ipython.notebooks/Cell2015_N7_A5SS_Model.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import scipy\n",
    "import scipy.sparse\n",
    "import scipy.stats\n",
    "import os\n",
    "import scipy.io as sio\n",
    "import scipy.ndimage\n",
    "import dnatools\n",
    "%matplotlib inline\n",
    "from pylab import *\n",
    "\n",
    "# Plotting Params:\n",
    "rc('mathtext', default='regular')\n",
    "fsize=20"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Make directory to save results:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "resultsdir = '../results/N7_A5SS_Model_Predictions/'\n",
    "if not os.path.exists(resultsdir):\n",
    "    os.makedirs(resultsdir)\n",
    "figdir = '../figures/N7_A5SS_Model_Predictions/'\n",
    "if not os.path.exists(figdir):\n",
    "    os.makedirs(figdir)\n",
    "    \n",
    "#Choose if you want to actually save the plots:\n",
    "SAVEFIGS = False"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data = sio.loadmat('../data/Reads.mat')\n",
    "# A5SS\n",
    "A5SS_data = data['A5SS']\n",
    "A5SS_reads = np.array(A5SS_data.sum(1)).flatten()\n",
    "A5SS_data = np.array(A5SS_data.todense())\n",
    "# Get minigenes with reads\n",
    "A5SS_nn = find(A5SS_data.sum(axis=1))\n",
    "A5SS_reads = A5SS_reads[A5SS_nn]\n",
    "A5SS_data = A5SS_data[A5SS_nn]\n",
    "A5SS_data = A5SS_data/A5SS_data.sum(axis=1)[:,newaxis]\n",
    "A5SS_seqs = pd.read_csv('../data/A5SS_Seqs.csv',index_col=0).Seq[A5SS_nn]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training the model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The first thing we have to do is create a giant matrix to represent 6-mers at each position. Our matrix have $4^6 \\cdot d$ columns, where d is the number of positions and n rows, where n is the number of data points:<br><br>\n",
    "$AAAAAA_{1,1},\\ ...\\ TTTTTTT_{1,1},\\ AAAAAA_{1,2},\\ ... TTTTTTT_{1,2},\\ ...\\ AAAAAA_{1,d},\\ TTTTTTT_{1,d}$<br>\n",
    "$AAAAAA_{2,1},\\ ...\\ TTTTTTT_{2,1},\\ AAAAAA_{2,2},\\ ... TTTTTTT_{2,2},\\ ...\\ AAAAAA_{2,d},\\ TTTTTTT_{2,d}$<br>\n",
    "...<br>\n",
    "$AAAAAA_{n,1},\\ ...\\ TTTTTTT_{n,1},\\ AAAAAA_{n,2},\\ ... TTTTTTT_{n,2},\\ ...\\ AAAAAA_{n,d},\\ TTTTTTT_{n,d}$<br><br>\n",
    "For example if AAAAAA occurs at position 1 in the first plasmid, then $AAAAAA_{1,1}$ will equal 1, while $AAAAAC_{1,1}$ through $TTTTTT_{1,1}$ will equal 0.<br><br>\n",
    "To score each splice site we are going to look at a 160nt window (-80nt and +80nt). We only want to look at variable sequences, so we will look at all 6-mers that overlap the randomized regions by at least 1nt. So we will pad the edges of our matrix with zeros. The fixed sequence between the random regions will be set to zero as well. We will use the scipy sparse matrix package to make this matrix."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "def make_padded_nmer_matrix(seqs,mer_len):\n",
    "    mer_dict = dict(zip(dnatools.make_mer_list(6),range(4**6)))\n",
    "    rows,cols,vals = [],[],[]\n",
    "    for i in xrange(len(seqs)):\n",
    "        cur_seq = A5SS_seqs[i]\n",
    "        for b in range(7-mer_len,32)+range(50-mer_len,75):\n",
    "            cur_mer = cur_seq[b:b+mer_len]\n",
    "            cur_mer_list_pos = mer_dict[cur_mer]\n",
    "            rows.append(i)\n",
    "            cols.append(cur_mer_list_pos+4**mer_len*(b+80))\n",
    "            vals.append(1)\n",
    "        if(i%10000)==0:\n",
    "            print i,\n",
    "            \n",
    "        # Add a zero to the bottom right of the matrix to ensure the\n",
    "        # correct size\n",
    "        rows.append(len(A5SS_data)-1)\n",
    "        cols.append(4**6*230-1)\n",
    "        vals.append(0)\n",
    "    X = scipy.sparse.csr_matrix((vals,(rows,cols)))\n",
    "    return X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000 110000 120000 130000 140000 150000 160000 170000 180000 190000 200000 210000 220000 230000 240000 250000 260000\n"
     ]
    }
   ],
   "source": [
    "X = make_padded_nmer_matrix(A5SS_seqs.values,6)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Save the matrix with the splice donor usage into a .mat file:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Only include the first 80 positions and the the last position (unspliced)\n",
    "Y = scipy.matrix(A5SS_data)\n",
    "Y = scipy.hstack((Y[:,:80],Y[:,-1:]))\n",
    "sio.savemat(resultsdir+'padded_6mer_matrix.mat',{'X':X,'Y':Y})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Testing the model with a small amount of data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that this is saved I have a script that we can use to train the model. (Warning: the script works, but is hacked together and probably very hard to understand. If there is interest, I can clean it up.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import MLR_8regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3.74247119891\n",
      "2.9732321692\n",
      "2.54643412892\n",
      "2.17820706676\n",
      "1.56332451642\n",
      "1.39551938998\n",
      "1.11694441434\n",
      "0.80798867898\n",
      "0.471788205254\n",
      "0.412724035669\n",
      "0.333173892018\n"
     ]
    }
   ],
   "source": [
    "# Example training on 1000 points for 10 steps:\n",
