Experiment_1.ipynb

{
 "metadata": {
  "name": "",
  "signature": "sha256:6e65051a068620b677905a7745ad85514cd03e43c8a33f414d569faded0971cc"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "heading",
     "level": 2,
     "metadata": {},
     "source": [
      "First Experiment"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Yi Bairen, 5th January 2015\n",
      "\n",
      "**Updated** 8th January 2015:\n",
      "*[Cross Validation on Mixture of Testing and Noise Data](#Cross-Validation-on-Mixture-of-Testing-and-Noise-Data)*."
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Introduction"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "In this experiment, a human plasma dataset is used, which comes from a search against NIST library by SpectraST, with 0.9 as threshold. All results are kept and no consensus procedure is conducted.\n",
      "\n",
      "The original data is stored in ``/data/yhu/splibs/MicroProt2/test.splib``.\n",
      "\n",
      "Special thanks to Yingwei Hu (yhu) for providing the data.\n",
      "\n",
      "---"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Data Preparation"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import numpy as np\n",
      "from scipy import sparse as sp"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 1
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "First we load the data from cache."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# numpy.ndarray\n",
      "names = np.load('test.names.npy')\n",
      "# Scipy.sparse.csr_matrix\n",
      "spectra = np.load('test.spectra.npy')[()].astype(np.double)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 2
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "This is the content of the array we loaded."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print 'This is an example of the name of the peptide:', names[0]\n",
      "print 'This is the shape of one single spectrum array:', spectra[0].todense().shape\n",
      "print 'The spectra is stored in sparse format:'\n",
      "print 'Sparseness: {:.2f}'.format(float(spectra.data.shape[0]) / (spectra.shape[0] * spectra.shape[1]))\n",
      "print repr(spectra)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "This is an example of the name of the peptide: DEM[147]PSPTFLTQVK/2\n",
        "This is the shape of one single spectrum array: (1, 2000)\n",
        "The spectra is stored in sparse format:\n",
        "Sparseness: 0.15\n",
        "<510003x2000 sparse matrix of type '<type 'numpy.float64'>'\n",
        "\twith 150753160 stored elements in Compressed Sparse Row format>\n"
       ]
      }
     ],
     "prompt_number": 3
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Note that for quick validation of the idea, information about the precursor is completely ignored. This experiment is not intended to serve as a searching software either.\n",
      "\n",
      "---"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Visualizing Spectra"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Let's inspect some spectra to make sure we did it right."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from matplotlib import pyplot as plt\n",
      "\n",
      "fig, axes = plt.subplots(nrows=1, ncols=4, sharex=True, sharey=True, figsize=(20, 4))\n",
      "for i, ax in enumerate(axes):\n",
      "    ax.annotate(names[i], (1200, 12000), fontsize=12)\n",
      "    ax.plot(spectra[i, :].T.todense())"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAABOYAAAECCAYAAABaEdgZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe1cb4b3d10>"
       ]
      }
     ],
     "prompt_number": 4
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "This looks pretty well. \n",
      "\n",
      "---"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Basic Investigation of the Dataset"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Next we would like to know more about the replicates, and the most important question: \n",
      "\n",
      "How many distinct peptides are there?"
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "_, counts = np.unique(names, return_counts=True)\n",
      "plt.plot(np.sort(counts)[::-1], figure=figure(figsize=(8, 4)))\n",
      "plt.title('Distribution of number of replicates for each peptide')"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 5,
       "text": [
        "<matplotlib.text.Text at 0x7fe1a4a4fd90>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAfoAAAENCAYAAADngqfoAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ncTvIYvoAAAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe1cb4b3610>"
       ]
      }
     ],
     "prompt_number": 5
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Here we can see that there are 6,000+ distinct peptides, and the number of replicates are not uniformly distributed. \n",
      "\n",
      "---"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Masking Original Data"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "To make sure the model is fed with enough training data, we mask the spectra so that rare peptide is excluded from training process."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "MASK_CONDITION = lambda counts: counts > 100\n",
