notebooks/Pulldown affinity feature extraction.ipynb

{
 "metadata": {
  "name": "",
  "signature": "sha256:bebc047d358a5d590da98762275ee23c2130ccd1da861786815d2c467e7bcbb7"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "The pulldown experiments let us identify proteins that are localised to the presynaptic active zone which we are interested in studying.\n",
      "Other than just identifying the proteins it also results in a couple of measures which can be useful for predicting whether proteins interact:\n",
      "\n",
      "* Protein abundances\n",
      "* Affinity measures\n",
      "\n",
      "This notebook looks at the second of these two:\n",
      "\n",
      "## Affinity measures\n",
      "\n",
      "The affinity measure which has already been extracted from this is based on [this paper][xie_construction_2011] by Xie et al.\n",
      "They call their co-complexed score a C2S score and has a probabilistic basis, which is ideal for our use.\n",
      "All we need to do is extract a value for each protein pair we're looking at and see what kind of coverage these values give us.\n",
      "\n",
      "Due to the convenient structure of the file we are using we can extract the feature with the functionality of `ocbio.extract`.\n",
      "Each protein pair is in a row in Entrez Gene ID format with the corresponding C2S value.\n",
      "\n",
      "[xie_construction_2011]: http://bioinformatics.oxfordjournals.org/content/27/13/i159.abstract"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "cd ../../"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "/data/opencast/MRes\n"
       ]
      }
     ],
     "prompt_number": 1
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "# Improving coverage\n",
      "\n",
      "Unfortunately, when running the ocbio.extract code naively we \n",
      "To improve the coverage we will first use the `zeromissinginternal` option in the FeatureVectorAssembler.\n",
      "This will replace any interactions between proteins which the database knows about but doesn't have a value stored with zeros.\n",
      "\n",
      "Secondly, we will interpolate the expected value for this feature based on the values of other features with better coverage.\n",
      "To do this, first we will generate a set of feature vectors over the bait and prey combinations.\n",
      "Using this, we will train a linear regression algorithm.\n",
      "\n",
      "Then, we can pickle the trained linear regressor and store the table to create the feature vectors it must use.\n",
      "These can then be passed as options for the parser in the code to generate the feature vectors for the classification task.\n",
      "\n",
      "## Loading regression features\n",
      "\n",
      "Loading the features which were used for classification in the [pipeline prototype notebook][protonotes]:\n",
      "\n",
      "[protonotes]: http://nbviewer.ipython.org/github/ggray1729/opencast-bio/blob/master/notebooks/Full%20Pipeline%20Prototype.ipynb"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import sys,csv"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 2
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "sys.path.append(\"opencast-bio/\")"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 3
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import ocbio.extract"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 4
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "reload(ocbio.extract)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 17,
       "text": [
        "<module 'ocbio.extract' from 'opencast-bio/ocbio/extract.py'>"
       ]
      }
     ],
     "prompt_number": 17
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "!git annex unlock datasource.proto.tab"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "unlock datasource.proto.tab (copying...) "
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "ok\r\n"
       ]
      }
     ],
     "prompt_number": 7
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "f = open(\"datasource.proto.tab\",\"w\")\n",
      "c = csv.writer(f,delimiter=\"\\t\")\n",
      "# the HIPPIE feature\n",
      "c.writerow([\"HIPPIE/hippie_current.txt\",\"HIPPIE/feature.HIPPIE.2.db\",\"protindexes=(1,3);valindexes=(4);zeromissing=1\"])\n",
      "# Gene Ontology features\n",
      "c.writerow([\"Gene_Ontology\",\"Gene_Ontology\",\"generator=geneontology/testgen.pickle\"])\n",
      "# iRefIndex features\n",
      "c.writerow([\"iRefIndex\",\"iRefIndex\",\"generator=iRefIndex/human.iRefIndex.Entrez.1ofk.pickle\"])\n",
      "# STRING features\n",
      "c.writerow([\"STRING\",\"STRING\",\"generator=string/human.Entrez.string.pickle\"])\n",
      "# ENTS summary feature\n",
      "c.writerow(['ENTS_summary','ENTS_summary','generator=ents/human.Entrez.ENTS.summary.pickle'])\n",
      "f.close()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 8
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "assembler = ocbio.extract.FeatureVectorAssembler(\"datasource.proto.tab\", verbose=True)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Using  from top data directory datasource.proto.tab.\n",
        "Reading data source table:\n",
        "\tData source: HIPPIE/hippie_current.txt to be processed to HIPPIE/feature.HIPPIE.2.db\n",
        "\tData source: Gene_Ontology to be processed to Gene_Ontology\n",
        "\tData source: iRefIndex to be processed to iRefIndex\n",