    "w,Wfull,w0,f,d = MLR_8regions.MLR(X[:1000],Y[:1000],reg_lambda=0.00001,mer_len=6,maxfun=10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The vector w contains the weights we care about. The first $4^6\\cdot8$ values contain the 6-mer weights. The remaining values contain bias terms for each position. Lets reshape the $4^6\\cdot8$ 6-mer weights and make a dataframe:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "W = pd.DataFrame(index=dnatools.make_mer_list(6),\n",
    "                 data=w[:4**6*8].reshape(4**6,-1))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If we look at the best possible sequence within each of our 8 regions, the regions overlapping the splice site (2,3,4,5), clearly start to look like splice donors, even though we trained on 1000 points for 10 steps. (Consensus splice site: CAG|GTAAGT)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    GGGTCG\n",
       "1    GAAGAA\n",
       "2    AAGGTA\n",
       "3    AGGTAA\n",
       "4    GGTAAG\n",
       "5    GTAAGG\n",
       "6    GGGGGG\n",
       "7    TTCGTG\n",
       "dtype: object"
      ]
     },
     "execution_count": 154,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "W.apply(argmax,axis=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training a full model with cross validation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we want to train on all the data with cross-validation. We will split the data into 10 partitions and use 8 for training, 1 for validation, and 1 for testing. We need the validation data to optimize the amount of regularization we use. We will use L1 regularization. I wrote a script to automate this:<br>\n",
    "../bin/MLR_8reg_cross_val_script.py<br>\n",
    "A list of commands is in:<br>\n",
    "../bin/Cross_val_lambda_opt.lst:<br>\n",
    "This take a long time. I ran all the commands in parallel on a cluster. The results are in a text file:<br>\n",
    "../results/N7_A5SS_Model_Predictions/Reg8_lambdas.txt<br>\n",
    "Let's load these"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "cv_df = pd.read_csv('../results/N7_A5SS_Model_Predictions/Reg8_lambdas.txt',sep='\\t',\n",
    "                    names=['Part_ind','Lambda','KL_div'])\n",
    "cv_df.Lambda = log10(cv_df.Lambda)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div style=\"max-height:1000px;max-width:1500px;overflow:auto;\">\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>Lambda</th>\n",
       "      <th>-7.0</th>\n",
       "      <th>-6.5</th>\n",
       "      <th>-6.0</th>\n",
       "      <th>-5.5</th>\n",
       "      <th>-5.0</th>\n",
       "      <th>-4.5</th>\n",
       "      <th>-4.0</th>\n",
       "      <th>-3.5</th>\n",
       "      <th>-3.0</th>\n",
       "      <th>-2.5</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Part_ind</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td> 0.244347</td>\n",
       "      <td> 0.242504</td>\n",
       "      <td> 0.243814</td>\n",
       "      <td> 0.243039</td>\n",
       "      <td> 0.304285</td>\n",
       "      <td> 0.250507</td>\n",
       "      <td> 0.274400</td>\n",
       "      <td> 0.328884</td>\n",
       "      <td> 0.479231</td>\n",
       "      <td> 0.766991</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td> 0.243765</td>\n",
       "      <td> 0.244759</td>\n",
       "      <td> 0.244747</td>\n",
       "      <td> 0.243620</td>\n",
       "      <td> 0.244544</td>\n",
       "      <td> 0.251401</td>\n",
       "      <td> 0.276902</td>\n",
       "      <td> 0.342629</td>\n",
       "      <td> 0.456405</td>\n",
       "      <td> 0.768611</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td> 0.241063</td>\n",
       "      <td> 0.242559</td>\n",
       "      <td> 0.242065</td>\n",
       "      <td> 0.240893</td>\n",
       "      <td> 0.242338</td>\n",
       "      <td> 0.249624</td>\n",
       "      <td> 0.270235</td>\n",
       "      <td> 0.335968</td>\n",
       "      <td> 0.470331</td>\n",
       "      <td> 0.737424</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td> 0.241540</td>\n",
       "      <td> 0.240981</td>\n",
       "      <td> 0.241015</td>\n",
       "      <td> 0.240701</td>\n",
       "      <td> 0.241897</td>\n",
       "      <td> 0.249771</td>\n",
       "      <td> 0.272750</td>\n",
       "      <td> 0.330091</td>\n",
       "      <td> 0.462111</td>\n",
       "      <td> 0.710388</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td> 0.239445</td>\n",
       "      <td> 0.240308</td>\n",
       "      <td> 0.239235</td>\n",
       "      <td> 0.239125</td>\n",
       "      <td> 0.240697</td>\n",
       "      <td> 0.249149</td>\n",
       "      <td> 0.271087</td>\n",
       "      <td> 0.340511</td>\n",
       "      <td> 0.473372</td>\n",
       "      <td> 0.735859</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td> 0.241926</td>\n",
       "      <td> 0.243398</td>\n",
       "      <td> 0.241734</td>\n",
       "      <td> 0.241911</td>\n",
       "      <td> 0.242930</td>\n",
       "      <td> 0.250345</td>\n",
       "      <td> 0.276985</td>\n",
       "      <td> 0.326082</td>\n",
       "      <td> 0.468734</td>\n",
       "      <td> 0.729471</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td> 0.242000</td>\n",
       "      <td> 0.241776</td>\n",