      "\n",
      "_, encoded, counts = np.unique(names, return_inverse=True, return_counts=True)\n",
      "mask = np.in1d(encoded, np.where(MASK_CONDITION(counts)))\n",
      "\n",
      "print '\\t'.join(['Count', 'Mask', 'Peptide Name'])\n",
      "print\n",
      "for _ in np.c_[counts[encoded[:10]], mask[:10], names[:10]]:\n",
      "    print '\\t'.join(_)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Count\tMask\tPeptide Name\n",
        "\n",
        "332\tTrue\tDEM[147]PSPTFLTQVK/2\n",
        "3438\tTrue\tESLSSYWESAK/2\n",
        "332\tTrue\tDEM[147]PSPTFLTQVK/2\n",
        "1485\tTrue\tTQQPQQDEM[147]PSPTFLTQVK/3\n",
        "1485\tTrue\tTQQPQQDEM[147]PSPTFLTQVK/3\n",
        "1485\tTrue\tTQQPQQDEM[147]PSPTFLTQVK/3\n",
        "1485\tTrue\tTQQPQQDEM[147]PSPTFLTQVK/3\n",
        "2106\tTrue\tTQQPQQDEMPSPTFLTQVK/3\n",
        "3\tFalse\tGGGFSSGGFSGGSFSR/2\n",
        "8\tFalse\tSSSKGSLGGGFSSGGFSGGSFSR/3\n"
       ]
      }
     ],
     "prompt_number": 6
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "We then mask both the peptides and spectra."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.preprocessing import LabelEncoder\n",
      "\n",
      "le = LabelEncoder()\n",
      "\n",
      "X = spectra[mask]\n",
      "y = le.fit_transform(names[mask])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 7
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print 'Number of noise peptides: {}'.format(sum(~MASK_CONDITION(counts)))\n",
      "print 'Number of noise spectra: {}'.format(sum(~mask))\n",
      "print\n",
      "print 'Number of masked peptide classes: {}'.format(np.max(y))\n",
      "print 'Number of masked spectra: {}'.format(y.shape[0])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "Number of noise peptides: 5467\n",
        "Number of noise spectra: 70549\n",
        "\n",
        "Number of masked peptide classes: 702\n",
        "Number of masked spectra: 439454\n"
       ]
      }
     ],
     "prompt_number": 8
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "print 'This is the final data we use to build our model:'\n",
      "\n",
      "print 'X (spectra) with shape:', X.shape\n",
      "print 'y (encoded peptide names) with shape:', y.shape"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "This is the final data we use to build our model:\n",
        "X (spectra) with shape: (439454, 2000)\n",
        "y (encoded peptide names) with shape: (439454,)\n"
       ]
      }
     ],
     "prompt_number": 9
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "---"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Stratified Split of Training and Testing Data"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "To make sure our model generalizes, cross-validation must be done. \n",
      "\n",
      "Here, 10% data is split as testing data."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.cross_validation import StratifiedShuffleSplit\n",
      "\n",
      "for train_index, test_index in StratifiedShuffleSplit(y, n_iter=1, test_size=0.1):\n",
      "    X_train, y_train = X[train_index], y[train_index]\n",
      "    X_test, y_test = X[test_index], y[test_index]"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 10
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "---"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Model Training"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Instead of canonical Logistic Regression, [Stochastic gradient descent (SGD)](http://scikit-learn.org/stable/modules/sgd.html) is used in favor of speed. \n",
      "\n",
      "SGD could easily be parallelized. Here 30 CPU cores is used for our training process.\n",
      "\n",
      "Note that scaling for each m/z is performed in advance to training, as SGD is sensitive to feature scaling."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "from sklearn.preprocessing import StandardScaler\n",
      "from sklearn.preprocessing import Normalizer\n",
      "from sklearn.linear_model import SGDClassifier\n",
      "from sklearn.pipeline import Pipeline\n",
      "\n",
      "scaler = StandardScaler(with_mean=False, copy=True)\n",
      "sgd = SGDClassifier(loss='log', shuffle=True, warm_start=False,\n",
      "                    alpha=1e-6, class_weight='auto',\n",
      "                    penalty='l2', l1_ratio=0,\n",
      "                    n_jobs=30)\n",
      "clf = Pipeline([\n",
      "    ('normalize', Normalizer()),\n",
      "    ('sgd', sgd)\n",
      "])"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 11
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "The training process takes roughtly 10 minutes. \n",
      "\n",
      "Set ``sgd.verbose`` larger than 0 to monitor the process."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "clf.fit(X_train, y_train)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 12,
       "text": [
