02.1_UV_melt.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "0161426b-24e5-48ad-b67a-cda34422eeed",
   "metadata": {},
   "source": [
    "Figure 2."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "512e6ac0-8e00-4f91-adcc-a0e98a7a18ef",
   "metadata": {},
   "source": [
    "## Import"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "53815b85-4e77-40ea-9df8-1d7a02250c5d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The autoreload extension is already loaded. To reload it, use:\n",
      "  %reload_ext autoreload\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 640x480 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "%run nnn/ipynb_init.py\n",
    "%matplotlib inline\n",
    "\n",
    "from nnn.uv import *\n",
    "import re"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "de914708-3130-4e1e-81c1-56ddb9b0a74d",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Prepare UV melting data\n",
    "\n",
    "Aggregate to SEQID level and salt adjust to array salt level\n",
    "\n",
    "`uv_df` - aggregated to cuvette level\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "08e64c08-f494-4a3c-848b-625a8d575433",
   "metadata": {},
   "outputs": [],
   "source": [
    "# copy pasted from command center for convenience\n",
    "# only one peak found by AnalyzePeaks\n",
    "hplc_clean_seqid = [\"WC68_stem_3p\", \"WC68_stem_5p\", \"WC1025_split_5p\", \\\n",
    "\"WC1025_split_3p\", \"WC1375_stem_5p\", \"WC1375_stem_3p\", \"DMM004_5p\", \\\n",
    "\"DMM004_3p\", \"DMM005_5p\", \"DMM005_3p\", \"DMM006_5p\", \"DMM007_5p\", \\\n",
    "\"DMM007_3p\", \"BC0\", \"BC1\", \"BC6\", \"BC7\", \"BC8\", \"BC9\", \"BC10\", \\\n",
    "\"BC11\", \"MMCNNN2328\", \"MMCNNN2331\", \"MMCNNN2334\", \"MMCNNN2335\", \\\n",
    "\"MMCNNN2336\", \"MMCNNN2337\", \"MMTM1325\", \"MMTM176\", \"MMTM178\", \\\n",
    "\"MMTM179\", \"MMTM181\", \"MMTM185\", \"MMTM189\", \"MMTM1917\", \"MMTM29\", \\\n",
    "\"MMTM397\", \"MMTM509\", \"MMTM877\", \"DMM003_3p_rc\", \"DMM004_3p_rc\", \\\n",
    "\"DMM005_3p_rc\", \"DMM006_3p_rc\", \"DMM007_3p_rc\", \"TETNN1003\", \\\n",
    "\"TETNN999\", \"TETNN963\", \"TETNN979\", \"TETNNN1696\", \"TETNNN2082\", \\\n",
    "\"TETNNN1570\", \"HP002\", \"HP003\", \"HP004\", \"HP005\", \"HP007\", \"HP008\", \\\n",
    "\"HP010\", \"HP011\", \"HP012\", \"HP014\", \"HP015\", \"HP016\", \"HP017\", \\\n",
    "\"HP018\", \"HP019\", \"HP021\", \"HP022\", \"HP023\", \"HP024\"]\n",
    "\n",
