SCDA_test.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h1>Sparse Convolutional Denoising Autoencoders for Genotype Imputation <span class=\"tocSkip\"></span></h1>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "toc": true
   },
   "source": [
    "<h1>Table of Contents<span class=\"tocSkip\"></span></h1>\n",
    "<div class=\"toc\"><ul class=\"toc-item\"><li><span><a href=\"#Introduction\" data-toc-modified-id=\"Introduction-1\"><span class=\"toc-item-num\">1&nbsp;&nbsp;</span>Introduction</a></span></li><li><span><a href=\"#Dataset\" data-toc-modified-id=\"Dataset-2\"><span class=\"toc-item-num\">2&nbsp;&nbsp;</span>Dataset</a></span><ul class=\"toc-item\"><li><span><a href=\"#Loading-data\" data-toc-modified-id=\"Loading-data-2.1\"><span class=\"toc-item-num\">2.1&nbsp;&nbsp;</span>Loading data</a></span></li><li><span><a href=\"#Preprocessing\" data-toc-modified-id=\"Preprocessing-2.2\"><span class=\"toc-item-num\">2.2&nbsp;&nbsp;</span>Preprocessing</a></span></li></ul></li><li><span><a href=\"#Method\" data-toc-modified-id=\"Method-3\"><span class=\"toc-item-num\">3&nbsp;&nbsp;</span>Method</a></span><ul class=\"toc-item\"><li><span><a href=\"#Load-model\" data-toc-modified-id=\"Load-model-3.1\"><span class=\"toc-item-num\">3.1&nbsp;&nbsp;</span>Load model</a></span></li><li><span><a href=\"#Prediction-on-test-data\" data-toc-modified-id=\"Prediction-on-test-data-3.2\"><span class=\"toc-item-num\">3.2&nbsp;&nbsp;</span>Prediction on test data</a></span></li></ul></li></ul></div>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Introduction\n",
    "\n",
    "This notebook demonstrates a case study of testing a SCDA model on yeast genotype dataset with 10% missing genotypes. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jchen67/Experiment/deeplearning/venv/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:516: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
      "/Users/jchen67/Experiment/deeplearning/venv/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:517: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
      "/Users/jchen67/Experiment/deeplearning/venv/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:518: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
      "/Users/jchen67/Experiment/deeplearning/venv/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:519: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
      "/Users/jchen67/Experiment/deeplearning/venv/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:520: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
      "/Users/jchen67/Experiment/deeplearning/venv/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:525: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n",
      "/Users/jchen67/Experiment/deeplearning/venv/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:541: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
      "/Users/jchen67/Experiment/deeplearning/venv/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:542: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
      "/Users/jchen67/Experiment/deeplearning/venv/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:543: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
      "/Users/jchen67/Experiment/deeplearning/venv/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:544: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
      "/Users/jchen67/Experiment/deeplearning/venv/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:545: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
      "/Users/jchen67/Experiment/deeplearning/venv/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:550: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Conv1D, MaxPooling1D, UpSampling1D, Dropout\n",
    "from tensorflow.keras.regularizers import l1\n",
    "from tensorflow.keras.utils import to_categorical\n",
    "\n",
    "from tensorflow.keras.models import load_model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Loading data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(877, 28220)"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "input_name = 'data/yeast_genotype_test.txt'\n",
    "df_ori = pd.read_csv(input_name, sep='\\t', index_col=0)\n",
    "df_ori.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>33070_chrI_33070_A_T</th>\n",
       "      <th>33147_chrI_33147_G_T</th>\n",
       "      <th>33152_chrI_33152_T_C</th>\n",
       "      <th>33200_chrI_33200_C_T</th>\n",
       "      <th>33293_chrI_33293_A_T</th>\n",
       "      <th>33328_chrI_33328_C_A</th>\n",
       "      <th>33348_chrI_33348_G_C</th>\n",
       "      <th>33403_chrI_33403_C_T</th>\n",
       "      <th>33502_chrI_33502_A_G</th>\n",
       "      <th>33548_chrI_33548_A_C</th>\n",
       "      <th>...</th>\n",
       "      <th>12048853_chrXVI_925593_G_C</th>\n",
       "      <th>12049199_chrXVI_925939_T_C</th>\n",
       "      <th>12049441_chrXVI_926181_C_T</th>\n",
       "      <th>12050613_chrXVI_927353_T_G</th>\n",
       "      <th>12051167_chrXVI_927907_A_C</th>\n",
       "      <th>12051240_chrXVI_927980_A_G</th>\n",
       "      <th>12051367_chrXVI_928107_C_T</th>\n",
       "      <th>12052782_chrXVI_929522_C_T</th>\n",
       "      <th>12052988_chrXVI_929728_A_G</th>\n",
       "      <th>12053130_chrXVI_929870_C_T</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SAMID</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>40_73</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40_74</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40_75</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40_76</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40_77</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 28220 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       33070_chrI_33070_A_T  33147_chrI_33147_G_T  33152_chrI_33152_T_C  \\\n",
       "SAMID                                                                     \n",
       "40_73                     1                     1                     1   \n",
       "40_74                     1                     1                     1   \n",
       "40_75                     2                     2                     2   \n",
       "40_76                     2                     2                     2   \n",
       "40_77                     1                     1                     1   \n",
       "\n",
       "       33200_chrI_33200_C_T  33293_chrI_33293_A_T  33328_chrI_33328_C_A  \\\n",
       "SAMID                                                                     \n",
       "40_73                     1                     1                     1   \n",
       "40_74                     1                     1                     1   \n",
       "40_75                     2                     2                     2   \n",
       "40_76                     2                     2                     2   \n",
       "40_77                     1                     1                     1   \n",
       "\n",
       "       33348_chrI_33348_G_C  33403_chrI_33403_C_T  33502_chrI_33502_A_G  \\\n",
       "SAMID                                                                     \n",
       "40_73                     1                     1                     1   \n",
       "40_74                     1                     1                     1   \n",
       "40_75                     2                     2                     2   \n",
       "40_76                     2                     2                     2   \n",
       "40_77                     1                     1                     1   \n",
       "\n",
       "       33548_chrI_33548_A_C  ...  12048853_chrXVI_925593_G_C  \\\n",
       "SAMID                        ...                               \n",
       "40_73                     1  ...                           2   \n",
       "40_74                     1  ...                           1   \n",