        "Pipeline(steps=[('normalize', Normalizer(copy=True, norm='l2')), ('sgd', SGDClassifier(alpha=1e-06, class_weight='auto', epsilon=0.1, eta0=0.0,\n",
        "       fit_intercept=True, l1_ratio=0, learning_rate='optimal', loss='log',\n",
        "       n_iter=5, n_jobs=30, penalty='l2', power_t=0.5, random_state=None,\n",
        "       shuffle=True, verbose=0, warm_start=False))])"
       ]
      }
     ],
     "prompt_number": 12
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "coef = sp.csr_matrix(sgd.coef_)\n",
      "\n",
      "print 'The model is essentially a set of coefficients, one for each peptide.'\n",
      "\n",
      "print 'Sparseness: {:.2f}'.format(float(coef.data.shape[0]) / (coef.shape[0] * coef.shape[1]))\n",
      "print repr(coef)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "The model is essentially a set of coefficients, one for each peptide.\n",
        "Sparseness: 0.75\n",
        "<703x2000 sparse matrix of type '<type 'numpy.float64'>'\n",
        "\twith 1055203 stored elements in Compressed Sparse Row format>\n"
       ]
      }
     ],
     "prompt_number": 13
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Let's inspect some of the coefficient of our model.\n",
      "\n",
      "It seems that to increase the discriminative power, most of the coefficients except few are biased towards negative."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "fig, axes = plt.subplots(nrows=2, ncols=3, sharex=True, sharey=False, figsize=(20, 8))\n",
      "\n",
      "for i in range(3):\n",
      "    index, = np.where(names == le.inverse_transform(i))\n",
      "    \n",
      "    axes[0,i].plot(coef[i, :].T.todense())\n",
      "    axes[0,i].set_title('Coefficients for {}'.format(names[index[0]]))\n",
      "    \n",
      "    \n",
      "    axes[1,i].plot(spectra[index[:10]].T.todense(), alpha=0.5)\n",
      "    axes[1,i].set_title('Example spectra of {}'.format(names[index[0]]))"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAABJcAAAHsCAYAAABv1ExNAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIZbs/wEsVkaDPwaIOwAAAABJRU5ErkJggg==\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe19a625f10>"
       ]
      }
     ],
     "prompt_number": 14
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "---"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Cross Validation on Testing Data"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "Next, we would like to cross validate our model by 10% test data."
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "``predict_proba`` output the full probability distribution, and then we can calculate the entropy.\n",
      "\n",
      "``predict`` output a single peptide class.\n",
      "\n",
      "The final probability is calculated based on two schemes: one is $e^\\text{-entropy}$, another is the top hit probability output by the model."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "y_pred = clf.predict_proba(X_test)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 15
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "import scipy.stats\n",
      "from sklearn.metrics import roc_curve, roc_auc_score\n",
      "\n",
      "entropy = scipy.stats.entropy(y_pred.T)\n",
      "true_discovery = (clf.predict(X_test) == y_test)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 16
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "fig, axes = plt.subplots(nrows=1, ncols=2, sharex=True, sharey=True, figsize=(12, 4))\n",
      "\n",
      "probs = {\n",
      "    'Probability': y_pred.max(axis=-1),\n",
      "    'Entropy score': np.exp(-entropy)\n",
      "}\n",
      "\n",
      "fig.text(0.5, -0.04, 'ROC for testing data', ha='center', va='center', fontsize=16)\n",
      "for ax, (scheme, prob) in zip(axes, probs.items()):\n",
      "    \n",
      "    fdr, tdr, _ = roc_curve(true_discovery, prob)\n",
      "    \n",
      "    ax.plot(fdr, tdr)\n",
      "    ax.set_xlabel('FDR')\n",
      "    ax.set_ylabel('TDR')\n",
      "    ax.set_title(scheme)\n",
      "    ax.annotate('Area: {:.4f}'.format(roc_auc_score(true_discovery, prob)), (0.25, 0.5), fontsize=16)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAtwAAAE0CAYAAAAfVuQbAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3XmYFNXZ9/HvDTgioqJoABGQAK7wKm6IIJpgolGzoI9H\njRI0iXtiiIoSRUWDgnEJWQyPMaiPiUSPETeMEheijkuCokYxalwwEAERlH1kmXr/qJpJT9OzNb2f\n3+e66hqq6nT1uXuK0/ecvqvaoihCRERERETyo02xOyAiIiIiUsmUcIuIiIiI5JESbhERERGRPFLC\nLSIiIiKSR0q4RURERETySAm3iIiIiEgeKeEWERGRgjOzv5rZlZt5jPFmNqsVbT/IsH20mdUmyyb7\n09oenrTrmW2fJUxKuGWzmdkdKYNV6vKzLI41b3MHYBER2Txp4/rnZvaumV1rZlvl8GmiZCmU64ED\nMmy/BegKjG7BMZ5L2i7ItLOxpF6kXbE7IBXjeeC4tG1rsjiOvolJRKQ01I3rVcDhwG+BbYEfFLFP\nWYuiaDWwOsP2tcBaM1vRgmOsBz7OQ/ekwmmGW3JlXRRFH6ctq+p2JrMkZ5hZtZmtNrO/mdnuKfvn\nmVkt0Au4MmVm5YrUJ0lmXWaZ2Slm9p6ZrTWzf5iZJfv3MrMnzGyNmS01s6lm1jHl8ePN7AMzG5fs\n/8TMrk3Z/x0zW2FmHdKed5aZ/bY1L4iZXW5mHyZ9nGdm12Vo800z+3vSZqGZ3WZmW6Ttf93MapJj\nXZLhGM29Jt3N7B4zW5bEe5+ZdW9NLCISpPXJWL4giqI/AH8ARtTtTMa1q81sopl9nIztv03Z3+z4\nBXzBzB5Nxq13zex/Uh5flYyJ7yXHWGBmN5hZVdox2pjZL81sedKPn6buNLOxLS0ZaYyZHZz2CW7P\ntP3jk/ewK4BeKe3eT/ZbEkf6e9phZrZBY3LlU8ItuWItaDMa+AlwMNAR+HnKvv2BbsQf091A/JFd\nV+DGtGNEQD/ge8CpwP8DpgKYWTvgIeBz4EDimZkvA79IO8YuwH7AEOBM4Idm9t1k371ALXB8fWBm\nvYBhdc/TEmY2gnjg/TGwG/Bt4KO0Nl8HpgOPAQOBY4FPgPbJ/n7An5KY+gPjgPFmNrIVr8nWwF+J\nfz/DgKHJYx6qS8hFRBqR/onj58AWaftHAdsRj7UHAy9Ci8cvA84AHgEGECf0f0weC/HM+ufA6cTj\n6GnAKcD4tH4dAmxFPO7/ELjAzE5P2f9L/vt+ku2nqC8lxzi+kf3XpzzHAv77HnYgQBRFEfG4/J20\nx40CHo+i6D9Z9kvKRRRFWrRs1gLcAWwAVqYt30hpUwucnbJ+AfBJhmN9AFzRzHOtAXbKsO+Y5Hl2\nTtk2ElgPdEzWxwMbga4pbaYAf09bfzJlfRwwt5WvyYXAp0BVE22eB+5rYv/1wFtp26YCf2vFa/J9\nYBmwRcq2rsnrdFCxzx0tWrSU5pKMK7OSfxtxUvspMC2lzTzgpUYe3+z4RTwZ8Fham7eAnzXRrxuB\n11LWxyf92jJl2y2pY3pa2w+aOPZpTe1P2hyejJ89G9nf6HMAOyfvR4cm61sBy4ETiv371pL/RTPc\nkiuzgX3SlifS2ryT8u9lwA5ZPtfbURQtybC9H7A0iqLUmeRXgbZA79TnjqJoUcr6XKBvyvpU4HAz\n65GsjwRub2Uf7yOuFXzXzG63uJxmm7Q2+wBPN3GMfsA/0ra9mmxP19hrsh9xzeUyM1tpZiuBfxHP\n8nyxBXGISLgOTcaMtcRj1V9oWL8dAc828tiWjF9Rhjavp7ZJxs6/m9nipC8/ALZOe8y7URR93tgx\nSkXy3jSTeFYb4vKcDcADReuUFIwumpRcqYmi6P1m2mzI0XN9upmPz/SRYn15RRRFL5nZG8B3zOxJ\n4mT9zlY9QRTNM7O+xLMhhxLPelxgZvumvTE018+Wln009Zq8BXwjw3Zd+CMiTZlNXMKxDvgoiqLa\nDG0aG3taOn6ltzHiGWTM7ATgJuJyxBeJE/+LgSNbcNxSvQB/KnCHmf2QuLzkrii+EFMqnGa4JVdy\nNbito2GNYCaNDeLvAp3TLj4ZSFxCknqhTGcz65qy3p941jfV74hnIUYCf46iqNXJaRRFNVEUPRZF\n0WXAt4Ddgb1SmrxKnJA35l3ieuxUAzP0FRp/TeYAuwLLoyh6P21Z1chjREQA1iZjxYJGku2mtGT8\nsgxtBgBvJ/8eBsyMomhqFEVzk0md3mz6ftPPzLZMWf9/xBMNrbUK6NBsq6ato+nJzBnEfzj8ABgO\n3LaZzydlQgm35MqWZtbFzLqmLJ2yOM47wFeSO2u0N7O2rXjsY8SJ9a1m1t/MDgOuAn6fllzWAr8x\nsz3M7HjiGZxb0o71B6AHcQ10a8tJ6u52ckbSj77EFzSuBN5LaTYR+KaZXWVme5rZvmZ2fUrpya3A\nFy2+9+1uZvYd4GTg5lZ0ZRqwCJhuZkPMrI+ZHZvctSSX99MVkcrT3Ax1U/tbOn4dbmbnmlk/M7uK\neIKg7k4nbwFDzWxYsv8KYHCG590a+JWZ7W5mJxKP6fXPU/eeRHyxftuU96r2aceZQzwhc2r6e5iZ\n7ZAco64U8gtJm23TjvE20M3MvmxmHdL+ECCZzb4TuAZ4I4qiVxt/CaWSKOGWXIiIB8GFxHfiqFua\nu41eplnxS4kT4reJLwS8LMNjMs6mR1G0gbh0Ykvg78R1cX8FfpTWdAHwGvBC0sdfkzbLEEXRp8RX\n1y8nnpForWXEF+BUEw/i/YFjoyiqv89rFEUziGv4vpa0eRzoAtQk+98BXBLT68AE4KdRFKWXtzT1\nmqwhniWaT/x6vEH8Ee1C4pkYEZFMWvKlNI3ub+H4FRGPwd8gruU+GXBRFM1L9t9CPG5NJy5v6UF8\nx5Eo7RjPEY+bs5P910VRdFdKm7r3pAuA7vz3vcql9fl94HzgZ8n+1PeF6cm2PyXP+fdkfXJa6PcT\nX3B6N/GM+T8zvDy3E8+C35Fhn1Qoi6LClDk5524gvmXZEu/9gGbaOuL/nBFwofc+m4RHZBNmNh4Y\nFUVR7xa0rQZmR1H047x3TEREgmBmRwB/Jr6j1ifF7o8URiFnuO8jvm1bk5xzVcAk4nskH8Gmfz2K\n5JWZ7WRmo4hn7Vv1ZTciIiKZmNlWZvZF4gnF+5Rsh6VgCbf3/gVgaQuaDgLmeu+XeO/nA/Odc/vk\nt3cSkJZ8TLqY+Mt3fhhFUaaPA0VERFrrEuLrlD4nvvOKBKQUbwvYBVjonDuLuA52EfE3EL5W1F5J\nRYii6CriCymbaqNrG0REJKeiKBrPpt+SKYEoxYQbAO/9LQDOueMo3ftpioiIiIg0qRQT7oXEM9p1\nuibbMnryySeVjItI2Ro+fHhLv9yoImjMFpFyl824XfSE2zk3EYi895cmm2YDezvndgLaA7t479O/\n+rWB/fbbL8+9LD1TZ3/E7PnL6dkp/TaiDUURrN1QS9/Ozd9yOQL6de5Aj05bbrLP0m572qGqDTtu\nXdWqPotIQ3PmzCl2F4qi1MfsD5at5bq/zqO2BX8azPu0huuO7suu2zc9FgNsXdWWqraqWBMpZ9mO\n2wVLuJ1zNxPfc3hH59x84Jzkdn9dSSkZ8d6vc86NJb6vJgR0YcHG2ohT757L6nUbm263cSPrI+MH\nh+zCN/baqUC9K67q6mqGDh1a7G4UVGgxhxavlK4lq9ex1RZtOX9Ij2bbtjWjR6ctMWt8wivEc1sx\nhyHEmLNVsITbe38ecF6G7adn2OYBX4h+lZKNUcTymg3cN7LJ25TzwgsvMHjwYNq300yJiEhLvbxg\nBR9+VtNsu1f+s5It2hq9d9CXsYpIbhS9pCQEz3/4Gfe9voQt2jZd8rNhY8SG2oittmj628y/PCy8\nvyZD/As6tJhDi1cK7/9eXkiXjlXs0GGLJtt123ZL+nfZOmfPG+K5rZjDEGLM2VLCnaXaKGLUPW82\nW/4BsGrdRobuuh1H77Fjs23P2HLnXHRPREQyOG7AF9jzC7lLpkVEWkIJd5aiKK7zu/fUpss/6nSs\nattkjV9rhFgzpZgrX2jxSm5srI14d+maFl3guHDluvx3KIMQz23FHIYQY86WEu7NtM2WeglFRIrl\nHwtXcdUT79OjmTs2AXTusAWdmyknERHJB2WLZSjEvyYVc+ULLV7JjY1RxJ5f2JqJX+tb7K40KsRz\nWzGHIcSYs6WEW0RESs6Vf3mfv81f3my72ggO6bVdAXokIpI9JdxlKMSaKcVc+UKLV5q2vGYDPzu6\nH3u34G4hbUr8uzpDPLcVcxhCjDlbSrhFRKQktTVoW+rZtIhIC+ibU8pQiH9NKubKF1q8Eo4Qz23F\nHIYQY86WEm4RERERkTxSwl2Gqquri92FglPMlS+0eCUcIZ7bijkMIcacLdVwp1m3oZaPVzf/5Qgb\na6MWfdGCiIiIiIQtmIT76fc/ZeXnzX8N+5/f+oR3l66l+7ZbNtkuImLX7Zv/ooV8CLFmSjFXvtDi\nlXCEeG4r5jCEGHO2gki4a6OIa56ax9F7dG62bb8dO/DDIT3Y8wvN34pKRERa7+NV66iN4k8Ja6OI\n2lqoJfmZbF+7vvkJEhGRchFEwg1gwOihPYvdjZwI8b6XirnyhRZvyEY//A5tzWhjYMnPtmaYQZtk\nvf0Wbei8dWV8DXuI57ZiDkOIMWcrmIRbRERKw7ST+xe7CyIiBaW7lJShEP+aVMyVL7R4JRwhntuK\nOQwhxpwtJdwiIiIiInlUEQn3g3OXNLtU0h38QrzvpWKufKHFK+EI8dxWzGEIMeZsVUQN9/zlNU3u\njyI4dWDXAvVGREREROS/LIrKe+73ySefjPbbb79id0NEpNXmzJnD8OHDrdj9KCSN2SJSzrIdtyui\npEREREREpFQp4S5DIdZMKebKF1q8Eo4Qz23FHIYQY86WEm4RERERkTxSDbeISJGohltEpLyohltE\nREREpAQp4S5DIdZMKebKF1q8Eo4Qz23FHIYQY86WEm4RERERkTxSDbeISJGohltEpLyohltERERE\npAQp4S5DIdZMKebKF1q8Eo4Qz23FHIYQY86WEm4RERERkTxSDbeISJGohltEpLyohltEREREpAQp\n4S5DIdZMKebKF1q8Eo4Qz23FHIYQY86WEm4RERERkTwqWA23c84BE4AIuNB7P6OJtlcCLlm9x3t/\ndWNtVQ8oIuVKNdwiIuWlpGu4nXNVwCRgCHAEMLmJtr2BkcAAYF9glHOuVyH6KSIiIiKSa4UqKRkE\nzPXeL/HezwfmO+f2aaTtCmA9sFWyrAOWF6ab5SHEminFXPlCi1fCEeK5rZjDEGLM2WpXoOfpAix0\nzp0FLAMWAd2A19Ibeu+XOud+Acwn/oPgQu/9ZwXqp4iIiIhIThX0oknv/S3e+3uT1YzF4865XYGz\ngV5AH2CMc65rU8dN/Qururq64tfTYy92fwqxPnTo0JLqTyHWhw4dWlL9Uby5Xw9VKf0OCrGeHnux\n+1OIdY3Zxe9PIdbrtpVKfwr9/7k1CnLRpHNuCDDWe//1ZH0W8CPv/T8ytD0ROMJ7f0ay/kfgTu/9\no5mOrQtwRKRc6aJJEZHyUtIXTQKzgb2dczs553oAu9Ql2865ic65a1Pavgcc6Jyrcs5tBewHvF+g\nfpaFzfkLq1wp5soXWrwSjhDPbcUchhBjzlZBEm7v/TpgLPAc8CQwOmV312Spa/sScD/wCvAScKv3\n/u1C9FNEREREJNcKdh/ufNHHkyJSrlRSIiJSXkq9pEREREREJEhKuMtQiDVTirnyhRavhCPEc1sx\nhyHEmLOlhFtEREREJI9Uwy0iUiSq4RYRKS+q4RYRERERKUFKuMtQiDVTirnyhRavhCPEc1sxhyHE\nmLOlhFtEREREJI9Uwy0iUiSq4RYRKS+q4RYRERERKUFKuMtQiDVTirnyhRavhCPEc1sxhyHEmLOl\nhFtEREREJI9Uwy0iUiSq4RYRKS+q4RYRERERKUFKuMtQiDVTirnyhRavhCPEc1sxhyHEmLOlhFtE\nREREJI9Uwy0iUiSq4RYRKS+q4RYRERERKUFKuMtQiDVTirnyhRavhCPEc1sxhyHEmLOlhFtERERE\nJI9Uwy0iUiSq4RYRKS+q4RYRERERKUFKuMtQiDVTirnyhRavhCPEc1sxhyHEmLOlhFtEREREJI9U\nwy0iUiSq4RYRKS+q4RYRERERKUFKuMtQiDVTirnyhRavhCPEc1sxhyHEmLOlhFtEREREJI9Uwy0i\nUiSq4RYRKS+q4RYRERERKUFKuMtQiDVTirnyhRavhCPEc1sxhyHEmLOlhFtEREREJI9Uwy0iUiSq\n4RYRKS+q4RYRERERKUFKuMtQiDVTirnyhRavhCPEc1sxhyHEmLOlhFtEREREJI8KVsPtnHPABCAC\nLvTez2ii7SDgVqAd8Lr3/sTG2qoeUETKlWq4RUTKS0nXcDvnqoBJwBDgCGByE23bAHcCZ3vv9wLO\nLUQfRURERETyoVAlJYOAud77Jd77+cB859w+jbTdH1jivX8ewHu/tEB9LBsh1kwp5soXWrwSjhDP\nbcUchhBjzla7Aj1PF2Chc+4sYBmwCOgGvJahbU9guXPu0eRxt3rvpxSonyIiIiIiOVXQiya997d4\n7+9NVhsrHm9PXHpyBnAYMNo517sQ/SsXQ4cOLXYXCk4xV77Q4pVwhHhuK+YwhBhztgqVcC8kntGu\n0zXZlski4E3v/QLv/UrgZWCPpg6e+pFGdXW11rWuda2XxXqoSul3oHWta13rrVnPVkHuUpJcNPkW\ncS13e+Ap732/ZN9EIPLeX5qsbwfMBQYAq4kT7uO99+9kOnaIV7xXV1cH91elYq58ocULuktJKEI8\ntxVzGEKMuaTvUuK9XweMBZ4DngRGp+zumix1bZcn+58C5gDTGku2RURERERKXcHuw50vIc6WiEhl\n0Ay3iEh5KekZbhERERGRUCnhLkObU7RfrhRz5QstXglHiOe2Yg5DiDFnSwm3iIiIiEgeqYZbRKRI\nVMMtIlJeVMMtIiIiIlKClHCXoRBrphRz5QstXglHiOe2Yg5DiDFnSwm3iIiIiEgeKeEuQ7n6VqfH\nH3+czp07s9tuu1HqtfzZxvzggw8yePBgdt55Zw4++GDuu+++rPtw7733csghh9CtWzf22msvxo4d\ny5o1axq0Wb58OWPHjmXfffdll1124eCDD+Y3v/lNgzbnnXcenTt3zriccMIJ9e2GDh3K4sWLOe+8\n89hjjz3o3r07Bx98MPfee2/WMZSy0L6tTMKhMbvlCj1mr1q1iosuuog99tiDnXfemWHDhjF9+vQG\nbV555RXOOuss9t13X7p3786BBx7I+PHjWbVqVYN21dXVGcf1m2++OesYSp3G7ZZrV+wOSPHMnDmT\n7bbbjmXLljF79mwOOuigYncpp5599lm+973vMWrUKG644QYefPBBzjrrLLbffnu+/OUvt+pYTzzx\nBGeffTannnoq1157LXPnzuXaa69l0aJF3HHHHfXtzj33XJ577jnGjx9Pv379ePrpp7n88svZsGED\n559/PgBjxozhu9/9boPjf/jhh5xxxhkcddRR9dtWrFjB0UcfjZlx9dVXs8suu/D++++zYcOG7F8U\nESlbGrNbrqVj9ujRo3niiScYP348PXv25Pe//z1nnnkmW2+9NUceeSQADz/8MAsWLOCCCy6gb9++\nvPnmm0yYMIEXX3yRRx99FLOG18/95S9/abC+yy67ZPeCSEXRXUrKUHV1dU7+qtxnn3049thj8d4z\natQoxo0bl4Pe5Uc2MX/rW99i+fLlzJo1q37bEUccQVVVFX/+859bdaxjjz2WlStX8vTTT9dvu/76\n65k0aRLPP/88u+++O6tXr6ZXr178+Mc/5rLLLmvQj08//bTBY9NNmjSJyZMn89Zbb9GpUycgTt4f\neughXnzxxSAG7Fyd1+VEdykJg8bslin0mD1v3jz2339/JkyYwDnnnANAbW0tgwYNolOnTjz++OMA\nLFu2jB122KHB8X/zm99w+eWX8/DDD3PIIYcA8Zh9zz33sHTp0lb1tZxp3G65rEtKnHPbO+cua76l\nlKK5c+eyYMECDj/8cIYMGcLMmTMztps0aRKdO3fmo48+4qSTTqJnz5706dOHiRMn1rdZsWIFF198\nMXvvvTfdunXjsMMOqx+o6syZM4dRo0YxYMAAunXrRv/+/fnJT36yyUdydRYvXsw777zDihUrsoqv\npqaGF154gaOPPrrB9mOOOYbZs2c3+ryNeeuttzaZTRo2bBhA/YBe98drx44dG7RLX08XRRF33303\nX/3qV+uTbYgHsqOOOiqIZFvKm94P8k9jdu7G7GeeeQaAt99+G4BBgwbVt2nTpg1DhgzhlVdeqY8l\nPdkG2G233YA4bpGWaDLhds5t5Zy7xjk3wzl3SbKti3Pu58C/gaOaerzkRy7+mpw5cybt2rVj8ODB\nHHroobz55pssWLCg0fajRo1ir7324s477+S6666jXbu4GmndunWMGDGC+++/n4suuog//vGPDBgw\ngFNOOYU33nij/vHvvfcevXr14pprrmH69OlcfvnlzJgxg3PPPTfj81199dUMHjyYRx55JKuY582b\nx4YNG+jTpw9A/cDZp08famtrmTdvXquOt379+vqY61RVVQHw/vvvA3FiPXLkSO68807+8Y9/sGrV\nKh577DFmzZpVP3uSybPPPsv8+fM56aST6rfV1NSwcOFC+vbty+jRo+nZsyc9e/bktNNOY8mSJa3q\ne7kIbZak3Oj9IHsas5tXyDH7vffeA+LXAmCLLbbYpF0URU0+50svvQTA3nvvXb+tZ8+eAOy5557s\ntNNOHHDAAUyZMqVV/S43Grdbrrka7l8Cg4DHgJOcc52AM4CZwGHe+zl57p/kycyZM9lnn33o2LEj\nhx56KBDXnaXXFtc58sgjueiiizbZfs899/Dqq6/ivWf48OEAHH744bzxxhvceOON3H777QCccMIJ\n9RcERlHEgQceyCeffMKVV17JypUr2WabbTY5tpltUhvXUp999hkA22yzDQ888ADf+973uOaaa9hj\njz0a7G+p3r178+abbzbY9sorrwCwcuXK+m033XQTP/rRj/jSl74EQLt27Zg0aVKDZDrdtGnT6Ny5\nM1/96lcb9L+2tpbf//739W+aixcv5oorruDMM8/k/vvvb1X/RXJA7wdFpDE792N27969gfjTgwED\nBtS3mzMnPpUbm61fsGABU6ZM4Zhjjqmf6Qbo27cv1157Lf3792fNmjVMnz6dcePG8emnn3LppZe2\nqv9SeZorKTkW+Kr3/mLga8AlwEjv/SkaXItnc+97uXTpUubMmVM/aO+222507dq10Y8oAU488cSM\n259++mk6derEYYcdxoYNG+qXgw46iFdffbW+3YoVK7jyyisZOHAg3bp1o0uXLlxxxRX1/Ul38803\n88knn9QnqpsTc8eOHenYsWPGN4iWGjlyJNXV1dxxxx189tlnPPfcc0yePJmtttqqQbuLL76YmTNn\n8otf/IIZM2Zw0UUXcemllzJt2rSMx125ciUzZsxgxIgRDWZjamtrAfj888+58847OfzwwznxxBO5\n7LLLeOaZZ5g7d27WsZQq3c+15On9IEsas1unUGN2//79GThwIDfccAOvvvoqy5Yt46abbuJf//oX\nQMY/Hmpqajj99NPp1KkTkydPbrCva9eunHXWWQwZMoSvfOUrTJkyhW9+85vcfPPNrF27NutYSpnG\n7ZZrLuHu6L1fBJD8XOm9fzT/3ZJ8evzxx6mtreWggw6ipqaGmpoaBg8eTHV1daODQvfu3TNuX7p0\nKZ999hldunRpsEydOpWPPvqovt0Pf/hDbr/9dr7//e/zwAMP8NRTT3HBBRcQRVFe7rqx3XbbAXFC\ne8QRR/Dhhx9yyimn1M9YpNZKt8Rpp53GyJEjGTNmDH369OGkk07ioosuoqqqih133BGAf/7zn9x2\n221MmDCBU089lcGDBzNmzBi+/e1vM27cuPokOtX999/P2rVrN3lz3HbbbYH4IqnUN4j9998fgHff\nfbdV/RfJAb0fFInG7PyM2RBf/Ni+fXuGDx9Ov379uPfeexkzZgwAnTt3bnDMKIo499xz+eCDD/De\nZ6ztTnfMMcdQU1NTXy8u4WqupKTKOXc1YI2sR977K/LWO8loc2um6mZFvv3tb2+y7+mnn25wa7o6\nbdpk/tusc+fO9OzZs8FtltLV1NTw6KOPMmbMGM4777z67Y899liL+9zamHv37k27du3qa/XqvPfe\ne7Rp04Zdd921Vcdr27YtkydP5uqrr+Y///kPPXr04LPPPuPHP/4x/fv3B+KLdCCu30u11157cccd\nd7Bo0SJ23nnnBvumTZtGv379SL9rQ8eOHdlpp50a7U+2H9uWMtUCljy9H2RJY3bzijFmQ/xpQXV1\nNf/+979Zt24dffv25Wc/+xkdOnSgb9++DY552WWX8fjjj3Pfffex++67b/KcmWKuxLE6lcbtlmsu\n4f4j0CNl/Z6UdQPK+56CAVq/fj2zZs3iy1/+MmPHjq3fXlNTw4gRI5g5c2bGwbsxw4YNY8aMGXTq\n1IlevXplbLNu3To2btzIlltuWb+ttraWBx54oNHBaPHixSxfvpyuXbvWz/a2Rvv27TnkkEPq3zTq\nPPLIIxx00EEZ7xyyYMEC1qxZQ48ePTYpFamz7bbb1vfnuuuua3Cv1rpk+o033mgwoL/55ptUVVXR\npUuXBsd69913mT17dqO39jr00EN