       "      <td> 0.241306</td>\n",
       "      <td> 0.240031</td>\n",
       "      <td> 0.241453</td>\n",
       "      <td> 0.248481</td>\n",
       "      <td> 0.272485</td>\n",
       "      <td> 0.321511</td>\n",
       "      <td> 0.461875</td>\n",
       "      <td> 0.767949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td> 0.237110</td>\n",
       "      <td> 0.237040</td>\n",
       "      <td> 0.237864</td>\n",
       "      <td> 0.236909</td>\n",
       "      <td> 0.238447</td>\n",
       "      <td> 0.246380</td>\n",
       "      <td> 0.270965</td>\n",
       "      <td> 0.340548</td>\n",
       "      <td> 0.484538</td>\n",
       "      <td> 0.717556</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td> 0.242899</td>\n",
       "      <td> 0.241052</td>\n",
       "      <td> 0.241599</td>\n",
       "      <td> 0.241463</td>\n",
       "      <td> 0.242656</td>\n",
       "      <td> 0.249630</td>\n",
       "      <td> 0.271548</td>\n",
       "      <td> 0.322897</td>\n",
       "      <td> 0.459984</td>\n",
       "      <td> 0.723547</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td> 0.245366</td>\n",
       "      <td> 0.243793</td>\n",
       "      <td> 0.243528</td>\n",
       "      <td> 0.244867</td>\n",
       "      <td> 0.245072</td>\n",
       "      <td> 0.253267</td>\n",
       "      <td> 0.275367</td>\n",
       "      <td> 0.340156</td>\n",
       "      <td> 0.470729</td>\n",
       "      <td> 0.702622</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Lambda        -7.0      -6.5      -6.0      -5.5      -5.0      -4.5  \\\n",
       "Part_ind                                                               \n",
       "0         0.244347  0.242504  0.243814  0.243039  0.304285  0.250507   \n",
       "1         0.243765  0.244759  0.244747  0.243620  0.244544  0.251401   \n",
       "2         0.241063  0.242559  0.242065  0.240893  0.242338  0.249624   \n",
       "3         0.241540  0.240981  0.241015  0.240701  0.241897  0.249771   \n",
       "4         0.239445  0.240308  0.239235  0.239125  0.240697  0.249149   \n",
       "5         0.241926  0.243398  0.241734  0.241911  0.242930  0.250345   \n",
       "6         0.242000  0.241776  0.241306  0.240031  0.241453  0.248481   \n",
       "7         0.237110  0.237040  0.237864  0.236909  0.238447  0.246380   \n",
       "8         0.242899  0.241052  0.241599  0.241463  0.242656  0.249630   \n",
       "9         0.245366  0.243793  0.243528  0.244867  0.245072  0.253267   \n",
       "\n",
       "Lambda        -4.0      -3.5      -3.0      -2.5  \n",
       "Part_ind                                          \n",
       "0         0.274400  0.328884  0.479231  0.766991  \n",
       "1         0.276902  0.342629  0.456405  0.768611  \n",
       "2         0.270235  0.335968  0.470331  0.737424  \n",
       "3         0.272750  0.330091  0.462111  0.710388  \n",
       "4         0.271087  0.340511  0.473372  0.735859  \n",
       "5         0.276985  0.326082  0.468734  0.729471  \n",
       "6         0.272485  0.321511  0.461875  0.767949  \n",
       "7         0.270965  0.340548  0.484538  0.717556  \n",
       "8         0.271548  0.322897  0.459984  0.723547  \n",
       "9         0.275367  0.340156  0.470729  0.702622  "
      ]
     },
     "execution_count": 148,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cv_df.groupby(['Lambda','Part_ind']).KL_div.mean().unstack('Part_ind').T"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Optimal $log_{10}\\lambda$ for each partition:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Part_ind\n",
      "0          -6.5\n",
      "1          -5.5\n",
      "2          -5.5\n",
      "3          -5.5\n",
      "4          -5.5\n",
      "5          -6.0\n",
      "6          -5.5\n",
      "7          -5.5\n",
      "8          -6.5\n",
      "9          -6.0\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "print cv_df.groupby(['Lambda','Part_ind']).KL_div.mean().unstack('Part_ind').apply(argmin)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we can use these optimized lambdas to retrain with the full training + validation (but not test) sets. The script to do this is:<br>\n",
    "../bin/MLR_8reg_cross_val_script_save_output.py<br>\n",
    "A list of calls to this script for each partition is located here:<br>\n",
    "../bin/Cross_val_save_8reg.lst<br>\n",
    "After these we have predictions for each partition of the cross-validation."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "### Plotting Predictions:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We first need to load all of the cross-validation test predictions:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 1 2 3 4 5 6 7 8 9\n"
     ]
    }
   ],
   "source": [
    "combined_predictions = np.zeros((len(A5SS_data),81))\n",
    "for i in range(10):\n",
    "    print i,\n",
    "    data = sio.loadmat('../results/N7_A5SS_Model_Predictions/Partition'+str(i)+'/Training_data.mat')\n",
    "    # Check if the data file is an old file that include the minigenes with no reads:\n",
    "    if len(data['Test'])==265137:\n",
    "        test_inds = find(data['Test'].ravel()[A5SS_nn]==1)\n",
    "        combined_predictions[test_inds] = data['Prediction'][A5SS_nn][test_inds]\n",