       "40_75                     2  ...                           1   \n",
       "40_76                     2  ...                           2   \n",
       "40_77                     1  ...                           2   \n",
       "\n",
       "       12049199_chrXVI_925939_T_C  12049441_chrXVI_926181_C_T  \\\n",
       "SAMID                                                           \n",
       "40_73                           2                           2   \n",
       "40_74                           1                           1   \n",
       "40_75                           1                           1   \n",
       "40_76                           2                           2   \n",
       "40_77                           2                           2   \n",
       "\n",
       "       12050613_chrXVI_927353_T_G  12051167_chrXVI_927907_A_C  \\\n",
       "SAMID                                                           \n",
       "40_73                           2                           2   \n",
       "40_74                           1                           1   \n",
       "40_75                           1                           1   \n",
       "40_76                           2                           2   \n",
       "40_77                           2                           2   \n",
       "\n",
       "       12051240_chrXVI_927980_A_G  12051367_chrXVI_928107_C_T  \\\n",
       "SAMID                                                           \n",
       "40_73                           2                           2   \n",
       "40_74                           1                           1   \n",
       "40_75                           1                           1   \n",
       "40_76                           2                           2   \n",
       "40_77                           2                           2   \n",
       "\n",
       "       12052782_chrXVI_929522_C_T  12052988_chrXVI_929728_A_G  \\\n",
       "SAMID                                                           \n",
       "40_73                           2                           2   \n",
       "40_74                           1                           1   \n",
       "40_75                           1                           1   \n",
       "40_76                           2                           2   \n",
       "40_77                           2                           2   \n",
       "\n",
       "       12053130_chrXVI_929870_C_T  \n",
       "SAMID                              \n",
       "40_73                           2  \n",
       "40_74                           1  \n",
       "40_75                           1  \n",
       "40_76                           2  \n",
       "40_77                           2  \n",
       "\n",
       "[5 rows x 28220 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_ori.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(877, 28220, 3)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# one hot encode\n",
    "test_X = to_categorical(df_ori)\n",
    "test_X.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Method"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W0822 20:37:20.972470 140734891664832 deprecation.py:506] From /Users/jchen67/Experiment/deeplearning/venv/lib/python3.6/site-packages/tensorflow/python/ops/init_ops.py:97: calling GlorotUniform.__init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Call initializer instance with the dtype argument instead of passing it to the constructor\n",
      "W0822 20:37:20.973669 140734891664832 deprecation.py:506] From /Users/jchen67/Experiment/deeplearning/venv/lib/python3.6/site-packages/tensorflow/python/ops/init_ops.py:1251: calling VarianceScaling.__init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Call initializer instance with the dtype argument instead of passing it to the constructor\n",
      "W0822 20:37:20.975331 140734891664832 deprecation.py:506] From /Users/jchen67/Experiment/deeplearning/venv/lib/python3.6/site-packages/tensorflow/python/ops/init_ops.py:97: calling Zeros.__init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Call initializer instance with the dtype argument instead of passing it to the constructor\n",
      "W0822 20:37:36.897370 140734891664832 deprecation.py:323] From /Users/jchen67/Experiment/deeplearning/venv/lib/python3.6/site-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n"
     ]
    }
   ],
   "source": [
    "# returns a compiled model\n",
    "SCDA = load_model('model/SCDA.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Prediction on test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# hyperparameters\n",
    "missing_perc = 0.1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(877, 28220, 3)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_X_missing = test_X.copy()\n",
    "test_X_missing.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def cal_prob(predict_missing_onehot):\n",
    "    # calcaulate the probility of genotype 0, 1, 2\n",
    "    predict_prob = predict_missing_onehot[:,:,1:3] / predict_missing_onehot[:,:,1:3].sum(axis=2, keepdims=True)\n",
    "    return predict_prob[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0/877, sample ID: 40_73, accuracy: 0.9975\n",
      "1/877, sample ID: 40_74, accuracy: 0.9975\n",
      "2/877, sample ID: 40_75, accuracy: 0.9982\n",
      "3/877, sample ID: 40_76, accuracy: 0.9979\n",
      "4/877, sample ID: 40_77, accuracy: 0.9972\n",
      "5/877, sample ID: 40_79, accuracy: 0.9968\n",
      "6/877, sample ID: 40_80, accuracy: 0.9979\n",
      "7/877, sample ID: 40_81, accuracy: 0.9986\n",
      "8/877, sample ID: 40_82, accuracy: 0.9986\n",
      "9/877, sample ID: 40_83, accuracy: 0.9979\n",
      "10/877, sample ID: 40_84, accuracy: 0.9989\n",
      "11/877, sample ID: 40_85, accuracy: 0.9968\n",
      "12/877, sample ID: 40_86, accuracy: 0.9975\n",
      "13/877, sample ID: 40_87, accuracy: 0.9989\n",
      "14/877, sample ID: 40_88, accuracy: 0.9979\n",
      "15/877, sample ID: 40_89, accuracy: 0.9979\n",
      "16/877, sample ID: 40_90, accuracy: 0.9961\n",
      "17/877, sample ID: 40_91, accuracy: 0.9982\n",
      "18/877, sample ID: 40_92, accuracy: 0.9972\n",
      "19/877, sample ID: 40_93, accuracy: 0.9993\n",
      "20/877, sample ID: 40_94, accuracy: 0.9989\n",
      "21/877, sample ID: 40_95, accuracy: 0.9979\n",
      "22/877, sample ID: 40_96, accuracy: 0.9968\n",
      "23/877, sample ID: 41_01, accuracy: 0.9989\n",
      "24/877, sample ID: 41_02, accuracy: 0.9989\n",
      "25/877, sample ID: 41_03, accuracy: 0.9979\n",
      "26/877, sample ID: 41_04, accuracy: 0.9975\n",
      "27/877, sample ID: 41_05, accuracy: 0.9979\n",
      "28/877, sample ID: 41_06, accuracy: 0.9979\n",
      "29/877, sample ID: 41_07, accuracy: 0.9961\n",
      "30/877, sample ID: 41_08, accuracy: 0.9993\n",
      "31/877, sample ID: 41_09, accuracy: 0.9982\n",
      "32/877, sample ID: 41_10, accuracy: 0.9993\n",
      "33/877, sample ID: 41_11, accuracy: 0.9993\n",
      "34/877, sample ID: 41_12, accuracy: 0.9993\n",
      "35/877, sample ID: 41_13, accuracy: 0.9996\n",
      "36/877, sample ID: 41_14, accuracy: 0.9993\n",
      "37/877, sample ID: 41_15, accuracy: 0.9965\n",
      "38/877, sample ID: 41_16, accuracy: 0.9982\n",
      "39/877, sample ID: 41_17, accuracy: 0.9957\n",
      "40/877, sample ID: 41_18, accuracy: 0.9982\n",
      "41/877, sample ID: 41_19, accuracy: 0.9982\n",
      "42/877, sample ID: 41_20, accuracy: 0.9986\n",
      "43/877, sample ID: 41_21, accuracy: 0.9961\n",
      "44/877, sample ID: 41_22, accuracy: 0.9968\n",
      "45/877, sample ID: 41_23, accuracy: 0.9975\n",
      "46/877, sample ID: 41_24, accuracy: 0.9982\n",
      "47/877, sample ID: 41_25, accuracy: 0.9954\n",
      "48/877, sample ID: 41_26, accuracy: 0.9975\n",
      "49/877, sample ID: 41_27, accuracy: 0.9996\n",
