notebooks/fourier_prior_example_pytorch.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## A standalone example of the Fourier-based prior (PyTorch)\n",
    "In this notebook, we will be running a simple example of the Fourier-based prior, in order to show how it can be applied. We will train a very simple model with and without the Fourier-based prior, and at the end we will show the importance scores on a random sample of input sequences using three different methods of computing attributions.\n",
    "\n",
    "The goal of this notebook is to present a very simple, standalone example of models trained with and without the Fourier-based prior. None of the code in this notebook will rely on libraries/repositories other than very standard and ubiquitous ones (e.g. PyTorch, NumPy, etc.).\n",
    "\n",
    "We'll be training a simple binary model to predict binding of the SPI1 transcription factor. For the sake of simplicity and efficiency, we'll be training with only one output task (i.e. single-task models), with a slightly simpler data processing. Thus, these results won't fully match those presented in the paper (the results in the paper can be reproduced from the other code/notebooks in this repository).\n",
    "\n",
    "For more results, see the corresponding paper [here](https://proceedings.neurips.cc/paper/2020/hash/1487987e862c44b91a0296cf3866387e-Abstract.html).\n",
    "\n",
    "Some of these cells can take awhile to run, and the entire notebook can take on the order of an 30 minutes to complete. Several intermediates are saved along the way, so feel free to run the notebook in pieces, commenting out certain cells that have already been run before."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/users/amtseng/miniconda3/envs/att-priors/lib/python3.7/site-packages/ipykernel_launcher.py:11: TqdmDeprecationWarning: This function will be removed in tqdm==5.0.0\n",
      "Please use `tqdm.notebook.tqdm` instead of `tqdm.tqdm_notebook`\n",
      "  # This is added back by InteractiveShellApp.init_path()\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "926b25fe4db84356b0ce4ce4c54a3e0b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=1.0, bar_style='info', max=1.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "|<bar/>| 0/? [00:00<?, ?it/s]"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import torch\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import scipy.special\n",
    "import scipy.stats\n",
    "import scipy.ndimage\n",
    "import sklearn.metrics\n",
    "import pyfaidx\n",
    "import tqdm\n",
    "tqdm.tqdm_notebook()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Download the data\n",
    "We'll need the following files to train our model:\n",
    "1. Locations of SPI1 binding in the human genome (using the hg38 annotation). We will download the called peaks from the [ENCODE project](https://www.encodeproject.org/). Specifically, we'll be fetching the IDR-thresholded optimal peaks (i.e. peaks that the replicates agreed upon the most), using these regions as our positive (binding) set. We will be using the experiment [ENCSR000BGQ](https://www.encodeproject.org/experiments/ENCSR000BGQ/), which measures SPI1 binding in the GM12878 cell line.\n",
    "\n",
    "2. The hg38 human reference genome. We will be downloading this from the UCSC genome portal.\n",
    "\n",
    "3. The hg38 chromosome sizes. We will also be downloading this from the UCSC genome portal."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Make a directory to store everything\n",
    "!mkdir -p prior_example/data\n",
    "!mkdir -p prior_example/models\n",
    "!mkdir -p prior_example/aux_code"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--2020-10-01 02:19:16--  https://www.encodeproject.org/files/ENCFF071ZMW/@@download/ENCFF071ZMW.bed.gz\n",
      "Resolving www.encodeproject.org (www.encodeproject.org)... 34.211.244.144\n",
      "Connecting to www.encodeproject.org (www.encodeproject.org)|34.211.244.144|:443... connected.\n",
      "HTTP request sent, awaiting response... 307 Temporary Redirect\n",
      "Resolving encode-public.s3.amazonaws.com (encode-public.s3.amazonaws.com)... 52.218.176.91\n",
      "Connecting to encode-public.s3.amazonaws.com (encode-public.s3.amazonaws.com)|52.218.176.91|:443... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 733979 (717K) [binary/octet-stream]\n",
      "Saving to: ‘prior_example/data/peaks.bed.gz’\n",
      "\n",
      "prior_example/data/ 100%[===================>] 716.78K  --.-KB/s    in 0.1s    \n",
      "\n",
      "2020-10-01 02:19:17 (6.88 MB/s) - ‘prior_example/data/peaks.bed.gz’ saved [733979/733979]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Download the peaks from ENCODE\n",
    "!wget https://www.encodeproject.org/files/ENCFF071ZMW/@@download/ENCFF071ZMW.bed.gz -O prior_example/data/peaks.bed.gz"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--2020-10-01 02:19:17--  https://hgdownload.soe.ucsc.edu/goldenPath/hg38/bigZips/hg38.fa.gz\n",
      "Resolving hgdownload.soe.ucsc.edu (hgdownload.soe.ucsc.edu)... 128.114.119.163\n",
      "Connecting to hgdownload.soe.ucsc.edu (hgdownload.soe.ucsc.edu)|128.114.119.163|:443... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 983659424 (938M) [application/x-gzip]\n",
      "Saving to: ‘prior_example/data/hg38.fasta.gz’\n",
      "\n",
      "prior_example/data/ 100%[===================>] 938.09M  92.2MB/s    in 10s     \n",
      "\n",
      "2020-10-01 02:19:27 (90.8 MB/s) - ‘prior_example/data/hg38.fasta.gz’ saved [983659424/983659424]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Download the hg38 reference genome, and unzip it\n",
    "!wget https://hgdownload.soe.ucsc.edu/goldenPath/hg38/bigZips/hg38.fa.gz -O prior_example/data/hg38.fasta.gz\n",
    "!gunzip prior_example/data/hg38.fasta.gz"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--2020-10-01 02:19:58--  https://hgdownload.soe.ucsc.edu/goldenPath/hg38/bigZips/hg38.chrom.sizes\r\n",
      "Resolving hgdownload.soe.ucsc.edu (hgdownload.soe.ucsc.edu)... 128.114.119.163\r\n",
      "Connecting to hgdownload.soe.ucsc.edu (hgdownload.soe.ucsc.edu)|128.114.119.163|:443... connected.\r\n",
      "HTTP request sent, awaiting response... 200 OK\r\n",
      "Length: 11672 (11K) [text/plain]\r\n",
      "Saving to: ‘prior_example/data/hg38.chrom.sizes’\r\n",
      "\r\n",
      "\r",
      "          prior_exa   0%[                    ]       0  --.-KB/s               \r",
      "prior_example/data/ 100%[===================>]  11.40K  --.-KB/s    in 0s      \r\n",
      "\r\n",
      "2020-10-01 02:19:58 (101 MB/s) - ‘prior_example/data/hg38.chrom.sizes’ saved [11672/11672]\r\n",
      "\r\n"
     ]
    }
   ],
   "source": [
    "# Download the hg38 chromosome sizes\n",
    "!wget https://hgdownload.soe.ucsc.edu/goldenPath/hg38/bigZips/hg38.chrom.sizes -O prior_example/data/hg38.chrom.sizes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "sys.path.append(\"prior_example/aux_code/\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "peaks_bed_path = \"prior_example/data/peaks.bed.gz\"\n",
    "reference_fasta_path = \"prior_example/data/hg38.fasta\"\n",
    "chrom_sizes_path = \"prior_example/data/hg38.chrom.sizes\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Prepare the training data\n",
    "We're going to create a rather simple data loader for our binary dataset. We will split each chromosome into windows of length 1000 bp (i.e. the input sequence length for our models), strided across the chromosome with a stride of 50 bp. A 1000 bp window will be considered a \"positive\" binding example if the central 200 bp region overlaps a peak summit, and will be considered a \"negative\" otherwise.\n",
    "\n",
    "We'll save these labels as big BED files: one for training and one for testing. Our test set will consist of chr1. Our training set will consist of all other canonical (non-scaffold) chromosomes other than chrY and chrM."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import the chromosome sizes, ignoring the small scaffolds\n",
    "chrom_sizes = {}\n",
    "with open(chrom_sizes_path, \"r\") as f:\n",
    "    for line in f:\n",
    "        chrom, size = line.strip().split()\n",
    "        if len(chrom) > 5 or chrom in (\"chrY\", \"chrM\"):\n",
    "            continue\n",
    "        chrom_sizes[chrom] = int(size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_chroms = [\"chr1\"]\n",
    "train_chroms = [chrom for chrom in chrom_sizes.keys() if chrom not in test_chroms]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import the peaks BED\n",
    "peaks_bed = pd.read_csv(\n",
    "    peaks_bed_path, sep=\"\\t\", header=None,  # Infer compression\n",
    "    names=[\n",
    "        \"chrom\", \"peak_start\", \"peak_end\", \"name\", \"score\",\n",
    "        \"strand\", \"signal\", \"pval\", \"qval\", \"summit_offset\"\n",
    "    ]\n",
    ")\n",
    "peaks_bed[\"summit\"] = peaks_bed[\"peak_start\"] + peaks_bed[\"summit_offset\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "input_length = 1000\n",
    "window_stride = 50\n",
    "center_overlap_length = 200"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_labels_bed_path = \"prior_example/data/train_labels.bed\"\n",
    "test_labels_bed_path = \"prior_example/data/test_labels.bed\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def write_chrom_labels(\n",
    "    chrom, chrom_size, summit_locs, window_length, window_stride,\n",
    "    window_center_length, labels_bed_fp\n",
    "):\n",
    "    \"\"\"\n",
    "    For a single chromosome, write its labels a BED file.\n",
    "    Arguments:\n",
    "        `chrom`: a single chromosome (e.g. \"chr1\")\n",
    "        `chrom_size`: size of the chromosome (e.g. 100000)\n",
    "        `summit_locs`: an iterable of locations of peak summits\n",
    "            in this chromosome\n",
    "        `window_length`: length of windows \n",
    "        `labels_bed_fp`: open file pointer of the file to write\n",
    "            the coordinates and labels\n",
    "    \"\"\"\n",
    "    coord_starts = np.arange(0, chrom_size - window_length, window_stride)\n",
    "    coord_ends = coord_starts + window_length\n",
    "    centers = coord_starts + (window_length // 2)\n",
    "    values = np.zeros_like(coord_starts)\n",
    "    for summit_loc in summit_locs:\n",
    "        delta = summit_loc - (window_length // 2)\n",
    "        values[np.abs(centers - summit_loc) < (window_center_length // 2)] = 1\n",
    "    for i in tqdm.notebook.trange(len(coord_starts), desc=(\"Writing \" + chrom)):\n",