    "arr = pd.read_csv('./data/models/raw/arr_v1_n=27732.csv', index_col=0)\n",
    "uv_oligos = pd.read_csv('./data/uv_melt/ECLTables/ECLOligos230502.csv').set_index('OligoID')\n",
    "\n",
    "agg_result_file='./data/uv_melt/uvmelt_agg_230502.csv'\n",
    "uv_df = pd.read_csv(agg_result_file, index_col=0)\n",
    "uv_df = uv_df[uv_df.SEQID.isin(hplc_clean_seqid)]\n",
    "uv_df = uv_df.query('is_hairpin').sort_values(by=['conc_uM', 'n_curve', 'curve_date'])\n",
    "\n",
    "pname_uv = ['dG_37_uv', 'dH_uv', 'Tm_uv']\n",
    "agg_list = [uv_df.groupby(['SEQID', 'Na_mM'])[p].apply(np.nanmedian) for p in pname_uv]\n",
    "agg_df = pd.DataFrame(agg_list[0]).reset_index()\n",
    "\n",
    "for i,p in enumerate(pname_uv):\n",
    "    agg_df[p] = agg_list[i].values\n",
    "\n",
    "agg_df = agg_df.set_index('SEQID').join(arr).join(uv_oligos['sequence'])\n",
    "\n",
    "def adjust_uv_param(i, agg_df):    \n",
    "    row = agg_df.iloc[i,:]\n",
    "    if row.Na_mM > 25:\n",
    "        adjusted_dict = util.get_Na_adjusted_param(\n",
    "            Na=0.083, from_Na=(row.Na_mM + 50)/1000,\n",
    "            dH=row.dH_uv, Tm=row.Tm_uv, seq=row.RefSeq)\n",
    "        for p in ['dG_37', 'dH', 'Tm']:\n",
    "            row[p+'_uv'] = adjusted_dict[p]\n",
    "    return row\n",
    "        \n",
    "for i in range(len(agg_df)):\n",
    "    agg_df.iloc[i,:] = adjust_uv_param(i, agg_df)\n",
    "\n",
    "agg_df.drop(columns=['Na_mM'], inplace=True)\n",
    "# agg_df.rename(columns=dict(sequence='RefSeq'), inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "67c81344-143a-425a-9365-ac6c1215dc10",
   "metadata": {},
   "outputs": [],
   "source": [
    "agg_df.to_csv('./data/models/processed/uv_hairpin_arr_salt_hplc_filtered.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "f1102132-eaf4-467b-9bd2-8b9060233ebe",
   "metadata": {},
   "outputs": [],
   "source": [
    "agg_result_file='./data/uv_melt/uvmelt_agg_230502.csv'\n",
    "uv_df = pd.read_csv(agg_result_file, index_col=0)\n",
    "\n",
    "pname_uv = ['dG_37_uv', 'dH_uv', 'Tm_uv']\n",
    "agg_list = [uv_df.groupby(['SEQID', 'Na_mM', 'is_hairpin'])[p].apply(np.nanmedian) for p in pname_uv]\n",
    "agg_df = pd.DataFrame(agg_list[0]).reset_index()\n",
    "\n",
    "for i,p in enumerate(pname_uv):\n",
    "    agg_df[p] = agg_list[i].values\n",
    "\n",
    "agg_df = agg_df.set_index('SEQID').join(arr)\n",
    "        \n",
    "for i in range(len(agg_df)):\n",
