Readbase API example

gistfile1.txt

{
 "metadata": {
  "name": "Map Bodymap STAR"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import urllib2\n",
      "import subprocess\n",
      "import os\n",
      "import json"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 16
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Query the database for Bodymap 2.0 files (`study_id=ERP000546`) that are single end (`library_layout=single`)\n",
      "Read all of the results (`.read()`), and split on newlines (the results are returned one uid by line)"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "uids = urllib2.urlopen('http://128.100.241.98/readbase/query?study_id=ERP000546&library_layout=single').read().split('\\n')"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 6
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Now request the JSON metadata for all the experiments. The uids are given as a comma-separated list in the URL and the result is automatically parsed by the JSON module."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "metadata_json = json.load(urllib2.urlopen('http://128.100.241.98/readbase/get-json?uid=' + ','.join(uids)))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 19
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Here I extract the run ids for all runs and all experiments (an experiment can have more than one run, so we need to iterate over all runs for the experiment)."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "run_ids = [run['run_id'] for experiment in metadata_json for run in experiment['runs']]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 29
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Load the genome into shared memory for the STAR mapping algorithm."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "subprocess.check_output(['/home/hannes/usr/bin/STAR', '--genomeDir', '/data/hg19/', '--genomeLoad', 'LoadAndExit'])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "pyout",
       "prompt_number": 33,
       "text": [
        "'Jan 31 17:14:32 ..... Started STAR run\\n'"
       ]
      }
     ],
     "prompt_number": 33
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Parameter for STAR\n",
      "star_parameters = [\n",
      "    '--genomeDir', '/data/hg19',\n",
      "    '--runThreadN', '32']"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 36
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Run STAR for each file in a new directory\n",
      "for run_id in run_ids:\n",
      "    os.chdir('/data/Bodymap/STAR_maps/')\n",
      "    if not os.path.exists(run_id): os.mkdir(run_id)\n",
      "    os.chdir(run_id)\n",
      "    print run_id\n",
      "    path = '/data/Bodymap/ERP000546_fastq/' + run_id + '.fastq'\n",
      "    subprocess.check_output(['/home/hannes/usr/bin/STAR'] + star_parameters + ['--readFilesIn'] + [path])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "ERR030865\n",
        "ERR030861"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030864"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030902"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030871"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030857"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030903"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030888"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030898"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030863"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030900"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030895"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030866"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030856"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030858"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030896"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030860"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030870"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030862"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030893"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030899"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030868"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030894"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030890"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030869"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030889"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030891"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030867"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030859"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030897"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030901"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "ERR030892"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n"
       ]
      }
     ],
     "prompt_number": 34
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Remove the index from the shared memory\n",
      "os.chdir('/data/Bodymap/STAR_maps/')\n",
      "subprocess.check_output(['/home/hannes/usr/bin/STAR', '--genomeDir', '/data/hg19/', '--genomeLoad', 'Remove'])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "pyout",
       "prompt_number": 35,
       "text": [
        "'Jan 31 19:12:47 ..... Started STAR run\\n'"
       ]
      }
     ],
     "prompt_number": 35
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [],
     "language": "python",
     "metadata": {},
     "outputs": []
    }
   ],
   "metadata": {}
  }
 ]
}