      "50/877, sample ID: 41_28, accuracy: 0.9972\n",
      "51/877, sample ID: 41_29, accuracy: 0.9986\n",
      "52/877, sample ID: 41_30, accuracy: 0.9979\n",
      "53/877, sample ID: 41_31, accuracy: 0.9982\n",
      "54/877, sample ID: 41_32, accuracy: 0.9993\n",
      "55/877, sample ID: 41_33, accuracy: 0.9989\n",
      "56/877, sample ID: 41_34, accuracy: 0.9979\n",
      "57/877, sample ID: 41_35, accuracy: 0.9979\n",
      "58/877, sample ID: 41_36, accuracy: 0.9975\n",
      "59/877, sample ID: 41_37, accuracy: 0.9986\n",
      "60/877, sample ID: 41_38, accuracy: 0.9989\n",
      "61/877, sample ID: 41_39, accuracy: 0.9986\n",
      "62/877, sample ID: 41_40, accuracy: 0.9993\n",
      "63/877, sample ID: 41_41, accuracy: 0.9982\n",
      "64/877, sample ID: 41_42, accuracy: 0.9979\n",
      "65/877, sample ID: 41_43, accuracy: 0.9965\n",
      "66/877, sample ID: 41_44, accuracy: 0.9986\n",
      "67/877, sample ID: 41_45, accuracy: 0.9975\n",
      "68/877, sample ID: 41_46, accuracy: 0.9975\n",
      "69/877, sample ID: 41_47, accuracy: 0.9972\n",
      "70/877, sample ID: 41_48, accuracy: 0.9989\n",
      "71/877, sample ID: 41_49, accuracy: 0.9986\n",
      "72/877, sample ID: 41_50, accuracy: 0.9954\n",
      "73/877, sample ID: 41_51, accuracy: 0.9961\n",
      "74/877, sample ID: 41_52, accuracy: 0.9972\n",
      "75/877, sample ID: 41_53, accuracy: 0.9979\n",
      "76/877, sample ID: 41_54, accuracy: 0.9986\n",
      "77/877, sample ID: 41_55, accuracy: 0.9968\n",
      "78/877, sample ID: 41_56, accuracy: 0.9975\n",
      "79/877, sample ID: 41_57, accuracy: 0.9961\n",
      "80/877, sample ID: 41_58, accuracy: 0.9989\n",
      "81/877, sample ID: 41_59, accuracy: 0.9986\n",
      "82/877, sample ID: 41_60, accuracy: 0.9989\n",
      "83/877, sample ID: 41_61, accuracy: 0.9982\n",
      "84/877, sample ID: 41_62, accuracy: 0.9979\n",
      "85/877, sample ID: 41_63, accuracy: 0.9993\n",
      "86/877, sample ID: 41_64, accuracy: 0.9996\n",
      "87/877, sample ID: 41_65, accuracy: 0.9989\n",
      "88/877, sample ID: 41_66, accuracy: 0.9986\n",
      "89/877, sample ID: 41_67, accuracy: 0.9975\n",
      "90/877, sample ID: 41_68, accuracy: 0.9989\n",
      "91/877, sample ID: 41_69, accuracy: 0.9989\n",
      "92/877, sample ID: 41_70, accuracy: 0.9975\n",
      "93/877, sample ID: 41_71, accuracy: 0.9982\n",
      "94/877, sample ID: 41_72, accuracy: 0.9986\n",
      "95/877, sample ID: 41_73, accuracy: 0.9986\n",
      "96/877, sample ID: 41_74, accuracy: 0.9979\n",
      "97/877, sample ID: 41_75, accuracy: 0.9993\n",
      "98/877, sample ID: 41_76, accuracy: 0.9986\n",
      "99/877, sample ID: 41_77, accuracy: 0.9968\n",
      "100/877, sample ID: 41_78, accuracy: 0.9986\n",
      "101/877, sample ID: 41_79, accuracy: 0.9989\n",
      "102/877, sample ID: 41_80, accuracy: 0.9979\n",
      "103/877, sample ID: 41_81, accuracy: 0.9979\n",
      "104/877, sample ID: 41_82, accuracy: 0.9986\n",
      "105/877, sample ID: 41_83, accuracy: 0.9989\n",
      "106/877, sample ID: 41_84, accuracy: 0.9975\n",
      "107/877, sample ID: 41_85, accuracy: 0.9979\n",
      "108/877, sample ID: 41_86, accuracy: 0.9993\n",
      "109/877, sample ID: 41_87, accuracy: 0.9982\n",
      "110/877, sample ID: 41_88, accuracy: 0.9993\n",
      "111/877, sample ID: 41_89, accuracy: 0.9965\n",
      "112/877, sample ID: 41_90, accuracy: 0.9986\n",
      "113/877, sample ID: 41_91, accuracy: 0.9979\n",
      "114/877, sample ID: 41_92, accuracy: 0.9965\n",
      "115/877, sample ID: 41_93, accuracy: 0.9979\n",
      "116/877, sample ID: 41_94, accuracy: 0.9979\n",
      "117/877, sample ID: 41_95, accuracy: 0.9996\n",
      "118/877, sample ID: 41_96, accuracy: 0.9975\n",
      "119/877, sample ID: 42_01, accuracy: 0.9996\n",
      "120/877, sample ID: 42_02, accuracy: 0.9986\n",
      "121/877, sample ID: 42_03, accuracy: 0.9979\n",
      "122/877, sample ID: 42_04, accuracy: 0.9979\n",
      "123/877, sample ID: 42_05, accuracy: 0.9979\n",
      "124/877, sample ID: 42_06, accuracy: 0.9979\n",
      "125/877, sample ID: 42_07, accuracy: 0.9989\n",
      "126/877, sample ID: 42_08, accuracy: 0.9989\n",
      "127/877, sample ID: 42_09, accuracy: 0.9975\n",
      "128/877, sample ID: 42_10, accuracy: 0.9961\n",
      "129/877, sample ID: 42_11, accuracy: 0.9979\n",
      "130/877, sample ID: 42_12, accuracy: 0.9972\n",
      "131/877, sample ID: 42_13, accuracy: 0.9968\n",
      "132/877, sample ID: 42_14, accuracy: 0.9965\n",
      "133/877, sample ID: 42_15, accuracy: 0.9975\n",
      "134/877, sample ID: 42_16, accuracy: 0.9979\n",
      "135/877, sample ID: 42_17, accuracy: 0.9989\n",
      "136/877, sample ID: 42_18, accuracy: 0.9982\n",
      "137/877, sample ID: 42_19, accuracy: 0.9982\n",
      "138/877, sample ID: 42_20, accuracy: 0.9965\n",
      "139/877, sample ID: 42_21, accuracy: 0.9989\n",
      "140/877, sample ID: 42_22, accuracy: 0.9979\n",
      "141/877, sample ID: 42_23, accuracy: 0.9972\n",
      "142/877, sample ID: 42_24, accuracy: 0.9972\n",
      "143/877, sample ID: 42_25, accuracy: 0.9979\n",
      "144/877, sample ID: 42_26, accuracy: 0.9982\n",
      "145/877, sample ID: 42_27, accuracy: 0.9986\n",
      "146/877, sample ID: 42_28, accuracy: 0.9982\n",
      "147/877, sample ID: 42_29, accuracy: 0.9996\n",
      "148/877, sample ID: 42_30, accuracy: 0.9975\n",
      "149/877, sample ID: 42_31, accuracy: 0.9968\n",
      "150/877, sample ID: 42_32, accuracy: 0.9996\n",
      "151/877, sample ID: 42_33, accuracy: 0.9982\n",
      "152/877, sample ID: 42_34, accuracy: 0.9961\n",
      "153/877, sample ID: 42_35, accuracy: 0.9975\n",
      "154/877, sample ID: 42_36, accuracy: 0.9982\n",
      "155/877, sample ID: 42_37, accuracy: 0.9989\n",
      "156/877, sample ID: 42_38, accuracy: 0.9986\n",
      "157/877, sample ID: 42_39, accuracy: 0.9986\n",
      "158/877, sample ID: 42_40, accuracy: 0.9982\n",
      "159/877, sample ID: 42_41, accuracy: 0.9979\n",
      "160/877, sample ID: 42_42, accuracy: 0.9972\n",
      "161/877, sample ID: 42_43, accuracy: 0.9965\n",
      "162/877, sample ID: 42_44, accuracy: 0.9982\n",
      "163/877, sample ID: 42_45, accuracy: 0.9975\n",
      "164/877, sample ID: 42_46, accuracy: 0.9982\n",
      "165/877, sample ID: 42_47, accuracy: 0.9982\n",
      "166/877, sample ID: 42_48, accuracy: 0.9972\n",
      "167/877, sample ID: 42_49, accuracy: 0.9986\n",
      "168/877, sample ID: 42_50, accuracy: 0.9975\n",
      "169/877, sample ID: 42_51, accuracy: 0.9975\n",
      "170/877, sample ID: 42_52, accuracy: 0.9986\n",
      "171/877, sample ID: 42_53, accuracy: 0.9975\n",
      "172/877, sample ID: 42_54, accuracy: 0.9982\n",
      "173/877, sample ID: 42_55, accuracy: 0.9986\n",
      "174/877, sample ID: 42_56, accuracy: 0.9979\n",
      "175/877, sample ID: 42_57, accuracy: 0.9989\n",
      "176/877, sample ID: 42_58, accuracy: 0.9961\n",
      "177/877, sample ID: 42_59, accuracy: 0.9986\n",
      "178/877, sample ID: 42_60, accuracy: 0.9989\n",
      "179/877, sample ID: 42_61, accuracy: 0.9968\n",
      "180/877, sample ID: 42_62, accuracy: 0.9957\n",
      "181/877, sample ID: 42_63, accuracy: 0.9972\n",
      "182/877, sample ID: 42_64, accuracy: 0.9986\n",
      "183/877, sample ID: 42_65, accuracy: 0.9979\n",
      "184/877, sample ID: 42_66, accuracy: 0.9965\n",
      "185/877, sample ID: 42_67, accuracy: 0.9989\n",
      "186/877, sample ID: 42_68, accuracy: 0.9972\n",
      "187/877, sample ID: 42_69, accuracy: 0.9972\n",
      "188/877, sample ID: 42_70, accuracy: 0.9975\n",
      "189/877, sample ID: 42_71, accuracy: 0.9957\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "190/877, sample ID: 42_72, accuracy: 0.9965\n",
      "191/877, sample ID: 42_73, accuracy: 0.9972\n",
      "192/877, sample ID: 42_74, accuracy: 0.9986\n",
      "193/877, sample ID: 42_75, accuracy: 0.9972\n",
      "194/877, sample ID: 42_76, accuracy: 0.9982\n",