    "    else:\n",
    "        test_inds = find(data['Test'].ravel()==1)\n",
    "        combined_predictions[test_inds] = data['Prediction'][test_inds]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtkAAADLCAYAAAC7zI8nAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nWFhYWFhYWFhYWFhYWFhYWAyK/x+WF9uc9kmaSwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x3638ae10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "Y = np.array(Y)\n",
    "a = 0.5\n",
    "psize = .1\n",
    "fsize = 13.85\n",
    "rasterized_on = True\n",
    "bins =50\n",
    "axlim = [0,1,0,1]\n",
    "hist_thresh = 300.\n",
    "\n",
    "fig = figure(figsize=(10*1,1.875*1))\n",
    "width=0.18\n",
    "height = 1\n",
    "#ax = fig.add_subplot(221)\n",
    "ax = axes([0,0,width,height])\n",
    "axis([0,1,0,1])\n",
    "sd = 0\n",
    "x,y = combined_predictions[:,sd],Y[:,sd]\n",
    "scatter(x,y,s=psize,edgecolor='None',alpha = a,rasterized=rasterized_on)\n",
    "z, xedge, yedge = np.histogram2d(y, x, bins=[bins, bins], \n",
    "                                    range=(axlim[:2],axlim[2:]))\n",
    "z = scipy.ndimage.zoom(z, 3)\n",
    "z[z<hist_thresh] = nan\n",
    "#contourf(np.log(z), bins, cmap=cm.jet, extent=axlim)\n",
    "ax.tick_params(labelsize=fsize)\n",
    "ax.set_xticks([0,0.5,1])\n",
    "ax.set_xticklabels(['0','0.5','1'])\n",
    "ax.set_yticks([0,0.5,1])\n",
    "ax.set_yticklabels(['0','0.5','1'])\n",
    "#title('$SD_1$',fontsize=fsize)\n",
    "#xlabel('Predicted',fontsize=fsize)\n",
    "#ylabel('RNA-seq',fontsize=fsize)\n",
    "text(1,0.03,'$R^2$=%.2f' %(scipy.stats.pearsonr(x,y)[0]**2),fontsize=fsize,ha='right')\n",
    "ax.set_title('$SD_1$',fontsize=fsize)\n",
    "ax.set_ylabel('RNA-seq',fontsize=fsize)\n",
    "ax.set_xlabel('Model Predictions',fontsize=fsize)\n",
    "\n",
    "ax = axes([0.25,0,width,height])\n",
    "axis([0,1,0,1])\n",
    "sd = 44\n",
    "x,y = combined_predictions[:,sd],Y[:,sd]\n",
    "scatter(x,y,s=psize,edgecolor='None',alpha = a,rasterized=rasterized_on)\n",
    "z, xedge, yedge = np.histogram2d(y, x, bins=[bins, bins], \n",
    "                                    range=(axlim[:2],axlim[2:]))\n",
    "z = scipy.ndimage.zoom(z, 3)\n",
    "z[z<hist_thresh] = nan\n",
    "#contourf(np.log(z), bins, cmap=cm.jet, extent=axlim)\n",
    "ax.tick_params(labelsize=fsize)\n",
    "ax.set_xticks([0,0.5,1])\n",
    "ax.set_xticklabels(['0','0.5','1'])\n",
    "ax.set_yticks([0,0.5,1])\n",
    "ax.set_yticklabels(['0','0.5','1'])\n",
    "#title('$SD_{2}$',fontsize=fsize)\n",
    "#xlabel('Predicted',fontsize=fsize)\n",
    "#ylabel('RNA-seq',fontsize=fsize)\n",
    "text(1,0.03,'$R^2$=%.2f' %(scipy.stats.pearsonr(x,y)[0]**2),fontsize=fsize,ha='right')\n",
    "ax.set_title('$SD_2$',fontsize=fsize)\n",
    "ax.set_xlabel('Model Predictions',fontsize=fsize)\n",
    "\n",
    "\n",
    "ax = axes([0.5,0,width,height])\n",
    "axis([0,1,0,1])\n",
    "sd = 79\n",
    "x,y = combined_predictions[:,sd],Y[:,sd]\n",
    "scatter(x,y,s=psize,edgecolor='None',alpha = a,rasterized=rasterized_on)\n",
    "z, xedge, yedge = np.histogram2d(y, x, bins=[bins, bins], \n",
    "                                    range=(axlim[:2],axlim[2:]))\n",
    "z = scipy.ndimage.zoom(z, 3)\n",
    "z[z<hist_thresh] = nan\n",
    "#contourf(np.log(z), bins, cmap=cm.jet, extent=axlim)\n",
    "ax.tick_params(labelsize=fsize)\n",
    "ax.set_xticks([0,0.5,1])\n",
    "ax.set_xticklabels(['0','0.5','1'])\n",
    "ax.set_yticks([0,0.5,1])\n",
    "ax.set_yticklabels(['0','0.5','1'])\n",
    "#title('$SD_{3}$',fontsize=fsize)\n",
    "#xlabel('Predicted',fontsize=fsize)\n",
    "#ylabel('RNA-seq',fontsize=fsize)\n",
    "text(1,0.03,'$R^2$=%.2f' %(scipy.stats.pearsonr(x,y)[0]**2),fontsize=fsize,ha='right')\n",
    "ax.set_title('$SD_{CRYPT}$',fontsize=fsize)\n",
    "ax.set_xlabel('Model Predictions',fontsize=fsize)\n",
    "\n",
    "\n",
    "ax = axes([0.75,0,width,height])\n",
    "axis([0,1,0,1])\n",
    "sd = 'other'\n",
    "x = sum(combined_predictions[:,5:35],axis=1) + sum(combined_predictions[:,48:77],axis=1)\n",
    "y = sum(Y[:,5:35],axis=1) + sum(Y[:,48:77],axis=1)\n",
    "scatter(x,y,s=psize,edgecolor='None',alpha = a,rasterized=rasterized_on)\n",
    "z, xedge, yedge = np.histogram2d(y, x, bins=[bins, bins], \n",
    "                                    range=(axlim[:2],axlim[2:]))\n",
    "z = scipy.ndimage.zoom(z, 3)\n",
    "z[z<hist_thresh] = nan\n",
    "#contourf(np.log(z), bins, cmap=cm.jet, extent=axlim)\n",
    "ax.tick_params(labelsize=fsize)\n",
    "ax.set_xticks([0,0.5,1])\n",
    "ax.set_xticklabels(['0','0.5','1'])\n",
    "ax.set_yticks([0,0.5,1])\n",
    "ax.set_yticklabels(['0','0.5','1'])\n",
    "#title('$SD_{3}$',fontsize=fsize)\n",
    "#xlabel('Predicted',fontsize=fsize)\n",
    "#ylabel('RNA-seq',fontsize=fsize)\n",
    "text(1,0.03,'$R^2$=%.2f' %(scipy.stats.pearsonr(x,y)[0]**2),fontsize=fsize,ha='right',)\n",
    "ax.set_title('$SD_{NEW}$',fontsize=fsize)\n",
    "ax.set_xlabel('Model Predictions',fontsize=fsize)\n",
    "\n",
    "if SAVEFIGS:\n",
    "    filename = 'A5SS_Library_Prediction_Scatter'\n",
    "    fig.savefig(figdir+filename+'.png', bbox_inches='tight',dpi=300)\n",
    "    fig.savefig(figdir+filename+'.eps', bbox_inches='tight',dpi=300)\n",