    "        labels_bed_fp.write(\"%s\\t%d\\t%d\\t%d\\n\" % (chrom, coord_starts[i], coord_ends[i], values[i]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6966fa489f154caca4784c559bb19ddf",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr10', max=2675929.0, style=ProgressStyle(descri…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9be87ffa9f854bab91c408b626b3b759",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr11', max=2701713.0, style=ProgressStyle(descri…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "05e64770125943bc9a3d645c08fe9b9f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr12', max=2665487.0, style=ProgressStyle(descri…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "824f281d02d54b59a4b56c587b3784b5",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr13', max=2287267.0, style=ProgressStyle(descri…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1870661d08274d659bac7e4d7af44cf7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr14', max=2140855.0, style=ProgressStyle(descri…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a42a0c8d05924cbda1b57a1c9ab7ab01",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr15', max=2039804.0, style=ProgressStyle(descri…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2ea4538cf1734f9e96ed1b060d6718fd",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr16', max=1806747.0, style=ProgressStyle(descri…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e95ab4e050b346be83d3bd024fcf67cd",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr17', max=1665129.0, style=ProgressStyle(descri…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "01536c999a6247d6822a932536c82d02",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr18', max=1607446.0, style=ProgressStyle(descri…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b29eb9ca7063428cb9d9df56eb6bcafb",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr19', max=1172333.0, style=ProgressStyle(descri…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "bb85f11faba4488db4edecb84e566a1a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr2', max=4843851.0, style=ProgressStyle(descrip…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9422b1a3283549da80ff80af40ee81d1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr20', max=1288864.0, style=ProgressStyle(descri…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "929d8cb60bd44ca5889d52622a6cc59c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr21', max=934180.0, style=ProgressStyle(descrip…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2ede3d4ca6404931b3ffa8011ad0a1fd",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr22', max=1016350.0, style=ProgressStyle(descri…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "dbaa3015e6304c06be2cee33617cb300",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr3', max=3965892.0, style=ProgressStyle(descrip…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "432c46c1a4794c0fade530356e0ac658",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr4', max=3804272.0, style=ProgressStyle(descrip…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "387ed93e6bf94093b1f1cd85bcf17bdc",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr5', max=3630746.0, style=ProgressStyle(descrip…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "feb462c1ecd34fe4bb76b90b5ef9c2a7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr6', max=3416100.0, style=ProgressStyle(descrip…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "31fa0aecceb24e50b6c09fd91827b99d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr7', max=3186900.0, style=ProgressStyle(descrip…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4d52fc1537be4978894c6f6b3092f885",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr8', max=2902753.0, style=ProgressStyle(descrip…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1926e605170b46e692f638f7a7d84e71",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr9', max=2767875.0, style=ProgressStyle(descrip…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2da1f2bfee6b4f6e8b7974a436c80897",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chrX', max=3120798.0, style=ProgressStyle(descrip…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6598d90e983c4781861a921a1f9fa163",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Writing chr1', max=4979109.0, style=ProgressStyle(descrip…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "# Create the label BEDs for each chromosome set\n",
    "\n",
    "# Clear the files first, if they already exist\n",
    "with open(train_labels_bed_path, \"w\") as f:\n",
    "    pass\n",
    "with open(test_labels_bed_path, \"w\") as f:\n",
    "    pass\n",
    "\n",
    "# Create the label BEDs\n",
    "with open(train_labels_bed_path, \"a\") as f:\n",
    "    for chrom in sorted(train_chroms):\n",
    "        summit_locs = peaks_bed[peaks_bed[\"chrom\"] == chrom][\"summit\"].values\n",
    "        write_chrom_labels(\n",
    "            chrom, chrom_sizes[chrom], summit_locs, input_length,\n",
    "            window_stride, center_overlap_length, f\n",
    "        )\n",
    "with open(test_labels_bed_path, \"a\") as f:\n",
    "    for chrom in sorted(test_chroms):\n",
    "        summit_locs = peaks_bed[peaks_bed[\"chrom\"] == chrom][\"summit\"].values\n",
    "        write_chrom_labels(\n",
    "            chrom, chrom_sizes[chrom], summit_locs, input_length,\n",
    "            window_stride, center_overlap_length, f\n",
    "        )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "def dna_to_one_hot(seqs):\n",
    "    \"\"\"\n",
    "    Converts a list of DNA (\"ACGT\") sequences to one-hot encodings, where the\n",
    "    position of 1s is ordered alphabetically by \"ACGT\". `seqs` must be a list\n",
    "    of N strings, where every string is the same length L. Returns an N x L x 4\n",
    "    NumPy array of one-hot encodings, in the same order as the input sequences.\n",
    "    All bases will be converted to upper-case prior to performing the encoding.\n",
    "    Any bases that are not \"ACGT\" will be given an encoding of all 0s.\n",
    "    \"\"\"\n",
    "    seq_len = len(seqs[0])\n",
    "    assert np.all(np.array([len(s) for s in seqs]) == seq_len)\n",
    "\n",
    "    # Join all sequences together into one long string, all uppercase\n",
    "    seq_concat = \"\".join(seqs).upper()\n",
    "\n",
    "    one_hot_map = np.identity(5)[:, :-1]\n",
    "\n",
    "    # Convert string into array of ASCII character codes;\n",
    "    base_vals = np.frombuffer(bytearray(seq_concat, \"utf8\"), dtype=np.int8)\n",
    "\n",
    "    # Anything that's not an A, C, G, or T gets assigned a higher code\n",
    "    base_vals[~np.isin(base_vals, np.array([65, 67, 71, 84]))] = 85\n",
    "\n",
    "    # Convert the codes into indices in [0, 4], in ascending order by code\n",
    "    _, base_inds = np.unique(base_vals, return_inverse=True)\n",
    "\n",
    "    # Get the one-hot encoding for those indices, and reshape back to separate\n",
    "    return one_hot_map[base_inds].reshape((len(seqs), seq_len, 4))\n",
    "\n",
    "def one_hot_to_dna(one_hot):\n",
    "    \"\"\"\n",
    "    Converts a one-hot encoding into a list of DNA (\"ACGT\") sequences, where the\n",
    "    position of 1s is ordered alphabetically by \"ACGT\". `one_hot` must be an\n",
    "    N x L x 4 array of one-hot encodings. Returns a lits of N \"ACGT\" strings,\n",
    "    each of length L, in the same order as the input array. The returned\n",
    "    sequences will only consist of letters \"A\", \"C\", \"G\", \"T\", or \"N\" (all\n",
    "    upper-case). Any encodings that are all 0s will be translated to \"N\".\n",
    "    \"\"\"\n",
    "    bases = np.array([\"A\", \"C\", \"G\", \"T\", \"N\"])\n",
    "    # Create N x L array of all 5s\n",
    "    one_hot_inds = np.tile(one_hot.shape[2], one_hot.shape[:2])\n",
    "\n",
    "    # Get indices of where the 1s are\n",
    "    batch_inds, seq_inds, base_inds = np.where(one_hot)\n",
    "\n",
    "    # In each of the locations in the N x L array, fill in the location of the 1\n",
    "    one_hot_inds[batch_inds, seq_inds] = base_inds\n",
    "\n",
    "    # Fetch the corresponding base for each position using indexing\n",
    "    seq_array = bases[one_hot_inds]\n",
    "    return [\"\".join(seq) for seq in seq_array]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create a data loader which returns one-hot encoded sequences\n",
    "# and labels\n",
    "class BinaryDataLoader:\n",
    "    def __init__(\n",
    "        self, labels_npy_path, reference_genome_path, batch_size,\n",
    "        reverse_complement=True, seed=20200930\n",
    "    ):\n",
    "        labels_table = pd.read_csv(\n",
    "            labels_npy_path, header=None, sep=\"\\t\",\n",
    "            names=[\"chrom\", \"start\", \"end\", \"value\"]\n",
    "        )\n",
    "        \n",
    "        self.coords = labels_table[[\"chrom\", \"start\", \"end\"]].values\n",
    "        labels = labels_table[\"value\"].values\n",
    "        self.pos_inds = np.where(labels)[0]\n",
    "        self.neg_inds = np.where(~labels)[0]\n",
    "        print(\"Positive coordinates: %d\" % len(self.pos_inds))\n",
    "        print(\"Negative coordinates: %d\" % len(self.neg_inds))\n",
    "        print(\"Total: %d\" % len(labels))\n",
    "        \n",
    "        self.reference_genome_path = reference_genome_path\n",
    "        \n",
    "        self.shuffle_rng = np.random.RandomState(seed)\n",
    "        self.reverse_complement = reverse_complement\n",
    "        \n",
    "    def shuffle_data(self):\n",
    "        self.shuffle_rng.shuffle(self.pos_inds)\n",
    "        self.shuffle_rng.shuffle(self.neg_inds)\n",
    "        \n",
    "    def __len__(self):\n",
    "        return int(np.ceil(len(self.pos_inds) / batch_size))\n",
    "    \n",
    "    def __getitem__(self, index):\n",
    "        \"\"\"\n",
    "        Returns batch of data: a B x L x 4 array of one-hot encoded\n",
    "        sequences, and a B-array of binary labels.\n",
    "        \"\"\"\n",
    "        batch_slice = slice(index * batch_size, (index + 1) * batch_size)\n",
    "        pos_coords = self.coords[self.pos_inds[batch_slice]]\n",
    "        neg_coords = self.coords[self.neg_inds[batch_slice]]\n",
    "        all_coords = np.concatenate([pos_coords, neg_coords])\n",
    "        \n",
    "        labels = np.ones(len(pos_coords) + len(neg_coords))\n",
    "        labels[len(pos_coords):] = 0\n",
    "        \n",
    "        genome_reader = pyfaidx.Fasta(self.reference_genome_path)\n",
    "        seqs = [\n",
    "            genome_reader[chrom][start:end].seq for\n",
    "            chrom, start, end in all_coords\n",
    "        ]\n",
    "        one_hot = dna_to_one_hot(seqs)\n",
    "        \n",
    "        if not self.reverse_complement:\n",
    "            return one_hot, labels\n",
    "        else:\n",
    "            return np.concatenate([one_hot, np.flip(one_hot, axis=(1, 2))]), \\\n",
    "                np.concatenate([labels, labels])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Positive coordinates: 158549\n",
      "Negative coordinates: 55641291\n",
      "Total: 55641291\n",
      "Positive coordinates: 16296\n",
      "Negative coordinates: 4979109\n",
      "Total: 4979109\n"
     ]
    }
   ],
   "source": [
    "batch_size = 64\n",
    "reverse_complement = True\n",
    "train_data_loader = BinaryDataLoader(\n",
    "    train_labels_bed_path, reference_fasta_path,\n",
    "    batch_size, reverse_complement\n",
    ")\n",
    "test_data_loader = BinaryDataLoader(\n",
    "    test_labels_bed_path, reference_fasta_path,\n",
    "    batch_size, reverse_complement\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define the model\n",
    "We'll be using the same binary model architecture defined in the paper."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "def place_tensor(tensor):\n",
    "    \"\"\"\n",
    "    Places a tensor on GPU, if PyTorch sees CUDA; otherwise, the returned tensor\n",
    "    remains on CPU.\n",
    "    \"\"\"\n",
    "    if torch.cuda.is_available():\n",
    "        return tensor.cuda()\n",
    "    return tensor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def smooth_tensor_1d(input_tensor, smooth_sigma):\n",
    "    \"\"\"\n",
    "    Smooths an input tensor along a dimension using a Gaussian filter.\n",
    "    Arguments:\n",
    "        `input_tensor`: a A x B tensor to smooth along the second dimension\n",
    "        `smooth_sigma`: width of the Gaussian to use for smoothing; this is the\n",
    "            standard deviation of the Gaussian to use, and the Gaussian will be\n",
    "            truncated after 1 sigma (i.e. the smoothing window is\n",
    "            1 + (2 * sigma); sigma of 0 means no smoothing\n",
    "    Returns an array the same shape as the input tensor, with the dimension of\n",
    "    `B` smoothed.\n",
    "    \"\"\"\n",
    "    # Generate the kernel\n",
    "    if smooth_sigma == 0:\n",
    "        sigma, truncate = 1, 0\n",
    "    else:\n",
    "        sigma, truncate = smooth_sigma, 1\n",
    "    base = np.zeros(1 + (2 * sigma))\n",
    "    base[sigma] = 1  # Center of window is 1 everywhere else is 0\n",
    "    kernel = scipy.ndimage.gaussian_filter(base, sigma=sigma, truncate=truncate)\n",
    "    kernel = place_tensor(torch.tensor(kernel))\n",
    "\n",
    "    # Expand the input and kernel to 3D, with channels of 1\n",
    "    # Also make the kernel float-type, as the input is going to be of type float\n",