56qmnWLNmDR06dABg9uzZABkHe5E80/tBEWjMzt+YnarugsfV\nq1dz1113ccwxxzQo8/vlL3/J1KlT+cMf/tCq+58/+uijbL311ptMxEh4mky4vfenFagf0gqbc9/L\nF154gZUrV3L88cfXlyfUGTx48CY37G/OSSedxO23386IESM4//zz6du3L5988gl/+9vfaNOmDddc\ncw3bbrstBx10EFOmTKF79+5ss8023HbbbaxevbrRb0u7+uqrufvuu/n1r3/NySefnFXMF1xwAccd\ndxwXXnj2ghNhAAAanklEQVQhI0aM4OGHH66/WCiTc845h+eff56HHnqIIUOGNNg3f/58pk2bxgEH\nHEC7du145JFHmDp1Ktddd139YH7AAQcwcOBArrzySmpqavjiF7/Iiy++yF133cVpp51G27ZtGxxz\n2rRpmBnOuYz9GTp0KA8//DAjR47k7LPPZvHixUyYMIEjjzyyIhPuEO/nWk70fpA9jdktU+gxG+Jv\no/z888/p3bs3S5Ys4Ve/+hVr165t8F0Kf/rTn7jqqqs444wz6NSpU/3EB8SlO3WTLfvvvz9nnnkm\nffv2ZcOGDTz00ENMnz6dq666qsEfL5VE43bLNftNk865rYHjgH2AbYEVwKvAdO/9mqYeK6Vn5syZ\ntG3btsEdMep87WtfY9y4cbz++uv1V2w3d9V5VVUVDz74IJMmTeLnP/85ixcvZocddmDgwIENrp6/\n9dZbGTNmDBdccAHt27fnhBNO4Nhjj2X06NGNHntzrniHeIb4d7/7HZMmTWLatGn06tWLW265pdFv\nLKt7vkzPWVVVRXV1NVOmTGH9+vXsueee/Pa3v+X444+vb9O2bVvuueceJk6cyOTJk/nkk0/o0aMH\nY8eObfCxLMDGjRu55557GDp0aKO1ln369OHuu+9mwoQJjBo1ig4dOnDUUUdx7bXXZv2aiGwOvR8U\nnsbs/I3ZEJfe/OpXv2L+/PlstdVWHHbYYfzud7+jR4//fpgza9YszIxbb72VW2+9tcHjL7nkEi6+\n+GIAunXrxs0338zSpUuJoog999yT//3f/62/24uErclvmnTO9QVmAeuA14BVQEfiwXYL4Eve+/ca\nPUABhPitZSJSGcrpmyZz9X6gMVtEylm243ZzM9zXA9O895ek73DOXZ/sP661TyoiImVH7wciIllq\n7raAhwGTGtk3CfhSbrsjLRHifS8Vc+ULLd4ypPeDLIV4bivmMIQYc7aam+HeEtjeObd9hn0GVOW+\nSyIiUoL0fiAikqXmEu6tAH3DRokJ8YpgxVz5Qou3DOn9IEshntuKOQwhxpyt5hLuN733/ZtpIyIi\nlU/vByIiWWquhjvzXfGlqEKsmVLMlS+0eMuQ3g+yFOK5rZjDEGLM2Wpuhtucc19sqoH3/v0c9kdE\nREqT3g9ERLLUXMLdgaZr9iKgbRP7JQ9CrJlSzJUvtHjLkN4PshTiua2YwxBizNlqLuFe7b3fpiA9\nERGRUqb3AxGRLDVXwy0lKMSaKcVc+UKLV8IR4rmtmMMQYszZai7h1ispIiKg9wMRkaxZFEXF7sNm\nefLJJ6P99tuv2N0QEWm1OXPmMHz4cCt2PwpJY7aIlLNsx22VlIiIiIiI5JES7jIUYs2UYq58ocUr\n4Qjx3FbMYQgx5mwp4RYRERERySPVcIuIFIlquEVEyku243Zz9+HOGeecAyYQfznChd77Gc203wZ4\nG7jRe39jAbooIiIiIpJzBSkpcc5VAZOAIcARwOQWPOwy4CXiBF1ShFgzpZgrX2jxSjhCPLcVcxhC\njDlbharhHgTM9d4v8d7PB+Y75/ZprLFzbndgJ+BlIKiPW0VERESkshSqpKQLsNA5dxawDFgEdANe\na6T9ROBHwHcL073yMnTo0GJ3oeAUc+ULLV4JR4jntmIOQ4gxZ6ugdynx3t/ivb83Wc1YKuKc+zrw\nTjITrtltERERESlrhUq4FxLPaNfpmmzL5CDgeOfcP4HzgIudcyc3dfDUGqLq6uqKX58yZUpJ9acQ\n63XbSqU/hVhPj73Y/VG8uV8PVSn9DgqxrjG7+P0pxHqIY9iUKVNKqj+FWM9WQW4LmFw0+RZxLXd7\n4Cnvfb9k30Qg8t5fmuFxVwIrvfc3NXbsEG8xVV1dHdzHOIq58oUWL+i2gKEI8dxWzGEIMeaS/mp3\n7/06YCzwHPAkMDpld9dkkRYK7eQGxRyC0OKVcIR4bivmMIQYc7baFeqJvPce8Bm2n97EY67Ka6dE\nRERERPJMX+1ehjanhqhcKebKF1q8Eo4Qz23FHIYQY86WEm4RERERkTwqyEWT+RTiBTgiUhl00aSI\nSHkp6YsmRURERERCpYS7DIVYM6WYK19o8Uo4Qjy3FXMYQow5W0q4RURERETySDXcIiJFohpuEZHy\nohpuEREREZESpIS7DIVYM6WYK19o8Uo4Qjy3FXMYQow5W0q4RURERETySDXcIiJFohpuEZHyohpu\nEREREZESpIS7DIVYM6WYK19o8Uo4Qjy3FXMYQow5W0q4RURERETySDXcIiJFohpuEZHyohpuERER\nEZESpIS7DIVYM6WYK19o8Uo4Qjy3FXMYQow5W0q4RURERETySDXcIiJFohpuEZHyohpuEREREZES\npIS7DIVYM6WYK19o8Uo4Qjy3FXMYQow5W0q4RURERETySDXcIiJFohpuEZHyohpuEREREZESpIS7\nDIVYM6WYK19o8Uo4Qjy3FXMYQow5W0q4RURERETySDXcIiJFohpuEZHyohpuEREREZESpIS7DIVY\nM6WYK19o8Uo4Qjy3FXMYQow5W0q4RURERETySDXcIiJFohpuEZHyohpuEREREZESpIS7DIVYM6WY\nK19o8Uo4Qjy3FXMYQow5W0q4RURERETySDXcIiJFohpuEZHyohpuEREREZESVLCE28Xecc697Zw7\ntol23Z1z1c65N5xzLzvnjihUH8tFiDVTirnyhRavhCPEc1sxhyHEmLNVkITbOVcFTAKGAEcAk5to\nvh44x3vfHxgB3JH3DoqIiIiI5ElBaridc4cCF3vvv56szwJGe+9fa8FjPwa6e+/XZ9qvekARKVeq\n4RYRKS/Zjtvt8tGZDLoAC51zZwHLgEVAN6DJhNs5dyTwcmPJtoiIiIhIqStUwg2A9/4WAOfccUCT\nU+vOua7ADcA3mjtudXU1Q4cOrf83UNHrr7/+Ouecc07J9KcQ63XbSqU/hVhPj73Y/VG8uV8Plcbs\n0upfPtbrtpVKfzSG5Wd9ypQpDBgwoGT6U4j1Dh06kI1ClZQMAcamlZT8yHv/j0batwceB37qvf9L\nU8cO8ePJ1DerUCjmyhdavKCSklCEeG4r5jCEGHO243ahEu4q4C1gENAeeMp73y/ZNxGIvPeXJusG\nTAOe8d5Pae7YIQ7eIlIZlHCLiJSXkq7h9t6vc86NBZ5LNo1O2d2VhuUlQ4DjgT2cc2cm277mvV+U\n/56KiIiIiORWQRJuAO+9B3yG7aenrVcDVYXqVzkK8SMcxVz5QotXwhHiua2YwxBizNnSN02KiIiI\niORRQWq480n1gCJSrlTDLSJSXrIdtzXDLSIiIiKSR0q4y1DqvT5DoZgrX2jxSjhCPLcVcxhCjDlb\nSrhFRERERPJINdwiIkWiGm4RkfKiGm4RERERkRKkhLsMhVgzpZgrX2jxSjhCPLcVcxhCjDlbSrhF\nRERERPJINdwiIkWiGm4RkfKiGm4RERERkRKkhLsMhVgzpZgrX2jxSjhCPLcVcxhCjDlbSrhFRERE\nRPJINdwiIkWiGm4RkfKiGm4RERERkRKkhLsMhVgzpZgrX2jxSjhCPLcVcxhCjDlbSrhFRERERPJI\nNdwiIkWiGm4RkfKiGm4RERERkRKkhLsMhVgzpZgrX2jxSjhCPLcVcxhCjDlbSrhFRERERPJINdwi\nIkWiGm4RkfKiGm4RERERkRKkhLsMhVgzpZgrX2jxSjhCPLcVcxhCjDlbSrhFRERERPJINdwiIkWi\nGm4RkfKiGm4RERERkRKkhLsMhVgzpZgrX2jxSjhCPLcVcxhCjDlbSrhFRERERPJINdwiIkWiGm4R\nkfKiGm4RERERkRKkhLsMhVgzpZgrX2jxSjhCPLcVcxhCjDlbSrhFRERERPJINdwiIkWiGm4RkfKi\nGm4RERERkRJUsITbxd5xzr3tnDs2V21DFGLNlGKufKHFK+EI8dxWzGEIMeZstSvEkzjnqoBJwCCg\nPTALmLG5bUVERERESl2hZrgHAXO990u89/OB+c65fXLQNkhDhw4tdhcKTjFXvtDilXCEeG4r5jCE\nGHO2CjLDDXQBFjrnzgKWAYuAbsBrm9lWRERERKSkFfSiSe/9Ld77e5PVJm+P0pq2oQmxZkoxV77Q\n4pVwhHhuK+YwhBhztgo1w72QeJa6Ttdk2+a2BeJbtISkQ4cOijkAocUcWrwhC+33HOK5rZjDEGLM\n2SrIfbiTCyHf4r8XQj7lve+X7JsIRN77S5trKyIiIiJSbgpSUuK9XweMBZ4DngRGp+zumiwtaSsi\nIiIiUlbK/psmRURERERKmb5pUkREREQkj5Rwi4iIiIjkUaHuUrJZnHMOmEB8e8ALvfeNfvNka9qW\nspbG4ZzrDtwDdAI+By7x3j9RsI7mSGt/b865bYC3gRu99zcWoIs518rzehBwK/H/2de99ycWppe5\n1cqYrwRcsnqP9/7qAnQxp5xzNwCnAku89wOaaVsRYxdozCaAMRvCG7c1Zlf+mA35G7dLfoY75ave\nhwBHAJNz0baUtTKO9cA53vv+wAjgjrx3MMey/L1dBrxEmd6jvZXndRvgTuBs7/1ewLkF6WSOtTLm\n3sBIYACwLzDKOderEP3MsfuAY5prVCljF2jMJoAxG8IbtzVmBzNmQ57G7ZJPuAnza+FbHIf3/mPv\n/evJv/8NVDnntihgX3OhVb8359