        "\tData source: STRING to be processed to STRING\n",
        "\tData source: ENTS_summary to be processed to ENTS_summary\n",
        "Initialising parsers.\n",
        "Database HIPPIE/feature.HIPPIE.2.db may be empty, must be updated, please run regenerate.\n",
        "Finished Initialisation."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n"
       ]
      }
     ],
     "prompt_number": 16
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "assembler.regenerate(verbose=True)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Regenerating parsers:\n",
        "\t parser 0\n",
        "Database file may be empty, regenerating at HIPPIE/feature.HIPPIE.2.db from HIPPIE/hippie_current.txt.\n",
        "Filling database."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "Parsed 169626 lines.\n",
        "\t parser 1\n",
        "Custom generator function, no database to regenerate.\n",
        "\t parser 2\n",
        "Custom generator function, no database to regenerate.\n",
        "\t parser 3\n",
        "Custom generator function, no database to regenerate.\n",
        "\t parser 4\n",
        "Custom generator function, no database to regenerate.\n"
       ]
      }
     ],
     "prompt_number": 12
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### Writing training vectors"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "assembler.assemble(\"forGAVIN/pulldown_data/pulldown.interactions.Entrez.tsv\",\n",
      "                   \"features/pulldown.interactions.interpolate.vectors.txt\", verbose=True)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Reading pairfile: forGAVIN/pulldown_data/pulldown.interactions.Entrez.tsv\n",
        "Checking feature sizes:"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "\t Data source HIPPIE/hippie_current.txt produces features of size 1.\n",
        "\t Data source Gene_Ontology produces features of size 90.\n",
        "\t Data source iRefIndex produces features of size 11.\n",
        "\t Data source STRING produces features of size 8.\n",
        "\t Data source ENTS_summary produces features of size 1.\n",
        "Writing feature vectors."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "Wrote 1699246 vectors.\n",
        "Matched 100.00 % of protein pairs in forGAVIN/pulldown_data/pulldown.interactions.Entrez.tsv to features from HIPPIE/hippie_current.txt\n",
        "Matched 100.00 % of protein pairs in forGAVIN/pulldown_data/pulldown.interactions.Entrez.tsv to features from Gene_Ontology\n",
        "Matched 100.00 % of protein pairs in forGAVIN/pulldown_data/pulldown.interactions.Entrez.tsv to features from iRefIndex\n",
        "Matched 100.00 % of protein pairs in forGAVIN/pulldown_data/pulldown.interactions.Entrez.tsv to features from STRING\n",
        "Matched 100.00 % of protein pairs in forGAVIN/pulldown_data/pulldown.interactions.Entrez.tsv to features from ENTS_summary\n"
       ]
      }
     ],
     "prompt_number": 17
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### Writing training labels"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "f = open(\"datasource.affinity.tab\",\"w\")\n",
      "c = csv.writer(f,delimiter=\"\\t\")\n",
      "# just the affinity feature\n",
      "c.writerow([\"affinityresults/results2/unique_data_ppi_coor_C2S_entrez.csv\",\n",
      "            \"affinityresults/results2/affinity.Entrez.db\",\"zeromissinginternal=1\"])\n",
      "f.close()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 18
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "!git annex unlock affinityresults/results2/affinity.Entrez.db"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 7
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "assembler = ocbio.extract.FeatureVectorAssembler(\"datasource.affinity.tab\", verbose=True)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Using  from top data directory datasource.affinity.tab.\n",
        "Reading data source table:\n",
        "\tData source: affinityresults/results2/unique_data_ppi_coor_C2S_entrez.csv to be processed to affinityresults/results2/affinity.Entrez.db\n",
        "Initialising parsers.\n",
        "Database affinityresults/results2/affinity.Entrez.db last updated 2014-06-25 12:06:47\n",
        "Finished Initialisation."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n"
       ]
      }
     ],
     "prompt_number": 8
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "assembler.assemble(\"forGAVIN/pulldown_data/pulldown.interactions.Entrez.tsv\",\n",
      "                   \"features/pulldown.interactions.interpolate.affinity.targets.txt\", verbose=True)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Reading pairfile: forGAVIN/pulldown_data/pulldown.interactions.Entrez.tsv\n",
        "Checking feature sizes:"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "\t Data source affinityresults/results2/unique_data_ppi_coor_C2S_entrez.csv produces features of size 1.\n",
        "Writing feature vectors."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "Wrote 1699246 vectors.\n",
        "Matched 92.24 % of protein pairs in forGAVIN/pulldown_data/pulldown.interactions.Entrez.tsv to features from affinityresults/results2/unique_data_ppi_coor_C2S_entrez.csv\n"
       ]
      }
     ],
     "prompt_number": 9
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "# Linear regression\n",
      "\n",