    "    input_tensor = torch.unsqueeze(input_tensor, dim=1)\n",
    "    kernel = torch.unsqueeze(torch.unsqueeze(kernel, dim=0), dim=1).float()\n",
    "\n",
    "    smoothed = torch.nn.functional.conv1d(\n",
    "        input_tensor, kernel, padding=sigma\n",
    "    )\n",
    "\n",
    "    return torch.squeeze(smoothed, dim=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "def binary_logits_to_probs(logit_pred_vals):\n",
    "    \"\"\"\n",
    "    Converts the model's predicted binary logits into probabilities via a\n",
    "    sigmoid on all values.\n",
    "    Arguments:\n",
    "        `logit_pred_vals`: a tensor/array containing the predicted logits\n",
    "    Returns a tensor/array of the same shape, containing the predictions as\n",
    "    raw probabilities by doing a sigmoid. If the input is a tensor, the output\n",
    "    will be a tensor. If the input is a NumPy array, the output will be a NumPy\n",
    "    array.\n",
    "    \"\"\"\n",
    "    if type(logit_pred_vals) is np.ndarray:\n",
    "        return scipy.special.expit(logit_pred_vals)\n",
    "    else:\n",
    "        return torch.sigmoid(logit_pred_vals)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "class BinaryPredictor(torch.nn.Module):\n",
    "\n",
    "    def __init__(self):\n",
    "        \"\"\"\n",
    "        Creates an binary TF binding site predictor from a DNA sequence.\n",
    "        \"\"\"\n",
    "        super().__init__()\n",
    "\n",
    "        # ReLU activation for the convolutional layers and attribution prior\n",
    "        self.relu = torch.nn.ReLU()\n",
    "\n",
    "        # Define the convolutional layers\n",
    "        depths = [4, 64, 64, 64]\n",
    "        conv_filter_sizes = [15, 15, 13]\n",
    "        self.conv_layers = torch.nn.ModuleList()\n",
    "        for i in range(3):\n",
    "            self.conv_layers.append(\n",
    "                torch.nn.Conv1d(\n",
    "                    in_channels=depths[i],\n",
    "                    out_channels=depths[i + 1],\n",
    "                    kernel_size=conv_filter_sizes[i],\n",
    "                    stride=1,\n",
    "                    padding=0  # No padding (AKA \"valid\")\n",
    "                )\n",
    "            )\n",
    "            self.conv_layers.append(self.relu)\n",
    "            self.conv_layers.append(\n",
    "                torch.nn.BatchNorm1d(depths[i + 1])\n",
    "            )\n",
    "\n",
    "        # Compute sizes of the convolutional outputs\n",
    "        conv_output_sizes = []\n",
    "        last_size = (4, 1000)\n",
    "        for i in range(3):\n",
    "            next_length = int(np.floor(\n",
    "                (last_size[1] - (conv_filter_sizes[i] - 1) - 1)\n",
    "            )) + 1\n",
    "            next_size = (64, next_length)\n",
    "            conv_output_sizes.append(next_size)\n",
    "            last_size = next_size\n",
    "\n",
    "        # Define the max pooling layer\n",
    "        self.max_pool_layer = torch.nn.MaxPool1d(\n",
    "            kernel_size=40,\n",
    "            stride=40,\n",
    "            padding=0  # No padding (AKA \"valid\")\n",
    "        )\n",
    "\n",
    "        # Compute size of the pooling output\n",
    "        pool_output_depth = conv_output_sizes[-1][0]\n",
    "        pool_output_length = int(np.floor(\n",
    "            (conv_output_sizes[-1][1] - (40 - 1) - 1) / 40\n",
    "        )) + 1\n",
    "        pool_output_size = (pool_output_depth, pool_output_length)\n",
    "        \n",
    "        # Define the fully connected layers\n",
    "        dims = [pool_output_size[0] * pool_output_size[1], 50, 15]\n",
    "        self.fc_layers = torch.nn.ModuleList()\n",
    "        for i in range(2):\n",
    "            self.fc_layers.append(\n",
    "                torch.nn.Linear(dims[i], dims[i + 1])\n",
    "            )\n",
    "            self.fc_layers.append(self.relu)\n",
    "            self.fc_layers.append(\n",
    "                torch.nn.BatchNorm1d(dims[i + 1])\n",
    "            )\n",
    "\n",
    "        # Map last fully connected layer to final outputs\n",
    "        self.out_map_fc = torch.nn.Linear(15, 1)\n",
    "\n",
    "        self.bce_loss = torch.nn.BCELoss()\n",
    "\n",
    "    def forward(self, input_seqs):\n",
    "        \"\"\"\n",
    "        Computes a forward pass on a batch of sequences.\n",
    "        Arguments:\n",
    "            `inputs_seqs`: a B x L x 4 tensor, where B is the batch size, and\n",
    "                L is the sequence length\n",
    "        Returns the LOGITS of each input as a B-tensor. Note that the logits\n",
    "        are returned in the order according to the input sequences.\n",
    "        \"\"\"\n",
    "        batch_size = input_seqs.size(0)\n",
    "        # PyTorch prefers convolutions to be channel first, so transpose the\n",
    "        # input\n",
    "        input_seqs = input_seqs.transpose(1, 2)  # Shape: B x 4 x L\n",
    "\n",
    "        # Run through convolutions, activations, and batch norm\n",
    "        x = input_seqs\n",
    "        for layer in self.conv_layers:\n",
    "            x = layer(x)\n",
    "        conv_output = x\n",
    "\n",
    "        # Perform max pooling\n",
    "        pooled = self.max_pool_layer(conv_output)\n",
    "\n",
    "        # Flatten\n",
    "        flattened = pooled.view(batch_size, -1)\n",
    "\n",
    "        # Run through fully connected layers, activations, and batch norm\n",
    "        x = flattened\n",
    "        for layer in self.fc_layers:\n",
    "            x = layer(x)\n",
    "        fc_output = x\n",
    "\n",
    "        # Run through last layer to get logits\n",
    "        logits = self.out_map_fc(fc_output)\n",
    "\n",
    "        return logits.view(-1)\n",
    "\n",
    "    def correctness_loss(self, true_vals, logit_pred_vals):\n",
    "        \"\"\"\n",
    "        Computes the binary cross-entropy loss.\n",
    "        Arguments:\n",
    "            `true_vals`: a B-tensor of true binary values\n",
    "            `logit_pred_vals`: a B-tensor containing the predicted LOGITS\n",
    "        Returns a tensor scalar that is the loss for the batch.\n",
    "        \"\"\"\n",
    "        # Convert logits to probabilities\n",
    "        probs = binary_logits_to_probs(logit_pred_vals)\n",
    "        return self.bce_loss(probs, true_vals)\n",
    "\n",
    "    def fourier_att_prior_loss(\n",
    "        self, status, input_grads, freq_limit, limit_softness,\n",
    "        att_prior_grad_smooth_sigma\n",
    "    ):\n",
    "        \"\"\"\n",
    "        Computes an attribution prior loss for some given training examples,\n",
    "        using a Fourier transform form.\n",
    "        Arguments:\n",
    "            `status`: a B-tensor, where B is the batch size; each entry is 1 if\n",
    "                that example is to be treated as a positive example, and 0\n",
    "                otherwise\n",
    "            `input_grads`: a B x L x 4 tensor, where B is the batch size, L is\n",
    "                the length of the input; this needs to be the gradients of the\n",
    "                input with respect to the output; this should be\n",
    "                *gradient times input*\n",
    "            `freq_limit`: the maximum integer frequency index, k, to consider for\n",
    "                the loss; this corresponds to a frequency cut-off of pi * k / L;\n",
    "                k should be less than L / 2\n",
    "            `limit_softness`: amount to soften the limit by, using a hill\n",
    "                function; None means no softness\n",
    "            `att_prior_grad_smooth_sigma`: amount to smooth the gradient before\n",
    "                computing the loss\n",
    "        Returns a single scalar Tensor consisting of the attribution loss for\n",
    "        the batch.\n",
    "        \"\"\"\n",
    "        abs_grads = torch.sum(torch.abs(input_grads), dim=2)\n",
    "\n",
    "        # Smooth the gradients\n",
    "        grads_smooth = smooth_tensor_1d(\n",
    "            abs_grads, att_prior_grad_smooth_sigma\n",
    "        )\n",
    "\n",
    "        # Only do the positives\n",
    "        pos_grads = grads_smooth[status == 1]\n",
    "\n",
    "        # Loss for positives\n",
    "        if pos_grads.nelement():\n",
    "            pos_fft = torch.rfft(pos_grads, 1)\n",
    "            pos_mags = torch.norm(pos_fft, dim=2)\n",
    "            pos_mag_sum = torch.sum(pos_mags, dim=1, keepdim=True)\n",
    "            pos_mag_sum[pos_mag_sum == 0] = 1  # Keep 0s when the sum is 0\n",
    "            pos_mags = pos_mags / pos_mag_sum\n",
    "\n",
    "            # Cut off DC\n",
    "            pos_mags = pos_mags[:, 1:]\n",
    "\n",
    "            # Construct weight vector\n",
    "            weights = place_tensor(torch.ones_like(pos_mags))\n",
    "            if limit_softness is None:\n",
    "                weights[:, freq_limit:] = 0\n",
    "            else:\n",
    "                x = place_tensor(\n",
    "                    torch.arange(1, pos_mags.size(1) - freq_limit + 1)\n",
    "                ).float()\n",
    "                weights[:, freq_limit:] = 1 / (1 + torch.pow(x, limit_softness))\n",
    "\n",
    "            # Multiply frequency magnitudes by weights\n",
    "            pos_weighted_mags = pos_mags * weights\n",
    "\n",
    "            # Add up along frequency axis to get score\n",
    "            pos_score = torch.sum(pos_weighted_mags, dim=1)\n",
    "            pos_loss = 1 - pos_score\n",
    "            return torch.mean(pos_loss)\n",
    "        else:\n",
    "            return place_tensor(torch.zeros(1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "def save_model(model, save_path):\n",
    "    \"\"\"\n",
    "    Saves the given model at the given path. This saves the state of the model\n",
    "    (i.e. trained layers and parameters), and the arguments used to create the\n",
    "    model (i.e. a dictionary of the original arguments).\n",
    "    \"\"\"\n",
    "    save_dict = {\n",
    "        \"model_state\": model.state_dict()\n",
    "    }\n",
    "    torch.save(save_dict, save_path)\n",
    "\n",
    "\n",
    "def restore_model(load_path):\n",
    "    \"\"\"\n",
    "    Restores a model from the given path. It will then restore the learned\n",
    "    parameters to the model.\n",
    "    \"\"\"\n",
    "    load_dict = torch.load(load_path)\n",
    "    model_state = load_dict[\"model_state\"]\n",
    "    model = BinaryPredictor()\n",
    "    model.load_state_dict(model_state)\n",
    "    return model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train the models\n",
    "We'll train two models, one with the Fourier-based prior, and the other without. Note that the first batch might take some time to load, while the reference Fasta is being indexed.\n",
    "\n",
    "While the model trained without the prior can converge in just 1 - 2 epochs, the model trained with the prior often requires a few more epochs (due to optimizing over multiple objectives). To keep it simple, we'll just train both models for 1 epoch each. This puts the Fourier-based prior at a slight disadvantage in the comparisons, but we will see that the interpretability of the model trained with the prior is still cleaner. We will also verify that the predictive performances between the two models are reasonably close."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "learning_rate = 0.001\n",
    "freq_limit = 150\n",
    "limit_softness = 0.2\n",
    "att_prior_grad_smooth_sigma = 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_model(\n",
    "    data_loader, model, num_epochs, learning_rate, use_prior=False\n",
    "):\n",
    "    \"\"\"\n",
    "    Trains the model for the given number of epochs.\n",
    "    \"\"\"\n",
    "    model.train()  # Switch to training mode\n",
    "    torch.set_grad_enabled(True)\n",
    "    \n",
    "    optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)\n",
    "    \n",
    "    for epoch_i in range(num_epochs):\n",