    "    agg_df.iloc[i,:] = adjust_uv_param(i, agg_df)\n",
    "\n",
    "agg_df.drop(columns=['Na_mM'], inplace=True)\n",
    "\n",
    "agg_df.to_csv('./data/models/processed/uv_arr_salt_not_hplc_filtered.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6e608e4f-163c-4595-9651-c7548f3e724e",
   "metadata": {},
   "source": [
    "### Data accounting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "6bb624c6-b9a2-4d5a-988f-836928442746",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_char(mystr):\n",
    "    return ''.join(re.split(\"[^a-zA-Z]*\", mystr))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "a6347243-22f7-4fef-b7d2-adc5df177c99",
   "metadata": {},
   "outputs": [],
   "source": [
    "uv_df = pd.read_csv(agg_result_file, index_col=0)\n",
    "\n",
    "# uv_df = uv_df.query('is_hairpin').sort_values(by=['conc_uM', 'n_curve', 'curve_date'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "54928b6d-f373-4cc1-96d0-9723acd91d1e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "70 sequences measured, of which:\n",
      "\t12 new hairpins, 58 in Array Melt dataset\n"
     ]
    },
    {
     "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>1</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>BC</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BNNN</th>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HP</th>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MMCNN</th>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MMCNNN</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MMTM</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TETNN</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TETNNN</th>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TRINNN</th>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WC</th>\n",
       "      <td>42</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         1\n",
       "0         \n",
       "BC       1\n",
       "BNNN     3\n",
       "HP      12\n",
       "MMCNN   10\n",
       "MMCNNN   2\n",
       "MMTM     2\n",
       "TETNN    2\n",
       "TETNNN  16\n",
       "TRINNN   8\n",
       "WC      42"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "n_measured_seq = len(np.unique(uv_df.SEQID))\n",
    "print('%d sequences measured, of which:'%n_measured_seq)\n",
    "\n",
    "# Count sequences in each category\n",
    "cat_count_df = pd.DataFrame(np.unique(uv_df.SEQID.apply(get_char), return_counts=True)).T.set_index(0)\n",