      "Scikit-learn has an implementation of [linear regression][linreg] we can train to interpolate missing values.\n",
      "Specifically, in this case we will be using the [ridge regression][ridge] model to avoid overfitting on this large dataset.\n",
      "We will also be using 10-fold cross-validation and looking at the mean squared error over the test set to estimate the effectiveness of the model.\n",
      "\n",
      "[linreg]: http://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LinearRegression.html\n",
      "[ridge]: http://scikit-learn.org/stable/modules/generated/sklearn.linear_model.Ridge.html#sklearn.linear_model.Ridge.predict"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import sklearn.linear_model"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 3
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### Loading the data"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "#load vectors into memory:\n",
      "#loading X vectors\n",
      "X = loadtxt(\"features/pulldown.interactions.interpolate.vectors.txt\")"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 4
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "#loading y targets\n",
      "y = loadtxt(\"features/pulldown.interactions.interpolate.affinity.targets.txt\",dtype=str)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 3
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "missinglabels = y == \"missing\""
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 4
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "#zero missing values\n",
      "y[missinglabels] = 0.0"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 5
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "#convert to float\n",
      "y = y.astype(np.float)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 6
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### Splitting the data"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import sklearn.utils"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 9
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "X,y = sklearn.utils.shuffle(X,y)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 10
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import sklearn.cross_validation"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 11
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "kf = sklearn.cross_validation.KFold(y.shape[0],10)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 12
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "for train,test in kf:\n",
      "    print train,test"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[ 169925  169926  169927 ..., 1699243 1699244 1699245] [     0      1      2 ..., 169922 169923 169924]\n",
        "[      0       1       2 ..., 1699243 1699244 1699245]"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " [169925 169926 169927 ..., 339847 339848 339849]\n",
        "[      0       1       2 ..., 1699243 1699244 1699245]"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " [339850 339851 339852 ..., 509772 509773 509774]\n",
        "[      0       1       2 ..., 1699243 1699244 1699245]"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " [509775 509776 509777 ..., 679697 679698 679699]\n",
        "[      0       1       2 ..., 1699243 1699244 1699245]"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " [679700 679701 679702 ..., 849622 849623 849624]\n",
        "[      0       1       2 ..., 1699243 1699244 1699245]"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " [ 849625  849626  849627 ..., 1019547 1019548 1019549]\n",
        "[      0       1       2 ..., 1699243 1699244 1699245]"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " [1019550 1019551 1019552 ..., 1189471 1189472 1189473]\n",
        "[      0       1       2 ..., 1699243 1699244 1699245]"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " [1189474 1189475 1189476 ..., 1359395 1359396 1359397]\n",
        "[      0       1       2 ..., 1699243 1699244 1699245] [1359398 1359399 1359400 ..., 1529319 1529320 1529321]\n",
        "[      0       1       2 ..., 1529319 1529320 1529321]"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " [1529322 1529323 1529324 ..., 1699243 1699244 1699245]\n"
       ]
      }
     ],
     "prompt_number": 13
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "### Fitting the model"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "scores = []\n",
      "for train,test in kf:\n",
      "    #split the data\n",
      "    X_train, X_test, y_train, y_test = X[train], X[test], y[train], y[test]\n",
      "    #train the classifier\n",
      "    linreg = sklearn.linear_model.LinearRegression()\n",
      "    linreg.fit(X_train,y_train)\n",
      "    #test the classifier\n",
      "    scores.append(linreg.score(X_test,y_test))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 29
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print scores"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[0.006851770404450952, 0.0074703763710419757, 0.0088518497911050931, 0.0091843298107785465, 0.0088906585310051245, 0.0084456750397341462, 0.0073650204069001246, 0.0071932408986311591, 0.0072068518622049327, 0.0085652637462354519]\n"
       ]
      }
     ],
     "prompt_number": 30
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "mean(scores)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 31,
       "text": [
        "0.0080025036862087506"
       ]
      }