    "        losses = []\n",
    "        if torch.cuda.is_available:\n",
    "            torch.cuda.empty_cache()  # Clear GPU memory\n",
    "        t_iter = tqdm.notebook.trange(\n",
    "            len(data_loader),\n",
    "            desc=(\"Epoch %d/%d: Loss: ---\" % (epoch_i + 1, num_epochs))\n",
    "        )\n",
    "        \n",
    "        data_loader.shuffle_data()\n",
    "        for i in t_iter:\n",
    "            input_seqs, output_vals = data_loader[i]\n",
    "            \n",
    "            input_seqs = place_tensor(torch.tensor(input_seqs)).float()\n",
    "            output_vals = place_tensor(torch.tensor(output_vals)).float()\n",
    "\n",
    "            # Clear gradients from last batch if training\n",
    "            optimizer.zero_grad()\n",
    "            \n",
    "            if use_prior:\n",
    "                input_seqs.requires_grad = True  # Set gradient required\n",
    "                logit_pred_vals = model(input_seqs)\n",
    "                # Compute the gradients of the output with respect to the input\n",
    "                input_grads, = torch.autograd.grad(\n",
    "                    logit_pred_vals, input_seqs,\n",
    "                    grad_outputs=place_tensor(\n",
    "                        torch.ones(logit_pred_vals.size())\n",
    "                    ),\n",
    "                    retain_graph=True, create_graph=True\n",
    "                    # We'll be operating on the gradient itself, so we need to\n",
    "                    # create the graph\n",
    "                )\n",
    "                input_grads = input_grads * input_seqs  # Gradient * input\n",
    "                input_seqs.requires_grad = False  # Reset gradient required\n",
    "                loss = model.correctness_loss(output_vals, logit_pred_vals) + \\\n",
    "                    model.fourier_att_prior_loss(\n",
    "                        output_vals, input_grads, freq_limit, limit_softness,\n",
    "                        att_prior_grad_smooth_sigma\n",
    "                )\n",
    "            else:\n",
    "                logit_pred_vals = model(input_seqs)\n",
    "                loss = model.correctness_loss(output_vals, logit_pred_vals)\n",
    "\n",
    "            loss.backward()  # Compute gradient\n",
    "            optimizer.step()  # Update weights through backprop\n",
    "            \n",
    "            losses.append(loss.item())\n",
    "            t_iter.set_description(\n",
    "                \"Epoch %d/%d: Loss: %6.4f\" % (epoch_i + 1, num_epochs, loss.item())\n",
    "            )\n",
    "        print(\"Average loss: %6.4f\" % np.mean(losses))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict_model(data_loader, model):\n",
    "    \"\"\"\n",
    "    Predicts data from the model, and returns the true values\n",
    "    and the predicted probabilities.\n",
    "    \"\"\"\n",
    "    model.eval()  # Switch to evaluation mode\n",
    "    torch.set_grad_enabled(False)\n",
    "    \n",
    "    true_vals, pred_vals = [], []\n",
    "    \n",
    "    t_iter = tqdm.notebook.trange(\n",
    "        len(data_loader), desc=\"Loss: ---\"\n",
    "    )\n",
    "    for i in t_iter:\n",
    "        input_seqs, output_vals = data_loader[i]\n",
    "        true_vals.append(output_vals)\n",
    "\n",
    "        input_seqs = place_tensor(torch.tensor(input_seqs)).float()\n",
    "        output_vals = place_tensor(torch.tensor(output_vals)).float()\n",
    "\n",
    "        logit_pred_vals = model(input_seqs)\n",
    "        loss = model.correctness_loss(output_vals, logit_pred_vals)\n",
    "\n",
    "        t_iter.set_description(\"Loss: %6.4f\" % loss.item())\n",
    "        pred_vals.append(\n",
    "            binary_logits_to_probs(logit_pred_vals.detach().cpu().numpy())\n",
    "        )\n",
    "    return np.concatenate(true_vals), np.concatenate(pred_vals)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "def show_performance(true_vals, pred_vals, acc_thresh=0.5):\n",
    "    \"\"\"\n",
    "    Shows accuracy, auROC, and auPRC.\n",
    "    \"\"\"\n",
    "    pos_mask = true_vals == 1\n",
    "    neg_mask = true_vals == 0\n",
    "    \n",
    "    pos_right = np.sum(pred_vals[pos_mask] > acc_thresh)\n",
    "    neg_right = np.sum(pred_vals[neg_mask] <= acc_thresh)\n",
    "    pos_acc = pos_right / np.sum(pos_mask)\n",
    "    neg_acc = neg_right / np.sum(neg_mask)\n",
    "    acc = (pos_right + neg_right) / len(true_vals)\n",
    "    \n",
    "    auroc = sklearn.metrics.roc_auc_score(true_vals, pred_vals)\n",
    "    \n",
    "    precis, recall, thresh = \\\n",
    "        sklearn.metrics.precision_recall_curve(true_vals, pred_vals)\n",
    "    auprc = sklearn.metrics.auc(recall, precis)\n",
    "    \n",
    "    print(\"Accuracy: %.2f%%\" % (acc * 100))\n",
    "    print(\"Positive accuracy: %.2f%%\" % (pos_acc * 100))\n",
    "    print(\"Negative accuracy: %.2f%%\" % (neg_acc * 100))\n",
    "    print(\"auROC: %.3f\" % auroc)\n",
    "    print(\"auPRC: %.3f\" % auprc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "device = torch.device(\"cuda\") if torch.cuda.is_available() \\\n",
    "        else torch.device(\"cpu\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "prior_model_path = \"prior_example/models/prior_model.pt\"\n",
    "noprior_model_path = \"prior_example/models/noprior_model.pt\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Instantiate new models\n",
    "prior_model = BinaryPredictor()\n",
    "prior_model = prior_model.to(device)\n",
    "\n",
    "noprior_model = BinaryPredictor()\n",
    "noprior_model = noprior_model.to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "865f3294b85c412c99c1d25e17e8d71f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Epoch 1/1: Loss: ---', max=2478.0, style=ProgressStyle(de…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Average loss: 0.2614\n"
     ]
    }
   ],
   "source": [
    "train_model(\n",
    "    train_data_loader, prior_model, 1, learning_rate, use_prior=True\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "save_model(prior_model, prior_model_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3ec662cb931c4fc59ea2eec3a1fd6587",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Epoch 1/1: Loss: ---', max=2478.0, style=ProgressStyle(de…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Average loss: 0.1869\n"
     ]
    }
   ],
   "source": [
    "train_model(\n",
    "    train_data_loader, noprior_model, 1, learning_rate, use_prior=False\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "save_model(noprior_model, noprior_model_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load in saved models\n",
    "prior_model = restore_model(prior_model_path)\n",
    "prior_model = prior_model.to(device)\n",
    "\n",
    "noprior_model = restore_model(noprior_model_path)\n",
    "noprior_model = noprior_model.to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1bb653f69c75415bbfe98320656a2836",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Loss: ---', max=255.0, style=ProgressStyle(description_wi…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e80ac4baaed548a4a71d337fae0af0b5",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, description='Loss: ---', max=255.0, style=ProgressStyle(description_wi…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "prior_true_vals, prior_pred_vals = predict_model(test_data_loader, prior_model)\n",
    "noprior_true_vals, noprior_pred_vals = predict_model(test_data_loader, noprior_model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Performance with prior\n",
      "Accuracy: 93.48%\n",
      "Positive accuracy: 92.66%\n",
      "Negative accuracy: 94.29%\n",
      "auROC: 0.982\n",
      "auPRC: 0.980\n",
      "\n",
      "Performance without prior\n",
      "Accuracy: 94.61%\n",
      "Positive accuracy: 96.46%\n",
      "Negative accuracy: 92.75%\n",
      "auROC: 0.987\n",
      "auPRC: 0.985\n"
     ]
    }
   ],
   "source": [
    "# Compare predictive performance\n",
    "print(\"Performance with prior\")\n",
    "show_performance(prior_true_vals, prior_pred_vals)\n",
    "print(\"\")\n",
    "print(\"Performance without prior\")\n",
    "show_performance(noprior_true_vals, noprior_pred_vals)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Compare interpretability\n",
    "We'll show the importance score tracks for a random sample of the test-set input sequences, to visually confirm the improved signal-to-noise ratio. More sophisticated methods of quantifying the improved interpretability can be found in the other notebooks of this repository (see paper for details).\n",
    "\n",
    "We will show the importance scores using input gradients, DeepSHAP, and _in silico_ mutagenesis.\n",
    "\n",
    "Note that SPI1 binding is a relatively simple task, because of the straightforward motif and binding mode. Although we will see an improvement in the interpretability of models trained with the Fourier-based prior, these improvements are still small compared to the improvements we get when we train on more complex tasks (see the paper for examples of these complex tasks)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To use DeepSHAP, we'll need to install the library. If you don't want to install this, then comment out all DeepSHAP-related cells (including the next four)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "# The code below will clone and install the DeepSHAP repository\n",
    "!git clone https://github.com/amtseng/shap.git prior_example/aux_code/shap\n",
    "!pip install prior_example/aux_code/shap"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "import shap"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--2020-10-01 02:45:02--  https://raw.githubusercontent.com/amtseng/fourier_attribution_priors/master/src/extract/dinuc_shuffle.py\n",
      "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 151.101.196.133\n",
      "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|151.101.196.133|:443... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 8137 (7.9K) [text/plain]\n",
      "Saving to: ‘prior_example/aux_code/dinuc_shuffle.py’\n",
      "\n",
      "prior_example/aux_c 100%[===================>]   7.95K  --.-KB/s    in 0s      \n",
      "\n",
      "2020-10-01 02:45:02 (37.4 MB/s) - ‘prior_example/aux_code/dinuc_shuffle.py’ saved [8137/8137]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Download the code for performing dinucleotide shuffles\n",
    "!wget https://raw.githubusercontent.com/amtseng/fourier_attribution_priors/master/src/extract/dinuc_shuffle.py -O prior_example/aux_code/dinuc_shuffle.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "from dinuc_shuffle import dinuc_shuffle"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We'll also need some code to visualize the importance score tracks."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--2020-10-01 02:45:02--  https://raw.githubusercontent.com/amtseng/fourier_attribution_priors/master/src/plot/viz_sequence.py\n",
      "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 151.101.196.133\n",
      "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|151.101.196.133|:443... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 7081 (6.9K) [text/plain]\n",
      "Saving to: ‘prior_example/aux_code/viz_sequence.py’\n",
      "\n",
      "prior_example/aux_c 100%[===================>]   6.92K  --.-KB/s    in 0s      \n",
      "\n",
      "2020-10-01 02:45:03 (34.0 MB/s) - ‘prior_example/aux_code/viz_sequence.py’ saved [7081/7081]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Download code for visualizing importance scores\n",
    "!wget https://raw.githubusercontent.com/amtseng/fourier_attribution_priors/master/src/plot/viz_sequence.py -O prior_example/aux_code/viz_sequence.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "import viz_sequence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "def interpret_input_grad(model, input_seq):\n",
    "    \"\"\"\n",
    "    Computes input gradient x input for an L x 4 one-hot\n",
    "    encoded sequence.\n",
    "    \"\"\"\n",