zuwE7Ay4AVqI+51pqY9yf+S/t5AO/90kJ1MsdaE/MK4sRkq2RZ\nBywvTDdzx3v/AtCS31eljF2gMTuEMRvCG7c1ZgcwZkP+xu1yKCkJ8Wvhs4rDOXck8LL3fn3+u5hT\nrY13IvAj4LuF6V5etCbmnsBy59yjyeNu9d5PKVhPc6fFMXvvlzrnfgHMJ54YuNB7/1khO1tglTJ2\ngcbsEMZsCG/c1pitMTtdq/4PlMMMNxDm18K3Jg7nXFfgBsr0oytoWbzOua8D7yR/TZbjLEkDLfwd\ntyf+yOoM4DBgdPLxXVlq4e95V+BsoBfQBxiTnOMVrVLGLtCYTQBjNoQ3bmvM1pidrqX/78sh4c7r\n18KXqFbF4ZxrD9xL/BflB3nuWz60Jt6DgOOdc/8EzgMuds6dnOf+5UNrYl4EvOm9X+C9X0n8kewe\nee5fPrQm5kHAbO/9yuTj2FeAgXnuXzFVytgFGrOh8sdsCG/c1pitMTtdq/7fl0NJyWxgb+fcTsR/\nNe7ivf8HbPq18E21LTMtjtk5Z8DtwDTv/V+K1eHN1OJ4vfeXA5cn+64EVnrv/1icbm+W1pzXLwE9\nnXPbA6uJL0p5rwh93lytifk94CfJRSltgf2A8YXvcn5U8NgFGrNDGLMhvHFbY3bAYzZs/vhV8jPc\nPsCvhW9NzMQfWx0PnOmceyVZyupjnFbGWxFaeV4vT/Y/BcwhfqN+p3C9zY1WxvwScD/xLMlLxDWQ\nbxeut7nhnLsZeB7Y3Tk33zl3bLKrIscu0JhNAGM2hDdua8wOY8yG/I3b+mp3EREREZE8KvkZbhER\nERGRcqaEW0REREQkj5Rwi4iIiIjkkRJuEREREZE8UsItIiIiIpJHSrhFRERERPJICbeIiIiISB6V\nwzdNiuScc+4O4GRgXcrmq4HrgLXA58DTwBjv/bsZHrMKmAGc771fW7COi4gESGO2lDvNcEuoIuA6\n7/02Kcv1yb4BQB/gHWCWc27r9McAA4FBwLhCd1xEJEAas6WsKeEWycB7/6n3/hLimZHjUnZZsn8R\n8BiwTxG6JyIiKTRmS6lTwi0hsxa0eY0MA7RzrivwFeDZXHdKREQy0pgtZUs13BIqAy5yzv0gWY+A\nvTO0WwVsm/aY84FtiD+qvC7vPRUREY3ZUtY0wy2hioDrvffbJ8sO3vuFGdp1BFakPWY74BjgDOdc\nlwL1V0QkZBqzpawp4ZaQteTjyX2AV9Mf471/FJgJjM1Dv0REZFMas6VsKeGWUDU1cJtzrrNz7npg\nC2B6I4+5Efi+c277fHRQRETqacyWsqaEW0IVJUsm/wDeJb7N1Je892syPcZ7P4d4JuXcPPZTREQ0\nZkuZsyhq7PwVEREREZHNpRluEREREZE8UsItIiIiIpJHSrhFRERERPJICbeIiIiISB4p4RYRERER\nySMl3CIiIiIieaSEW0REREQkj5Rwi4iIiIjkkRJuEckpM5tnZrXJUmNm/zSziWa2TVo7M7NxZvZv\nM1trZq+a2bcaOebBZvaomX1qZivM7O9mdvJm9HEnM3sgOV6tmT2V7bFyycxOM7NRBX7O0Wb2zRa0\nOzx5rYYVol/Z2Nw+Jo+/Mtf9EhFRwi0iuRYBjwMHA18B7gTOBx5NazceuAL4BXAkMAf4k5kdltrI\nzL4KPAN8DpwCfBN4Gvj6ZvRxHHA48N2kn6XyVc+nJUshjSZ+TZvzMvFr9Up+u1NUhwNKuEUk59oV\nuwMiUnEMWBpF0d+T9WfNbAtgvJkNiqLob2bWAbgI+F0URTcCmNmzwCHA5cQJNcnjpgJPR1GUOvs9\ny8zab0Yf9wJejaLo/s04RiWx5hpEUbQS+Htz7UREZFOa4RaRQngp+blr8nMIsBUps95RFEXAX4Bh\nZlaVbP4q0J14FryBKIpqWtuJulIXYHjyPLWZSkrMrL+ZPWZmK5Plz2a2Z4bjnZY8/v+Z2R/M7DMz\nW25mU7Ps1zDgsJR+1WZoOzIpv1lrZh+Z2SQza5vWZncze9DMPjaz1UlZzyUp+3dNOX5PYFTKc76f\ndqwJqf1J/wQi7XUYbmazkud8y8yOTmtnZnZt0q+VZvZ/ZvazTHG28HU71cz+lbwWzwD9MrQ5ycye\nMrPFSYnTXDM7P63NX5M+XJGsp8Y7LKXdgWb2p6QMqsbM5pvZZDPrmE3/RSQcmuEWkULonvz8KPm5\nW/JzXlq794nHpS8CbwGDku0vkRsHE8/m/oa49KWulGRFXQMz+wLwV2AR8J2k/U+Bv5rZXlEULc1w\n3KnEJRcnAVsTz9Tnol8NmNkFwA3ALcSfEOwOXAu0BcakNH0YWA+cDSwF9gT6p+z/KOU570/6/tNk\n3+dpTzsFeAjYH7g56V9jfp30bxIwAZhmZrtEUbQq2X8xcAkwkfhTjJHAOc0cMyMzO5S4XOmPxK/X\nYOCmDE33If5D7ibi3/NBwPVmFkVR9KukzTnANsAZwPeIX5s6/0z5927AB8lzfkz8B+Q1QA/g+NbG\nICLhUMItIvnQJpl1rQL2Ay4lrv2tTvZ3Sn6uSnvcqrT9XZKfS3LRqboyFzNbCdSmlL2k+i6wAzAk\niqK3k/ZziROv7wLXZ3jMa1EUnZ2yfl+u+2Vm2wJXA/dEUXROsvmJ5HWeZGbXRlH0qZntCPQFLoqi\naHrS7um051tHUh5iZp8DSxp5LYii6D/Af5IyoOZMjqJoanLcDcS1/AcSlwC1Ay4E7o6iaFzS/nEz\n+xfxH1itNQaYF0XRKSnH2pW0Gvgoin5S928zawM8DxwKjAJ+lbT5Z7L/6GS9sdfiLuCupK0BLwA7\nESfw2yRlNyIim1BJiYjkgyOeYV0NPAvMAr6UlI2UugOBD+uSbYDk3x8k+zL5fQH6NRjoAHgza1e3\nAC8C7YEBSbtlwALgPDM73cz6JclhITyX8u95yc+uyc8ewI7AzLTHPEkLasgz2Jf4vEr1eHqjJP67\nzew/xOfkOuILbr/Q2ic0s22TEpj3gZrkWDcku3dq7fFEJBxKuEUkH2YCBxDf9eFe4lKLASn7P0t+\npte+dkzb/3Hys5DJzHbEZRjplvLfmfd08/PXnXo7Jj/vI0706pYXiUsydgGIoqiW+O4ws4GfAW8D\n883s1AL0cUXKvzcmP7dIftZ9WrEs7THp6y21U4bHNvi9JbXVfwEGEpeyHEL8R9NjZPcJ723E5Se/\nJr4O4ADikhLL8ngiEggNECKSD59GUTQHwMyqicsXbjWzAVEUbQT+lbTbFXg95XG9iWch6y7cq/to\n/0DiuuRCWE5ckpFuRzatOa+zIW+9+a+6ZHIUMDfD/nl1/0hm5E8CMLOBwGRgqpk9HkXR4jz3szF1\nz9s5bfsOm3G89GOlrw8GegGHRVH0bN1GM9uqtU9m8V1xvglcHUXRTSnbN+f2lCISCM1wi0iuNSgb\nSWZcxwB7AKcnm58D1gJH1bVL6muPBJ5JaowhnilfDPwo/Uls824L2JS/Az3NbI+U59qD+I+DfN8W\nbwXxxXuZvACsAXpFUTQnw5JxpjiKoleAXxLPNHfN0KSp58yl+cSfWBxZtyEpdRlOFhdNEt+3/Utp\n5TJfSWtTl1jX39HGzHoS3yUnkxVJm60z7NuS+OLU1GO1IS6fKodSKREpIs1wi0iubVKPG0XRrGSm\ne5yZ/V8URavN7CbgYjN7jzh5+g7Qh/jOGnWPW2dmZwLTzexB4rtz1BDfLrAn8O1c9jNxO/EfCH8y\nsyv4711KPkn25dMrwNfM7JTk37VRFL0FEEXRcjO7CphgZtsBTxEnensD/wMMi6JovZn1Av5AfHHf\nO8D2wFXAv4E3Mzznq8A3k5nad4F1URS9B2Dx7Rn3S9rtlfzc28zWJX16saWBRVG0IfmdTzSzD4gv\n5DyV7Ge4rye+CPcuM/s/4ruP/E9am+eJP7H4ZfLadSL+wqWFxBf0pqv7Up8rzewPxCU7H0ZRtDZ5\n/Z8HLjCzBcTJ+TnEZVCFqpEXkXIVRZEWLVq05GwhvrhwWobtw4nres9O1tsQ3/f438RJ9GvAiEaO\neShxLe5nxHcyeRk4eTP6OAt4qon9A4hn11cly6PAXhnanZbE1DNHr922xInyx0AtsDFDm28Tz7Sv\nJk4mZyevo6Uc4zbiZHsN8e0N7wN2a+Q5dwFmEJes1ALvp+zbNdlWt2xM/XdTr0PKY7+Tss2Ib2P4\nMbCS+HaKNwDLsny9Tk7irCH+NtLvJ/0YltJmGPFtJdcQlzJ9H7gV+KiRY44HPiQubUo/Vk/i0qbl\nxJ+83Eh855qNjb2+WrRo0RJFUf0ALSIiUnBmdj/QNYqiwcXui4hIvqikRERECsLM9iOeoa8mLsk4\njPgWfZtTGiQiUvKUcIuISKGsIq4JP534IsS3gVFRFPmi9kpEJM9UUiIiIiIikke6LaCIiIiISB4p\n4RYRERERySMl3CIiIiIieaSEW0REREQkj5Rwi4iIiIjkkRJuEREREZE8+v8EBsGWkxxtFwAAAABJ\nRU5ErkJggg==\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe19a625b90>"
       ]
      }
     ],
     "prompt_number": 17
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "---"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Cross Validation on Mixture of Testing and Noise Data"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "However, confusions within peptides known at training time is not the major problem of peptide identification. \n",
      "\n",