      "195/877, sample ID: 42_77, accuracy: 0.9975\n",
      "196/877, sample ID: 42_78, accuracy: 0.9986\n",
      "197/877, sample ID: 42_79, accuracy: 0.9972\n",
      "198/877, sample ID: 42_80, accuracy: 0.9979\n",
      "199/877, sample ID: 42_81, accuracy: 0.9979\n",
      "200/877, sample ID: 42_82, accuracy: 0.9975\n",
      "201/877, sample ID: 42_83, accuracy: 0.9975\n",
      "202/877, sample ID: 42_84, accuracy: 0.9989\n",
      "203/877, sample ID: 42_85, accuracy: 0.9982\n",
      "204/877, sample ID: 42_86, accuracy: 0.9965\n",
      "205/877, sample ID: 42_87, accuracy: 0.9979\n",
      "206/877, sample ID: 42_88, accuracy: 0.9989\n",
      "207/877, sample ID: 42_89, accuracy: 0.9975\n",
      "208/877, sample ID: 42_90, accuracy: 0.9986\n",
      "209/877, sample ID: 42_91, accuracy: 0.9982\n",
      "210/877, sample ID: 42_92, accuracy: 0.9979\n",
      "211/877, sample ID: 42_93, accuracy: 0.9982\n",
      "212/877, sample ID: 42_94, accuracy: 0.9968\n",
      "213/877, sample ID: 42_95, accuracy: 0.9982\n",
      "214/877, sample ID: 42_96, accuracy: 0.9979\n",
      "215/877, sample ID: 43_01, accuracy: 0.9972\n",
      "216/877, sample ID: 43_02, accuracy: 0.9986\n",
      "217/877, sample ID: 43_03, accuracy: 0.9979\n",
      "218/877, sample ID: 43_04, accuracy: 0.9982\n",
      "219/877, sample ID: 43_05, accuracy: 0.9961\n",
      "220/877, sample ID: 43_06, accuracy: 0.9979\n",
      "221/877, sample ID: 43_07, accuracy: 0.9986\n",
      "222/877, sample ID: 43_08, accuracy: 0.9972\n",
      "223/877, sample ID: 43_09, accuracy: 0.9965\n",
      "224/877, sample ID: 43_10, accuracy: 0.9989\n",
      "225/877, sample ID: 43_11, accuracy: 0.9986\n",
      "226/877, sample ID: 43_12, accuracy: 0.9972\n",
      "227/877, sample ID: 43_13, accuracy: 0.9957\n",
      "228/877, sample ID: 43_14, accuracy: 0.9993\n",
      "229/877, sample ID: 43_15, accuracy: 0.9982\n",
      "230/877, sample ID: 43_16, accuracy: 0.9975\n",
      "231/877, sample ID: 43_17, accuracy: 0.9989\n",
      "232/877, sample ID: 43_18, accuracy: 0.9957\n",
      "233/877, sample ID: 43_19, accuracy: 0.9979\n",
      "234/877, sample ID: 43_20, accuracy: 0.9989\n",
      "235/877, sample ID: 43_21, accuracy: 0.9975\n",
      "236/877, sample ID: 43_22, accuracy: 0.9982\n",
      "237/877, sample ID: 43_23, accuracy: 0.9972\n",
      "238/877, sample ID: 43_24, accuracy: 0.9989\n",
      "239/877, sample ID: 43_25, accuracy: 0.9996\n",
      "240/877, sample ID: 43_26, accuracy: 0.9965\n",
      "241/877, sample ID: 43_27, accuracy: 0.9979\n",
      "242/877, sample ID: 43_28, accuracy: 0.9961\n",
      "243/877, sample ID: 43_29, accuracy: 0.9979\n",
      "244/877, sample ID: 43_30, accuracy: 0.9982\n",
      "245/877, sample ID: 43_31, accuracy: 0.9989\n",
      "246/877, sample ID: 43_32, accuracy: 0.9982\n",
      "247/877, sample ID: 43_33, accuracy: 0.9993\n",
      "248/877, sample ID: 43_34, accuracy: 0.9961\n",
      "249/877, sample ID: 43_35, accuracy: 0.9986\n",
      "250/877, sample ID: 43_36, accuracy: 0.9975\n",
      "251/877, sample ID: 43_37, accuracy: 0.9965\n",
      "252/877, sample ID: 43_38, accuracy: 0.9965\n",
      "253/877, sample ID: 43_39, accuracy: 0.9993\n",
      "254/877, sample ID: 43_40, accuracy: 0.9957\n",
      "255/877, sample ID: 43_41, accuracy: 0.9972\n",
      "256/877, sample ID: 43_42, accuracy: 0.9996\n",
      "257/877, sample ID: 43_43, accuracy: 0.9986\n",
      "258/877, sample ID: 43_44, accuracy: 0.9982\n",
      "259/877, sample ID: 43_45, accuracy: 0.9972\n",
      "260/877, sample ID: 43_46, accuracy: 0.9996\n",
      "261/877, sample ID: 43_47, accuracy: 0.9993\n",
      "262/877, sample ID: 43_48, accuracy: 0.9972\n",
      "263/877, sample ID: 43_49, accuracy: 0.9986\n",
      "264/877, sample ID: 43_50, accuracy: 0.9989\n",
      "265/877, sample ID: 43_51, accuracy: 0.9993\n",
      "266/877, sample ID: 43_52, accuracy: 0.9982\n",
      "267/877, sample ID: 43_53, accuracy: 0.9979\n",
      "268/877, sample ID: 43_54, accuracy: 0.9972\n",
      "269/877, sample ID: 43_55, accuracy: 0.9975\n",
      "270/877, sample ID: 43_56, accuracy: 0.9986\n",
      "271/877, sample ID: 43_57, accuracy: 0.9979\n",
      "272/877, sample ID: 43_58, accuracy: 0.9993\n",
      "273/877, sample ID: 43_59, accuracy: 0.9968\n",
      "274/877, sample ID: 43_60, accuracy: 0.9972\n",
      "275/877, sample ID: 43_61, accuracy: 0.9982\n",
      "276/877, sample ID: 43_62, accuracy: 0.9986\n",
      "277/877, sample ID: 43_63, accuracy: 0.9982\n",
      "278/877, sample ID: 43_64, accuracy: 0.9972\n",
      "279/877, sample ID: 43_65, accuracy: 0.9996\n",
      "280/877, sample ID: 43_66, accuracy: 0.9989\n",
      "281/877, sample ID: 43_67, accuracy: 0.9989\n",
      "282/877, sample ID: 43_68, accuracy: 0.9986\n",
      "283/877, sample ID: 43_69, accuracy: 0.9982\n",
      "284/877, sample ID: 43_70, accuracy: 0.9986\n",
      "285/877, sample ID: 43_71, accuracy: 0.9972\n",
      "286/877, sample ID: 43_72, accuracy: 0.9979\n",
      "287/877, sample ID: 43_73, accuracy: 0.9989\n",
      "288/877, sample ID: 43_74, accuracy: 0.9975\n",
      "289/877, sample ID: 43_75, accuracy: 0.9965\n",
      "290/877, sample ID: 43_76, accuracy: 0.9975\n",
      "291/877, sample ID: 43_77, accuracy: 0.9979\n",
      "292/877, sample ID: 43_78, accuracy: 0.9961\n",
      "293/877, sample ID: 43_79, accuracy: 0.9979\n",
      "294/877, sample ID: 43_80, accuracy: 0.9993\n",
      "295/877, sample ID: 43_81, accuracy: 0.9979\n",
      "296/877, sample ID: 43_82, accuracy: 0.9989\n",
      "297/877, sample ID: 43_83, accuracy: 0.9957\n",
      "298/877, sample ID: 43_84, accuracy: 0.9975\n",
      "299/877, sample ID: 43_85, accuracy: 0.9982\n",
      "300/877, sample ID: 43_86, accuracy: 0.9954\n",
      "301/877, sample ID: 43_87, accuracy: 0.9972\n",
      "302/877, sample ID: 43_88, accuracy: 0.9965\n",
      "303/877, sample ID: 43_89, accuracy: 0.9979\n",
      "304/877, sample ID: 43_90, accuracy: 0.9986\n",
      "305/877, sample ID: 43_91, accuracy: 0.9975\n",
      "306/877, sample ID: 43_92, accuracy: 0.9982\n",
      "307/877, sample ID: 43_93, accuracy: 0.9982\n",
      "308/877, sample ID: 43_94, accuracy: 0.9979\n",
      "309/877, sample ID: 43_95, accuracy: 0.9979\n",
      "310/877, sample ID: 43_96, accuracy: 0.9986\n",
      "311/877, sample ID: 44_01, accuracy: 0.9986\n",
      "312/877, sample ID: 44_02, accuracy: 0.9982\n",
      "313/877, sample ID: 44_03, accuracy: 0.9975\n",
      "314/877, sample ID: 44_04, accuracy: 0.9982\n",
      "315/877, sample ID: 44_05, accuracy: 0.9986\n",
      "316/877, sample ID: 44_06, accuracy: 0.9972\n",
      "317/877, sample ID: 44_07, accuracy: 0.9965\n",
      "318/877, sample ID: 44_08, accuracy: 0.9972\n",
      "319/877, sample ID: 44_09, accuracy: 0.9982\n",
      "320/877, sample ID: 44_10, accuracy: 0.9975\n",
      "321/877, sample ID: 44_11, accuracy: 0.9982\n",
      "322/877, sample ID: 44_12, accuracy: 0.9986\n",
      "323/877, sample ID: 44_13, accuracy: 0.9961\n",
      "324/877, sample ID: 44_14, accuracy: 0.9986\n",
      "325/877, sample ID: 44_15, accuracy: 0.9989\n",
      "326/877, sample ID: 44_16, accuracy: 0.9986\n",
      "327/877, sample ID: 44_17, accuracy: 0.9979\n",
      "328/877, sample ID: 44_18, accuracy: 0.9968\n",
      "329/877, sample ID: 44_19, accuracy: 0.9989\n",
      "330/877, sample ID: 44_20, accuracy: 0.9986\n",
      "331/877, sample ID: 44_21, accuracy: 0.9982\n",
      "332/877, sample ID: 44_22, accuracy: 0.9989\n",
      "333/877, sample ID: 44_23, accuracy: 0.9982\n",
      "334/877, sample ID: 44_24, accuracy: 0.9979\n",
      "335/877, sample ID: 44_25, accuracy: 0.9972\n",
      "336/877, sample ID: 44_26, accuracy: 0.9986\n",
      "337/877, sample ID: 44_27, accuracy: 0.9982\n",
      "338/877, sample ID: 44_28, accuracy: 0.9979\n",
      "339/877, sample ID: 44_29, accuracy: 0.9979\n",