    "    assert input_seq.shape == (input_length, 4)\n",
    "    torch.set_grad_enabled(True)\n",
    "    input_seq_np = input_seq\n",
    "    input_seqs = place_tensor(torch.tensor([input_seq])).float()\n",
    "    input_seqs.requires_grad = True\n",
    "    pred_logits = model(input_seqs)\n",
    "\n",
    "    model.zero_grad()\n",
    "    input_grads, = torch.autograd.grad(pred_logits, input_seqs)\n",
    "    return input_grads.cpu().numpy()[0] * input_seq_np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "def interpret_deepshap(model, input_seq):\n",
    "    \"\"\"\n",
    "    Computes DeepSHAP scores for an L x 4 one-hot\n",
    "    encoded sequence.\n",
    "    \"\"\"\n",
    "    assert input_seq.shape == (input_length, 4)\n",
    "    torch.set_grad_enabled(True)\n",
    "    \n",
    "    class WrapperModel(torch.nn.Module):\n",
    "        def __init__(self, inner_model):\n",
    "            super().__init__()\n",
    "            self.inner_model = inner_model\n",
    "        def forward(self, input_data):\n",
    "            return torch.unsqueeze(self.inner_model(input_data), dim=1)\n",
    "\n",
    "    # Create wrapper model whose sole output is B x 1 tensor, which DeepSHAP\n",
    "    # requires\n",
    "    wrapper_model = WrapperModel(model)\n",
    "    \n",
    "    def bg_func(input_seq):\n",
    "        if not input_seq:\n",
    "            return place_tensor(torch.zeros((10, input_length, 4))).float()\n",
    "        else:\n",
    "            return place_tensor(torch.tensor(\n",
    "                dinuc_shuffle(input_seq[0].cpu().numpy(), 10)\n",
    "            )).float()\n",
    "\n",
    "    explainer = shap.DeepExplainer(\n",
    "        model=wrapper_model,\n",
    "        data=bg_func\n",
    "    )\n",
    "\n",
    "    input_seqs = torch.tensor([input_seq]).float().cuda()\n",
    "\n",
    "    # We'll hide some of the internal DeepSHAP messages just for aesthetic purposes\n",
    "    scores = explainer.shap_values(input_seqs)[0]\n",
    "    return scores * input_seq"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "def interpret_ism(model, input_seq):\n",
    "    \"\"\"\n",
    "    Computes in silico mutagenesis for an L x 4 one-hot\n",
    "    encoded sequence.\n",
    "    \"\"\"\n",
    "    assert input_seq.shape == (input_length, 4)\n",
    "    torch.set_grad_enabled(False)\n",
    "    mutations = np.tile(input_seq, (len(input_seq) + 1, 1, 1))\n",
    "    inds = np.arange(len(input_seq))\n",
    "    mutations[(inds + 1, inds)] = 0  # First one is the original\n",
    "\n",
    "    mutations = place_tensor(torch.tensor(mutations)).float()\n",
    "    pred_logits = model(mutations).detach().cpu().numpy()\n",
    "    return np.expand_dims(pred_logits[0] - pred_logits[1:], axis=1) * input_seq"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Pick a few random positive sequences from the test set\n",
    "rng = np.random.RandomState(20200930)\n",
    "rand_inds = rng.choice(test_data_loader.pos_inds, size=5, replace=False)\n",
    "pos_coords = test_data_loader.coords[rand_inds]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr1:208342950-208343950\n",
      "-------------------------------\n",
      "With prior:\n",
      "Input gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DeepSHAP\n",
      "Warning: unrecognized nn.Module: BinaryPredictor; using regular gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ISM\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No prior:\n",
      "Input gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DeepSHAP\n",
      "Warning: unrecognized nn.Module: BinaryPredictor; using regular gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABIYAAACMCAYAAAAJO6J1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nO3dd3gU1f4G8Pdk03tCSAKEXpSiqIBSlCoKiAS9FizY8HIVbFxFFBUCCFixoXixYgPB/hMLoCA2pIMgBpBeUighve77++NsdhMIfSFA3s/z5Mns2TMzZ9qZM9+dOWNIQkREREREREREqh6fyi6AiIiIiIiIiIhUDgWGRERERERERESqKAWGRERERERERESqKAWGRERERERERESqKAWGRERERERERESqKAWGRERERERERESqKN/KmGlMTAzr1atXGbMWERERERERETkjLVmyZBfJ6kczTqUEhurVq4fFixdXxqxFRERERERERM5IxpjNRzuOHiUTEREREREREamiFBgSEREREREREamiFBgSEREREREREamiFBgSEREREREREamiFBgSEREREREREamiFBgSEREREREREamiFBgSEREREREREamiFBgSEREREREREamijjswZIypbYyZa4xZY4xZbYy53xsFExERETkZhnw3BPd+e29lF0NERESkUnjjjqFiAA+SbAqgLYDBxphmXpiuiIiIyAn34h8vYuLCiZVdDBEREZFKcdyBIZI7SS51DWcBWAOg1vFOV0RERERERERETiyv9jFkjKkH4HwAf3hzuiIiIiIiIiIi4n1eCwwZY0IBfArgAZKZFXw/0Biz2BizOD093VuzFRERERERERGRY+SVwJAxxg82KPQhyc8qykNyMsnWJFtXr17dG7MVEREREREREZHj4I23khkAbwFYQ3LC8RdJREREREREREROBm/cMdQBQH8AXY0xy11/vbwwXREREREREREROYF8j3cCJH8BYLxQFhEREREREREROYm8+lYyERERERERERE5fSgwJCIiIiIiIiJSRSkwJCIiIiIiIiJSRSkwJCIiIiIiIiJSRSkwJCIiIiIiIiJSRSkwJCIiIiIiIiJSRSkwJCIiIiIiIiJSRSkwJCIiIiIiIiJSRSkwJCIiIiIiIiJSRSkwJCIiIiIiIiJSRSkwJCIiIlVT7naATvdHZ5lhERERkapCgSERERGpevYsAb5IQMmKEe6kopIiBI8NxtBZQyuxYCIiIiInlwJDIiIiUvVs+z8APijYMsOdVFBSgLziPMz6Z1bllUtERETkJFNgSERERKqe7V8BcKIga6M7qaC4AADghB4pExERkapDgSERERGpejKTAQAFPoHupIISV2DIqcCQiIiIVB0KDImIiEjVUpQJOAuB4AQUOulOLiwpBACUsKSySiYiIiJy0nklMGSMedsYk2aMWeWN6YmIiIicMLnbgfCzgb5bUVCrrzu59FGyYmdxZZVMRERE5KTz1h1D7wLo4aVpiYiIiJw4BelAXFeAREFcF0+y61Gy0juHRERERKoCrwSGSM4HsMcb0xIRERE5ofLTgZh2gDEoCG7gTs4pzAEA5BXnVVbJRERERE469TEkIiIipy2n0+l+BOyIFaQBoTYgVGgc7uSsgiwAQHZhttfKJyIiInKqO2mBIWPMQGPMYmPM4vT09JM1WxERETmDjfppFALHBh4+Y1n5aUBgdQDAjn1b3MkbM+yr6/OL821CSSHgLALi4wFj7F98vFfKLSIiInKq8D1ZMyI5GcBkAGjdujUPk11ERETksI7pDWL5aYB/FADg711/u5PXpK/x5HGWAF/WARwhQGqqJ73ssIiIiMgZ4KQFhkRERES8rcRpA0MkYYw5dGZnMTDvCiA/FfANBQDkl+lPKKswy5M39QegOBcozvF6mUVEREROJd56Xf1UAL8DOMsYs80YM8Ab0xURERE5lNL+gI6oX6Ad3wJp84HsDQCcAIDcMv0TlZvGlhlAcZYCQyIiInLG88odQyRv8MZ0RERERI5GaTAnsyATYQFhh878z5uAM98+HkYbGCr7BrLSt5IBAHYvAurfZvNjmpdLLSIiInLq0FvJRERE5LSVUZABANhXsO/wmfcsBurdBAREA7DdHeaVdjQNIKfs3UGFe4B27wBtpwBxcZ70ssMiIiIiZwAFhkREROS0tStnFwB7x9AhOUsAHz+g/QdAtQsA2sBQQXGhO0vp28h8ASCqJVCYAZTkAikpnumsnAF8Xss+liYiIiJyBlBgSERERE5bqTn2LWH78g9zx1DeDqDWlbbPIOOL0juG8ks8fQzlFuUCAGr5AohoATiCbDCprMX32GktGuwOLomIiIiczvRWMhERETltpWTbu3ky8jMOnTFnExB5HuAbUhoTAgA4t33lHi5wdURd0xdgWBMYRwDAMoGhSACZyUBoQyAvBTjcW9BERERETgO6Y0hEREROW6WvmC8NEB1U3g4gvIkdNj4AbFCnWZm4j9PVIXWggQ3+uPO61AdQ93qgz3qg4e3HX3gRERGRU4ACQyIiInLa25G149AZnIVAcIIdNr7uu31yytw9tCdvDwDAzwDwq+ANZzUANBlshxsPPr4Ci4iIiJwiFBgSERGR097O7J2HzuAsBByBdtj4oLQJlOn0ZCntfNrP4MC+hQCgLoCIZnY4rOFxlVdERETkVKHAkIiIiJz2tmZuLZ9AAiWeN47BWejqdBr2biFzYGCoyFkEAPADPHnLqh3jSS/TabWIiIjI6UyBIRERETntrUhZ4flAAj9fDXwcAJTYu4DgLAaM44DxssoEhpxOzwf3E2YfOewfAETV8ASEWOy9wouIiIhUIgWGRERE5LSXWZDp+bBzFrB3BZBwVZnHxxwAXIGff94Eimyn1dll31Dm+r4IAFjkTnWPFxCAcq80ExERETkD6HX1IiIiclrKzPcEg0pY4vli6wyg0xdAeFNPmo+/vWuoVOEeICAa2WXuGCpVRAAlRQd+4VdBv0MiIiIipzndMSQiIiKnFSedMKMMvln3TcUZsjfY182X7UDaxw9wlgn25NnX2+fsFxgK8g1CIQEUZeIAxa7A0mfxwOcJx74AIiIiIqcQBYZERETktLIpYxMAYFX6KgBAgCOgfIbQBkDZO4gAe8dQfornc+bfAMq/rh4AIgIikE+A2esPnHGBq3+h/FSgJOdYiy8iIiJySlFgSERERE4rK1NXAgBWp60GALSMa1k+gyMUMPs99hUQA2SVCfak/QQAyNsvMBQVFIUdxQAy1x745rHMXTbAFBh3vIsgIiIicspQYEhERERODGex7QQ6d4dXJld7Qm2M/mk0vlv/HQDgp802uBMfFl/xCJ/FAx+H2uGQekDGSs9bylJmAwBy93uULDooGjuKAbNvlSdvqW077Gvur06ByBmtpNAGUvc/BkrFxwPG2L/4gxx/IiJy2lBgSERERDzoBIq98JiUswj4KRFInQusSgJK8g6ZfeJE4OqrgV69gJ9+OvD7guICbMvahud/fx7TVk0DAOzN3wsAqBFaAwBQXNq5dEm2fatY2Ue+QurYt5WVvqUsPxX4yAfFsI2hIN8gAED1kOooBpCf/gfgF1a+EJtQ/q4jkTNR7g5gXi9gy3Tgpz7lO20XEZEzklfeSmaM6QHgJQAOAG+SfMob0xURES/JXAvsWw2ENQIiz6ns0sipKnUusPljIOoCoO51gH+kTY+PB1JT7XBcHJByiDtmSvN2BPDZUiDqHMBZADiCDjrKkiWAwwG89569AaGogheCzds0D8B+r6V3iQiIcH8XHRQNZG2wr6cPjLMBIABwBAA5m4AN75YZ0z5HFmwAf78g5BXnIS7EPia222kQ/NvN5We0A8CG94ALnjn48kvVlrkO2LfKdn4edW5ll+bYrBwBdPwM8A0FzroP8KngciElxR6spcOwx+3UqUBaGtCkCXDllZ4sIiJyajvuwJAxxgHgVQDdAWwDsMgY8xXJv4532iIip6sNezegyFmEs6qdVdlFAda+BoQ1Bmr2so/1FGUDfqGVXSqP4jygOAvwjyr/Fik5sXI2A5nJts+cahcBvkHA+jeAi94AfEOOblqF+4D0n+0r3um6uyAGQGQze1E5zdU59I0lFY6+bBlw1VVA6CF2y3u/vbfc5/iQeKTk2AvSyEAbwHIHhkLrAdkbgb5bPPMGgIgWwILbD5h2kI+Bv+uOoajAKADAGoaj9uZp5TNuAbD+dSB77cELeppzOp3w8dEN5RVaO8lGOgLjgbSfgaYPAsE1y3w/GQira+vajD9Pbl17NMHbsorzgMy/ABggopm9o64kHyCBz2vaZbg++4jmPbj2Nxi7uAdCQg38/RUUOkB+mg0s+/jZYHlCIhB+FpAyF9i73K4wv3CgekcbyA5O0EqsArZtA1avBvz9gQsvBEKO8vQr4i3euGPoQgDrSW4AAGPMNACJAE5aYCgjPwMhfiFIyU7Bd+u/Q+d6ndG4WuOD5ieB5GQgx3V3eWAg0LQpcLB20JYtQEaGHe/9Dc9gyrpnMeCCAXjq0pN3Y5TT6cSU5VOQU5yDO867AzuzdyLMPwyxobGHHrFgD1CQBjidgI8DCK5jG/8uJDHom0FYtG0RXun1CtrVbnfwaRXnALuXAHDaE5VvOBB9PrBrIVC4FwDtIwhR59qT2REs0+aMzYgIikC0X6DnMQNHEOAbfPiVcoRyi3IR4AiAw8fhmq/9NauoyC5GRAQQFnbgeEUlRfD18YUB7HIBtm+Joz1J710J5GwEjK+dTlANoFrrg2bfnbsbu3J3oXF048M2zhdtX4TlKctxTb12iCrZa6dvfICQuvaxicJ9QHEuANrGnn8UCODRHx7F2t1rkdQpCWPmj4Gfww9vdn0cwSy0eX38bGOlpMD+8ukssdMNiAHCyx9bpOcNzr6+Fa+eguICGGPg7/A/unW3P9I2rJyFAHwA//Dyj3o4S4D1k+3yG4d9ZXVCYvmG+xFw0onM/ExEBEbAHGp7ZyYDWz8HGt0JlBRgw5IkbK1xFegIRpf3ugAAvuj3BRLPSgTydwFrXwHq9gNJnP9OV6zITMXITiOR1GmEXTZj7Ho+YLGJ7MJshPqHYvrq6ViVtgqD2wxGbGgsSLr37YNK/w1oNBD4IsF194SPvUDP2QIUZcFu80AgrEGF8wdg65LCDLuPOfyBoFo2b0ke4CxCidMgMycYTvjCx8ceU75HcoZZ9jBw9hB7XJR9lbizyLWdAfgEVPyLNWDLs/MHwLh6EGYJENfF87jQkSCBon2u+Ru7TzkCbGPd6epfo6QQiLnQlvMY5OcD69fbesfHB6heHai53245bdU0PDLnEVzRuDee7jQRhYV2lwgJsQ3GgxV99247fUfeP6iGxfAPCgZAu31q9a54xC0zbJ87NS6znwv3AQiC+wnzlUlA6jyg+zz7uYK7A+z8CZYUwmfJA0C7d2zizisBh58NoqT/DMS0A7BfRz77adwY+PVXoGdPu0wVNYzX7VkHXx9f9+Nit513G5761Z6Ho4JsMCdxaiJyinIw97JHUHt+XyC6VfmJ1L8Z2LMYKMm1HVSX5AJwIiSsLvxcnVWH+NuZf1cYhsuCQoDibFdn06lAGgD/CGDbl/aYOYyikiLkFeUhPDD8sHmP26aP7bHpF4HCrK1YmJuDuFo9kLm5AQpdh1JQENCypd2Uy3cuR05xDtrFngOf4n0AiH/PGo43V36ImOAYpD6UCp/96oNiZzFmrJoBf19/9GzUEwu2LUDdyLpoGN3w0EXL2IQBXw5A3ci6mNx7MnwdFRzPxxrYOAFSU22w0hh7vEZFAa3PzwdytwItx9jzS0LigSe89PlAo3cPrGsPYnPGZmzP3I76kfXR/YPuKHYWY+6tc1EjzNYz01dPx5Pzn0S/5v0wvOPwA8b/809g1y57zHQoBAIOyHEY+9YAO74Fmv7Xfi7Ktvf++0XYRzHJcm/fy8gAsl0xosBAG/stK/+cNggJNQj2XhPuiKSlAe+8A3TubNt4c+cC990HhAblA/np9rxgHEBg9SM7N2RvAPJ2utp+PkDE2UBAteMv6PJHgbZv2XONj589p+alAHuX2vPgnmX2fB3eyOYvyrSBoqOQmgqsXWvXg8MB1KoF1K9fPg9J5BXlIcgvyNPGKcqyZSldV8G1vNoWP5R1u9dh+urp6FizB3zTW8HpusyIirLXZyfbsp3LMPbnsehSrwsGXzjYnZ5dmI29eXtRK6yW14LnM2faevnyy+12uX/mw9iY+Tcmtb0NCYGu6zVnCRDR1N7xXaooy15fwLWPBlSz++nRKq13D1Lnfvn3l1i6cykGthqIWuG1ANg2/cy1M1E9uDouqXeJO++fqX/i1y2/okf9PnDk1URxsd0HY2KA4GDbVtm3z+YNCLBtoLJVqNPpxIy/ZiCvOA83trgR/r6uhk/udqA4BySQkx+CAt+aMMYgNPTgbaMjsXAhkJVlqznSBuVCQ4GdO+Fuf0VG2v0wLc2uJqfTtm0bNLDb7Vh9s/YbzN8yH7e3vB0vLHgBKdkpeO2K11Az/MivV7IKsgAAYaXX0HRdvzmCjj2gTPK4/gBcA/v4WOnn/gAmHmqcVq1akSRvGX8JkQQiCbyxz3KWKk1DEsidP5AkF3zyhfv7n8f+q8K8pX91h9YinxhOAvzuqh7uvKHBeSTJwkLywQfpkfYLSfLXj6a6k1qNinBP75prbFp+Ub477fLHa5AZa0iS3z8z3D1e3rsBnukuGOAZ3rXIlv29Ke6kc0eF2Ok94eAd/fd68ia/av8tXOlO6j6mUYXLyjtbl+7P5LC7SYA/dbnYPd6Wd7vagWXDufWL+8nivHLrrfHQBPfw/Xf/h9zyObl7SYXz6vk03Mv8wIjG7vS177u2x4qRFY7XfkgLd3l+HO9ZJ+WWY+VokuSq5blcssR+f/dzfdzfr3+xDjlypF3OWxq6l7n0+xb9+zIry47XZUytitfV8kftdp63j9nZNu/ciWPd5flv//nMySlfNr8kkKk/kST/+Pwbd94/nunt2bT/d6MdyN7MgkVP2LTVc8ndS0mSi9982L1+okZ5ynOja7SikiJ3WuuHGrqHvx91gWef6Bfp2c5rnidJzvtjjjvvGyO7uucRUnaZF9xpy/NxdXdam6EN3MP1kwI8eX++9oDteENSLFlSSJK8ZoRnX9k4xbP8d9/tKWbp9+2GepajUZJ/hdtj0iQ7zspfkplnd0v+9Nxdnol9iAOm22IcyOxNNu8Tl7i//2vC+XagMIslv//bvRy8ry8JcHX9s915Zz87zD38xn/Hk1u/InfMIhfdQ5LMys1yz6/jw57lmP/CNQeU57zxoPPvl0iSb03w1GdNRoR51uv0cDtS1gaWrBhJkvz1ozbu7x99sL1nXV7/B0myuKSYSAJb/a9VuXV27kP1PdtxRJRnHp/Getbbn2NIkntT95C5O0mSU577hVw2nJwRzR0T4z3H0oKBnnW1/VuS5PLP3iX32rqn7H7OH7p78q4aR5JMX7vM/fWDt/ziHq4x2njKNt+13vauJDe8R7J8fT73xWFkzhZyxYjy+/lau4NsXbORLMq1eV/wbLuMN13rNX8XuXy4p2yu+vznj2a48y57tqNnmb/21EdMnkiS3LLkJ3fSvMc7euYx8zqSZO4fY/hv126Vlp3mXrYON9zkzvvLS/e6h9u29cyiNG/C9de5035+zlMPOhz2v9Pp9Gzrextyk93N+dDL/d3pS54+yyYW53PPLHuspKeTn35qkzfOfs2uD5LzR1ziXuYnbphErnuD3Polufh+kmRJ6u8sSZ5E5u7gvNGJ7vLc3msOuXAQuf4tcnZnz4JsmkYCXNHEs/66janrLtsvL99vt/HiIeSyYSTAHAQya84AcsEADnrCcyz5xSSTv95MzqjGPr1y7Hy+bMiH/rOed91Fdhh8O6OeiuLMtTM5/XG7MkvryfE/j3dP57t13zHwcVuHzflnDi96xFNHIWwLObc3t0+xn3/b8hv9Rvtx1jeZ5LQgzhz7CLnvb/JD0CSBPT/oyd4f9iaSwHeWvUMkgePGlZCf1WLqe23Igj0kwBFIIjd9TH4Ipr1e3b0urr+2kJzTlfy/Zu52y7i3HnaXp1lSMNu+0Zat/teK3S78m5zdiZzTle++9Cf5ZSNOeLkJkQT+tOknjn+wk3u6UaPAGz+5kWHjwniTa3fLziaH3r+Lr79O3j7sAfc8fproylD2XHxfA3ddu3HlWnLECBJgnyE1PMfozHPsOv4ktsK6usNQT/ujbpLfgXnuaczRo4pIkot+XMsJE8jffiP/b4SnPi93Ll44qNw2RRJ45/0Xk3FxJMAmSX5s92Y71p5Qm/y+zMG0wbafVv+8yJ00/9k7PN9PLdP+cu3nW7eU8LFH8vj11+SUx152HxPrnm5I0nPcBT0ZZMu2aRpJcuJT692T+uWZm9zjvfvgCHLb/9m6c5mtj35b9xuRBEY/Fc22N11PLnuEnBHNeqMcjBwf6V7G0HGhdh5DricBTuzqORc3K3vO+MTuVyXOEnfaRaNC3O3Id8d/w1mzyM2byat6pnqWed6VnvrTVZ8v/GKmu+x/PNjGUw9+Xtu1M21m7h9PkHlpTP2wm93PSY5P2m3r/IWDmP56dXbtSl54IXnLLZ7ZcecPJMAFrdq4NjL49djHOXIk+d135AuDJ/Htt8nPPyefecZmKVufdx/bnM1fbc7/fv9ftmmx2T3Z/v3LzMNVny+Y9j6ZMpck2XpUuHu9NG1qs61cSb5kT8Xl9vNfB7j2n0/j3O3znBzy8fvWcPZs8ttvyVWrbJay7fN/prjq62WPcfPn95HFBfZ8utouyKfvbeIi126YOKaZZ9u52i1paeR/Hyj0LIerff59m8ttXZy5jn8929RuDx9w6VsPkOl/kL/cwMIVT5NFOVz9Wh/39pp9S1f3pFo0zfZMt7Q+3zGP3PgRSfKxuxa6v173YjP3NCo6tsu2z5e85GkDutst9By7gY952otDHvCcUzdNqH1A+3w1mvJVeynDuZ8udk+r82hP+7xsu6U0LeSGPtyxwzXexLHu6e54Jf6Qy3HD0AvInXNIkvOeuu2A6folwd1O+H3lr+70To94zmHnNNp4wHhIAvngLSTAcb0957BWSWEVlqO0fV52P3/nyc85cyb555/kHf9aeeA8/tOSL0wo4o4d5APP3+VOX/D6dZ7t8cNlnm1+T08S4M72sWRxPkny+TvGce1a+/VVl23w5B1wvqdOGPpvEuDPF7d3r8uVSc3d9S5vjPfkXTeZJDlj1kee9vBjNSpc5or+2vb1tIeajQpyp//Lc/nuWZf/9bTPu4yu507fM93VFispJpfZ67c1n7/MnbZZy3mTPfV5kzLnpdJ2y8aN5NhRmVy6lHz2iTVMTrbjXT7GM4++fQ8sj/+Adtxlm1F8bug8T9nap1e8f2y17dklM+e4v7+ui623SkrI4XfM4aJF5I8/kg+7LgeXr13mWVdl2i3X9ljinkbPnhWsq9s9iReNivGU4Zsy14vOYs94wGIebVznaEc4YALAtRUEhl6pIN9AAIsBLK5Tpw5JcmfqTn52/5X8uhH4Qd2b+dGEJZz34lJOu7knH+4GTmoFvtslgQSY5+fPWd2akAAzqodx5TP9yOfP5tprm3NvALgtFFzf4Sw6AaYFGV7xVFdmhAcyo0kwN08dTH5wCUddEs7ttVpze0IbvtDFn7t9q3OXXxzf61rbzsPXn9/0O4+8qCFXJkXzmvujeXNf8LnODu72rc70gJrcNPA2EuDuGB+umHg1+dETzHqjPv96+UbytZZ8pkOQ+8B6rWM19/D7XeuQAHP9AjjrUrscf8WG8v7bY/loZ8MB3d/hT1/s4pK5GZx6a2Oyf0cWvRDJqd0bkABTwkK4o3MLFjjADW0SmOsL7gkEzxvZjBnhgcwL9+XlYy9hRngg91ULYmZYAAlwl4niZ50akjUDmTLtMjI3hVz+GD+44yyO6AR+3NSHWy+x6y3vKdfBuWIk/7ytDV9rBb5/LvhZj9oc2Ql8vzn47MW1yHAfbnzenzGPORg8HHyjSTV+0rkRGQouvKUVPz0b/LYBOPC2aDa+B3you697OfZFh3L5k1eTzzbhxmubMdcB7goEn+4YQ8aHMveJII6++iH+ct2L/L9/jeaQu87mfZeDf9SO4pMT+zM5MZZNH2vkXubPbm3Oj5qDXyfE8qXbX+e8e2dw6q2v8KUBrfl2S3BV2wSmBoMbI8CP+tcn7+hK5yvhfLV7LNNC6nJrWD3OSbyQBPj02ZP4x28F3LmtmL/1v43bQ8HkaPCl7nXJOlHMHxXIZU/2IB+7lhkv1Obc3i1IgD8+2oT84SXyq8uY/n83kVmbWfxVay6Z0Jt8dxiz/teIH13agAwFl9wYwaQhF/LlNuD9XQbxxwe+4A8PfsFhd7Zkn37ga638+E+ns5nvADdEh3NuDzuPusMT3Mv8RidbeRcacMPFjVhiwM0BoZzWsT4Z4eCa/hfwgxbgjKbguIujyNhgFr0GtnnIn/6Pg/0T/bi1UzMWGzA5JpzpXc9jvgMc3T6MDHOQ0f6cPqAFR3UEZzRx8PmOCWSkL7dNCObLd9ay+8XN1ciXB5LvtOWzXaO5K6I+0yMa8Pu7+3BOPXB+bX9+3bs+Z9cHf6sdxjWXn8OtYeDKDvU4qJetyEZfdhlXtP8PF7f9D5/617+5pOuDnN9pCOdccQEJlDuWfu9Uj/PqgL/XAid0qUXGhTDnzUDO7tnYHo+hkZzX4xwS4Oh2EWQQyBDwktFtmBEeyNxIzz6YVSeUm6bcTU5py5JfXQ2I5U/wp8dbkjNGseTDTlzW7zzu8wNXxAYw6aZYDu8C/nVJPLd/+gj5YUcOGRRHMwKMGQp+07oBGQzmBDu48d+32vVaLZTpnVuy0Ad87uIQ93I82a4aGQSWBIKb7ryVBLg9KIif9WpG3h7Ij4cO4vQ3N/GXGdtZ7THPyXRh50bcEwCujnZwS8emLDHgquphXNqjKddHlZ/Hy5fXJJvWYMkYP37QrR4JMBvB/KB9AzIITA+L5pMfPcLkxFg+2SGSjPQjqwW469qCJg7mLHqeTFnKjA9bce6Vdh8c3SGUjGIWtiMAABWeSURBVPInqwexTVJLZoQHMj/chy9fWo+sFc7iJAc/ufJspofV47Pt49xlf7VjtC1bMLh0Qnfyi/HMe6cRZ/Ww9eAeRHLaJQ3JkPL7+XNd4tzT/dC1HJmBIZxzezvyvzHl9o8x7aLIYJCh8NTnQf5cMLQLedfl3PNMApPeuJnJibF8tls4eX5dclA43+xcgwRYBB++3K0uWTOMmZND+EOvs0iAI9sFk4FgcYgP7+k+kj+M/plznpzHix6MZcij4N09Q/jmv+7gig53c3X/a7n35wnkzEQ+fHFH/t5nHH9PHMfHB5zLLreA4y4BJ3WPYXpATW4JrM2Zt3Uh7wrjiA4x3Nj4Um5o3J3Jfa+gE+Bef3DiZTWYHlaPa2Pq88N7u3Nqc3BFvGs/9wH/mngRuXku+dtAvtwtkukhdbgzqB6fv6Q2GQrumxDKEf+7lsl9q5NLHyKdTnLFSL7eqTpd93by6fa1XPtuIGf3aUMCTEc1Tru4MRkINnm8gXt7jJ90HTPCA5kT5ce+z17GjPBAbqoZxvk3N+PKWDCpXSjpZ7dzq5HnMCM8kIXhhk+3TyD9wS0Rgez0SE1eMBBceXUwixcPI/es4C9DAshZl5O/3cFF19RmLgL50KU+RBI4sDcY8Wg4R75wPZMTY7mx3wV0Asz1BTe2t9todwC4pYOtBzaHB3Jn53NY5ANu8IvhtEsa8+94ux++1tr+H9+uJukA8+HP9y5rTRpwZ3gY0wbeyUIfcFOb+u55jG1XnTRgsfG0KfaGh/KDLo1IX3ssjZ36KJMTY/ndXUHkeyHk1Gh2HdOeGeGB3B3jzz/7NeeeQLBbfweRZINQ+z5uSRZlk4VZ3PtOHLn3T774WjCRBA7rBoY+FsJHn7+GyYmxrP+oJ7A+unNzJjfvy6Sak7htm5MlJeSWlJ1sOjyasQ+Bb50Vxekd7bG06aa2LDbg3kBwYrcg7nJUZ6qjOq98+lJmhAdye3wIt153PvMd4KgOQWR0ADkJ3NrH7h8ZAeDEaxI49FJwWlN/PnljHB/rAs6rF860bucz1xf8p219FhtwVUAtPnT3P9y+pZhD7i90B/r37drHH/q2JQFuOz+BRT5gjh+YHuJw74Ovd4nhyE7gN40C+cgzify697lsO9ZT9718iSdY/Hbnmvbc53B42hRBoXxiwrVMToxl3LAY3j35FiYnxrr38+cwhHv3OOl0kjn7cjjsnfv5de9zufDVhpzXpyUJsOOgQPf8nhregxnhgSyKBv93WRTTUY3bguI5YsoQJifG2v3cATII/F+nWFsv+XsuEmoPCuYH7eqTweADPTwX0ufc73fAPPZF+/Kfa1u4zhlhTOvSkgUOcMJloeTl55IPx3LjVefZejIC/H1cd3LiPRzd73X3xXNmeho/S2zFdBNTrm20O9SPBJjv68eHHr+cGeGB3BUbzpFv3MTkxFimhvi41+vYdtXJILA4sIL9PAC8rtp3TNnpZF4e+XznYO4KTmB6SB33fp4VHcCHXunLr3ufy7VTqnHmlbHchDocN85WOyQ5qnsX/nnRHVx+0e3cE+pPArwt0eFeJw+1O5e/9hrD3/uO59gujbipUVdubNTNXZ8XdAykc/ss0lnCRe+1ZquHg9nuDnBEh0b86/wb+Nf5N/C1S/25B5Hc5RPp3s/3RIRy+hVNyEmeNsXfaMLJk13XUE7yxR5nMyX+XG5KaMVZPe0+sRuRnNGlARntw20fdiOzt5JLh3He463JyUO4ZvwlfPrhhfxl+na+88xcXjakFi+7GUwNMbadgRD+56rvuXR+Flf8lu1pn4cH8/0uDclgcHd4JEd98DCTE2Ntu8XP7lfj2sWRQWBhkIOJz3a326NeKP98vh/5YjMObX8+f+81igv6POmpz2uAu2c/QKYuJ79qwc971mS6I5abQxP45IfDmJwYyxf+XY917we79QcXd2nMXYHgmmhP2ygDYfyka0MyNqh8m6JvHHvcBD7REfzwypp8vRU4v04QH3n2KiYnxnLX/6I4+u1by7XPd4ZV49M3f8R5Ly3jN2MXcNK1g7n8knv4es+r3Nu87Dz+adeYToB7gsBJ3SOY7hvH7QHxHP7OfUxOjGX669Ec8/btTE6M5cKbWnJRPLioBvhLt8b8Mwbc2M2fJZu/JEsKuW96W657407y9Qv4bf9G7HM9OKgX+PW9TclxtzJ7chxbPRTIyGHgsG5BvHxwGDvdCn5dtzE/vHscF132KP+66io6XfVg2fr87+taMMcXnF8nhL0GR7Dv9eAzfasx5FGw/n1wt8/LnjMerD/FXSfeefF/+c2r//CX6dv578HnMuAxsPnd4OTugdyLCKYEhHPrwNtYbMBNflGc3qkhGQp+9p/a5Ov3klPasf3oVu7j/PmOCWSEg2/d8QDn/lDIPbtLeG+HPvz+nq8495Fvec6Isw+4RsysFsihr15l68HXGnDUW7bObDi8rjvvW662Ub7DcFPHJiwx4MrYEA69NZZDLwWTWyew0AfM8gefvLY6zx4E3ngVuL5jczoB7g4DP+1an2khdTmnbgLbPhrPrv3BZzsHcFdADaYH1+bMQVdyanNwbh1H+WvdLucw3wG+0y+aHGOvrcdfXI2sFkBntQC+ellNpofV49bIhnzkrcH8uve5nPxCTYY+4cNqQ8E01/nlF7TnNBvn55pVBXx22K+c/94Gvjt6Fv8z6Cw+cDn4bGcwA+FMRzV+ck8vDu8Cvncu+OqlMUwLrsMtofX4wQ03c13T3lzbvAs/u6Il0wITOHpIW0YOAy8YCP56cSNbRzl8+X032xbZFRTND/rdzuUX3MobrrE3fmRlkfdc9TF//ng75324hcv/1Y9OgBsjArnD1W6ZWbMP33/yR/6YNI+PdWzLpZ0f4JLOQzj7um5cWR1cGevHD3p14caGXbmgaVeuva4HdwWCYy4OJeNCyFv9WbzQ8wNqZQWG2gH4vsznRwE8eqhxSu8YkhOoNAIM2OEj4XSSe1eRqT+TJUUntnwip4K8VHLLZ+TmGfYXwVNIYXEhe7zfg9+t+66yi3Kg0l8HATssp7YNH5FrXrJ3XBRmHvl4B9vOx7v951zq/jWVX9S3d7IW5bp/bZ+6ciqRBFZ7upr9NYx0NVeOXkpWCpEEjvlpjL0Y+fW5Y5rOIQHkt63JrI1kUY79hbd0/bhk5GUQSWD1MaDz75ft+XZGtL1DuLiAP7xhL5bqvWh/zSxdx9tHDnFfSDldV9krVpBjx5K5uWRqaoUlqjQlJeTcueSMGfZOkZKSCjIdpn2SV5jnXuYSZ0UTqABQ4X6ZkUGOH09OmkQ+/zzdF2j727BnA3MKc455384pzOH8TfPd26jU1JVTOXv9bJLkypSVzCrIOuJpHkpBgb1b8GDLc9g24KG+32/fvftuuu+yPhpvvkm+8go5ZQqZWbbacc17WZwnoPbighePfgbHITnZ7qPTp9unCLzmWNrecmYqU5ekD32af/11kPpQvKvMMZhVvT6XLrV3FO5XNZ80Tie5ZQu5du3JL8OxBIaMHe/YGWN8AawF0A3AdgCLANxIcvXBxmndujUXL158XPMVERGR09DCu4DznwO+qg80vBM4Z6Tt9+MjA9xIfJX8FRKnJQIA/H38UfBEgX1e/hjaK/nF+QgaG4QhFw3BC3+8gDevfBMDLhjg3eXxM8CP19i+Q3wCgI8DgZtc35Up86RFk+DPQgzwTwVaPAF8XgPo/isQ2Rx/vWPQfItnkhzpGe/r5K9RK7wWzq9xvjstIwNYsMD2xXTxxWde/7S7c3fD18cXEYERh8+clASMGgWMHGmH5fiU7kyufXfwYOCZZ05Mh7hmlJ1X8RPFh+8rT0REjpgxZgnJg3dqWwHf450pyWJjzD0Avoftsu7tQwWFREREpArz8bMdY5NAVEsbFPos3v11p7qd3MNlgyHHItDXdjS7PXs7AM/by7yqGoD4brYz60N0SHp3m7vtwMJBQNo8oOV4wHUNfnaZl/HVi6hXbrzeZx3YgXhkJNCjx/EV+1RWLfgoOvpNSlJA6ATavdt2OO1w2A6nvemTaz/B+j3rFRQSETkFHHdgCABIfgPgG29MS0RERM5geSmAsxjo9qN9M14ph70loexdIvdeeO/+Yx+Trfu2AvC8vcyrAmHfAunjUy7AdVCtXwE2vgtk/m1fCx7aED79coCxdvndASSRU8DEicB119k3+c6ZU/FbXI/Vv5r9y3sTExGR4+KVwJCIiIjIEfENBQp32zuHCtJs2tUVv5q8eWxzr8xy/Z71AE7QHUO+KP/qekcIgJyD5QZ8HEBD1+Nss9oDkS2AkjzEhcQhNScVbRPaer+MIkciPr78cEoKYmKAH3+svCKJiMjJcfB7nkVERES8rflwYHZH29fQzu+AoizAWVQuS+OoxgCAs6qd5ZVZpuemAwByi3K9Mr1yCgGU5Nnhq1OA67OPfNwOU4FVo4HkV3BRrQsBAM2qN/N+GUVEREQOQYEhEREROXnCGwPd5wON7wbydwE5m4Di8gGbmJAYAECQX5BXZ10jtIZXpwcAyAew72/7eNzRCqkLdP8Z6DYb8a6yxQTHeLd8IkcqJaX0nVp2WEREqgw9SiYiIiInV2h9+xdUA/j+Ipt2vSc4NPnKyfh23bdemVWAIwAFJQUAgITwBK9Ms5xdAFLnAI0HAk5fIKG+5zvX4zhH4qWeL2FYh2HeL5+IiIjIYSgwJCIiIpUjrhPQZRaQ77pTwfWq7BaxLdAitoVXZhETHIPtWfatZAG+AV6ZZjkEUJQN7FkKhNYDUlM935UdPoxA30A0iG7g9eKJiIiIHI4eJRMREZHKE3sxUOcad1DI29rXbn9CpltOk0HAvF7A12cDcXGe9LLDIiIiIqco3TEkIiIiZ6w6EXVO/Ezq9gN2LwLytgM7p52wIJeIiIjIiaDAkIiIiJy6Sl+hfRT99ZR1QvoV2p8xQKsJJ34+IiIiIieAHiUTERGRU1dpPz1H0V9PWTXDanqxMCIiIiJnHgWGRERE5NRV2k/PMfbXEx0UDQAIDwj3VolEREREzih6lExEREROXcfw+FhZkYGRAID4kHhvlEZERETkjKM7hkREROSMVRoYigvVG8JEREREKqLAkIiIiJyxSgNDUUFRlVwSERERkVOTAkMiIiJyxooIiAAAhPurjyERERGRiigwJCIiImcsP4cfACDQN7CSSyIiIiJyajquwJAx5lpjzGpjjNMY09pbhRIRERHxJoePo7KLICIiInJKOt47hlYBuBrAfC+URURERMTrDAza125f2cUQEREROSUd1+vqSa4BAGOMd0ojIiIi4mXOkc7KLoKIiIjIKUt9DImIiIiIiIiIVFGHvWPIGDMHQHwFXz1G8ssjnZExZiCAgQBQp06dIy6giIiIiIiIiIicGIcNDJG81BszIjkZwGQAaN26Nb0xTREREREREREROXZ6lExEREREREREpIo63tfVX2WM2QagHYCZxpjvvVMsERERkVNcfHz5/yIiIiKnoeN9K9nnAD73UllERERETh+pqeX/i4iIiJyG9CiZiIiIiIiIiEgVpcCQiIiIiIiIiEgVpcCQiIiIiIiIiEgVpcCQiIiIiIiIiEgVpcCQiIiIiIiIiEgVpcCQiIiIyLGIiyv/X0REROQ0dFyvqxcRERGpslJSKrsEIiIiIsdNdwyJiIiIiIiIiFRRCgyJiIiIiIiIiFRRCgyJiIiIiIiIiFRRCgyJiIiIiIiIiFRRCgyJiIiIiIiIiFRRCgyJiIiIiIiIiFRRCgyJiIiIiIiIiFRRCgyJiIiIiIiIiFRRCgyJiIiIiIiIiFRRhuTJn6kx6QA2uz7GANh10gshImcq1Ski4m2qV0TE21SviIg3la1T6pKsfjQjV0pgqFwBjFlMsnWlFkJEzhiqU0TE21SviIi3qV4REW863jpFj5KJiIiIiIiIiFRRCgyJiIiIiIiIiFRRp0JgaHJlF0BEziiqU0TE21SviIi3qV4REW86rjql0vsYEhERERERERGRynEq3DEkIiIiIiIiIiKVoNICQ8aYHsaYZGPMemPMI5VVDhE5vRhj3jbGpBljVpVJizbGzDbGrHP9j3KlG2PMy656ZqUx5oLKK7mInIqMMbWNMXONMWuMMauNMfe70lWviMgxMcYEGmMWGmNWuOqVUa70+saYP1z1ysfGGH9XeoDr83rX9/Uqs/wicmoyxjiMMcuMMV+7PnutTqmUwJAxxgHgVQA9ATQDcIMxpllllEVETjvvAuixX9ojAH4g2RjAD67PgK1jGrv+BgKYdJLKKCKnj2IAD5JsCqAtgMGuNonqFRE5VgUAupJsCeA8AD2MMW0BPA3gBVe9shfAAFf+AQD2kmwE4AVXPhGR/d0PYE2Zz16rUyrrjqELAawnuYFkIYBpABIrqSwichohOR/Anv2SEwFMcQ1PAdC3TPp7tBYAiDTG1Dg5JRWR0wHJnSSXuoazYBtctaB6RUSOkat+yHZ99HP9EUBXAJ+40vevV0rrm08AdDPGmJNUXBE5DRhjEgBcAeBN12cDL9YplRUYqgVga5nP21xpIiLHIo7kTsBe5AGIdaWrrhGRI+a61fp8AH9A9YqIHAfXIx/LAaQBmA3gHwAZJItdWcrWHe56xfX9PgDVTm6JReQU9yKAhwE4XZ+rwYt1SmUFhiqKVun1aCLibaprROSIGGNCAXwK4AGSmYfKWkGa6hURKYdkCcnzACTAPi3RtKJsrv+qV0TkoIwxvQGkkVxSNrmCrMdcp1RWYGgbgNplPicA2FFJZRGR019q6aMcrv9prnTVNSJyWMYYP9ig0IckP3Mlq14RkeNGMgPAPNg+zCKNMb6ur8rWHe56xfV9BA58bF5Eqq4OAPoYYzbBdsPTFfYOIq/VKZUVGFoEoLGrF21/AP0AfFVJZRGR099XAG51Dd8K4Msy6be43iLUFsC+0kdDREQA9zP6bwFYQ3JCma9Ur4jIMTHGVDfGRLqGgwBcCtt/2VwA17iy7V+vlNY31wD4kaTuGBIRAADJR0kmkKwHGzv5keRN8GKdYiqrzjHG9IKNcjkAvE1ybKUUREROK8aYqQA6A4gBkApgJIAvAEwHUAfAFgDXktzjuuCbCPsWs1wAt5NcXBnlFpFTkzHmYgA/A/gTnuf2h8P2M6R6RUSOmjHmXNiOXx2wP8RPJznaGNMA9tf+aADLANxMssAYEwjgfdg+zvYA6EdyQ+WUXkROZcaYzgAeItnbm3VKpQWGRERERERERESkclXWo2QiIiIiIiIiIlLJFBgSEREREREREamiFBgSEREREREREamiFBgSEREREREREamiFBgSEREREREREamiFBgSEREREREREamiFBgSEREREREREamiFBgSEREREREREami/h9Q4YtSg35QzwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ISM\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "===============================\n",
      "chr1:234880850-234881850\n",
      "-------------------------------\n",
      "With prior:\n",
      "Input gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DeepSHAP\n",
      "Warning: unrecognized nn.Module: BinaryPredictor; using regular gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ISM\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No prior:\n",
      "Input gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DeepSHAP\n",
      "Warning: unrecognized nn.Module: BinaryPredictor; using regular gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ISM\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "===============================\n",
      "chr1:110259400-110260400\n",
      "-------------------------------\n",
      "With prior:\n",
      "Input gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DeepSHAP\n",
      "Warning: unrecognized nn.Module: BinaryPredictor; using regular gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ISM\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No prior:\n",
      "Input gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DeepSHAP\n",
      "Warning: unrecognized nn.Module: BinaryPredictor; using regular gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ISM\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "===============================\n",
      "chr1:88647100-88648100\n",
      "-------------------------------\n",
      "With prior:\n",
      "Input gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DeepSHAP\n",
      "Warning: unrecognized nn.Module: BinaryPredictor; using regular gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ISM\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No prior:\n",
      "Input gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DeepSHAP\n",