    "    fig.savefig(figdir+filename+'.pdf', bbox_inches='tight',dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAOAAAAA+CAYAAAAyGRe+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYRiGYRiGYRzr/A9MPvjU8taQBQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x37aa6f90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = figure(figsize=(3,1))\n",
    "ax = fig.add_axes([0.05, 0.80, 0.9, 0.15])\n",
    "norm = mpl.colors.Normalize(vmin=0, vmax=1)\n",
    "cb1 = mpl.colorbar.ColorbarBase(ax, cmap=cm.gray_r,\n",
    "                                   norm=norm,\n",
    "                                   orientation='horizontal')\n",
    "ax.xaxis.set_tick_params(labelsize=14)\n",
    "ax.xaxis.set_ticks(arange(0,1.2,0.2))\n",
    "ax.xaxis.set_ticklabels(arange(0,101,20))\n",
    "ax.tick_params(size=2)\n",
    "ax.set_xlabel('Splice Site Use (%)',fontsize=14)\n",
    "if SAVEFIGS:\n",
    "    name = 'SS Prediction Colorbar'\n",
    "    fig.savefig(figdir+name+'.png', bbox_inches='tight',dpi=300)\n",
    "    fig.savefig(figdir+name+'.eps', bbox_inches='tight',dpi=300)\n",
    "    fig.savefig(figdir+name+'.pdf', bbox_inches='tight',dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVsAAABUCAYAAAAs7pSMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x37aa7e50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = figure(figsize=(5,1.25))\n",
    "ax = fig.add_axes([0.05, 0.80, 0.9, 0.15])\n",
    "norm = mpl.colors.Normalize(vmin=-0.4, vmax=.4)\n",
    "cb1 = mpl.colorbar.ColorbarBase(ax, cmap=cm.RdBu_r,\n",
    "                                   norm=norm,\n",
    "                                   orientation='horizontal')\n",
    "ax.xaxis.set_tick_params(labelsize=14)\n",
    "ax.xaxis.set_ticks(arange(0,1.01,1/4.))\n",
    "ax.xaxis.set_ticklabels([-1,-0.5,0,0.5,1])\n",
    "ax.set_title('Nucleotide Contribution to Isoform',fontsize=14)\n",
    "ax.set_xlabel('$(log\\ odds-ratio)$',fontsize=14)\n",
    "ax.tick_params(size=2)\n",
    "if SAVEFIGS:\n",
    "    name = 'Nucleotide_Contribution_Colorbar'\n",
    "    fig.savefig(figdir+name+'.png', bbox_inches='tight',dpi=300)\n",
    "    fig.savefig(figdir+name+'.eps', bbox_inches='tight',dpi=300)\n",
    "    fig.savefig(figdir+name+'.pdf', bbox_inches='tight',dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "seqs = pd.Series(data=A5SS_seqs.values,index=range(len(A5SS_seqs)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "w = data['Mer_scores'][:4**6*8]\n",
    "w_mat = w.reshape(4**6,size(w)/4**6)\n",
    "mer6_dict = dict(zip(dnatools.make_mer_list(6),range(4**6)))\n",
    "pos_map = dict(zip(range(-80,80),[0]*40+[1]*(40-9)+[-1]*6+[2,3,4,5]+[-1]*5+[6]*29+[7]*45))\n",
    "def score_seq(seq,sd):\n",
    "    scores = np.zeros(90)\n",
    "    for b in range(2,32)+range(45,75):\n",
    "        rel_pos = b-sd\n",
    "        cur_mer = seq[b:b+6]\n",
    "        cur_pos_ind = pos_map[rel_pos]\n",
    "        if(cur_pos_ind>=0):\n",
    "            scores[b+4:b+6+4] += w_mat[mer6_dict[cur_mer],cur_pos_ind]\n",
    "    return scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def convertT2U(seq):\n",
    "    sU = ''\n",
    "    for b in seq:\n",
    "        if b=='T':\n",
    "            sU += 'U'\n",
    "        else:\n",
    "            sU += b\n",
    "    return sU\n",
    "\n",
    "def make_nt_contribution_fig_4(num,savefig=False):\n",
    "    pred_li = np.array([0.]*3+list(combined_predictions[num,:80])+[0.]*7)\n",
    "    data_li = np.array([0.]*3+list(Y[num,:80])+[0.]*7)\n",
    "    new_sds = set(find((np.array(pred_li)>0.05) | (np.array(data_li)>0.05)))-set([3,47,82,83])\n",
    "    all_sds = sorted(list(new_sds.union(set([3,47,82]))))\n",
    "    seq = seqs.values[num]\n",
    "    \n",
    "    all_weights = []\n",
    "    for sd in all_sds:\n",
    "        temp_weight_row = np.zeros(90)\n",
    "        if(sd==3):\n",
    "            temp_weight_row = score_seq(seq,0)\n",
    "        elif(sd==47):\n",
    "             temp_weight_row = score_seq(seq,44)\n",
    "        elif(sd==82):\n",
    "            temp_weight_row = score_seq(seq,79)\n",
    "        else:\n",
    "            temp_weight_row = score_seq(seq,sd-4)\n",
    "        all_weights.append(temp_weight_row)\n",
    "    all_weights = np.array(all_weights)\n",
    "    \n",
    "    fig = figure(figsize=(13.5,len(all_sds)/1.75))\n",
    "    for i in range(len(all_sds)):\n",
    "        ax = fig.add_axes((0,(len(all_sds)-i-1)/float(len(all_sds)),0.71,0.5/len(all_sds)))\n",
    "        ax.imshow(np.array([all_weights[i],all_weights[i]]),interpolation='None',cmap=cm.RdBu_r,aspect=0.75,vmin=-1,vmax=1)\n",
    "        ax.set_xticklabels([]);\n",
    "        ax.set_yticklabels([]);\n",
    "        ax.tick_params(size=0)\n",
    "        cur_row =i\n",
    "        cur_pos = all_sds[i]\n",
    "        ax = fig.add_axes((0,(len(all_sds)-i-1+0.4)/float(len(all_sds)),0.71,0.5/len(all_sds)))\n",
    "        ax.set_xticklabels([]);\n",
    "        ax.set_yticklabels([]);\n",
    "        ax.tick_params(size=0)\n",
    "        box('off')\n",
    "        rect1 = matplotlib.patches.Rectangle((0,0.1), cur_pos, .5, color=(0.2,0.2,0.2))\n",
    "        ax.add_patch(rect1)\n",
    "        ax.plot([cur_pos,90],[0.35,0.35],color=(0.2,0.2,0.2),linewidth=2)\n",