      "340/877, sample ID: 44_30, accuracy: 0.9982\n",
      "341/877, sample ID: 44_32, accuracy: 0.9968\n",
      "342/877, sample ID: 44_33, accuracy: 0.9986\n",
      "343/877, sample ID: 44_34, accuracy: 0.9979\n",
      "344/877, sample ID: 44_35, accuracy: 0.9986\n",
      "345/877, sample ID: 44_36, accuracy: 0.9975\n",
      "346/877, sample ID: 44_37, accuracy: 0.9982\n",
      "347/877, sample ID: 44_38, accuracy: 0.9982\n",
      "348/877, sample ID: 44_39, accuracy: 0.9982\n",
      "349/877, sample ID: 44_40, accuracy: 0.9979\n",
      "350/877, sample ID: 44_41, accuracy: 0.9954\n",
      "351/877, sample ID: 44_42, accuracy: 0.9982\n",
      "352/877, sample ID: 44_43, accuracy: 0.9986\n",
      "353/877, sample ID: 44_44, accuracy: 0.9975\n",
      "354/877, sample ID: 44_45, accuracy: 0.9979\n",
      "355/877, sample ID: 44_46, accuracy: 0.9979\n",
      "356/877, sample ID: 44_47, accuracy: 0.9979\n",
      "357/877, sample ID: 44_48, accuracy: 0.9975\n",
      "358/877, sample ID: 44_49, accuracy: 0.9979\n",
      "359/877, sample ID: 44_50, accuracy: 0.9989\n",
      "360/877, sample ID: 44_51, accuracy: 0.9972\n",
      "361/877, sample ID: 44_52, accuracy: 0.9993\n",
      "362/877, sample ID: 44_53, accuracy: 0.9986\n",
      "363/877, sample ID: 44_54, accuracy: 0.9982\n",
      "364/877, sample ID: 44_55, accuracy: 0.9986\n",
      "365/877, sample ID: 44_56, accuracy: 0.9975\n",
      "366/877, sample ID: 44_57, accuracy: 0.9979\n",
      "367/877, sample ID: 44_58, accuracy: 0.9972\n",
      "368/877, sample ID: 44_59, accuracy: 0.9979\n",
      "369/877, sample ID: 44_60, accuracy: 0.9972\n",
      "370/877, sample ID: 44_61, accuracy: 0.9986\n",
      "371/877, sample ID: 44_62, accuracy: 0.9982\n",
      "372/877, sample ID: 44_63, accuracy: 0.9982\n",
      "373/877, sample ID: 44_64, accuracy: 0.9989\n",
      "374/877, sample ID: 44_65, accuracy: 0.9957\n",
      "375/877, sample ID: 44_66, accuracy: 0.9972\n",
      "376/877, sample ID: 44_67, accuracy: 0.9989\n",
      "377/877, sample ID: 44_68, accuracy: 0.9989\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "378/877, sample ID: 44_69, accuracy: 0.9996\n",
      "379/877, sample ID: 44_70, accuracy: 0.9986\n",
      "380/877, sample ID: 44_71, accuracy: 0.9996\n",
      "381/877, sample ID: 44_72, accuracy: 0.9982\n",
      "382/877, sample ID: 44_73, accuracy: 0.9989\n",
      "383/877, sample ID: 44_74, accuracy: 0.9975\n",
      "384/877, sample ID: 44_75, accuracy: 0.9993\n",
      "385/877, sample ID: 44_76, accuracy: 0.9972\n",
      "386/877, sample ID: 44_77, accuracy: 0.9975\n",
      "387/877, sample ID: 44_78, accuracy: 0.9965\n",
      "388/877, sample ID: 44_79, accuracy: 0.9979\n",
      "389/877, sample ID: 44_80, accuracy: 0.9996\n",
      "390/877, sample ID: 44_81, accuracy: 0.9996\n",
      "391/877, sample ID: 44_82, accuracy: 0.9972\n",
      "392/877, sample ID: 44_83, accuracy: 0.9982\n",
      "393/877, sample ID: 44_84, accuracy: 0.9972\n",
      "394/877, sample ID: 44_85, accuracy: 0.9979\n",
      "395/877, sample ID: 44_86, accuracy: 0.9986\n",
      "396/877, sample ID: 44_87, accuracy: 0.9996\n",
      "397/877, sample ID: 44_88, accuracy: 0.9972\n",
      "398/877, sample ID: 44_89, accuracy: 0.9972\n",
      "399/877, sample ID: 44_90, accuracy: 0.9982\n",
      "400/877, sample ID: 44_91, accuracy: 0.9979\n",
      "401/877, sample ID: 44_92, accuracy: 0.9993\n",
      "402/877, sample ID: 44_93, accuracy: 0.9989\n",
      "403/877, sample ID: 44_94, accuracy: 0.9982\n",
      "404/877, sample ID: 44_95, accuracy: 0.9986\n",
      "405/877, sample ID: 44_96, accuracy: 0.9982\n",
      "406/877, sample ID: 45_01, accuracy: 0.9996\n",
      "407/877, sample ID: 45_02, accuracy: 0.9979\n",
      "408/877, sample ID: 45_04, accuracy: 0.9972\n",
      "409/877, sample ID: 45_05, accuracy: 0.9982\n",
      "410/877, sample ID: 45_06, accuracy: 0.9993\n",
      "411/877, sample ID: 45_07, accuracy: 0.9979\n",
      "412/877, sample ID: 45_08, accuracy: 0.9975\n",
      "413/877, sample ID: 45_09, accuracy: 0.9982\n",
      "414/877, sample ID: 45_10, accuracy: 0.9943\n",
      "415/877, sample ID: 45_11, accuracy: 0.9982\n",
      "416/877, sample ID: 45_12, accuracy: 0.9982\n",
      "417/877, sample ID: 45_13, accuracy: 0.9986\n",
      "418/877, sample ID: 45_14, accuracy: 0.9979\n",
      "419/877, sample ID: 45_15, accuracy: 0.9996\n",
      "420/877, sample ID: 45_16, accuracy: 0.9982\n",
      "421/877, sample ID: 45_17, accuracy: 0.9986\n",
      "422/877, sample ID: 45_18, accuracy: 0.9982\n",
      "423/877, sample ID: 45_19, accuracy: 0.9972\n",
      "424/877, sample ID: 45_20, accuracy: 0.9965\n",
      "425/877, sample ID: 45_21, accuracy: 0.9982\n",
      "426/877, sample ID: 45_22, accuracy: 0.9982\n",
      "427/877, sample ID: 45_23, accuracy: 0.9986\n",
      "428/877, sample ID: 45_24, accuracy: 0.9975\n",
      "429/877, sample ID: 45_25, accuracy: 0.9979\n",
      "430/877, sample ID: 45_26, accuracy: 0.9982\n",
      "431/877, sample ID: 45_27, accuracy: 0.9982\n",
      "432/877, sample ID: 45_28, accuracy: 0.9972\n",
      "433/877, sample ID: 45_29, accuracy: 0.9975\n",
      "434/877, sample ID: 45_30, accuracy: 0.9986\n",
      "435/877, sample ID: 45_31, accuracy: 0.9982\n",
      "436/877, sample ID: 45_32, accuracy: 0.9982\n",
      "437/877, sample ID: 45_33, accuracy: 0.9989\n",
      "438/877, sample ID: 45_34, accuracy: 0.9972\n",
      "439/877, sample ID: 45_35, accuracy: 0.9961\n",
      "440/877, sample ID: 45_36, accuracy: 0.9982\n",
      "441/877, sample ID: 45_37, accuracy: 0.9975\n",
      "442/877, sample ID: 45_38, accuracy: 0.9989\n",
      "443/877, sample ID: 45_39, accuracy: 0.9986\n",
      "444/877, sample ID: 45_40, accuracy: 0.9986\n",
      "445/877, sample ID: 45_41, accuracy: 0.9972\n",
      "446/877, sample ID: 45_42, accuracy: 0.9968\n",
      "447/877, sample ID: 45_43, accuracy: 0.9979\n",
      "448/877, sample ID: 45_44, accuracy: 0.9961\n",
      "449/877, sample ID: 45_45, accuracy: 0.9979\n",
      "450/877, sample ID: 45_46, accuracy: 0.9972\n",
      "451/877, sample ID: 45_47, accuracy: 0.9982\n",
      "452/877, sample ID: 45_48, accuracy: 0.9972\n",
      "453/877, sample ID: 45_49, accuracy: 0.9986\n",
      "454/877, sample ID: 45_50, accuracy: 0.9975\n",
      "455/877, sample ID: 45_51, accuracy: 0.9975\n",
      "456/877, sample ID: 45_52, accuracy: 0.9979\n",
      "457/877, sample ID: 45_53, accuracy: 0.9968\n",
      "458/877, sample ID: 45_54, accuracy: 0.9975\n",
      "459/877, sample ID: 45_55, accuracy: 0.9996\n",
      "460/877, sample ID: 45_56, accuracy: 0.9965\n",
      "461/877, sample ID: 45_57, accuracy: 0.9989\n",
      "462/877, sample ID: 45_58, accuracy: 0.9957\n",
      "463/877, sample ID: 45_59, accuracy: 0.9979\n",
      "464/877, sample ID: 45_60, accuracy: 0.9982\n",
      "465/877, sample ID: 45_61, accuracy: 0.9993\n",
      "466/877, sample ID: 45_62, accuracy: 0.9975\n",
      "467/877, sample ID: 45_63, accuracy: 0.9979\n",
      "468/877, sample ID: 45_64, accuracy: 0.9979\n",
      "469/877, sample ID: 45_65, accuracy: 0.9961\n",
      "470/877, sample ID: 45_66, accuracy: 0.9979\n",
      "471/877, sample ID: 45_67, accuracy: 0.9989\n",
      "472/877, sample ID: 45_68, accuracy: 0.9986\n",
      "473/877, sample ID: 45_69, accuracy: 0.9993\n",
      "474/877, sample ID: 45_70, accuracy: 0.9975\n",
      "475/877, sample ID: 45_71, accuracy: 0.9979\n",
      "476/877, sample ID: 45_72, accuracy: 0.9989\n",
      "477/877, sample ID: 45_73, accuracy: 0.9975\n",
      "478/877, sample ID: 45_74, accuracy: 0.9968\n",
      "479/877, sample ID: 45_75, accuracy: 0.9979\n",
      "480/877, sample ID: 45_76, accuracy: 0.9975\n",
      "481/877, sample ID: 45_77, accuracy: 0.9989\n",
      "482/877, sample ID: 45_78, accuracy: 0.9979\n",
      "483/877, sample ID: 45_79, accuracy: 0.9986\n",
