      "  GO:1990820 response to mitotic DNA integrity checkpoint signaling\n",
      "  GO:0007089 traversing start control point of mitotic cell cycle\n",
      "  GO:0072435 response to mitotic G2 DNA damage checkpoint signaling\n",
      "  GO:0051445 regulation of meiotic cell cycle\n",
      "  GO:2001033 negative regulation of double-strand break repair via nonhomologous end joining\n",
      "  GO:0008361 regulation of cell size\n",
      "  GO:1902845 negative regulation of mitotic spindle elongation\n",
      "  GO:0001100 negative regulation of exit from mitosis\n",
      "  GO:1903380 positive regulation of mitotic chromosome condensation\n",
      "  GO:0072429 response to intra-S DNA damage checkpoint signaling\n",
      "  GO:0072434 signal transduction involved in mitotic G2 DNA damage checkpoint\n",
      "  GO:0098783 correction of merotelic kinetochore attachment, mitotic\n",
      "  GO:0031138 negative regulation of conjugation with cellular fusion\n",
      "  GO:1904537 negative regulation of mitotic telomere tethering at nuclear periphery\n",
      "  GO:1905785 negative regulation of anaphase-promoting complex-dependent catabolic process\n",
      "GO:0005575 cellular_component\n",
      "  GO:0000775 chromosome, centromeric region\n",
      "  GO:0044732 mitotic spindle pole body\n",
      "  GO:0035974 meiotic spindle pole body\n",
      "  GO:0005829 cytosol\n",
      "  GO:0000790 nuclear chromatin\n",
      "  GO:0000307 cyclin-dependent protein kinase holoenzyme complex\n",
      "  GO:1990023 mitotic spindle midzone\n"
     ]
    }
   ],
   "source": [
    "## Report again, this time non-redundant\n",
    "print_by_group(roots, nr_anns, rootmap)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "slim_nodes = ont.extract_subset('goslim_pombe')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "slim_map = ont.create_slim_mapping(subset='goslim_pombe', relations=['subClassOf', 'BFO:0000050'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "GO:1901990 regulation of mitotic cell cycle phase transition\n",
      "  GO:0010971 positive regulation of G2/M transition of mitotic cell cycle\n",
      "  GO:0045842 positive regulation of mitotic metaphase/anaphase transition\n",
      "  GO:0031031 positive regulation of septation initiation signaling\n",
      "  GO:0007089 traversing start control point of mitotic cell cycle\n",
      "  GO:0001100 negative regulation of exit from mitosis\n",
      "  GO:0072434 signal transduction involved in mitotic G2 DNA damage checkpoint\n",
      "GO:0000070 mitotic sister chromatid segregation\n",
      "  GO:0098783 correction of merotelic kinetochore attachment, mitotic\n",
      "GO:0008150 biological_process\n",
      "  GO:1905168 positive regulation of double-strand break repair via homologous recombination\n",
      "  GO:1903465 positive regulation of mitotic cell cycle DNA replication\n",
      "  GO:1903467 negative regulation of mitotic DNA replication initiation\n",
      "  GO:0010971 positive regulation of G2/M transition of mitotic cell cycle\n",
      "  GO:0045842 positive regulation of mitotic metaphase/anaphase transition\n",
      "  GO:0031031 positive regulation of septation initiation signaling\n",
      "  GO:1902424 negative regulation of attachment of mitotic spindle microtubules to kinetochore\n",
      "  GO:1990820 response to mitotic DNA integrity checkpoint signaling\n",
      "  GO:0007089 traversing start control point of mitotic cell cycle\n",
      "  GO:0072435 response to mitotic G2 DNA damage checkpoint signaling\n",
      "  GO:0051445 regulation of meiotic cell cycle\n",
      "  GO:2001033 negative regulation of double-strand break repair via nonhomologous end joining\n",
      "  GO:0008361 regulation of cell size\n",
      "  GO:1902845 negative regulation of mitotic spindle elongation\n",
      "  GO:0001100 negative regulation of exit from mitosis\n",
      "  GO:0004693 cyclin-dependent protein serine/threonine kinase activity\n",
      "  GO:1903380 positive regulation of mitotic chromosome condensation\n",
      "  GO:0072429 response to intra-S DNA damage checkpoint signaling\n",
      "  GO:0072434 signal transduction involved in mitotic G2 DNA damage checkpoint\n",
      "  GO:0098783 correction of merotelic kinetochore attachment, mitotic\n",
      "  GO:0031138 negative regulation of conjugation with cellular fusion\n",
      "  GO:1904537 negative regulation of mitotic telomere tethering at nuclear periphery\n",
      "  GO:1905785 negative regulation of anaphase-promoting complex-dependent catabolic process\n",
      "GO:0023052 signaling\n",
      "  GO:0072434 signal transduction involved in mitotic G2 DNA damage checkpoint\n"
     ]
    }
   ],
   "source": [
    "## Report, with breakdown by slim\n",
    "print_by_group(slim_nodes, nr_anns, slim_map)"
   ]
  }
 ],
 "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": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}