    "        axis([0,90,0,1])\n",
    "        if(i==0):\n",
    "            ax.set_xticks(arange(90)+0.5)\n",
    "            ax.set_xticklabels('GAGG'+convertT2U(seq))\n",
    "            for tick in ax.get_xaxis().get_major_ticks():\n",
    "                tick.set_pad(-5.)\n",
    "            ax.xaxis.tick_top()\n",
    "            ax.tick_params(labelsize=10)\n",
    "        if(all_sds[i]==3):\n",
    "            ylab = '$SD_1$'\n",
    "        elif(all_sds[i]==47):\n",
    "            ylab = '$SD_2$'\n",
    "        elif(all_sds[i]==82):\n",
    "            ylab = '$SD_{CRYPT}$'\n",
    "        else:\n",
    "            ylab = '$SD_{NEW}$'\n",
    "            \n",
    "        text(-0.55,0,ylab,fontsize=fsize,rotation=0,ha='right')\n",
    "    ax = fig.add_axes((0.72,0,0.15,0.92))\n",
    "    bars = ax.barh(arange(len(all_sds)),pred_li[all_sds[::-1]],height=0.3,color='white')\n",
    "    ax.barh(arange(len(all_sds))+0.35,data_li[all_sds[::-1]],height=0.3,color='gray')\n",
    "    axis([0,1,-0.05,len(all_sds)-0.25])\n",
    "    ax.spines['top'].set_visible(False)\n",
    "    ax.spines['right'].set_visible(False)\n",
    "    ax.get_xaxis().tick_bottom()\n",
    "    ax.get_yaxis().tick_left()\n",
    "    ax.set_xticks(arange(0,1.01,0.25))\n",
    "    ax.set_xticklabels([0,0.25,0.5,0.75,1])\n",
    "    ax.tick_params(size=0,labelsize=fsize)\n",
    "    ax.set_yticklabels([])\n",
    "    ax.set_xlabel('Isoform Fraction',fontsize=fsize)\n",
    "    if savefig:\n",
    "        name = 'ExamplePredictionEffectsnt/Individual_Nucleotide_Contributions' + str(num)\n",
    "        fig.savefig(figdir+name+'.png', bbox_inches='tight',dpi=300)\n",
    "        fig.savefig(figdir+name+'.eps', bbox_inches='tight',dpi=300)\n",
    "        fig.savefig(figdir+name+'.pdf', bbox_inches='tight',dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def make_nt_contribution_fig_5(num,savefig=False):\n",
    "    pred_li = np.array([0.]*3+list(combined_predictions[num,:80])+[0.]*7)\n",
    "    data_li = np.array([0.]*3+list(Y[num,:80])+[0.]*7)\n",
    "    new_sds = set(find((np.array(pred_li)>0.05) | (np.array(data_li)>0.05)))-set([3,47,82,83])\n",
    "    all_sds = sorted(list(new_sds.union(set([3,47,82]))))\n",
    "    seq = seqs.values[num]\n",
    "    \n",
    "    all_weights = []\n",
    "    for sd in all_sds:\n",
    "        temp_weight_row = np.zeros(90)\n",
    "        if(sd==3):\n",
    "            temp_weight_row = score_seq(seq,0)\n",
    "        elif(sd==47):\n",
    "             temp_weight_row = score_seq(seq,44)\n",
    "        elif(sd==82):\n",
    "            temp_weight_row = score_seq(seq,79)\n",
    "        else:\n",
    "            temp_weight_row = score_seq(seq,sd-4)\n",
    "        all_weights.append(temp_weight_row)\n",
    "    all_weights = np.array(all_weights)\n",
    "    \n",
    "    scaling_factor = 0.8\n",
    "    fig = figure(figsize=(13.5,len(all_sds)/1.75*scaling_factor))\n",
    "    for i in range(len(all_sds)):\n",
    "        cur_row =i\n",
    "        cur_pos = all_sds[i]\n",
    "        ax = fig.add_axes((0,(len(all_sds)-i-1+0.1)/float(len(all_sds))*scaling_factor,0.71,0.85/len(all_sds)*scaling_factor))\n",
    "        ax.set_xticklabels([]);\n",
    "        ax.set_yticklabels([]);\n",
    "        ax.tick_params(size=0)\n",
    "        box('off')\n",
    "        rect1 = matplotlib.patches.Rectangle((0,0.1), cur_pos, .5, color=(0.2,0.2,0.2))\n",
    "        ax.add_patch(rect1)\n",
    "        ax.plot([cur_pos,90],[0.35,0.35],color=(0.2,0.2,0.2),linewidth=2)\n",
    "        axis([0,90,0,1])\n",
    "        if(i==0):\n",
    "            ax.set_xticks(arange(90)+0.5)\n",
    "            ax.set_xticklabels('GAGG'+convertT2U(seq))\n",
    "            for tick in ax.get_xaxis().get_major_ticks():\n",
    "                tick.set_pad(-5.)\n",
    "            ax.xaxis.tick_top()\n",
    "            ax.tick_params(labelsize=10)\n",
    "        if(all_sds[i]==3):\n",
    "            ylab = '$SD_1$'\n",
    "        elif(all_sds[i]==47):\n",
    "            ylab = '$SD_2$'\n",
    "        elif(all_sds[i]==82):\n",
    "            ylab = '$SD_{CRYPT}$'\n",
    "        else:\n",
    "            ylab = '$SD_{NEW}$'\n",
    "            \n",
    "        text(-0.55,0,ylab,fontsize=fsize,rotation=0,ha='right')\n",
    "    ax = fig.add_axes((0.72,0,0.15,0.92*scaling_factor))\n",
    "    bars = ax.barh(arange(len(all_sds)),pred_li[all_sds[::-1]],height=0.3,color='white')\n",
    "    ax.barh(arange(len(all_sds))+0.35,data_li[all_sds[::-1]],height=0.3,color='gray')\n",
    "    axis([0,1,-0.05,len(all_sds)-0.25])\n",
    "    ax.spines['top'].set_visible(False)\n",
    "    ax.spines['right'].set_visible(False)\n",
    "    ax.get_xaxis().tick_bottom()\n",
    "    ax.get_yaxis().tick_left()\n",
    "    ax.set_xticks(arange(0,1.01,0.25))\n",
    "    ax.set_xticklabels([0,0.25,0.5,0.75,1])\n",
    "    ax.tick_params(size=0,labelsize=fsize)\n",
    "    ax.set_yticklabels([])\n",
    "    ax.set_xlabel('Isoform Fraction',fontsize=fsize)\n",
    "    if savefig:\n",
    "        name = 'ExamplePredictionEffectsnt/Individual_Nucleotide_Contributions' + str(num)\n",
    "        fig.savefig(figdir+name+'.png', bbox_inches='tight',dpi=300)\n",