      "484/877, sample ID: 45_80, accuracy: 0.9982\n",
      "485/877, sample ID: 45_81, accuracy: 0.9972\n",
      "486/877, sample ID: 45_82, accuracy: 0.9986\n",
      "487/877, sample ID: 45_83, accuracy: 0.9968\n",
      "488/877, sample ID: 45_84, accuracy: 0.9975\n",
      "489/877, sample ID: 45_85, accuracy: 0.9979\n",
      "490/877, sample ID: 45_86, accuracy: 0.9979\n",
      "491/877, sample ID: 45_87, accuracy: 0.9975\n",
      "492/877, sample ID: 45_88, accuracy: 0.9975\n",
      "493/877, sample ID: 45_89, accuracy: 0.9982\n",
      "494/877, sample ID: 45_90, accuracy: 0.9979\n",
      "495/877, sample ID: 45_91, accuracy: 0.9979\n",
      "496/877, sample ID: 45_92, accuracy: 0.9968\n",
      "497/877, sample ID: 45_93, accuracy: 0.9965\n",
      "498/877, sample ID: 45_94, accuracy: 0.9975\n",
      "499/877, sample ID: 45_95, accuracy: 0.9972\n",
      "500/877, sample ID: 45_96, accuracy: 0.9972\n",
      "501/877, sample ID: 46_01, accuracy: 0.9972\n",
      "502/877, sample ID: 46_02, accuracy: 0.9986\n",
      "503/877, sample ID: 46_03, accuracy: 0.9979\n",
      "504/877, sample ID: 46_04, accuracy: 0.9986\n",
      "505/877, sample ID: 46_05, accuracy: 0.9972\n",
      "506/877, sample ID: 46_06, accuracy: 0.9961\n",
      "507/877, sample ID: 46_07, accuracy: 0.9982\n",
      "508/877, sample ID: 46_08, accuracy: 0.9975\n",
      "509/877, sample ID: 46_09, accuracy: 0.9989\n",
      "510/877, sample ID: 46_10, accuracy: 0.9986\n",
      "511/877, sample ID: 46_11, accuracy: 0.9968\n",
      "512/877, sample ID: 46_12, accuracy: 0.9961\n",
      "513/877, sample ID: 46_13, accuracy: 0.9972\n",
      "514/877, sample ID: 46_14, accuracy: 1.0000\n",
      "515/877, sample ID: 46_15, accuracy: 0.9993\n",
      "516/877, sample ID: 46_16, accuracy: 0.9972\n",
      "517/877, sample ID: 46_17, accuracy: 1.0000\n",
      "518/877, sample ID: 46_18, accuracy: 0.9972\n",
      "519/877, sample ID: 46_19, accuracy: 0.9986\n",
      "520/877, sample ID: 46_20, accuracy: 0.9979\n",
      "521/877, sample ID: 46_21, accuracy: 0.9975\n",
      "522/877, sample ID: 46_22, accuracy: 0.9975\n",
      "523/877, sample ID: 46_23, accuracy: 0.9986\n",
      "524/877, sample ID: 46_24, accuracy: 0.9986\n",
      "525/877, sample ID: 46_25, accuracy: 0.9989\n",
      "526/877, sample ID: 46_26, accuracy: 0.9979\n",
      "527/877, sample ID: 46_27, accuracy: 0.9979\n",
      "528/877, sample ID: 46_28, accuracy: 0.9968\n",
      "529/877, sample ID: 46_29, accuracy: 0.9982\n",
      "530/877, sample ID: 46_30, accuracy: 0.9979\n",
      "531/877, sample ID: 46_31, accuracy: 0.9989\n",
      "532/877, sample ID: 46_32, accuracy: 0.9982\n",
      "533/877, sample ID: 46_33, accuracy: 0.9982\n",
      "534/877, sample ID: 46_34, accuracy: 0.9986\n",
      "535/877, sample ID: 46_35, accuracy: 0.9982\n",
      "536/877, sample ID: 46_36, accuracy: 0.9986\n",
      "537/877, sample ID: 46_37, accuracy: 0.9968\n",
      "538/877, sample ID: 46_38, accuracy: 0.9993\n",
      "539/877, sample ID: 46_39, accuracy: 0.9979\n",
      "540/877, sample ID: 46_40, accuracy: 0.9982\n",
      "541/877, sample ID: 46_41, accuracy: 0.9986\n",
      "542/877, sample ID: 46_42, accuracy: 0.9972\n",
      "543/877, sample ID: 46_43, accuracy: 0.9982\n",
      "544/877, sample ID: 46_44, accuracy: 0.9972\n",
      "545/877, sample ID: 46_45, accuracy: 0.9972\n",
      "546/877, sample ID: 46_46, accuracy: 0.9993\n",
      "547/877, sample ID: 46_47, accuracy: 0.9975\n",
      "548/877, sample ID: 46_48, accuracy: 0.9986\n",
      "549/877, sample ID: 46_49, accuracy: 0.9986\n",
      "550/877, sample ID: 46_50, accuracy: 0.9780\n",
      "551/877, sample ID: 46_51, accuracy: 0.9968\n",
      "552/877, sample ID: 46_52, accuracy: 0.9986\n",
      "553/877, sample ID: 46_53, accuracy: 0.9986\n",
      "554/877, sample ID: 46_54, accuracy: 0.9989\n",
      "555/877, sample ID: 46_55, accuracy: 0.9986\n",
      "556/877, sample ID: 46_56, accuracy: 0.9982\n",
      "557/877, sample ID: 46_57, accuracy: 0.9989\n",
      "558/877, sample ID: 46_58, accuracy: 0.9979\n",
      "559/877, sample ID: 46_59, accuracy: 0.9982\n",
      "560/877, sample ID: 46_60, accuracy: 0.9972\n",
      "561/877, sample ID: 46_61, accuracy: 0.9972\n",
      "562/877, sample ID: 46_62, accuracy: 0.9989\n",
      "563/877, sample ID: 46_63, accuracy: 0.9986\n",
      "564/877, sample ID: 46_64, accuracy: 0.9972\n",
      "565/877, sample ID: 46_65, accuracy: 0.9989\n",
      "566/877, sample ID: 46_66, accuracy: 0.9965\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "567/877, sample ID: 46_67, accuracy: 0.9979\n",
      "568/877, sample ID: 46_68, accuracy: 0.9986\n",
      "569/877, sample ID: 46_69, accuracy: 0.9972\n",
      "570/877, sample ID: 46_70, accuracy: 0.9954\n",
      "571/877, sample ID: 46_71, accuracy: 0.9982\n",
      "572/877, sample ID: 46_72, accuracy: 0.9975\n",
      "573/877, sample ID: 46_73, accuracy: 0.9982\n",
      "574/877, sample ID: 46_74, accuracy: 0.9968\n",
      "575/877, sample ID: 46_75, accuracy: 0.9989\n",
      "576/877, sample ID: 46_76, accuracy: 0.9968\n",
      "577/877, sample ID: 46_77, accuracy: 0.9986\n",
      "578/877, sample ID: 46_78, accuracy: 0.9989\n",
      "579/877, sample ID: 46_79, accuracy: 0.9982\n",
      "580/877, sample ID: 46_80, accuracy: 0.9989\n",
      "581/877, sample ID: 46_81, accuracy: 0.9965\n",
      "582/877, sample ID: 46_82, accuracy: 0.9986\n",
      "583/877, sample ID: 46_83, accuracy: 0.9660\n",
      "584/877, sample ID: 46_84, accuracy: 0.9972\n",
      "585/877, sample ID: 46_85, accuracy: 0.9972\n",
      "586/877, sample ID: 46_86, accuracy: 0.9961\n",
      "587/877, sample ID: 46_87, accuracy: 0.9979\n",
      "588/877, sample ID: 46_88, accuracy: 0.9979\n",
      "589/877, sample ID: 46_89, accuracy: 0.9982\n",
      "590/877, sample ID: 46_90, accuracy: 0.9954\n",
      "591/877, sample ID: 46_91, accuracy: 0.9972\n",
      "592/877, sample ID: 46_92, accuracy: 0.9982\n",
      "593/877, sample ID: 46_93, accuracy: 0.9975\n",
      "594/877, sample ID: 46_94, accuracy: 0.9965\n",
      "595/877, sample ID: 46_95, accuracy: 0.9982\n",
      "596/877, sample ID: 46_96, accuracy: 0.9979\n",
      "597/877, sample ID: 47_02, accuracy: 0.9975\n",
      "598/877, sample ID: 47_03, accuracy: 0.9982\n",
      "599/877, sample ID: 47_04, accuracy: 0.9979\n",
      "600/877, sample ID: 47_05, accuracy: 0.9975\n",
      "601/877, sample ID: 47_06, accuracy: 0.9957\n",
      "602/877, sample ID: 47_07, accuracy: 0.9982\n",
      "603/877, sample ID: 47_08, accuracy: 0.9986\n",
      "604/877, sample ID: 47_09, accuracy: 0.9975\n",
      "605/877, sample ID: 47_10, accuracy: 0.9979\n",
      "606/877, sample ID: 47_11, accuracy: 0.9986\n",
      "607/877, sample ID: 47_12, accuracy: 0.9975\n",
      "608/877, sample ID: 47_13, accuracy: 0.9979\n",
      "609/877, sample ID: 47_14, accuracy: 0.9975\n",
      "610/877, sample ID: 47_15, accuracy: 0.9975\n",
      "611/877, sample ID: 47_16, accuracy: 0.9982\n",
      "612/877, sample ID: 47_17, accuracy: 0.9979\n",
      "613/877, sample ID: 47_18, accuracy: 0.9982\n",
      "614/877, sample ID: 47_19, accuracy: 0.9975\n",
      "615/877, sample ID: 47_20, accuracy: 0.9979\n",
      "616/877, sample ID: 47_21, accuracy: 0.9993\n",
      "617/877, sample ID: 47_22, accuracy: 0.9979\n",
      "618/877, sample ID: 47_23, accuracy: 0.9972\n",
      "619/877, sample ID: 47_24, accuracy: 0.9986\n",
      "620/877, sample ID: 47_25, accuracy: 0.9982\n",
      "621/877, sample ID: 47_26, accuracy: 0.9989\n",
      "622/877, sample ID: 47_27, accuracy: 0.9975\n",
      "623/877, sample ID: 47_28, accuracy: 0.9979\n",
      "624/877, sample ID: 47_29, accuracy: 0.9979\n",
      "625/877, sample ID: 47_30, accuracy: 0.9727\n",
      "626/877, sample ID: 47_31, accuracy: 0.9975\n",
      "627/877, sample ID: 47_32, accuracy: 0.9982\n",