      "Recall the spectra we excluded previously, we can use them to imitate the real scenario.\n",
      "\n",
      "That is, using these spectra to represent noise or unidentified spectra."
     ]
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "X_noise = spectra[~mask, :]\n",
      "y_noise = clf.predict_proba(X_noise)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 18
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "true_discovery = np.hstack([clf.predict(X_test) == y_test,\n",
      "                            np.zeros(y_noise.shape[0])]\n",
      "                           ).astype(np.bool)\n",
      "\n",
      "entropy = scipy.stats.entropy(np.vstack([y_pred, y_noise]).T)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 19
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "figure(figsize=(12, 4))\n",
      "\n",
      "fig, (ax1, ax2) = plt.subplots(1, 2, sharey=True, figsize=(12, 4))\n",
      "ax1.plot(y_pred[0, :])\n",
      "ax1.set_title('Probability distribution of True ID')\n",
      "\n",
      "ax2.plot(y_noise[0, :])\n",
      "ax2.set_title('Probability distribution of False ID')"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 21,
       "text": [
        "<matplotlib.text.Text at 0x7fe199291910>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe199392a10>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAssAAAENCAYAAAD5USjvAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkiRJUh//F5p4WUHV+r43AAAAAElFTkSuQmCC\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe199392f50>"
       ]
      }
     ],
     "prompt_number": 21
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "fig, axes = plt.subplots(nrows=1, ncols=2, sharex=True, sharey=True, figsize=(12, 4))\n",
      "\n",
      "probs = {\n",
      "    'Probability': np.vstack([y_pred, y_noise]).max(axis=-1),\n",
      "    'Entropy score': np.exp(-entropy)\n",
      "}\n",
      "\n",
      "fig.text(0.5, -0.04, 'ROC for mixture data', ha='center', va='center', fontsize=16)\n",
      "for ax, (scheme, prob) in zip(axes, probs.items()):\n",
      "    \n",
      "    fdr, tdr, _ = roc_curve(true_discovery, prob)\n",
      "    \n",
      "    ax.plot(fdr, tdr)\n",
      "    ax.set_xlabel('FDR')\n",
      "    ax.set_ylabel('TDR')\n",
      "    ax.set_title(scheme)\n",
      "    ax.annotate('Area: {:.4f}'.format(roc_auc_score(true_discovery, prob)), (0.25, 0.5), fontsize=16)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAtwAAAE0CAYAAAAfVuQbAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/wdxuEPZWE6XigAAAABJRU5ErkJggg==\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe1991e6250>"
       ]
      }
     ],
     "prompt_number": 22
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "fig, axes = plt.subplots(nrows=1, ncols=2, sharex=True, sharey=True, figsize=(18, 3))\n",
      "\n",
      "probs = {\n",
      "    'Probability': np.vstack([y_pred, y_noise]).max(axis=-1),\n",
      "    'Entropy score': np.exp(-entropy)\n",
      "}\n",
      "\n",
      "true_mask = true_discovery[:y_test.shape[0]]\n",
      "\n",
      "fig.text(0.5, -0.04,\n",
      "         'Score distribution of logistic regression output probabilities',\n",
      "         ha='center', va='center', fontsize=16)\n",
      "for ax, (scheme, prob) in zip(axes, probs.items()):\n",
      "    \n",
      "    ax.plot(fdr, tdr)\n",
      "    ax.hist([prob[:y_test.shape[0]][true_mask], \n",
      "             prob[:y_test.shape[0]][~true_mask], \n",
      "             prob[-y_noise.shape[0]:]],\n",
      "            bins=50,\n",
      "            stacked=False,\n",
      "            label=['True ID', 'False ID', 'Noise ID'],\n",
      "            color=['b', 'r', 'k'])\n",
      "    ax.legend(loc='best')\n",
      "    ax.set_title(scheme)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAABCUAAADwCAYAAAA3tN4dAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEgAAAECFUAIAAACoeABeeL5+u1TvjwAAAABJRU5ErkJggg==\n",
       "text": [
        "<matplotlib.figure.Figure at 0x7fe1991d59d0>"
       ]
      }
     ],
     "prompt_number": 23
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "---"
     ]
    },
    {
     "cell_type": "heading",
     "level": 3,
     "metadata": {},
     "source": [
      "Conclusion"
     ]
    },
    {
     "cell_type": "markdown",
     "metadata": {},
     "source": [
      "1. It is clear that false identification is rare within known peptides.\n",
      "\n",
      "2. Entropy based score does pretty well on separating noise identification on right hand side, high scoring area. However, the false negative rate is also high (true IDs that scored low in left hand side).\n",
      "\n",
      "3. In overall, probability score output by the model has better discriminative power.\n",
      "\n",
      "\n",
      "**Side note**\n",
      "\n",
      "If you would like to reproduce the result yourself, the chances are pretty good that the result will slightly change, as the training/testing split is done at random."
     ]
    }
   ],
   "metadata": {}
  }
 ]
}