{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Loading obsolete GO terms   \n",
    "Obsolete GO terms are loaded upon request.\n",
    "\n",
    "Relevant GODag arguments include:\n",
    "  * load_obsolete=True\n",
    "  * optional_attrs={'consider', 'replaced_by'}\n",
    "\n",
    "## 1) Download GO DAG file, go-basic.obo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "requests.get(http://purl.obolibrary.org/obo/go/go-basic.obo, stream=True)\n",
      "  WROTE: go-basic.obo\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from goatools.base import download_go_basic_obo\n",
    "fin_dag = download_go_basic_obo(\"go-basic.obo\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2) Read GO terms into a GODag instance: load_obsolete=True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "go-basic.obo: fmt(1.2) rel(2020-09-10) 50,377 GO Terms\n"
     ]
    }
   ],
   "source": [
    "from goatools.obo_parser import GODag\n",
    "\n",
    "godag = GODag(fin_dag, load_obsolete=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3) Load the optional GODag attributes, *consider* and *replaced_by*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "go-basic.obo: fmt(1.2) rel(2020-09-10) 50,377 GO Terms; optional_attrs(consider replaced_by)\n"
     ]
    }
   ],
   "source": [
    "godag = GODag(fin_dag, optional_attrs={'consider', 'replaced_by'}, load_obsolete=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3a) How to suppress the GODag output messages\n",
    "Set argument: `prt=None`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "godag = GODag(fin_dag, optional_attrs={'consider', 'replaced_by'}, load_obsolete=True, prt=None)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4) What GO term replaced the obsolete GO terms?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2,971 GO terms are obsolete\n",
      "1 GO:0042142 was replaced by GO:0046911\n",
      "2 GO:0006502 was replaced by GO:0018342\n",
      "3 GO:0006503 was replaced by GO:0018343\n",
      "4 GO:0006504 was replaced by GO:0018344\n",
      "5 GO:0006510 was replaced by GO:0004176\n"
     ]
    }
   ],
   "source": [
    "goterms_obsolete = set(o for o in godag.values() if o.is_obsolete)\n",
    "print('{N:,} GO terms are obsolete'.format(N=len(goterms_obsolete)))\n",
    "\n",
    "num = 1\n",
    "for goterm in goterms_obsolete:\n",
    "    if goterm.replaced_by[:3] == 'GO:':\n",
    "        print('{I} {OLD} was replaced by {NEW}'.format(\n",
    "            I=num, OLD=goterm.item_id, NEW=goterm.replaced_by))\n",
    "        if num == 5:\n",
    "            break\n",
    "        num += 1\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5) What GO terms were considered?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 Given GO:0006894, consider {'GO:0055107', 'GO:0055108'}\n",
      "2 Given GO:0045757, consider {'GO:0030837', 'GO:0030835'}\n",
      "3 Given GO:0052108, consider {'GO:0044117', 'GO:0044115'}\n",
      "4 Given GO:0006815, consider {'GO:0006814', 'GO:0006813'}\n",
      "5 Given GO:0006510, consider {'GO:0051603', 'GO:0016485', 'GO:0006508'}\n"
     ]
    }
   ],
   "source": [
    "num = 1\n",
    "for goterm in goterms_obsolete:\n",
    "    if goterm.consider:\n",
    "        print('{I} Given {OLD}, consider {NEW}'.format(\n",
    "            I=num, OLD=goterm.item_id, NEW=goterm.consider))\n",
    "        if num == 5:\n",
    "            break\n",
    "        num += 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Copyright (C) 2019-present, DV Klopfenstein and Haibao Tang. All rights reserved."
   ]
  }
 ],
 "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": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}