      "628/877, sample ID: 47_33, accuracy: 0.9989\n",
      "629/877, sample ID: 47_34, accuracy: 0.9975\n",
      "630/877, sample ID: 47_35, accuracy: 0.9972\n",
      "631/877, sample ID: 47_36, accuracy: 0.9986\n",
      "632/877, sample ID: 47_37, accuracy: 0.9965\n",
      "633/877, sample ID: 47_38, accuracy: 0.9972\n",
      "634/877, sample ID: 47_39, accuracy: 0.9972\n",
      "635/877, sample ID: 47_40, accuracy: 0.9989\n",
      "636/877, sample ID: 47_41, accuracy: 0.9993\n",
      "637/877, sample ID: 47_42, accuracy: 0.9979\n",
      "638/877, sample ID: 47_43, accuracy: 0.9979\n",
      "639/877, sample ID: 47_44, accuracy: 0.9957\n",
      "640/877, sample ID: 47_45, accuracy: 0.9979\n",
      "641/877, sample ID: 47_46, accuracy: 0.9996\n",
      "642/877, sample ID: 47_47, accuracy: 0.9989\n",
      "643/877, sample ID: 47_48, accuracy: 0.9989\n",
      "644/877, sample ID: 47_49, accuracy: 0.9986\n",
      "645/877, sample ID: 47_50, accuracy: 0.9975\n",
      "646/877, sample ID: 47_51, accuracy: 0.9972\n",
      "647/877, sample ID: 47_52, accuracy: 0.9979\n",
      "648/877, sample ID: 47_53, accuracy: 0.9996\n",
      "649/877, sample ID: 47_54, accuracy: 0.9982\n",
      "650/877, sample ID: 47_55, accuracy: 0.9989\n",
      "651/877, sample ID: 47_56, accuracy: 0.9986\n",
      "652/877, sample ID: 47_57, accuracy: 0.9989\n",
      "653/877, sample ID: 47_58, accuracy: 0.9975\n",
      "654/877, sample ID: 47_59, accuracy: 0.9979\n",
      "655/877, sample ID: 47_60, accuracy: 0.9968\n",
      "656/877, sample ID: 47_61, accuracy: 0.9982\n",
      "657/877, sample ID: 47_62, accuracy: 0.9965\n",
      "658/877, sample ID: 47_63, accuracy: 0.9979\n",
      "659/877, sample ID: 47_64, accuracy: 0.9965\n",
      "660/877, sample ID: 47_65, accuracy: 0.9989\n",
      "661/877, sample ID: 47_66, accuracy: 0.9979\n",
      "662/877, sample ID: 47_67, accuracy: 0.9972\n",
      "663/877, sample ID: 47_68, accuracy: 0.9968\n",
      "664/877, sample ID: 47_69, accuracy: 0.9986\n",
      "665/877, sample ID: 47_70, accuracy: 0.9972\n",
      "666/877, sample ID: 47_71, accuracy: 0.9982\n",
      "667/877, sample ID: 47_72, accuracy: 0.9975\n",
      "668/877, sample ID: 47_73, accuracy: 0.9982\n",
      "669/877, sample ID: 47_74, accuracy: 0.9989\n",
      "670/877, sample ID: 47_75, accuracy: 1.0000\n",
      "671/877, sample ID: 47_76, accuracy: 0.9989\n",
      "672/877, sample ID: 47_77, accuracy: 0.9989\n",
      "673/877, sample ID: 47_78, accuracy: 0.9982\n",
      "674/877, sample ID: 47_79, accuracy: 0.9979\n",
      "675/877, sample ID: 47_80, accuracy: 0.9972\n",
      "676/877, sample ID: 47_81, accuracy: 0.9996\n",
      "677/877, sample ID: 47_82, accuracy: 0.9989\n",
      "678/877, sample ID: 47_83, accuracy: 0.9979\n",
      "679/877, sample ID: 47_84, accuracy: 0.9968\n",
      "680/877, sample ID: 47_85, accuracy: 0.9986\n",
      "681/877, sample ID: 47_86, accuracy: 0.9993\n",
      "682/877, sample ID: 47_87, accuracy: 0.9989\n",
      "683/877, sample ID: 47_88, accuracy: 0.9975\n",
      "684/877, sample ID: 47_89, accuracy: 0.9982\n",
      "685/877, sample ID: 47_90, accuracy: 0.9989\n",
      "686/877, sample ID: 47_91, accuracy: 0.9975\n",
      "687/877, sample ID: 47_92, accuracy: 0.9972\n",
      "688/877, sample ID: 47_93, accuracy: 0.9989\n",
      "689/877, sample ID: 47_94, accuracy: 0.9979\n",
      "690/877, sample ID: 47_95, accuracy: 0.9982\n",
      "691/877, sample ID: 47_96, accuracy: 0.9965\n",
      "692/877, sample ID: 48_02, accuracy: 0.9968\n",
      "693/877, sample ID: 48_03, accuracy: 0.9979\n",
      "694/877, sample ID: 48_04, accuracy: 0.9982\n",
      "695/877, sample ID: 48_05, accuracy: 0.9982\n",
      "696/877, sample ID: 48_06, accuracy: 0.9989\n",
      "697/877, sample ID: 48_07, accuracy: 0.9975\n",
      "698/877, sample ID: 48_08, accuracy: 0.9982\n",
      "699/877, sample ID: 48_09, accuracy: 0.9982\n",
      "700/877, sample ID: 48_10, accuracy: 0.9989\n",
      "701/877, sample ID: 48_11, accuracy: 0.9989\n",
      "702/877, sample ID: 48_12, accuracy: 0.9982\n",
      "703/877, sample ID: 48_13, accuracy: 0.9989\n",
      "704/877, sample ID: 48_14, accuracy: 0.9979\n",
      "705/877, sample ID: 48_15, accuracy: 0.9986\n",
      "706/877, sample ID: 48_16, accuracy: 0.9993\n",
      "707/877, sample ID: 48_17, accuracy: 0.9986\n",
      "708/877, sample ID: 48_18, accuracy: 0.9993\n",
      "709/877, sample ID: 48_19, accuracy: 0.9989\n",
      "710/877, sample ID: 48_20, accuracy: 0.9986\n",
      "711/877, sample ID: 48_21, accuracy: 0.9965\n",
      "712/877, sample ID: 48_22, accuracy: 0.9982\n",
      "713/877, sample ID: 48_23, accuracy: 0.9979\n",
      "714/877, sample ID: 48_24, accuracy: 0.9975\n",
      "715/877, sample ID: 48_25, accuracy: 0.9972\n",
      "716/877, sample ID: 48_26, accuracy: 0.9993\n",
      "717/877, sample ID: 48_27, accuracy: 0.9993\n",
      "718/877, sample ID: 48_28, accuracy: 0.9975\n",
      "719/877, sample ID: 48_29, accuracy: 0.9975\n",
      "720/877, sample ID: 48_30, accuracy: 0.9989\n",
      "721/877, sample ID: 48_31, accuracy: 0.9986\n",
      "722/877, sample ID: 48_32, accuracy: 0.9982\n",
      "723/877, sample ID: 48_33, accuracy: 0.9993\n",
      "724/877, sample ID: 48_34, accuracy: 0.9982\n",
      "725/877, sample ID: 48_35, accuracy: 0.9993\n",
      "726/877, sample ID: 48_36, accuracy: 0.9979\n",
      "727/877, sample ID: 48_37, accuracy: 0.9965\n",
      "728/877, sample ID: 48_38, accuracy: 0.9986\n",
      "729/877, sample ID: 48_39, accuracy: 0.9975\n",
      "730/877, sample ID: 48_40, accuracy: 0.9982\n",
      "731/877, sample ID: 48_41, accuracy: 0.9986\n",
      "732/877, sample ID: 48_42, accuracy: 0.9979\n",
      "733/877, sample ID: 48_43, accuracy: 0.9982\n",
      "734/877, sample ID: 48_44, accuracy: 0.9993\n",
      "735/877, sample ID: 48_45, accuracy: 0.9975\n",
      "736/877, sample ID: 48_46, accuracy: 0.9972\n",
      "737/877, sample ID: 48_47, accuracy: 0.9989\n",
      "738/877, sample ID: 48_48, accuracy: 0.9975\n",
      "739/877, sample ID: 48_49, accuracy: 0.9989\n",
      "740/877, sample ID: 48_50, accuracy: 0.9982\n",
      "741/877, sample ID: 48_51, accuracy: 0.9979\n",
      "742/877, sample ID: 48_52, accuracy: 0.9968\n",
      "743/877, sample ID: 48_53, accuracy: 0.9972\n",
      "744/877, sample ID: 48_54, accuracy: 0.9982\n",
      "745/877, sample ID: 48_55, accuracy: 0.9979\n",
      "746/877, sample ID: 48_56, accuracy: 0.9986\n",
      "747/877, sample ID: 48_57, accuracy: 0.9979\n",
      "748/877, sample ID: 48_58, accuracy: 0.9993\n",
      "749/877, sample ID: 48_59, accuracy: 0.9982\n",
      "750/877, sample ID: 48_60, accuracy: 0.9982\n",
      "751/877, sample ID: 48_61, accuracy: 0.9979\n",
      "752/877, sample ID: 48_62, accuracy: 0.9975\n",
      "753/877, sample ID: 48_63, accuracy: 0.9993\n",
      "754/877, sample ID: 48_64, accuracy: 0.9975\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "755/877, sample ID: 48_65, accuracy: 0.9982\n",
      "756/877, sample ID: 48_66, accuracy: 0.9986\n",
      "757/877, sample ID: 48_67, accuracy: 0.9982\n",
      "758/877, sample ID: 48_68, accuracy: 0.9986\n",
      "759/877, sample ID: 48_69, accuracy: 0.9965\n",
      "760/877, sample ID: 48_70, accuracy: 0.9975\n",
      "761/877, sample ID: 48_71, accuracy: 0.9968\n",
      "762/877, sample ID: 48_72, accuracy: 0.9989\n",
      "763/877, sample ID: 48_73, accuracy: 0.9975\n",
      "764/877, sample ID: 48_74, accuracy: 0.9979\n",
      "765/877, sample ID: 48_75, accuracy: 0.9993\n",
      "766/877, sample ID: 48_76, accuracy: 0.9979\n",
      "767/877, sample ID: 48_77, accuracy: 0.9975\n",
      "768/877, sample ID: 48_78, accuracy: 0.9975\n",
      "769/877, sample ID: 48_79, accuracy: 0.9972\n",
      "770/877, sample ID: 48_80, accuracy: 1.0000\n",
      "771/877, sample ID: 48_81, accuracy: 0.9972\n",