     ],
     "prompt_number": 31
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "According to the [documentation][]:\n",
      "\n",
      "> The coefficient R^2 is defined as (1 - u/v), where u is the regression sum of squares `((y_true - y_pred) ** 2).sum()` and v is the residual sum of squares `((y_true - y_true.mean()) ** 2).sum()`. Best possible score is 1.0, lower values are worse.\n",
      "\n",
      "So, this isn't a very good regression algorithm.\n",
      "Trying an [ensemble method][ensemble], such as an [AdaBoost][] regressor.\n",
      "\n",
      "[adaboost]: http://scikit-learn.org/stable/modules/generated/sklearn.ensemble.AdaBoostRegressor.html#sklearn.ensemble.AdaBoostRegressor\n",
      "[ensemble]: http://scikit-learn.org/stable/modules/ensemble.html#regression\n",
      "[documentation]: http://scikit-learn.org/stable/modules/generated/sklearn.linear_model.Ridge.html#sklearn.linear_model.Ridge.score"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import sklearn.ensemble"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 15
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "adascores = []\n",
      "for train,test in kf:\n",
      "    #split the data\n",
      "    X_train, X_test, y_train, y_test = X[train], X[test], y[train], y[test]\n",
      "    #train the classifier\n",
      "    ada = sklearn.ensemble.AdaBoostRegressor()\n",
      "    ada.fit(X_train,y_train)\n",
      "    #test the classifier\n",
      "    adascores.append(ada.score(X_test,y_test))\n",
      "    break"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 17
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Had to break the above loop after one iteration as it was taking too long to run."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "adascores"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 18,
       "text": [
        "[-1.9735555022882036]"
       ]
      }
     ],
     "prompt_number": 18
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "So the regression sum of squares much be three times larger than the residual sum of squares.\n",
      "Obviously, that's worse than the above."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "Xycov = cov(X.T,y.T)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 24
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print Xycov.shape"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "(112, 112)\n"
       ]
      }
     ],
     "prompt_number": 28
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print Xycov[:,-1][Xycov[:,-1]>1]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "[ 1.87683432  1.41172834  1.56386886  1.88611184  4.12909717  2.59874547]\n"
       ]
      }
     ],
     "prompt_number": 38
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "So several of the features are strongly correlated with this feature.\n",
      "Worth checking that this actually was the case.\n",
      "\n",
      "Trying a [ridge regression][ridge] classifier, with different values for alpha:\n",
      "\n",
      "[ridge]: http://scikit-learn.org/stable/modules/generated/sklearn.linear_model.Ridge.html#sklearn.linear_model.Ridge.score"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "ridgescores = []\n",
      "for train,test in kf:\n",
      "    #split the data\n",
      "    X_train, X_test, y_train, y_test = X[train], X[test], y[train], y[test]\n",
      "    #train the classifier\n",
      "    ridge = sklearn.linear_model.Ridge(alpha=1.0)\n",
      "    ridge.fit(X_train,y_train)\n",
      "    #test the classifier\n",
      "    ridgescores.append(ridge.score(X_test,y_test))\n",
      "    break"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 123
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "ridgescores"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 47,
       "text": [
        "[0.0084991392124383891]"
       ]
      }
     ],
     "prompt_number": 47
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Still not better than simple linear regression."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "plot(X[:,-2][:100],y[:100])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 49,
       "text": [
        "[<matplotlib.lines.Line2D at 0x3e47750>]"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD9CAYAAABHnDf0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
       "text": [
        "<matplotlib.figure.Figure at 0x42317d0>"
       ]
      }
     ],
     "prompt_number": 49
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import time"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 61
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "N = 100"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 125
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "XN = X[:,:][N-100:N]\n",
      "plot(XN/amax(XN,axis=0))\n",
      "plot(y[N-100:N]/max(abs(y[N-100:N])), label=\"target\")\n",
      "ypred = ridge.predict(X[N-100:N,:])\n",
      "plot(ypred/max(abs(ypred)))\n",
      "legend()\n",
      "N += 100"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAD9CAYAAAC1DKAUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
       "text": [
        "<matplotlib.figure.Figure at 0xf77d0e10>"
       ]
      }
     ],
     "prompt_number": 191
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "amax(XN,axis=0)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 159,
       "text": [
        "array([ 0.,  0.,  0.,  0.])"