    "        fig.savefig(figdir+name+'.eps', bbox_inches='tight',dpi=300)\n",
    "        fig.savefig(figdir+name+'.pdf', bbox_inches='tight',dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAIoAAABECAYAAABeZa2gAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSZIkSZIk2c/8D+eQJrLJ1yMxAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1830fd90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = figure(figsize=(0.4,0.25))\n",
    "ax = fig.add_subplot(111)\n",
    "ax.bar([1],[1],color='gray',label='RNA-Seq')\n",
    "ax.bar([1],[1],color='w',label='Predictions')\n",
    "leg = ax.legend(bbox_to_anchor=(0.5,0.5))\n",
    "leg.get_frame().set_alpha(0)\n",
    "box('off')\n",
    "ax.tick_params(labelsize=0,size=0)\n",
    "ax.axis([100,101,100,101])\n",
    "ax.grid('off')\n",
    "ax.spines['top'].set_visible(False)\n",
    "ax.spines['right'].set_visible(False)\n",
    "ax.spines['bottom'].set_visible(False)\n",
    "ax.spines['left'].set_visible(False)\n",
    "ax.set_xticklabels([])\n",
    "ax.set_yticklabels([])\n",
    "\n",
    "if SAVEFIGS:\n",
    "    name = '/ExamplePredictions/Legend'\n",
    "    fig.savefig(figdir+name+'.png', bbox_inches='tight',dpi=300)\n",
    "    fig.savefig(figdir+name+'.eps', bbox_inches='tight',dpi=300)\n",
    "    fig.savefig(figdir+name+'.pdf', bbox_inches='tight',dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA5cAAACGCAYAAABJ935GAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAFpBJREFUeJzt3X2UHWV9wPHvhUShkbwiCWpCFG2qEATXNyAqaOWlWgUR\nlGrbYPClaostQjmnFlawtdqeIketbQ+VRRCwHNFiBCXyJh6CyCUkBIyoDRo1QCQb2EJoEnL7xzPb\nnVzuzsze586duzffzzlzkjvPPC/zzNzZ/e0z8wxIkiRJkiRJkiRJkiRJkiRJkiRJkiRJkiRJkiRJ\nkiRJkiRJkiRJUl+ZC1wB3AesBn4EnJxKPwHYCSxqI29W+kLg3qbyBoEzC5Y9D7gKWAusAb6XamNW\n2Xn15tWd1y7I7rOsdhdJH6/sIvuV1a7Y9JjzJGufY/s75ljGpMe2Oy89pr+z8sccq7z8MdeDvLqr\n/F7m5c2qu8xraF56zLHMK7vMc7DMYx1bd7tlL6S663fePsW2rcz+bnefy9ynvPxV/kxbSPv7nZc3\nL73d39mKtluq3B5VN6BH7Al8F7gOOAh4OXAS4QI16lRgefLvRPIWKbtZYwJlfydp18HAIcBfAc8t\nUHZeWlbdRfcpq8+y2l1kv8YrO2+/iuSNSY85T8bb5070d7vHMiY9tt1F0tvt76z8sd+7TvRJWvP3\nsszzpN3vZdFrUbv9HdNnWekxx7JI2WWeg2Ud605cE2LP0bRuXL/baddE2lZmf2fVWzS90/uUl7/K\nn2njmeh5lpV3vPRO/85WJF1SRf6A8IUfz3OAB4EFwI8nmDcvfSHP/EvUeYS/ROXlPR64oc2ys9LI\nqTuvXZDdZ3ntzkvPKnsh2fuVlTc2PeY8ydrn2P6OOZYx6bHtzkuP6e+s/LHfu5g+WUj2+VvmeRLz\nvczLm1V3mdfQvPSYY5lXdpnnYJnHOrbumLIXUs31u0h/xrStzP6O6ZOFlLNPeXVX+TMN4vY7L29W\neszvbEXSpZ7gyGWwGFiVkf52wsXql8Am4BUTyJuXHtOuQ4B6m2XH1F1kn7L6LK/deelZZefJyxuT\nHnOeZO1zbH/HHMuY9Nh256XH9HdW/tjvXWyfZCnzPIn5Xha5FrXb37F9liXmWOYp8xws81jH1h17\njmYp6/rdiXOsqv6O7ZMsMdeLrPxV/kwrIqbPspT5O5vUMwwug+bbCr4A3APcmXw+Fbg6+f/V7Hqb\nRF7eiaany20373htyyq7OU9W3Xntgon12XhtGE87ZY+uz8obm97J86RV28fLG1N3p8/fTp4neekx\n/Z2Vv9Pfu5j+bi6zm+dJXvlF0/LqLvMampcecywnWnaZ52Anj3VZxyMmb9nX7yL9GdO2Mvs7pk/K\n2qe8uqv8mZbV9pj9LuN6M5GyJfWY4wn36KfNAdYDs4EnCLdIrCf8JesXBfMWSd83KTvtn4H3t1l2\nWlbZWWl57c5rVzt9RsH0vLKz9isvb0x6GedJ0bwxdceev2WeJ2X2d1b+2O9dTJ+0870sUi+U+73M\ny9vp/p5In030GptW5vW7zD6JOdZlXBOKll3V9TuvXbFtK6u/Y/ukrH3Kq7vKn2kQt99lXm9iypbU\nY/YkzCj2x6l1LyBcWD4AfKlp+1uA1xXIWyQdwj30hyf/n064x3/RBMp+Typ9MbCkQNl5aVl157Wr\naJ+N1+6s9Lyys/YrL29MeqfOk1b73Kn+budYxqTHtrvM/s7KH/u9i+kTKPa9LPM8aed7mZe3E/3d\nbp9lpcccy6Jll3kOlnGsO3VNaKdsqOb6XaRdMW0rq79j+6Ssfcqru8qfaaPa3e+svHnpsb+zFUmX\n1EPmAVcC9wN3AbcB7wVuAo5p2vbPgS8WyFs0/eXA7YRbN9YC75tA3v2BryX5VhP+KvaSgmVnpeXV\nnZVWpM/y2j1eepGyx9uvvLwx6Z04T7L6JLa/2z2Wsekx7S6zv/Pyx37vYvo773tZ5nnS7vcyLy22\nv2P7LCs95ljmlV3mOVjmsY49HjFlV3H9LtKumLZBOf0du89l7lNMf5f9My12v8u83sSULUmSJEmS\nJEmSJEmSAKhV3QBJkiRJ5TnvvPMag4ODVTdDk1ytVsuNHQ0uJUmSpP7WaDR8a4niFAkufc+lJEmS\nJCmawaUkSZIkKdqUqhtQtYGBgceBfUoqfqRer08vqWxJkiRJ6hmOXJYXWJZdtiRJkiT1DINLSZIk\nSVK0Xgsu5wJDwMPAdmATcCNwZJI+BOxMlm3AQ8AKYBm9ty+SJEmStNvotYBsOfAi4ARgAXAscD0w\nM7XN9cA84ADgOOAm4ELgOmDPbjZWkiRJkhT00oQ+84ABwijlymTdRuDupu22AY+k0u8B7iCMcJ4G\nXFx6SydgYGDAlwpJkiSp4+r1uu+sV0/ppZHLx4CtwInA1AnmvRlYk+SVJEmSJHVZL41cbgU+BHwe\n+DCwCrgd+CohcMyzDlhcWuva5F+UJEmSJO0OemnkEuAy4HnAycANhFtk7wZOL5jfW1AlSZKkJrVa\nrSvL7Nmzq95VVajXgkuAJwmT9lwALCHMEHt+gXwvBdanPl8DPApc3eH2SZIkSZNKo9HoyjI8PFz1\nrqpCvRhcNlsDTE99bjU6eTThlthrUusuAv60xHZJkiRJkhK99MzlHOBy4FKgDjxFmD32LODa1HZ7\nEd6HOSX59zjgHMKrSC5NbXcrcFSBekeAfeKanlm2JEmSJPW9XgouR4C1wNnAIsKMsRuAq4Bzk20a\nhGByI7ADGAZWA2cAl7RTab1en56/lSRJkiQpS7/PZHoU8BHCBEGSJEnS7qjRaHRn3starUa36lJ3\n1Wq13NhxMjxzGcMzW5IkSZK6oN+Dy34fmZUkSZKkntBLz1x22neBw4BphGc3TyBMFCRJkiTtVgrc\n0dgRs2bN6ko96k2O7EmSJEn9rWvPXKp/+cylJEmSJKkrDC4lSZIkSdEMLiVJkiRJ0QwuJUmSJEnR\nDC4lSZIkSdEMLiVJkiRJ0QwuJUmSJEnRDC4lSZKkPler1Sa8zJw5s+pma5KZUnUDqjYwMPA4sE/V\n7ZAkPcNIvV6fXnUjJKkfDA4OdiWPdm+OXBpYSlKv8vosSdIkYnApSZIkSYrWa8HlXGAIeBjYDmwC\nbgSOTNKHgJ3Jsg14CFgBLKP39kWSJEmSdhu9FpAtB14EnAAsAI4FrgfSTxNfD8wDDgCOA24CLgSu\nA/bsZmMlSZIkSUEvTegzDxggjFKuTNZtBO5u2m4b8Egq/R7gDsII52nAxaW3VJLUFQMDA42q2yBJ\nvaper9eqboOU1kvB5WPAVuBE4C7CbbFF3QysSfIaXEqSJEkp7cz8OmPGjM43RH2tl4LLrcCHgM8D\nHwZWAbcDXyUEjnnWAYtLa50kqev8q7wkdUaj4Y0gKl+vPXN5GfA84GTgBsItsncDpxfM77dGkiRJ\nalKr1SbNMnv27Kq7S23qteAS4EnCpD0XAEsIM8SeXyDfS4H1yf/nAz8A7gXuB97T8VZKkiRJk0Sj\n0Zg0y/DwcNXdpTb1YnDZbA0wPfW51ejk0YRbYq9JPm8jjHYuBt4AfBZfxi1JkiRJpemlZy7nAJcD\nlwJ14CnC7LFnAdemttuL8D7MKcm/xwHnEF5FcmmyzcPJAuFdmRuBfYGRUvdAkiRJknZTvRRcjgBr\ngbOBRcBUYANwFXBusk2DEExuBHYAw8Bq4AzgknHKfSUhIF0/TvoIjmpKUi/yD4KSJE0i/T4L3xzg\nVmAZ8MOK2yJJkiRVoTGZZout1WrObtuDarVabuw4GZ65bNezCc9gfhoDS0mSJEkqVb8GlzXCLLM3\nEd6TKUmSJEkqUb/eFruEcDvsasb28VRgXWUtkiRJkqoxqe4xnTVrFps3b666