      "772/877, sample ID: 48_82, accuracy: 0.9989\n",
      "773/877, sample ID: 48_83, accuracy: 0.9979\n",
      "774/877, sample ID: 48_84, accuracy: 0.9972\n",
      "775/877, sample ID: 48_85, accuracy: 0.9982\n",
      "776/877, sample ID: 48_86, accuracy: 0.9986\n",
      "777/877, sample ID: 48_87, accuracy: 0.9979\n",
      "778/877, sample ID: 48_88, accuracy: 0.9972\n",
      "779/877, sample ID: 48_89, accuracy: 0.9989\n",
      "780/877, sample ID: 48_90, accuracy: 0.9972\n",
      "781/877, sample ID: 48_91, accuracy: 0.9993\n",
      "782/877, sample ID: 48_92, accuracy: 0.9993\n",
      "783/877, sample ID: 48_93, accuracy: 0.9979\n",
      "784/877, sample ID: 48_94, accuracy: 0.9989\n",
      "785/877, sample ID: 48_95, accuracy: 0.9979\n",
      "786/877, sample ID: 48_96, accuracy: 0.9982\n",
      "787/877, sample ID: 49_02, accuracy: 0.9968\n",
      "788/877, sample ID: 49_03, accuracy: 0.9968\n",
      "789/877, sample ID: 49_04, accuracy: 0.9989\n",
      "790/877, sample ID: 49_05, accuracy: 0.9979\n",
      "791/877, sample ID: 49_07, accuracy: 0.9979\n",
      "792/877, sample ID: 49_08, accuracy: 0.9986\n",
      "793/877, sample ID: 49_09, accuracy: 0.9979\n",
      "794/877, sample ID: 49_10, accuracy: 0.9979\n",
      "795/877, sample ID: 49_11, accuracy: 0.9993\n",
      "796/877, sample ID: 49_13, accuracy: 0.9972\n",
      "797/877, sample ID: 49_14, accuracy: 0.9965\n",
      "798/877, sample ID: 49_15, accuracy: 0.9996\n",
      "799/877, sample ID: 49_16, accuracy: 0.9982\n",
      "800/877, sample ID: 49_17, accuracy: 0.9989\n",
      "801/877, sample ID: 49_18, accuracy: 0.9972\n",
      "802/877, sample ID: 49_20, accuracy: 0.9982\n",
      "803/877, sample ID: 49_21, accuracy: 0.9986\n",
      "804/877, sample ID: 49_23, accuracy: 0.9975\n",
      "805/877, sample ID: 49_24, accuracy: 0.9982\n",
      "806/877, sample ID: 49_25, accuracy: 0.9993\n",
      "807/877, sample ID: 49_26, accuracy: 0.9993\n",
      "808/877, sample ID: 49_27, accuracy: 0.9982\n",
      "809/877, sample ID: 49_28, accuracy: 0.9989\n",
      "810/877, sample ID: 49_29, accuracy: 0.9968\n",
      "811/877, sample ID: 49_32, accuracy: 0.9986\n",
      "812/877, sample ID: 49_33, accuracy: 0.9989\n",
      "813/877, sample ID: 49_34, accuracy: 0.9975\n",
      "814/877, sample ID: 49_35, accuracy: 0.9986\n",
      "815/877, sample ID: 49_37, accuracy: 0.9986\n",
      "816/877, sample ID: 49_38, accuracy: 0.9979\n",
      "817/877, sample ID: 49_39, accuracy: 0.9961\n",
      "818/877, sample ID: 49_40, accuracy: 0.9986\n",
      "819/877, sample ID: 49_41, accuracy: 0.9968\n",
      "820/877, sample ID: 49_42, accuracy: 0.9982\n",
      "821/877, sample ID: 49_43, accuracy: 0.9965\n",
      "822/877, sample ID: 49_44, accuracy: 0.9989\n",
      "823/877, sample ID: 49_45, accuracy: 0.9979\n",
      "824/877, sample ID: 49_46, accuracy: 0.9982\n",
      "825/877, sample ID: 49_47, accuracy: 0.9979\n",
      "826/877, sample ID: 49_48, accuracy: 0.9975\n",
      "827/877, sample ID: 49_49, accuracy: 0.9982\n",
      "828/877, sample ID: 49_50, accuracy: 0.9989\n",
      "829/877, sample ID: 49_51, accuracy: 0.9975\n",
      "830/877, sample ID: 49_52, accuracy: 0.9986\n",
      "831/877, sample ID: 49_53, accuracy: 0.9968\n",
      "832/877, sample ID: 49_54, accuracy: 0.9986\n",
      "833/877, sample ID: 49_55, accuracy: 0.9979\n",
      "834/877, sample ID: 49_56, accuracy: 0.9989\n",
      "835/877, sample ID: 49_57, accuracy: 0.9982\n",
      "836/877, sample ID: 49_58, accuracy: 0.9979\n",
      "837/877, sample ID: 49_59, accuracy: 0.9989\n",
      "838/877, sample ID: 49_60, accuracy: 0.9982\n",
      "839/877, sample ID: 49_62, accuracy: 0.9989\n",
      "840/877, sample ID: 49_63, accuracy: 0.9993\n",
      "841/877, sample ID: 49_64, accuracy: 0.9982\n",
      "842/877, sample ID: 49_65, accuracy: 0.9982\n",
      "843/877, sample ID: 49_66, accuracy: 0.9975\n",
      "844/877, sample ID: 49_67, accuracy: 0.9982\n",
      "845/877, sample ID: 49_68, accuracy: 0.9979\n",
      "846/877, sample ID: 49_70, accuracy: 0.9986\n",
      "847/877, sample ID: 49_71, accuracy: 0.9968\n",
      "848/877, sample ID: 49_72, accuracy: 0.9989\n",
      "849/877, sample ID: 49_73, accuracy: 0.9979\n",
      "850/877, sample ID: 49_74, accuracy: 0.9989\n",
      "851/877, sample ID: 49_75, accuracy: 0.9986\n",
      "852/877, sample ID: 49_76, accuracy: 0.9986\n",
      "853/877, sample ID: 49_78, accuracy: 0.9982\n",
      "854/877, sample ID: 49_79, accuracy: 0.9972\n",
      "855/877, sample ID: 49_80, accuracy: 0.9979\n",
      "856/877, sample ID: 49_81, accuracy: 0.9989\n",
      "857/877, sample ID: 49_82, accuracy: 0.9986\n",
      "858/877, sample ID: 49_83, accuracy: 0.9979\n",
      "859/877, sample ID: 49_84, accuracy: 0.9972\n",
      "860/877, sample ID: 49_86, accuracy: 0.9986\n",
      "861/877, sample ID: 49_87, accuracy: 0.9979\n",
      "862/877, sample ID: 49_88, accuracy: 0.9965\n",
      "863/877, sample ID: 49_89, accuracy: 0.9982\n",
      "864/877, sample ID: 49_90, accuracy: 0.9972\n",
      "865/877, sample ID: 49_91, accuracy: 0.9989\n",
      "866/877, sample ID: 49_92, accuracy: 0.9972\n",
      "867/877, sample ID: 49_94, accuracy: 0.9989\n",
      "868/877, sample ID: 49_95, accuracy: 0.9972\n",
      "869/877, sample ID: 49_96, accuracy: 0.9986\n",
      "870/877, sample ID: 50_02, accuracy: 0.9989\n",
      "871/877, sample ID: 50_03, accuracy: 0.9965\n",
      "872/877, sample ID: 50_10, accuracy: 0.9982\n",
      "873/877, sample ID: 50_11, accuracy: 0.9982\n",
      "874/877, sample ID: 50_26, accuracy: 0.9972\n",
      "875/877, sample ID: 50_27, accuracy: 0.9979\n",
      "876/877, sample ID: 50_35, accuracy: 0.9986\n"
     ]
    }
   ],
   "source": [
    "avg_accuracy = []\n",
    "for i in range(test_X_missing.shape[0]):\n",
    "    # Generates missing genotypes\n",
    "    missing_size = int(missing_perc * test_X_missing.shape[1])\n",
    "    missing_index = np.random.randint(test_X_missing.shape[1],\n",
    "                                      size=missing_size)\n",
    "    test_X_missing[i, missing_index, :] = [1, 0, 0]\n",
    "\n",
    "    # predict\n",
    "    predict_onehot = SCDA.predict(test_X_missing[i:i + 1, :, :])\n",
    "    # only care the missing position\n",
    "    predict_missing_onehot = predict_onehot[0:1, missing_index, :]\n",
    "    \n",
    "    # calculate probability and save file.\n",
    "    predict_prob = cal_prob(predict_missing_onehot)\n",
    "    pd.DataFrame(predict_prob).to_csv('results/{}.csv'.format(df_ori.index[i]),\n",
    "                                      header=[1, 2],\n",
    "                                      index=False)\n",
    "\n",
    "    # predict label\n",
    "    predict_missing = np.argmax(predict_missing_onehot, axis=2)\n",
    "    # real label\n",
    "    label_missing_onehot = test_X[i:i + 1, missing_index, :]\n",
    "    label_missing = np.argmax(label_missing_onehot, axis=2)\n",
    "    # accuracy\n",
    "    correct_prediction = np.equal(predict_missing, label_missing)\n",
    "    accuracy = np.mean(correct_prediction)\n",
    "    print('{}/{}, sample ID: {}, accuracy: {:.4f}'.format(\n",
    "        i, test_X_missing.shape[0], df_ori.index[i], accuracy))\n",
    "\n",
    "    avg_accuracy.append(accuracy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The average imputation accuracyon test data with 0.1 missing genotypes is 0.9979: \n"
     ]
    }
   ],
   "source": [
    "print('The average imputation accuracy' \\\n",
    "      'on test data with {} missing genotypes is {:.4f}: '\n",
    "    .format(missing_perc, np.mean(avg_accuracy)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.0"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": true,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": true
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}