      "Warning: unrecognized nn.Module: BinaryPredictor; using regular gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABIYAAACMCAYAAAAJO6J1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOzdd3hUxR7G8e9JJ4UQaihSBESkFxUUFbGhoKBX9Ip6wV6xN2w0QRSUoqAU6QLSe+9IDb33EiAEAklIb7tz/zhhEyRIIIEQeD/Ps8+ePXV2d3Z2zu/MmbGMMYiIiIiIiIiIyI3HLa8TICIiIiIiIiIieUOBIRERERERERGRG5QCQyIiIiIiIiIiNygFhkREREREREREblAKDImIiIiIiIiI3KAUGBIRERERERERuUF55MVBixYtasqXL58XhxYRERERERERuS6tX7/+lDGm2KVskyeBofLly7Nu3bq8OLSIiIiIiIiIyHXJsqzDl7qNbiUTEREREREREblBKTAkIiIiIiIiInKDUmBIREREREREROQGpcCQiIiIiIiIiMgNSoEhEREREREREZEblAJDIiIiIiIiIiI3KAWGRERERERERERuUAoMiYjcADaEbWDKzil5nQwREREREbnGeOR1AkRE5Mp7YMQDRCdHYzqYvE6KiIiIiIhcQ9RiSETkBmBQQEhERERERM6nwJCIyA1AgSEREREREcmKAkMiIjcAYxQYEhERERGR8+VKYMiyrCGWZZ20LGtbbuxPRERyl9M48zoJIiIiIiJyDcqtFkPDgKa5tC8REcllCgyJiIiIiEhWciUwZIxZBkTmxr5ERCSHYvZC7L5zZjmMI48SIyIiIiIi1zINVy8icj0xBmbXBMsTnolxzXY61WJIRERERETOd9U6n7Ys63XLstZZlrUuIiLiah1WROTGkngMDOBMOWd2mknLm/SIiIiIiMg17aoFhowxA40x9Y0x9YsVK3a1DisicmM5HQLuXuDmndcpERERERGRfEDD1YuIXE8i10FqLKTFER4bTo8VPfI6RSIiIiIicg3LreHqxwCrgCqWZR21LOuV3NiviIhcoth9pN9LRr+Qfny24LO8TpGIiIiIiFzDcqXzaWPMc7mxHxERyaH4ULDUGFRERERERLJHZw8iIteT5JPwyFp4aEWWQ9SnOdUJtYiIiIiIZFBgSETkepIWD4XrQdEGOI09RL0xxrU4OS05r1ImIiIiIiLXIAWGRESuJwVKQWoMpMaS6kwFcD0DJDsUGBIRERERkQy50seQiIhcA5wOOzCU3lIoKTUJgMTURNcqSWlJeZI0ERERERG5NikwJCJyvUg9A75lwPIAjCsIlJiWERjSrWQiIiIiIpKZAkMiIteLlCjwLgbuXmCMKyCUkJrgWkW3komIiIiISGYKDImIXC9So8G7CLh5gXESnxoPnHsrmVoMiYiIiIhIZup8WkTkerCrD6xqCz7F7deWGzHJMcC5t5KpjyEREREREclMLYZERPI7Y2Bbl4xbydKdDQzpVjIREREREbkQtRgSEcnvoreAMwlwgqe/a/aZpDMAxCXHuebpVjIREREREclMgSERkfwucn3GtOXpmjwbGDqTfMY1Ty2GREREREQkM91KJiKS38XshEI1wa8iWO4Zs1PsW8mikqJc89THkIiIiIiIZKbAkIhIfhe7D+r0gNQYsDIagqY4UgA4FX/KNU+3komIiIiISGa6lUxEJL9LDIMit0Na/Dkths46EXfCNX0g6sDVTJmIiIiIiFzjFBgSEcnv3LzAkQyWR5aBoaOxR13T209uv5opExERERGRa5wCQyIi+Z2HPxgHWBZg7HmjMwJE0UnRrum41DhERERERETOUmBIRCQ/Mwa8guy+hSx3O0AEgNO1SubAUEJqwlVOoIiIiIiIXMsUGBIRyc+MAwqUAjfvfwSGMsQkx7im41Pir2bqRERERETkGqfAkIhIfuZMBc8Au58hyx2M87xVTiVkjEoWn6rAkIiIiIiIZFBgSEQkPzNp6UEh64KBoYS0jNvHElJ0K5mIiIiIiGRQYEhEJI+1mdyG4j2KZ29lRwrsHwpxh+zXZwNDkB4YSj1vE2emYFGSIymHqRURERERkeuJAkMiInls+p7pRCREZG/lHd/D2tdhcVO742nnPwJDKXZ/QonnNxwCoKBXQUaOhLvvhtDQXEi8iIiIiIjkawoMiYjksWRHcvZWdDpgZw8IqAjJEem3j7lldDi9rAUkhQEQeX4f1PYu4ovy5puwejX4++dC4kVEREREJF9TYEhEJI9daAj5Uj+VYtruaRkzYnZA8SbQfBfU/cmeZ3mAM8WeTomCxBMARF2gxdCpjXdxxx0weDAULpxb70BERERERPIrBYZERK4RqY5z+wc6HneccdvHZcyI2gxV2tnT5Z6zn908IHOLo5TT4EjBpL+0sM7ZZ9zuOxg+HF54IZcTLyIiIiIi+ZICQyIi14iIhAh6rujJ2G1jSUqzO4mOS4nLWCF6MwTVtqfPdiJtedrTzjT7dfJpcKbgnr7JrUVvPecYpa3bKVIEPD2v4BsREREREZF8wyOvEyAiIrYTcSf4dMGnuFvu7Hp3FwCHzxzOWCElBryC7GkrvSWQmwcknwRnMngXh8TjYNJITG8yVNC7oGtzC4s6lYuRlnY13o2IiIiIiOQHajEkIpKHHM6MXqJPxNv9AzmMgyNnjgCwM2InxhjiU+LtoelNelTHSm8TZLlB0klwpsJ/TkDcAbDcXYGhQj6FXPv39fSlWBFPPHRJQERERERE0ikwJCKSh47FHHNNbw7f7JreH7UfsEcs23ZyG/7f+5PiSAaT3qv0uAAYnR4cSo2Ds30Jxe4FzwAS0lcL8gly7dPPyw8fbzfc3bk2OVJg3Xuw8AGI2ZPXqRERERERuSEoMCQicoUdPw5r14Ix5y9bG7bWNZ15BLK1xzLmbz5hB4yik2IzWgphgPTojyMB3H3s4ewxcHzBBVsMORwmy3RcE7Z9Bx7+UPFV8NKQaXIDSI2BTe1haydIS8zr1IiIiMgNSoEhkeuQMZCaevH15MpbvBiefhoGDoTo6POX74vc55o+20oIYOKOia7p2ftmAxAaG5YpMJSJh5/d6fTZAcgWP+QKDPl6+bpW8/fyJz7RgcNx/i4uV9++ULUqNGwIJ07kYEfGCafXQPWv4Kb/gE/RXEvjBTnTIHwRnPw766idyJW27GkoVBN8y9r9hEn+Fxxs9wFnWfa0iIhIPqDAkFwVqamwZ0/WJ8aSu7ZvhyZN4OWXL/x5jxkD//0vdOlCvuuI+Ngx6NcPxo4lVwMcBw9C27bwv//ZeTW3fP89zJsHffpAUND5y309MwI3MUkxrunIpEjX9KQdkwAIObXf7k/on/zKw4EhGbeZgetWMl+PcwNDew4k5tp3vnkzHD0KW7bApEng5ZWDnaWegeAHwL0AuOdkR9lkDKx6EWJ2wqlVkBab411u2ACjR8PhwxdfV4TorVDmcbjpSSj3LHgVuvg2V8jKlTBuHJzMongRERGR61+uBIYsy2pqWdZuy7L2WZb1RW7sU64fJ07As8/Ctm3QrRskJeV1ivLA6XWw/w+IWHXFD9W+PcyYAb/9BoWyOM+YPNkOBo0eDU89BQkJOTveyZPQuzf07AmHDuVsXxdz/Dh8+CG0bAnly0Nc3EU3yZa4OPjgA/szGzgQihXLtNAYSEu4rBYlSUnQoIEdMPHzy3qdFEeKazrRce6tJN7u3vZ+0oemX3wmCk6HnBMAAiCwKmzrAn/5pM9wy2gxlCnwFOAVwKZjOyhQ4JLfSpamToVvvwVPTzuvZRX4yra0RPAMtDvTnhTMhu5388kn8OmnsHNn7qT3HKfXQqnH7dvWbnkHPAtefBuw88GOnyDkXVj1Mn8s/ZJHRz3K4MGwbh3ccUfWv7vcMnbbWObsm3PZ2xtjaDulLUdjjp67oGPHjFYOHTvmKI15KjUWordDyrlR8YgIO4h5TTUMi90HwQ/Zt4FmCuBebd99B0eOQM2aOQzuii083H4+m9nUckhERPKBHI9NY1mWO9APeAg4CoRYljXNGLMjp/uWbAgOzrh/o0SJjApJZsmRkBIFBUrmSeWzRw8YOdI+MX74YfBJP3edNMlu3RIUZFfY27eHwEAYs3UMz096nugvolkftp49p/fwRv03Lvv4fx/+m/KFylMmsEwuvaNLtPsXCKwON7e1T+pT48DT/4ocyhi45Rbw8ABv76zXmTIFBgwANzcoWxYCAjJtHH/QPkEPqGifrFxEVBR8/rkdUHFzu4JBv9Pr4NQq/hpei/7976FoUYtixS5wEpMaAyeXgzMFCtcFv3IX3f369XYA4mzA5Gwe5XQIhE6Ako/AqTVQ9cNsfS5nGZNxrn0hZ28lK1OwzHkn643LN2bu/rmu15uSwez8EeumFufupGBVO12pMVD2vxA2i0QrGUjGzzMjIhXgFUBCsaWMGXMXbdtm+21cUHx8xmeW4xNKryC79U5aAvuOl2dm0iK6dLEXXc5tkUlJMGKEHfRzc7OD0yVLZlohIRSK3gHu3hmdeLfORhO0o1Ps7Sq/DpYbr3a1f8uF9zoY8Ls7/jn4aZ88CcuX2/mmQQMo848iK8WRwnMTnwPA+a0T698y1gVM2z2N4ZuHExEfweRWMwkNBV9fKNWxI3TqdOHISVoCRG8DN08oVAPcrsGh7Y7Ngpjddiuc0InQ6EscJ07x/m3zadmnCQUL2uVdYOBl7Ds7/7WXup2bF5wNBo92Z02ik+5+LZn47ETcrKvToNvphLAw+OgjOx/IJcj83bq52R8m2N+zXLqEY3bLUb9y9u3R+czZ/3sRyR1xKXF4uHng45H9erfkTG7UPO4A9hljDhhjUoCxQIuLbJN7UqIgaitEbYa4Q9m7HGiccHQahI6H0Al8MbElHp09GLN1zPmrGruFxejR9u03z/bpgdXJ4rP5n130MIeiDvHoyEeZsH2C3ZdFyhn7amZ6GnfvhunT7ceMGfYVzQtKOGZXeo/NtJ8Tj5+z2IEbG3wbsWwZ/P13plsZNnxkf0Y+JcDK3Yr8jpM7mLdvHuYin3lyMq7hsT3CxsPBUSQdXsqa1YbPP4e334bvhwUTWMgi1d2i9aTWGAy39buNJiOa8ObMNxm8YfC5O3WmgiM56+/bkQwxeyF6O/2Xd+aeYfdwU++bzrlNJyI+gum7p5Pm/Pd7asLC7MDWrFl2PujZ8zJuvTq9GordBZNLw8IHLxoUOpVwii0ntthv0zhxGicHow5S+ufSvDjxxYwVnWkQtQlOroRTqyHxBJYFoaH2ifTZOmpYTBhpCcdhT384NBZfxwEcDvtzs6z0Y6QlQMhbdoe/ARWZuHMygd8HMmDdgH9N65498PzzdiDFywsKZrPRxQWd2Zn+u5wIH7ayE1jasn/flV7Du1QDEhLs95blbWQpZ2Dz11CqKZR+3M732RAYaF8xT0mxP7uvFnxFwe8LEr25C9TuBsFNoOrHlxQUAjtosnKlvd/EC/Qru/XkVgBuKXILwDmBnNrBtV3T79z+DntTwXlmB/z1j0pz8XvBkQRu3lCvFziTiHfY0RQ/r4x1C3oXhErz+OADKHfxeNm/SkhJoE4dWLrUfn/JOe0exaMAnNkOzmR2lF5Oiyc9KFDAzlv/lq+cTjvYOX68/Rg3zk7Lp5/CSy/BK6/Au+9yfispvwp2Cz5HEnYn3uk/mJjdGXkwdDyc2fWPN37EbqHlGcCCoRm/5eiEy7wV7chkODCCQ39PZcgQBy1bGlq2hMKZ+t7eEr6Fqbum8t6s99LftBuvfb2Jv/6y33uvXnZrOoB6A+phdbJYcGBBRl8nmVortPyrJQCzFsTx11+GkiXB3R3OnDk/actDl2N1snh7ZEM4MAyK1Odw9G0MH+5kzBgYPBiGDIHVq+3b6C6UxwE7sHQ6BCJWQMRq+7ea2w79CVXehYBKUOEFwGI3Vbj3m/to0sRuzXVZQaELMMYwYfsEVoSuuLQNz7bOqt4cdvaFtHhSnU4aHIUpu6fQfHTzy0/U2X1ns9WXZdn/aWf/SuNT4umzug97T+296LbG2MHhuLjcua33YNRBlh9efsE6RXJaMosPLiY57er2xZSSlkLDPxpSo39NevVNYdw4u04Qn53WtmcDgdkNJGKXYyNH2vXO9Y93vLyWfJlbAPr7nz99LbQKPLEM9g2C0El2X1uWG/hXuLS6aux+OLEUTiyxyxfnZWTES+kTKot1f/sNZs+GffsyXRxzpsHh8XBoDBweaz87zs23cXGwdSts3Gi3io2Pt8vQs9k/JgbmzIGZM+38djktZzdutC/CTp4MEyeemw2dTueFN8zG8rOMMUzeOZklB5dccJ3Q6FDm7J2D85+tndN9t+w7rE4WT4x54pz5kYmRhJ4JveB+Oy/pjHtnd4ZuHJqttOZUt+XdaDqqKZEJkRdfOYecTic9V/ak75q+JKUm8euaX88ZlCTXZPGfceyY3UXA1q0XOTe9gmbtmUXA9wEU6FqAuJRcuj0gt5xak34+PhOOL7R/7xcSvR3C5tiP8EXpdc7clZyWTOelnV39keZEbgSGSgNHMr0+mj7vopaMmAARKwFY80Mz1/xTU54B7MLG6mRhdbKo8cY9/DkqjZ074alOt7vms+lLCKqBI7A6QX2q8MioppT+sA6DfzvD0qXQ4a1lrv0ee6iynflLuNuV+7Kt2LviTX7YNhWHcdB9xFeu/f72xYOAXVCvXAmtW8PDLU4zLtoOCI1cOMy17iefPOw6xvoeD7nSXqFvBeYcmEPfGa3sE3ivQNbNWQlffwqWRbePQ3n8cXj8cXhzkzfF+6e/p42fAhC2cTEk2qX40dnfQOnHYM0rhP/vZfAtZb+XcQ0B+ItnKbJ4AvfeCxsPdKb8MHtfu3cftlt/7OxJy4cPwJo3YeX/6NVpJyxvxcQhLbh/UD0WH1xMs29quN5TaoCn68/v7Dyrg4VXZw/uGHQHd7SvTLXfqvHIn48wsP2jrvdf90d/vlr4FRX6VICtHQF4/6XdvPKygy1bIHTNbKjwAt7RizmxdzunTtkn4laqfZIwtYqdJT3dPPGKyLjFxnfikIyMs/oV+6q1uzds7wbA+NmjKP9zKd6a8RYhI56CgpWhUDXeWdTB3sbhzoQfFzFoEHTstZHiPYvzxNgneLJ9dX791T65uaNjCdd7/eXjUbBvEBtnzeOuu+Cxx6By0Go+eX0PHh7wyvf3cc+Qe2g8rDEHB93tStqUHgPo1g3af7vHta9XJpyyW50YJy/+0h82fg7r3uPV/2yCDZ9wcmFHfvtxD+PGQbuu3SnWoxi1fq9Fva+L497ZHffO7tzc92bCYsMYt30UCwYMgP1DiJn/qt1pafG7WDZtMxSwgyDP3NaVr7+2KwFWJ4vSvUrz3G+loPg9UP6/fHbHY7z4gmHxYnioe2XcO7sT0M2PxCQDXoU4s6kjT09sTUxKDG+O7sLm2TNh2ZP8+tkE2NUXlj5Bsy7FmbhjIv1CX2Por6Hs3g2bdkcS0NWHJ8Y8wX8+vw8S7SupgzrNggPDIXQ8kX/Zn1VMDPTvuo6xY2FI3wPw6ctgWRz74x0o2wrWvcOm4nvTzzwg8eROwOKFuj/w7pvxLF8OH3f/imf/epLpu6cztsNj9hdguREXlwbGSdLmbynyQxAvTXmJ+l/VtAtyoG+PSNjWFbZ24fFGW2Ddh9RKfJX5Y9YwcSL0Hr2Jbiu6EZsSyydzwjHHZkJaAm3uH+v6nuOn3uuanj9wBB98AD9+uZUNi3fRvDmM+Pw71/KP7/+YFi3gs/apVOtZkLHbxnJLrxrMHHcYlv2Hw8e2A1C9eHUA7r4pIz9VKlzJNV2/VH0Iq0N48e/BkUSHiV3giYMQ8jaNGvtAsUbgTKbtW8Hg7kOcww4IBXoH4u20K9pBBYIY12AJlSrZrfSccxvz9dd2f1Spm3/i9deh99erXUPFt28PnFoLy5/GMbsh338PrVrZ+crvez/e2u3Od9+e4bvv4Ic3hroqGC/4j4QNn0PI27Cgsf0TdDoyypKX7+bUqt8g5F3adq5OzxU96f53d17o9THMrE1jxwN83zmaPXtgz+Zw5o1dR4sW0Pqjtq59vNBiOmAHwcPC7HS1agXP1P4Vb2+4rewe1q0zFCwIb72RSqHQDjiWPc0jPQoTciyEF5b2ZvO8+bCnP0sjS3NrxE38uvZXVk9s58qDDYa+gtW7ql0uL0n/jyr7DEl/t4OY3fQLK+z6noKDXqNvXwgJgchxjQC7hU/RbgVoO6UtdXoGEjJvM48/DoM+/tn1vTpDJ8PN/8MzcR/xsQ6MsbAs8OuR/t/3WXlqDahFy79aMmBDeqA2KZDipg7PPguPNEuke5IfP2z+gErflmdD+AYA5v3SxXWMac0fZdAg6DX4GAAPlH8AEopw8812S86DyyYR6GtXvDq02wC7emF2/8q9Q+18vjFsNU4P+za/Pz7tTZuXvHjuOVgx/wgvv2y3bloyagohIXZlkhWtAUhOSHYFeA9/2hCK3A7F7ubpgW0YvG087895H8ds+z8sJQUGdl1O//4wtv8Ghvx6lNdeg/k/daBPH2jWDII7e2X8969qC0BsVKwdXANWHXrArqhtas+q7+6D8HBu5gDTJ4Rz/DjsPHqchr+UY+vJrTzX+XFXPl/Su73rs1re7xv7Fq/kKEZ3+xN29SFqRU9Yb1+cmFurOaz6GY5Op3hnH1pNaMUDwxqxblgnuwIYux/efQQsiy1VasI++/bhJT9/DG++CZZFWLl59sGiYc7fpWBVW4ZGZESsQmf7we5+sH8ID3UuTuNhjSnXqxxvPLsF1r8P697nr9/Ww7In6drnQd6b9jKrj65m+LdPpP8GYW6gj12/it4K6z8EYNOqEwz6PYGtW+GFzi2wOlm4dbYomPQ1I0bYAT7/7/35YO4H1Op/CxwcCUCPEd+6PvdV/e93pfP55+384+8P7lvsesuS1Qtd6zb+opJr+v5OQbw7613q/F4Hlj0JQGJqomt57c8qcHPfm7l32L3c821J1/z69XH9lny6+tBkRBOqflnRtbxJm3dg7+9wYDgzf+hgty6NPwLLWwFwKPIQt/xckmEbh9H0qztc24379iHX+8hcnp9dXvGzMq7pqh2CWX10NdvCd7BlozvPPGPXCfxmN7HzuYcnHLT/H1be2QCO2VfmVvRo7dpv0syMY7Ctq/0cuYGQRftp3hyW/96FTp3g9ddhwQJ48UW73lnvy0N2Pvfxg1dq2fsdNcq1q1U9nnZNp825L+MYz9jl/7ZKt4GfHRVf8ux9rvuvozcMzQhwrHkNgKjwSDg2A4DlwzNOtNcN/ISff7YHGjhbngOwqxcAWxb8DXEH7WP0+sS1OGx4o4x1Q94B4OThcDhu5/+jf4+ESq9B9BYio93tVu3bf4CXb8tI255fAdg4Z7F9gRT4vdNC2PcHHBhpf/cl7oMV/4VDo8HNneP7j8InbcCy+PjFyq7vcWm3hhnpWZJxznG21Z6pDr9+voc334Txn73tWvxB2y3EnL2umN4XXWixMvDiY2BZbBo6n0cfhcqV4aEehSn5U0lq9vCFtBgo/5z9H5oYBu7eHFy5gF96pzBoELR5LooaNaBOHbubBT8/+wKGtflLAHp1P0XTpnbZ53voR6pWtZPw43sTYf9Q2D+Uhb//BqETcB4cbwfGgd5Dv6ZEj+I0HNyQH74/zVNP2ecZYWunsWkTzJ9vn1u5d3Gn8mdlsTpZBP0QRNOvq7g+q8KdPHDv4m6XtZnL89DxdhoGfY7VycL7O29Kd/LlqXFPcf+I+/Ht7GaXK50s2NUbgD+m96dcn3I8OvpRGn1V1nWMvb/fDkBscizfLP4GgMXbF8GHz4Jl8V7r8hT5sQjlepejYcciru3KdPOm4PcFsTpZdFjaAadx8tPozq7lX3euQ8+edn4d1b4rAwbAH38Ar9YFy2JvsYKudd/8+AHX9/xYg/UZeSKLfN5t9Dd8tegr5u6fy+uv9OFoegPvmJZlM/Jrej7fuH296xit338Mjs+F5EhCfngwY7+L0s8dkyPZNb0/HBjJkEFvUfSHIO4Zeg+3dCrCp/M/5f0571OgWwHazWnHnYPvdJXnxw7HEbqwL+zqw9Kf2rl2e1ungIy6VhaPRz6s56qfL/vpbVdAKPK2BYBdhwkJsW8t9jWhroYGS37PqLds6ple1jhTmfLTQPu/YldfjiweAKET+LX/KxT+IYi6A+ryQIeMsrRZpp8dm+x8PmvZNNfyjz7I+H9pNsZe2d+C8Bk/QPxhUs8cdK371Ad32ndgAD9+OId96eO43NbdlzoD6lClu7d9QQvYE7IVIjcC8PqPD7v28WKLTBd1ssjnm8f86Fr8YJdyWJ0sinax4MRiKN2MmMXv039IEUb+6cGA7hvsVtVAmx/uy6ir7O1nX6xe3RYOjrDr5yfDXP+vnX57x7Xu9E6NXcfL6rur8cnNruleXzR1revT1YcOSzrw2J+P8dfid3GkXv6VmtxoQpJVw8nzLvdYlvU68DpA2bJlAahQ0RPjSMLa/A34lCLpvRfx+WUUUT/dS9GIbTj39aNCsg8nrCQ891ai9nseVK0KD1Z5lpU7tmElJDHg89W8EVyP1KrhRJsU5h2Yh+fRZ6jfJpDataGQWzCje+6mbFF3fqpWmKG/z8O/yycMfXk0rxXuzc0tohgbWpUnRuxkZL14RjQOIu1UFI4jp3AWDcL/dDR1m5RiWdnGOHwDGH1XXZ4csYExNeN45QkwFhw/cZjY154kYPAU9t9ZmnqfWVjAiKcr0Xz6PpZVgmUHenOvOUpw+eMcavoewXEv0rN+M+a9/BzeG8K5554STCp0BN9U2NNxGbdMs/DzDWDn189R1cNBROwBYv5uw201TvNejfoMGvQnfp0/YtimwrxcNIhn0qYy66MXOHS4OBQKw6ORhQPDtz/F0HFxK6ruX8SodxZC3dngSOKtiCfhzukkb72VJREnWDKiCbcml6O45Y0zKZl4p8XZbjLuivJnXUAcBRPhVICDkLAQbkkpzbCYOtw6dSPhNU6SumcGHmHj2JgYz8a/7WDNqo1x3NmkMBW9Y/jk7gYcr+XHGbcAkle0porUBGMAACAASURBVMbRVTyR2pnDpe9kM4U43e9/tO4wnocj4pgRWoMHh29m1O3eVK95HxX+XMqqOqk4iwXhdiqa8U81olXwI1B2P2EhJSk18WsK3FqQw/+N4ff1vzMrtD6LDr5FRa8ZTNlWjUYztzPR9y5KjWpJ06YQF1+Zwp9W44Hx22lfZhDv2mUFJ4a+RPyB36kQe4Z27yZCpddoltCRBV92ZOGWW0hyT6XUfZEUHj6HxOcK83eKfcXgp6SH6XpgIZveHYHXO0P44H44dsqXnkMgzYKNB46x+uPpNDgQQf/PB2Kqz8FKOsnvX/wGNTuxdoYbte8pQMOGUGPn41T4bTTVZm5l9a0FefC2Irhv2EV17/I8suQQRxtArdEVIfhB3LeNZenLz3Gf4yDBFRIJC/GmeL9epIVt4e35wwntU5YfX6lG+ZXb2VMEJs/9iid9tlOhxUG6+jxIaBNPWjcK5N6GVXFbv5NfJ63nU8uHQH9v/qxUjSdHbWdFo0iqvxgCtSbwxNaqOH174HbfNGYtsZg13q6QLjGTOHFrNU76GeI+TWb6numUM6U5FDKB4IlreL7pabh5JmzpSM+dRfmuWGE2J9eh3JiFNGsGZ04VYfv6W6g6LIjw72KJOvgC1eucptuJCgw6sBC/ljUZ0m4j7yT7EOBu0bXKWE40Lkn8o/sY53aYcbum8FBMXZ4pGoTb6WgG1LqTj40vKf7uRD6czLDNwwgygezYsIRbn3+UVzrejLl1OZZJZfhbD0KtRViWG6+4lYXWVSjhZhj7Sm2aD9vEuOAoVu7+k7sPt+C5+x5h9vB78U9LZsnxEnz1amFCkqtxrPcivvsOIsNL8nQrX77vCUv+uo8Tm/+iSK9uJIZvZeCCSWxfWo5fv4rluYnP4eZwo3Ghu6DYZ4zx30jB3XUp37Iv5ZrcxrOn/En+E9beW5H/dH6NGiUh/P7qPFHnJfzLFmfLztWUPunkkYOzOPbDJEpX3UjT6FXwxCYoDLcXaQZTYvjM251HH7+T+2q0otddZWhU7RZuatmLk3fDxJrPELr+OO0+asiKtAQCCrnz+YspjNyaSHChYjz+4EwqtGrAYynVSb55I96+cbz9XTsiKydRvY4nVcIrU3btXsIKO3n2pns43qUo7mVSOfbeq5Q4VJN+TwzF1JiKhUWf3m/S7rnCpJJA4JvunHF30K3cTorceg/4v8DsWUUYvsA+qex6oB68EUrBIqF8ULU2R6tUJtTzJra+PYhJk6Df6Hv5a/9wDFAwfCVrFt1PUDEPik+tx9JeDTDuHnh3KcyddxfmzYAYQtY8xqxJTj58PQxTbjTmtvLMW1yAHT/dwdFAuHfq49Tq9THuhbzZ/Vgy7Wa3wzu8JiEhz1Oj/ima7PNncznwdMC67seof7/9dzepwTO0jruVb54P5ouDjajS/W/C79xHQkIbIr4KZ3KHInxXrDBJbgmcfjuZ4ZuH4x8fzJfdqvDrAFg09m6ObxhNsb4/0GNzMu0LBlM6qCD3pfVmSYFqON08uPuDQsTGRFM/IoVPS9Sm6pRNLK1ZlacqVcb8OY2N9UuztMRdRBf05uQLCfRZ04dCaQWZebwed49cz8KaEYTP+pmg3j2YX2Aov7wGUJqybe+kReeFrLwJnD/ey9xlpUjxTyW0zGmKv/ci3/znJ7j1T6xJwYzbXp6Hpx1ifSmYO2U8j/IGHdsmMOupbRTY6U/bAseY+3ojfLdEEFLkTT5KP/cd1T+A1i0L4+aIZ+PQz6k1KojkUclspS01guYz24Qzcbp9MvrkyTtp7PkLE2Ylcs9rn1K1KgwdUJHK1f159iUY0e95fHxT+fNPT377uBpznJvwToPEn8dRIGY43u4ebJr0FTXrvkJgcmlWPVSNhvv/wr357UTvmY1/myq0OX0b+8vUYn9hJ6vfC6XmbzUJTizG/nUzqfBsA4JKlyV+9z78fp9AhcGNoUB7WPsWz73yEhRvjc9wP7ZEfUH1okHc/vA+qNAatnSka6GWVD0cy5l5s7mt1wEoVA22dOTvRrW4668Qiiae5sD66ZR670UCnNuIePo1gg7V5Md9gfxcNAjLkcihsVth4HRauPtQtvGd3DtsDfF9D0CV8bClI2lRKSw1SynmDv0+GAS1foCkEzx1+h24axyrV5ZixsaFRA4byrFytWi+fTKBbWpw963roNgXsKUj0zal0bxpYQ6l3sctUydTowY8F/4/3OZtxmv3Ye5I6kq1tydzjNIMalObJlM3sbIc7AucT6WK4dwfPY22sTfhu+MIIZEplONrSiVP4r5TwaxqfAfGzYOoF1N49OHCVPRx8uh/i+I8fopbEtI4fps/zpg49nrHsiSkHwAbpkLd+yyMJ/h/4Ua8m5PCkckMTqjFrVM3M+V2Pw5XL4DHmUSe8n+VJe3t/5oep26lyYxdzKkcw3f3u5GMkyFNN0PlfrClI7WKLYUinWBLRxKazcQ31iK1MOx9D16a9hJlHMFYxq6g7tkaSeK7rSnQbww/d3mar14tjOVI5Nn/FSfy9ElKx8Zz982lCNgdRjGK0Lx8PQLHLiCkThDLijQhNaAwxz8LoHXRIDufb9xKraJBBJ+O4MjakZRsUxOvpL1E7ZlFQJua9N5Sms/a2sdY99uH1Lu/MCe9fPig9H6+7wmdvnyD0ZMMvr4WU1sUZVnJJjgCChH+kQ/PFg3Cy5lI+Gd9CB77FMGlKuB4/x3cI2Nwr3M70S1m49/tC9YcdXD3o3YZtXT0p9xTNIiiSVHsH/slFZ7pQtCuaCK2jieoTU2+LhlM195D8Ov8Ebu/20S1GRY+nt6ceLohJcY8TnCZisQsn0LBP6YxoG0yg4barcJWN42kwUP2MUL/cxdlJ35EsaASRD1Rm6Dhcyl0a03i9zyD3+/j2dKoIqXa2usmFSyAT0x//LwLcOKpOykx5hEi3apzqs2z1D69gV8PuvHtl15Q1J9fmzfn7aJRWI5Ejn41lpsmtKN4kZJENa9F0PA5tBnzEFSaB1s68tsHC3mreA147DS7h62gynQLfx8/tk36ktuGB1HvUBq+N7vjcDj4JcpQ68AC/Dt/zODadXm9lf19hAz8gPpvDeCn8Ddo2sSftq/BptlPkfDOc/j2H0vMw0tITjY4HDCnx6s8+tlICrrFsf1Bu97SP/ExFrR6FbMinkKtChBeOIxwYEjbEbwc2A0KwJm2zQkssoUu3TwZMsO+77r57ZuZ/85WfDaE0qjgVlY/WA/j5k4x30VUmmrxla87M/a+RMA+X6K94onb8jT+Ayfy2arXoOJACGnHA40rQdmnsTZ1YMMhB7VvL0xAVS9OPhDByYQI7ix2M3Pbfk5aSBQ77vmBdk3t07b3p1agxKaDHC10hn23Q3RSNI6IQIqU8MIkp/DFjmKcTA4n0QMSf5pCgVgLdzc3jrdqRMm/nqFc7RLQ0g7YPnagCi8ULkvCzPlMvK8IC/1PEZgMByetosJtiTQ+PprfT1Tn9snbmFHHhxI1SmMdOsaOPTFUvvV3AmJmMHb/bbQYvYMFFZPZ1KMqNUcF8fQOuLl4DUrN38qimgXZcUssXomp+DstjgXYAbo/D9zGU3/uYGTdGH550J+k2Dh27J/OpFkQGws//PAV36X3RDHe0ZD/TD2Ej08yHsaun5+KOEzMqy0o+Mc03qn0GPNH/IJXbDzeU87P5y2Di1HjifrcMWodg6qcIvRQGr4F3OhcrRo/r4jDciSyq0sIVadblPX1ofmbRbHCTtEyYI8d9Nzdl/2Td3L7F/Z+1z5UkzsesOBuuHXWbih4C4EL+3M6KZq/Q//m0ROVGe5VD2vyQnbWKcddqw6zrRjsXm/n86LPeOE1Jh6AEjNqEPv6kwQMmsIbzYMZWTIObwdU9CjGeo8IApOgsl8JfA+coETyCVf93KfMSU7vmEJgm5ps7B/BA89a1MPi0BNPsbx7eZKNgx31WxC/IpCI8LLEvv4UAYMm0+39Bnb9/Ke3aflRe6jwIix/hptqtYfCdQhy60FUUjRR4RtpElEO7+JuuKc6eTilFctb34/l7kZwrJ3PqxcqSOuXSuK/5ziVUo+TtO4PfGIWMm1LFRrM2c2xYCg98Dj4lcPa+A11EgM4kxhLqdBwzrzxXwKHzsSjRndSXnuEtDQ4E2MIS96EJ7Ch5yzqFppJkWIbiehbgWKjFnB3w3JMedALk5SCf8QKNq6oQ0AhNxK67abmrHPzuV+hm9l/OIyKnnO41+nOZqcXHrEpDHprKq8VHcHCUrVp8HJt6taFiNDSRL7fhsIj5vJgw5tZ2yQAEx1L73bL+aD4rdDyNFuHbKPGTAuvAh5EjnmLwo6S1EtdS1GHF6kpKYRuSyCp3fP4/Dqajq1LE3PqGDFesLxaAY5biRSNPIOfrzspOIjYE0b8W63w+30Co5uWo97aw+woBo2OLMX9/ixGL84uY0yOHkBDYG6m1+2B9v+2Tb169cyV4HA6zNqja02qI9U4ncbExRmTnJyDHZYoYQzYz/ld7H5jDoy0p8cFGnN6vT39J8atI4aOmBm7Z1x0N2mONHMk+sgFlyelJpmmo5qayTsnG9Ohg/35gT2dUxf5PubunWvazWpn0hxp5y1zOIwZO9aYgQONGTDAmIQEe/7hw/brP/6wn1NT0zc4NtuYfYON2TvYmNSELI/3x/o/zICQAa7X+6lgOnWy89wF893ad41JSzRmzdvGHBhhjCPZmE3fmONjnzSrVxsTEmKn9YKcTmPC5htzcLQxB8cYk5bpQGc/79z4rM9a96ExjhR7elZdY1JijZlYwswcaOcZOmLOJJ5xrZ6QkmDm7ptrHE5HRno8Mebnt40JX2LvKz1PHD9uzN9/G7Ny5UXe8wUkpya70jBv37ws80dCSoLpt6afiU+Oz1j+fQVjUmKMWfexMZu/NsaRZszmDsbMu+/SEgBmFK3N3r32y+7djdm1y56Oj89YJ7ONxzcaOmImLXgj/ftPMeZP7MflmF3v8rc1xtx+u53WxERjypa1v7Knn86U9soYM6m0MStam3vvOmMSEoxJSsrGjkPes/P55g7GBHlmlANnv5uVbVz5auwAT9f3OHHHxPN29cUXGcdMTMzmGzt7vLNWtrW/84klTNHOGXl3Q9gG1yo9/u5hArsFmsiEyGwe5F+czWtgEksVN7+t/c182ynZhIXZi2NjTa7+v0QnRpv3Zr1nohKisly+ebMxo0bZj4t+hvuHGbPvD2PGBRkz1u/yEnSRsn/poaWu7+BYzDH7PUQb062bncYHH8woi5980v4vP0+mz9i4uZ2fxy6gSt8qho6YZqOa/Xs6k6ON2fCpMeGLjPnTsh9npZwxJu6wMclZf945FnfImL0Djdk3xJjUeLPowCJzYN8EY45MN8bpMGZxc3sdY8zv/ezP0berr6Hjv5QF//xN5LaLfOcOp8M0+7OZ6b68e+4eN/6oMQdG2Y/M/9WZ8ke/xv6Gjpj9kfsvvJ8r/fn88xhgvqO9OXDAfvnii+nlQnZlzv9np6+QHj2MmTjRrp9ctn+mNyeSo40Jm2vM0Rnn1oGy62we/Ud+feGFjP+aS/ou/iE1LTXLeqgxxhw4YEyfPsYMGmTMr78ak5JyGQdICDPm6ExjjkyzP4uo7cZEbTPm228z6lcdPjJbwrect+mZM8acPm1XJbMtN7+7q2DwYPt/ZP16Y5YssT/rP/7IqPdfc5wOY04uNyZ0sjGhk034maNmfdj6i293KsSYbd8bE70j/T/KLfvHzOn5Webtu75nnzMdnWVMQrgxkRuMidpmohIizcGog5e+71wQFhNmklKzU1nNGafTmBUrjJk0yX64ziGvEcA6c6lxnUvd4Lwd2K2ODgAVAC9gM1Dt37a5UoEh+Repccasfs1+/qugMTH77Pl/upmTwzGLDizK2/RdD8BERBjz11/GjBt3gWDHxq/twE78MWO2djNm50/G7Pn9qic12zZ9ZZ9QG2PMzJp2/plYwpixfubuP+42HRZ1uLT95XKwcOquqeajOR9lb+WzlRsvjHm5rDHbuhmz9En7JC8tuxGHTMCcIcC0bm0Ht5o1y6jkXSgw5LK9u31cYy4/MORIM2aM92UHhsLD7aBLQoJ9Ut6smR00cJ2EgzGPYsyBESY+3pj27dOyH2hf84Yx0Tvt4E9Wlcpjc4zZ9Yudn06udAUJsvLllxkVumxV7LLKY8v+48q7Mwd7/+vxckUW7/nNNzMqDcnJ5vq68HAZ6g2oZ1pPaJ3lsj//NGb2bDs/vviinT//1SWcuByPPW7uGHiHSUy9jN98XnKk2gHOE38bs+EzY5JOGWOMGdY/I9AZ+H1g1tvm9kWa683V+nz+ERh6JWiCKxDxySfGbNx4eRdJrshFoSxcUjDhWnSR73nwYGPmzrXL6f3/EkOU/OGyfkv5jdNp120c11hEQq4JlxMYsuztcsayrMeA3oA7MMQY0/Xf1q9fv75Zt25djo8rl+j4Qtg3ADDgWwZqdQUsu9NXyTnLyl7n55EbIfEYlHoUcLM7Q3fLQbO/KynhKKx7z84ru3+BIvWh3H/zdGjlXJUSDTt+gORoqN/b7rcqOzKNRuMsHszORcfp1g2GDbOHbk9IAN+b09fJagSjzd9AjW/tvrIuZUSszOKPwIxbwZEA1TtAzY6XtPn27bB3L7RsCYsX2331tG1rd0IcGIidnz+woEccJ077MmFcKu+088zezo0Tjky0888t79rv85/O7LQ7Ig6oyBGfSlhYWY4cuGMH9OsHP/9sd9hfs+YlvU3biaVwcBSUeoy0Mo/j2cWT+iXrE/J6yGXs7PK0bw9ff233IxEbm2k0QMnS+PF2vly/HhYtsj+3G54xEL3Z7ssobj9U/YTw2HBK9q0IwNNVn2b8M+PzOJGSpbOj/gF06ACdOvFawBj6n/4vnp52mTB7tl0uf/nlhUcVlStr2jR7ZMhKlez+niyNMiYi+ZRlWeuNMfUvaZvcCAxdKgWGrgH7h8P+wWBS4cFl4J7TsaZvYJc7jHF+ER8KYbPtIWRTY+yOI2//Bdxv4IBiFt95r17w4INQvbo9ssi/BoZ2/gzln7c7DE9LuLxA26nVsLipPbwvQOtLK8s3bLCHR2/aFEaNgnr1oGpVe5QvNzfsGvHAm6HtZuKS/enxo4Ovv3HHM5uxody0bJk98kulStC582VW1pMi7A5Si9TndGIUfp5++HhevSFQN2+GoUPtANfOnVCt2lU7dL62fTt88AE88QS0a3fx9W8Yu/tB6DgwqVir7A6u5zw/h0cqPZLHCZNssSz68i53r/uFOnXs0ax8r5PrLSIikvcuJzCUu+OXS/5RsY39ELkYv7JQ+Y2M1+Weybu0XCuyCP699BI88IB9tXf2bBj8bwHC4Cawpz9U+xzMZY4ekHgcvIpC9a9dIxleioAA+6Qb7EDW2eHR3TKPVVm0Ipg0/OYGs3TBVj7/ohjGgNdVjiPfe6/9yBGfYvYDKOJbJOeJukS1asEtt0DZsnDrrTBv3j8+a8lStWr2qJFbt2YKWgpUecd+AKyyI6U1StTIwwTJpXqJYTR6+RdefdUeiWzq1LxOkYiI3MhUxRLJqfDws3etX3+thSTbChWyW95MmWIHMf61MWZQbTi1Elb9D+bfc3kHTD4NN7WAqp9cfN0slCkD06fboxfHxZF1SyAf+xK2lXyCR276mU8+geHDLy+5Am+/DUeP2ieBCnBk30032cOD6zPLWqmAUgCU9C+ZxymRbAkOBiCAOGaG1SEiAt5/P4/TJCIiNzxVs0REcknVqjByJPzvf9m43anxDCjZFG7/7fIOlhIJxe/LXr9WWShQAKKjYd06iImxhyI+j7snYIFPCT5t0R8PDzswlJZ2eUkWkdx3e6nbAbDUIUq+U8b9OJ07Q5MmeZ0SERG50elWMhGRvODuDZVevfztk07CTU/nqHfM7t3hnnsgNdXuePo8qcmAgafC8QD6tr7sQ4nIFTLlv1PyOglyKdSyWERErkFqMSQikh8lnQCfojnaRd26dj9DPXvCtm32vLi4TCskxuZo/yIiIiIicu1TYEhEJD9KigCPnI/hXb68PWT9nDmQkPCPflyO7QD3qzdyl4iIiIiIXH0KDImI5EfORHDmTmc/N90E48ZBaChMnJhpwfaTkBZ3we1ERERERCT/U2BIRCQ/sjzBmZo7u7LgoYegRg14661MC/YDu/rkyjFEREREROTapMCQiEh+5FkQTFZDiV2eHj3A3x/q1Mk0cxew62eYUjbXjiMiIiIiItcWjUomIpIfuRcADEwKzpXdlSwJUVH/mBkGFK4Hp1bmyjFEREREROTao8CQiEh+dHaY+qQTV/Y4jcbDselgnGCpkamIiIiIyPVGgSERkXzJujqHKVACKr16dY4lIiIiIiJXnS7/iojkR1ey9U5wcNbTIiIiIiJy3VFgSEQkX7K4Yq2GTpzIelpERERERK47CgyJiORLV+lWMhERERERua4pMCQiki8Z+8mnBLj75W1SREREREQk31Ln0yIi+dlT4bm/zxIlMm4hK1Ei9/cvIiIiIiLXDAWGRETyI+PE1Woot4VfgWCTiIiIiIhck3QrmYiIiIiIiIjIDUqBIRGRfOkKtRYSEREREZEbigJDIiL5kXHmdQpEREREROQ6oMCQiEh+lBaLhqwXEREREZGcUmBIRCQ/So0HN/e8ToWIiIiIiORzCgyJiORLDnDzzutEiIiIiIhIPqfAkIhIfuThh24lExERERGRnFJgSEQkP/IuDmkxeZ0KERERERHJ5xQYEhHJjwoEQ9LJvE6FiIiIiIjkcwoMiYjkR95FIXIjGJPXKRERERERkXxMgSERkfzIqzCcXAaW+hkSEREREZHLp8CQiEh+5F0UjkyGFa3zOiUiIiIiIpKPKTAkIpIf+ZYCZyIcHpPXKRERERERkXxMgSERkfyoQGlwpOR1KkREREREJJ/LUWDIsqxWlmVttyzLaVlW/dxKlIiIXIRPCXDzyOtUiIiIiIhIPpfTFkPbgKeAZbmQFhERyS7LgqC6eZ0KERERERHJ53IUGDLG7DTG7M6txIiIyCUIfiivUyAiIiIiIvmc7kMQEcmvqrSDAiXBGA1bLyIiIiIil+WigSHLshYAwVks+soYMzW7B7Is63XgdYCyZctmO4EiInIBXoFQ6ZW8ToWIiIiIiORjFw0MGWMezI0DGWMGAgMB6tevb3JjnyIiIiIiIiIicvk0XL2IiIiIiIiIyA0qp8PVP2lZ1lGgITDTsqy5uZMsERERERERERG50nLU+bQxZjIwOZfSIiIiIiIiIiIiV5FuJRMRERERERERuUEpMCQiIiIiIiIicoNSYEhERERERERE5AalwJCIiIiIiIiIyA3KMsZc/YNaVgRwOP1lUeDUVU+EiFyvVKaISG5TuSIiuU3liojkpsxlSjljTLFL2ThPAkPnJMCy1hlj6udpIkTkuqEyRURym8oVEcltKldEJDfltEzRrWQiIiIiIiIiIjcoBYZERERERERERG5Q10JgaGBeJ0BErisqU0Qkt6lcEZHcpnJFRHJTjsqUPO9jSERERERERERE8sa10GJIRERERERERETyQJ4FhizLampZ1m7LsvZZlvVFXqVDRPIXy7KGWJZ10rKsbZnmFbYsa75lWXvTn4PS51uWZfVNL2e2WJZVN+9SLiLXIsuybrIsa7FlWTsty9puWdb76fNVrojIZbEsy8eyrLWWZW1OL1c6pc+vYFnWmvRy5S/LsrzS53unv96Xvrx8XqZfRK5NlmW5W5a10bKsGemvc61MyZPAkGVZ7kA/4FHgNuA5y7Juy4u0iEi+Mwxo+o95XwALjTGVgYXpr8EuYyqnP14HfrtKaRSR/CMN+NgYUxVoALyTXidRuSIilysZaGKMqQXUBppaltUA+AHolV6uRAGvpK//ChBljKkE9EpfT0Tkn94HdmZ6nWtlSl61GLoD2GeMOWCMSQHGAi3yKC0iko8YY5YBkf+Y3QIYnj49HGiZaf4IY1sNFLIsq+TVSamI5AfGmOPGmA3p07HYFa7SqFwRkcuUXj7Epb/0TH8YoAkwIX3+P8uVs+XNBOABy7Ksq5RcEckHLMsqAzQDBqe/tsjFMiWvAkOlgSOZXh9NnycicjlKGGOOg32SBxRPn6+yRkSyLb2pdR1gDSpXRCQH0m/52AScBOYD+4FoY0xa+iqZyw5XuZK+/Mz/27ufF5vCOI7j72+DKGUy7IakLGxkqVhMstJkNWqKTP4DCyk2StnKwpaNpKYIf8BQlsjCwk6S1MxiwkJZ8LE4Z8bF+DVuzkz3/dqce55zFs/qU/dzzvMcYOT/zljSCncZOAN8bs9H6GOmdFUMLdVW+Xk0Sf1m1kj6I1W1EbgFnEry/le3LjFmrkj6RpJPSfYCozSrJXYvdVt7NFck/VRVjQNzSZ70Di9x67Izpati6DWwred8FHjT0VwkrX6zC0s52uNcO27WSPqtqlpLUwrdSHK7HTZXJP2zJG+BBzR7mA1X1Zr2Um92LOZKe30TPy6blzS49gNHquolzTY8B2neIOpbpnRVDD0CdrW7aK8DJoF7Hc1F0up3D5hqf08Bd3vGT7RfEdoHvFtYGiJJsLhG/yrwPMmlnkvmiqRlqaqtVTXc/t4AHKLZv+w+MNHe9n2uLOTNBDCTxDeGJAGQ5GyS0SQ7aLqTmSTH6GOmVFeZU1WHaVquIeBakoudTETSqlJVN4ExYAswC5wH7gDTwHbgFXA0yXz7h+8KzVfMPgAnkzzuYt6SVqaqOgA8BJ7xdd3+OZp9hswVSX+tqvbQbPw6RPMgfjrJharaSfO0fzPwFDie5GNVrQeu0+xxNg9MJnnRzewlrWRVNQacTjLez0zprBiSJEmSJElSt7paSiZJkiRJkqSOWQxJkiRJkiQNKIshSZIkSZKkAWUxJEmSJEmSNKAshiRJkiRJkgaUxZAkSZIkSdKAjb/QDgAAABlJREFUshiSJEmSJEkaUBZDkiRJkiRJA+oLAWBrhzYF828AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ISM\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "===============================\n",
      "chr1:35506700-35507700\n",
      "-------------------------------\n",
      "With prior:\n",
      "Input gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DeepSHAP\n",
      "Warning: unrecognized nn.Module: BinaryPredictor; using regular gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ISM\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "No prior:\n",
      "Input gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DeepSHAP\n",
      "Warning: unrecognized nn.Module: BinaryPredictor; using regular gradients\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ISM\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x144 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "===============================\n"
     ]
    }
   ],
   "source": [
    "# Show the importance scores\n",
    "genome_reader = pyfaidx.Fasta(reference_fasta_path)\n",
    "center_slice = slice(300, 700)\n",
    "for chrom, start, end in pos_coords:\n",
    "    print(\"%s:%d-%d\" % (chrom, start, end))\n",
    "    print(\"-------------------------------\")\n",
    "    \n",
    "    input_seq = dna_to_one_hot([genome_reader[chrom][start:end].seq])[0]\n",
    "    \n",
    "    print(\"With prior:\")\n",
    "    print(\"Input gradients\")\n",
    "    viz_sequence.plot_weights(\n",
    "        interpret_input_grad(prior_model, input_seq)[center_slice],\n",
    "        subticks_frequency=100\n",
    "    )\n",
    "    print(\"DeepSHAP\")\n",
    "    viz_sequence.plot_weights(\n",
    "        interpret_deepshap(prior_model, input_seq)[center_slice],\n",
    "        subticks_frequency=100\n",
    "    )\n",
    "    print(\"ISM\")\n",
    "    viz_sequence.plot_weights(\n",
    "        interpret_ism(prior_model, input_seq)[center_slice],\n",
    "        subticks_frequency=100\n",
    "    )\n",
    "    print(\"No prior:\")\n",
    "    print(\"Input gradients\")\n",
    "    viz_sequence.plot_weights(\n",
    "        interpret_input_grad(noprior_model, input_seq)[center_slice],\n",
    "        subticks_frequency=100\n",
    "    )\n",
    "    print(\"DeepSHAP\")\n",
    "    viz_sequence.plot_weights(\n",
    "        interpret_deepshap(noprior_model, input_seq)[center_slice],\n",
    "        subticks_frequency=100\n",
    "    )\n",
    "    print(\"ISM\")\n",
    "    viz_sequence.plot_weights(\n",
    "        interpret_ism(noprior_model, input_seq)[center_slice],\n",
    "        subticks_frequency=100\n",
    "    )\n",
    "    print(\"===============================\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}