GWpS5LbYfg0uJUmS\nJAWT6plL9abd/ZlLSZIkSVKXGFxKkiRJkqIZXEqSJEmSohlcSpIkSZKiGVxKkiRJkqIZXEqSJEmS\nohlcSpIkSZKiTam6AZIkSZLKVeAVhZrEZsyYwZYtW6puhsGlJEmS1O8GBwerboJK1CvHd7cPLgcG\nBh4H9qm6HZIkSYoyUq/Xp1fdCGl35jOXBpaSJEn9wN/ppIoZXEqSJEmSohUNLucCQ8DDwHZgE3Aj\ncGSSPgTsTJZtwEPACmDZBOqQJEmSJE1SRZ+5XA5sBU4AHgT2B94IzExtcz1wGrAnIRg9HrgQOBl4\nC/B0R1osSZIktTAwMNCoug3dVK/XnQJWPaXICTkP+A1hlHLlONsMATOAE5vWH00Y4fwAcHF7TSzX\n7nYRkiRJUn+YQHDp77t9rhuvIqkVeJ9NkZHLxwijlicCdxFuiy3qZmBNkrcng0tJkiT1B0fyxtdo\nGF+qfEWeh9wKfIgw+jgM3AZ8BjikYB3rgIXtNE6SJElSvFqt1tfL7Nmzq+5iUfyZy8uArwNvAF4J\nHAucSQg6i4xI+qcSSZIkqSL9PnJZ4I5NdcFEZnJ9kjBpzwXAEsJzlucXyPdSYH3q83zgm8Ba4F7g\nasYmBtoBrAIeAL5KCH6/Bbwjlf+DwJebtv8pcCnwLOCeZN1GYEPy/xuL7qQkSZIkaeKKjly2sgZ4\nd+pzqz+HHA0sBj6Xqu964DzCSCjAm4FZwBbCbbeHEYLeFcApwBnAt5Nlb8KI6RFJ3vT2NwInAYcm\naecRXpnyL+3voiRJkiSpiCLB5RzgcsLIYB14ChgAzgKuTW23F+EVJFOSf48DzgGuS/JCuJ12A2OB\nJYQgstlO4E7CKOcVwH8mZc0GLgJ+22L7O4AXNq0vMj4+AuxTYDtJkiT1rpGqGyDt7ooElyOEW1jP\nBhYBUwkB4lXAuck2DUIwuZFwq+owsJow6nhJqqyDCbet5tmL8Hznp5LPnybc3joCfKzF9nsTRkn/\noUDZu6jX69MnmkeSJEmStKsiweU2wihlltOSJU/ek8QzCUHkAYTXmFyXrH+KMHq5qamM0e2nAssJ\nz3JKkiRJkros5pnLdtwHfDQjfQvhGcp9gFsJM9PelaTt5JnB6ej2kiRJkqQKTWS22E74LrAAODG1\n7vd55rOSI4Rbbj+FJEmSpChVv4ey7GXWrFlVd7HofnC5g/Bs5lLCc5xrgNOBzUl6emRyOWFioPTI\nZPPIZd5ttv39Qh9JkiQp361VN6Bsw8PDlQe4/b4AgxUfZkmSJEmSJEmSJEmSJEmSJEmSJEmSJEmS\nJEmqzEeAB4GtwB3Aq3qgng8CPwAeB54AbgEOb9pmiPAawvRyXRfbuLRF/U92sf5bWtS/kzDh6aih\nFumxfdTK64FvAb9O6nh7CXVIkiRJ6mHvIgRvJwEHABcS3tDw3IrruYTwtoiXAC8GvgT8DzC/aZtv\nA/ullhldbONSYEtT/TH9NtH6ZzXV/TJgO/AnqW063UfjOQ44HziBEFy+rYQ6JEmSJPWwHwL/mPpc\nA34O/HWP1VMDHiUEnKOGgG90onGJibZxKTBcYf3NPgY8BuydWjdEZ/uoiNzgstvvuZQkSZJUrmcB\nrwBuTq1rJJ+bb0Gtup5phKBpc2pdAzgK2AT8EriY9kcO223jc5K6HyGMEB7W5frTlgFXEm6pTZdx\nFJ3po44xuJQkSZL6y77AnsBDTesfBub1WD2fAX5FeK5v1HLgFOBI4APAAHBjUlc32ng/8EfAMYSR\nuieAW4EDu1R/2quBgwjBY1on+6hjplRZuSRJkqSuafRYPR8HTiWMwG1Prf966v8PAGuAXwDH0rlJ\na7LaeGeyjHoXsJYwGdHZXag/bRlh/+9qWt+NPpowRy4lSZKk/vJb4GlgbtP6/YGNPVLPmcAnCMHQ\nmpxtf0MY6VvQ5TaOagCrCJPxdLP+acC7gf8oUE9MH3WMwaUkSZLUX7YBdeDo1LoaYYRwZQ/UcxZw\nLiGw/FGBeuYSZkP9dRfbmFYDDiHcvtvN+k8mPLN5eYF6YvpIkiRJksZ1CuEVH+8AXkh4/cWjdH7S\nl7x6vgL8fWr7s4Gnku3npZZpSfo04ALCJDjPB5YQ3gu5DpjapTaeQwgG5wMHE1778QTwe12qf9Rt\nwBUt1pfRR+OZBhyaLDsJM9ceyq6vjpEkSZLU5z4CPEgI5lYCr6qgnpuBL6c+ryfcJrqzaTk3Sd8L\nWEGYBXUHYbbWrxBuI+1WGz9HmIF1O/A4cAPw2i7WD7CI0C9valFWWX3UylGMHaP0cWturyRJkiRJ\nkiRJkiRJkiRJkiRJkiRJkiRJkiRJkiRJkiRJkiRJkiRJkiRJkiRJkiRJlVoMrAK2A3dX3JZesx74\ni6obIUmSJEllGAK+0cHyvgGsAOYDszpYbrsWAjtbLF8usc6lwHCL9XOAvUustxRTqm6AJEmSpEmh\nkSydcgBwLbAhooxnA//bmeb8vzcB96U+b22xzRTGgs8yPFpSuZIkSZJUuSF2Hbl8L/ATwm2tjwHf\nB34nSasB5xICx6cIt78em8rbPDp4brJ+MXAT8CTwW+DfgGkt2vAJ4NfAA4QgdSdwMnBbkvdHwEuA\nI5K6twK3AHMz9m9hUs4hLdKWEkYY/xD4cbLPzwdem5S7JdnPO4AlTXlnJvvxELANWAe8DTgqox8e\nBM5IlbEA+C9ghNDXXwP2S6UPJvu5NMn7JPBNYHrG/kqSJElSJYYYCy4XAE8TAqAFwO8CyxgLBP8W\n2Ay8NUkfJARWhybpc4F7gc8SgqRpwGzgEeAy4MWEwHAdcHVTGx4nBGsHAosYCwrXAK9P8t4K1AmB\n6gBwEHA/2be4jpbz8hZpSwkjpDcm+3AgsBdwNPDOJO8i4F8Jgebobb57ACsJwe6RhID0WOAYYCrh\nucotSR/sx1hwnn7mcmrSDzcAL0vadztwZ6p9g0m/XEk4Fq8hBN//lLG/kiRJklSJIcaCy1cTArH5\nLbarAZuAv2xafxvw76nPqxgbqQP4KCG4nJpa9yZCEPu8VBs2JHWMWpi05T2pde9I1h2RWncm8N8t\n2ttczhOEEcLR5WBCcLmTEEBm2YNwS+s7k8/HADto3U8w/jOX6eDyrYSR0uem0g9M2nN48nmQEFzu\nldrm7+jyREl7dLMySZIkSX2hTggW7yWMLP4ZYeQRwgjcHMLoWtrthEBtPAcl5W5PrVtJCCQPSq1b\nTetnP9PPSW5O/r0/te5Rwi2qeU4hjA6OLj9J1m9N/X/U/sClhFtRH0uWmYzdfnso8DPinis9CPg5\nIWAfNfo53Z/rCbfmjtpEsf3tGCf0kSRJkjRRTwNvAF4HvBF4P3AB4RnEkTbLbLDriOToumbbW6yD\n1pPrNK9rLr+VDbQe4dzWYt2VwLOA04FfEvrlB8CeBeqJ1dw3rfa/yP52jCOXkiRJkopqDmhuAz4J\nvJIwsnci8DBh1OyIpm2PIIx0juc+4BWEYC2dp8Guo5K95HDgIuB7hMmFnmDX21fvITwD+oJx8j9N\nfgB4H/Ainnlb7H7A2ok3uTwGl5IkSZKKGg2EXgN8nDCz6nzgJELw89Mk/fOEGV3fQpjN9TzCc5pf\nbFHWqCsIgeTFhIDscOALwNeB33R4PzrlZ8D7CM9iHgZczq6vRlkB3AVcQ5jQ5wXAm5MFwijpdMIM\ns/vS+t2W30nquZyxCX0uJUwStLKjexPJ4FKSJElSEen3XD5GCJBuJtxC+mngbwivv4AwmcxFhNlT\nHwDeTniNx5qm8tKGCRP4PD/Z7luEWV+XjtOG5ra1u65oequ0pYzNfHsl8CXCpETpPMcn6d8k9NXn\nGBud/T4hUPx2ku+sFnXsIMwwuxX4YZLnV4SJftL1NLev0+8llSRJkiRJkiRJkiRJkiRJkiRJkiRJ\nkiRJkiRJkiRJkiRJkiRJkiRJkiRJkiRJkiRJkiRJkiRpkvg/QCyHwrRniP4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x36de4f90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA5cAAACGCAYAAABJ935GAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1990ddd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA5cAAACgCAYAAACYAphtAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSZIkSZLGiP8HxLiqzha9toAAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x362b8cd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i in np.random.choice(test_inds,3):\n",
    "    make_nt_contribution_fig_5(i,SAVEFIGS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 204,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA5cAAADYCAYAAAB/cxfBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1bd4e710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA5cAAACvCAYAAABpVCq4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x186f5a10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA5cAAACvCAYAAABpVCq4AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/wOsb6K7mIwxHQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x38219550>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i in np.random.choice(test_inds,3):\n",
    "    make_nt_contribution_fig_4(i,SAVEFIGS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}