       ]
      }
     ],
     "prompt_number": 159
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "len(amax(X,axis=0))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 156,
       "text": [
        "111"
       ]
      }
     ],
     "prompt_number": 156
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "So it looks like there is not much point to this approach.\n",
      "Would be as useful to simply fill with an average value."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print \"Average value of pulldown affinity score over pulldown data: {0}\".format(mean(y))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Average value of pulldown affinity score over pulldown data: -0.578366553592\n"
       ]
      }
     ],
     "prompt_number": 7
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import pickle"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 8
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "f = open(\"affinityresults/results2/affinity.pulldown.average.pickle\",\"wb\")\n",
      "pickle.dump([mean(y)],f)\n",
      "f.close()"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 9
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "# Checking training set coverage\n",
      "\n",
      "An alternative to trying to fill in the missing values prior to training is to train two models and compute the weighted average of their results.\n",
      "One of these models would be trained without the pulldown features over the whole training set.\n",
      "The other would be trained with the pulldown features over a subset of the training data.\n",
      "These models would then be tested with different weights for the averaging over the subset.\n",
      "\n",
      "As the dataset we are using is large this subset may still be large enough to produce viable results, especially if we use [leave-one-out cross validation][loo].\n",
      "\n",
      "Before we can try this though, we have to check what the coverage of these features over the training set is.\n",
      "\n",
      "[loo]: http://en.wikipedia.org/wiki/Cross-validation_(statistics)#Leave-one-out_cross-validation"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "assembler = ocbio.extract.FeatureVectorAssembler(\"datasource.affinity.tab\", verbose=True)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Using  from top data directory datasource.affinity.tab.\n",
        "Reading data source table:\n",
        "\tData source: affinityresults/results2/unique_data_ppi_coor_C2S_entrez.csv to be processed to affinityresults/results2/affinity.Entrez.db\n",
        "Initialising parsers.\n",
        "Database affinityresults/results2/affinity.Entrez.db last updated 2014-06-25 12:06:47\n",
        "Finished Initialisation."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n"
       ]
      }
     ],
     "prompt_number": 18
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "## DIP\n",
      "\n",
      "By generating vectors we can see what the coverage is over the DIP training set:"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "assembler.assemble(\"DIP/human/training.nolabel.positive.Entrez.txt\",\n",
      "                   \"features/training.DIP.positive.affinity.txt\", verbose=True)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Reading pairfile: DIP/human/training.nolabel.positive.Entrez.txt\n",
        "Checking feature sizes:\n",
        "\t Data source affinityresults/results2/unique_data_ppi_coor_C2S_entrez.csv produces features of size 1.\n",
        "Writing feature vectors\n"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Wrote 4857 vectors.\n",
        "Matched 3.71 % of protein pairs in DIP/human/training.nolabel.positive.Entrez.txt to features from affinityresults/results2/unique_data_ppi_coor_C2S_entrez.csv\n"
       ]
      }
     ],
     "prompt_number": 8
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print \"Number of positive vectors available: {0}\".format(int(0.0371*4857))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Number of positive vectors available: 180\n"
       ]
      }
     ],
     "prompt_number": 9
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "assembler.assemble(\"DIP/human/training.nolabel.negative.Entrez.txt\",\n",
      "                   \"features/training.DIP.negative.affinity.txt\", verbose=True)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Reading pairfile: DIP/human/training.nolabel.negative.Entrez.txt\n",
        "Checking feature sizes:"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "\t Data source affinityresults/results2/unique_data_ppi_coor_C2S_entrez.csv produces features of size 1.\n",
        "Writing feature vectors."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".\n"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Wrote 3061800 vectors.\n",
        "Matched 2.28 % of protein pairs in DIP/human/training.nolabel.negative.Entrez.txt to features from affinityresults/results2/unique_data_ppi_coor_C2S_entrez.csv\n"
       ]
      }
     ],
     "prompt_number": 10
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print \"Number of negative vectors available: {0}\".format(int(0.0228*3061800))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Number of negative vectors available: 69809\n"
       ]
      }
     ],
     "prompt_number": 11
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "## HIPPIE\n",
      "\n",
      "We can do the same for the [HIPPIE training set][hippie]:\n",
      "\n",
      "[hippie]: http://nbviewer.ipython.org/github/ggray1729/opencast-bio/blob/master/notebooks/Training%20and%20test%20featureset%20generation.ipynb"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "assembler.assemble(\"HIPPIE/hippie_current.pairs.txt\",\n",
      "                   \"features/training.HIPPIE.positive.affinity.txt\", verbose=True)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Reading pairfile: HIPPIE/hippie_current.pairs.txt\n",
        "Checking feature sizes:"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "\n",
        "\t Data source affinityresults/results2/unique_data_ppi_coor_C2S_entrez.csv produces features of size 1.\n",
        "Writing feature vectors."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".."
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        ".\n"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Wrote 169626 vectors.\n",
        "Matched 8.96 % of protein pairs in HIPPIE/hippie_current.pairs.txt to features from affinityresults/results2/unique_data_ppi_coor_C2S_entrez.csv\n"
       ]
      }
     ],
     "prompt_number": 19
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print \"Number of positive HIPPIE interactions available: {0}\".format(int(0.0896*169626))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Number of positive HIPPIE interactions available: 15198\n"
       ]
      }
     ],
     "prompt_number": 20
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "So the above may be a viable strategy."
     ]
    }
   ],
   "metadata": {}
  }
 ]
}