    "n_hp = cat_count_df.loc['HP', 1]\n",
    "print(f'\\t{n_hp} new hairpins, {n_measured_seq - n_hp} in Array Melt dataset')\n",
    "cat_count_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "cc442d50-7ea0-495f-ae0b-1aa49571da0c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "14 sequences measured, of which:\n",
      "\t8 new hairpins, 6 in Array Melt dataset\n"
     ]
    },
    {
     "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>1</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>BC</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HP</th>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MMCNNN</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MMTM</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>TETNNN</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        1\n",
       "0        \n",
       "BC      1\n",
       "HP      8\n",
       "MMCNNN  2\n",
       "MMTM    1\n",
       "TETNNN  2"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "uv_df = uv_df[uv_df.SEQID.isin(hplc_clean_seqid)]\n",
    "\n",
    "n_measured_seq = len(np.unique(uv_df.SEQID))\n",
    "print('%d sequences measured, of which:'%n_measured_seq)\n",
    "\n",
    "# Count sequences in each category\n",
    "cat_count_df = pd.DataFrame(np.unique(uv_df.SEQID.apply(get_char), return_counts=True)).T.set_index(0)\n",
    "n_hp = cat_count_df.loc['HP', 1]\n",
    "print(f'\\t{n_hp} new hairpins, {n_measured_seq - n_hp} in Array Melt dataset')\n",
    "cat_count_df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c8d657f1-a0ef-4b57-983d-4dd1aee0e6c1",
   "metadata": {},
   "source": [
    "## Plot direct comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "93f09007-4920-4472-b123-9d199e5b17ac",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 501.969x137.795 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbwAAACpCAYAAABQ3IhDAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAAA1H0lEQVR4nO3deViU9f74/+cMy8CwC4gsarnhnoqaa5mmhibmvoGa/sLyY8vXq0xPnlLLLOtcldrno6fcYlMxcjlhmno8ZrlduOaCu4YgKAqyDjBz//5Q5oig3gOzz/txXV3JOPfcL+TFvOZ9v9/3662QJElCEARBEOyc0tIBCIIgCII5iIInCIIgOARR8ARBEASHIAqeIAiC4BBEwRMEQRAcgih4giAIgkMQBU8QBEFwCKLgCYIgCA5BFDxBEATBIYiCJwhWKCYmhmnTpgHQt29fFixYUOPzUlJSCA8P5/bt2+YMTxBskih4giAIgkMQBU8QBEFwCM6WDkAQHF1RURGLFi1ix44dKBQKpk6daumQBMEuiYInCBY2c+ZMjh07xqxZs/D19WXp0qVcvHiRnj176p8jSRIVFRXVjtXpdOYMVRBsmih4gmBBZ8+eZc+ePXz11VcMGjQIgPbt29OvX78qz0tMTCQxMdESIQqC3RAFTxAs6MiRIwA899xz+sfq169Phw4dqjwvMjKyxkude/bsYdmyZSaNURDshSh4gmBBd+/excXFBU9PzyqPBwYGUlRUpP+6Xr16tGvXrtrx58+fN3mMgmAvxCpNQbAgX19fysvLuXv3bpXH8/LyLBOQINgxUfAEwYK6du0KwI4dO/SP5efnc+zYMQtFJAj2S1zSFAQLatKkCVFRUXz66adoNBpCQkJYsWJFjSsyBUGoG1HwBMHCFi5cSL169Vi6dCnl5eWMHDmSoKAgSktLLR2aINgVhSRJkqWDMERxcTFnz56lZcuWqNVqS4cjWCGRI4IcIk8cj83N4Z09e5aIiAjOnj1r6VCEWtJqtfz444/8+OOPJnl9kSP24ejRoya95ULkie3Lycnhm2++kf18myt4gm0rLy9n/fr1/Pnnn2KeSnikP/74gy1btogcER4pIyODNWvWVFvh/DhiDk8wm9LSUpKSkrh27RodOnRgyJAhlg5JsDKSJLF792727dtHYGAg0dHRlg5JsEKXLl1i3bp1SJLEuHHjZB8nu+DpdDrS09NRKBSEh4ejUChqFahgv86fP09ycjL5+fk4Ozvj7e3N+PHjadiwISUlJaxdu5bs7Gy6devGgAEDRA45oMflCMC2bds4fPgwoaGhjB8/XsytCdWcO3eODRs24OLiwrhx42jUqJHsY2UVvC1btnDmzBkaNGiAk5MTmzdvplOnTvTv37/WQQv256+//uJvf/ub/mtJkti1axcNGzbEzc2N+vXr06ZNG3r16iWKnYN6XI4AhIWFkZuby5gxY3B1dbVUmIIVCwgIIDAwkFdeeYWgoCCDjpVV8Lp06UJUVFSVxzIyMgw6kWD/Tp8+zenTp6s8FhMTgyRJKBQKhg0bJgqdg6spR6ZNm0ZRUREeHh60b9+edu3aiTwRqsnLy8PX15d69eoRGxtbqxyRVfCWLFlS7bFFixYZfDLBvg0bNgyAgoIC4N4+bytWrOCZZ55h4MCB4k1MqJYjLi4uxMXFUVZWxpQpU3B1dRV5IlRReRXg0KFDTJ48mZCQkFrniKyCt2jRIsrKyti3bx8AvXr1qtXJBPsWGhrKsmXLuHPnDvn5+Vy/fp22bdvStGlTS4cmWIkHc0Sj0ZCVlUWjRo3o0aMHLi4ulg5PsDI6nY6ff/6ZI0eOEBYWhp+fX51eT/ZtCfPnz6e8vJzS0lIWL15cp5MK9quwsJARI0bomyJPnDiRZs2aWToswYoUFhYyY8YM6tevT25uLv369aN///5iZCdUUXm/7pEjR2jatCkxMTG4u7vX6TVlr9J0c3Nj4MCBwH/38BKEh924cYOPP/4YV1dXwsPD9YsRBKFSQUEB77zzDhqNhsaNG4srRkI1Wq2WpKQkLl68SOvWrRk2bBjOznW/i072KzRt2pQ5c+YgSVK1zSkFodL777/PqlWrGDhwIJ07d7Z0OIIV+vDDD/nmm29o164dkZGRlg5HsEJOTk40aNAAHx8fBg8ejFJpnB4psgveoEGDaNGiBTqdTlx6EKqQJIm7d+/i4+PD0qVLq7QOE4ubhEr5+fn4+Pgwf/58tFote/fuZe/evSJHBL3CwkLc3NxwdnamX79+AEatN7IL3rx58+jduzdKpRKFQkGXLl2MFoRguyonlc+cOcPUqVMJCwtj+vTplg5LsDKnT58mJSWFwYMHixwRanT79m3i4uIIDg5m1KhRJhlYyS544eHhREREiJtBBT2tVktKSgqnT5+mSZMmeHl5cfbsWd5++219hwzx6V1IS0vj559/xsPDg5CQEJEjQjXZ2dnEx8dTVFRE7969TXYVUXbBO3HiBGVlZTg5OQEwY8YMkwQk2IaysjI2bNhQbVK5WbNmvPXWW5YOT7AS+/btY9euXfj5+RETE4Ofn5/IEaGKv/76i8TERP1ekK1btzbZuQwa4cXGxtZ5Wahg+8rKyoiLiyMjI4NOnTpVmVQ+cuSI+PQuALB7925+++036tevT3R0NF5eXoDIEeG/Ll++TGJiIgqFgvHjx9OkSROTnk92wcvLy2PBggX6r0WSOi4XFxcCAgJo3Lgx/fr1019+WL58OW+88QYRERFGWUIs2LYGDRrQqFEjxo4dq/+gLHJEeJCfnx/+/v68/PLLhIWFmfx8sjJu2bJlhIaG8uyzz4o+dw7s7t27eHp6olQqiYqKqpYHr732GkeOHGHVqlWUlZURHBzMiBEjLBStYAkVFRWUlpbi6elJ69atadWqVZU8ETkiwH9X7Pr6+jJt2jSz1RRZBW/GjBmUlZVx6NAhli9fjlarFXN4DqZyUjk8PJyXX365xgStbEhQeZ9mTk6OOUMULKysrIz169dTUFDAlClTcHNzq5YnIkccmyRJ7Nu3j7179xITE0OjRo3MOoCSPcITHNeDk8qPu8aekZEhRv8Oqri4mMTERK5fv/7Y1dwiRxyXJEn8+uuv7N+/n6CgIOrVq2f2GGQVvMoO55WJKkmSQSc5fPgw69ato6ysjLfeeousrCy2b9+OQqFg/vz5+pWfgvW5cOECGzZsAHjipPKwYcPYunUr+/fv1+fIK6+8IvtcIk9sU0FBAXFxcdy8eZNevXrRt2/fRxY1kSOOSafTsXXrVo4dO0bDhg0ZP348bm5uZo9DVsELDQ3l888/5/Tp0/j4+JCTk8O6detkn6SwsJDFixdz7tw5fv/9d/7880++/PJLNm3axIEDB+jZs2e1Y3Jycrh582a1xy9evCj7vLYuS6Mhq6ys2uPBrq4Eq1QmP/+ZM2fYuHEjrq6uTJgwQdak8r59++jUqRMdO3YkKSnJoPMZmiciRywvPz+fNWvWkJeXR//+/enRo8cTjzFnjoDIE0vT6XQkJydz9uxZmjVrxujRoy22M4bsZVJlZWWMHj2apk2bkpKS8sTnr1mzhv379wPQuXNnunfvTnx8PDNnziQ9PR2AwMBAbt26VePx69evr/FSamlpqdyQbd6KzEzmX71a7fGPGjdm3tNPm/z89erVIyAggBEjRlC/fn3ZxzVq1Ihdu3Zx9+7dJz63LnkicsTy1Go1Pj4+PPfcc3Ts2FH2cebKERB5YmlKpZKgoCCcnZ155ZVXLDoKl13wysvLCQgI4Pvvv6eoqOiJz588eTKTJ08G7n0KXLhwITNmzMDf319/z9bNmzcf+UY6ZswY+vbtW+3x06dPM3r0aLlh27RpISFEBQQw8tQpADa2aQPcG+GZiiRJFBQU4O3tTVBQEK+//rpBcy4zZ85EpVKRm5srq8l4XfJE5Ijl3L17F29vb1xcXJg0aZLV5giIPLGU4uJiXFxccHFx4fnnnweM2xezNhSSzAm5vLw8dDqdPtENmXCcM2cOOTk5+Pn58eKLL+Lm5sbu3bvRarV8/PHHBnXCPnLkCBEREaSlpdGpUyfZx9myNocOAXCqa1eTnqdyUvnIkSNMmTLFoFFdpXnz5jFp0iSys7PZuXMnc+fOlX2ssfLEEXPEnM6fP8+GDRsYMGBArXrqWkOOgMgTU7p79y5xcXH4+fkxduxYo+12UFeyR3gLFizggw8+QKlUsnDhQv7xj3/IPklNN6n36dNH9vGOrkCrJVujYXtuLgP9/U1yjocnlb29vWv1Ou7u7jz99NM8/fTT/PrrrwYdK/LE+p08eZJNmzahUqkICQmp1WuIHLFvubm5xMXFkZ+fT0REhNUUOzCg4Pn4+ODv74+fn59oL2ZGK7Oy+EujASDy5Em+Cw9nanCwUc9RUVHBjz/+aJRJ5caNGzN37ly0Wi1t27Y1apyCZR0+fJjU1FS8vLyIiYkhMDCwVq8jcsR+ZWVlER8fT0lJCUOHDrW6vVNlF7yoqCg+/PBDtFotw4cPN2VMwn0ZpaXE3p+UB5CAaenpDPTzI8xIS3orKipITEzk8uXLtG3bts6Tyu3bt2fs2LFotVri4uKMEqNgeb///js7d+6kXr16xMTE4OvrW+vXEjlinzIyMoiPj6eiooLRo0fTsmVLS4dUjeyCd/HiRRYsWEBGRgZLliwhIiLClHEJwPmSEnQPPaYFLpSUGK3gOTk54e/vj7+/P5GRkXW+/LB582aOHTvGgQMHGDlypFFiFCwvICCAkJAQxo0bh6enZ51eS+SIffL29sbHx4eXXnqJp82wirw2ZBc8Z2dn3n33Xby9vXn//fdNGZNwX3N3d5RQpeg5Ac2McEm5sLAQtVqNUqlk0KBBQN1XUP3000+0bNmS1atXM3jwYG7fvl3nOAXL0el0FBcX4+npSXh4OC1atBA5IlRTuZDR29vb4FXd5ibr4/ycOXM4ePAgp06d4vr163z55ZemjksAwtzc+Gd4uP5rJbAiPLzOo7vc3Fy+//57tmzZgiRJKBQKoyRpaGgooaGhzJ07l44dOxIaGlrn1xQso6KiguTkZFavXk1xcTFgnCXlIkfsy6FDh1iyZAmXLl0CLH/bwZPIGuF98MEH1S5jFBYW1vnShvBkU4ODmX/lCtkaDVvatavzKs0bN24QHx9PcXExTz31lFET9OjRo3To0IE2bdrg5OTEkSNHWL16Na+++qrRziGYnkajYd26dVy5coV27dqhMmJXH5Ej9kGSJPbu3cuePXv0UyK2QFbB27RpE0VFRTRv3hwnJydOnz5NQEAAo0aNMnV8AuDl5ISXWl3nYnf16lWSkpJMNqk8bdo0du/ezddff41CoaBnz57ijczGFBUVkZCQQFZWFl26dCEyMtKoH4pEjtg+SZL45ZdfOHToEMHBwUyYMAEPDw9LhyWLrIIXHR3N7du3SUtLQ6vVMmbMGIt0uhZq7+LFi6xbtw6lUsmECRNMNqnct2/fGrtaCNavsLCQtWvXcuvWLZ577jn69OljkktUIkdslyRJbN68mePHj9O4cWPGjRtn1CsApiZ70Uq9evXo37+/KWMRHlLZPLpEd2/ZypGCAqB2zaMrN1scNmxYrW8YFuybm5sb3t7eRERE0K1bN0uHI1ghhUJBQEAALVq0YOTIkRZrAl1bsgueYH4PN4+OSEsDDGseXVBQgJeXFwEBAbzxxhsm7XowZ86cao/V1BlDsC6VOeLs7MyECRNEjgjVaDQaFAoFrq6u9OrVC51OZ1UdVOSSVfAqk/Ty5cv6S2EiSU2vsnn0w+Q0j66cVP7jjz+YPHkywcHBJk/QypyYM2eOyA8bceXKFZKSknj++efp0aOHyBGhmqKiIuLj43F3d2fChAk4OTnZZLEDmQVPJKllBKtUtdr3TpIktm/fzsGDB2nQoEGt+2IK9i09PZ2NGzeiVCrFZW6hRvn5+cTFxZGbm0ufPn1sttBVEpc07YxOp2PLli36SeWxY8eabWfhn376Cbi3GnTTpk2AYbtZC+Zz4sQJNm3ahLu7O9HR0QQbuT/ro4gcsR23bt0iLi6Ou3fvEhkZSVcT79ZiDgZd0kxLS9P/WYz0rI9WqyU5OZn09HSLTCpX7oj+zjvvWP0NqI6ssgm0t7c3MTExBNRw2dxURI7Yhhs3bhAXF0dJSQnDhg2jffv2lg7JKAy6pClYN6VSiZ+fH+3btycqKsrsOwunpKRQUFCASqVCoVCg1WprtV+aYFp+fn4EBQUxbtw4fHx8zHpukSO2wdPTEy8vL4YOHUqLFi0sHY7RyB7hnThxgvDwcBQKBRcuXGDz5s2mjk2Qqbi4GJVKhZOTEwMGDAAs0+Jn0aJFfPbZZ8yePRuATz/91OwxCDWTJImioiI8PT1p1qwZTZo0sch8jMgR61a5YtfT05PY2Fibn7N7mOwR3uLFi5k1axYAn3/+uUmDEuSrnFRu0KABI0aMsPhlouLiYnbu3IkkSZSVlVk0FuEerVbLli1buHbtGlOmTMHLy8uib2QiR6zTsWPH+Pnnnxk9ejTNmze3u2IHBixaUavV+k9josuKdXhwUrlLly4WL3Zwr+/qgQMHAPjb3/5m4WiE8vJyNm7cyLlz5wgPD7eKzZtFjlifAwcOsH37dnx8fOz6/V12wZsxYwbFxcUoFAqr+KVxdJmZmSQkJFjdpLJKpeL555+3dBgCUFpaSlJSEteuXeOZZ54hKirKKj61ixyxHpIk8e9//5vffvuNgIAAYmJi7Po2JoNuS1Cr1aaKQzDAtWvXSEhIQKfTMWbMGMIf2EJIEODeZcO4uDhu3LjBs88+y8CBA63iCoBgXbZt28bhw4cJCQlhwoQJdv8eL7vg7dq1i379+gGQlJTEuHHjTBaU8HhFbm7kBQby3HPPURQcXKcem8b2wQcfUF5erl8hKlb4WoZKpcLT05MXXniB3r17W1WxEzliPfz9/Xn66acZM2aMTTWBrq0nFrycnBy++uorTp8+zc6dO1EoFDRq1MgcsQkPqdyDcH1REYubNWNxZiZkZur/3pAem6bSsGFDXn/9dYvG4MgKCwvx8PDAycmJcePGWcUlzIeJHLGs8vJytFotbm5uPPvss3Tp0sUq88QUnljw6tevz6JFi8jOzuavv/5Cp9NZ1adFR3HgwAF2795NTEyMvsfmyFOnANjYpg0gr8emqR06dIj09HR9dxfx6d18rl+/TkJCAt26deO5556z2jcxkSOWU1paSmJiIgqFgpiYGJydna02T0xB9iXNTz/9lF69euHs7IxCoRA3i5rJw5PKPj4+eN/vsel+P1E7eXlZOMr/Wr58Ofv27QOgV69eFo7GcVy+fJl169ah0+nM1iastkSOWEZhYSHx8fFkZ2fTvXt3szemsAayS3uLFi2Iiopi2LBhovedmUiSRGpqKr/99hshISG8+uqrVVZQFWi1XCguZnturgWjrGr+/PmUl5dTWlrK4sWLLR2OQzh79iwJCQn6T+3Nmze3dEiPJXLE/PLy8li9ejXZ2dn07duX/v37O+SVOtkjvJs3bzJv3jz91+IyhGlJksRPP/3EyZMneeqppxg7dmyVSeWVWVn8pdEAEHnyJN+FhzPVCj7Zu7m5MXDgQACOHDli4Wjs3/Hjx9m8eTNqtZro6GgaNGhg6ZCeSOSIed26dYsffviBgoICBg8eTOfOnS0dksXILnjjx4/n+PHjNG3alKCgIFPGJHCvNZivry/h4eGMHDkSZ+f//qgySkuJTU/Xfy0B09LTGejnR9gjdkao3D39YcZe2dm0aVPmzJmDJEl06NDBaK8r1MzHxwd/f3/Gjh2Lv7+/pcORReSIeanVatRqNQMGDKBt27aWDseiZBe81atX4+3tTa9evVi0aBFLliwxZVwOq7S0FGdnZ5ydnXnhhReQJKnapPL5khJ0Dx2nBS6UlDyy4D28e3olY67sPHv2LM2aNaNZs2aAZfp5OoIH+2I+9dRTJt/J3phEjphP5aputVptl30xa0N2wfPw8EClUuHr62vWLWccSeWksq+vL6NHj0apVNb4htDc3R0lVCl6TkCzx3TAMcfKzpycHG7fvq3/uqKiQixuMjKdTkdqairnz59nypQp+Pj42NQbmcgR8zhz5gwpKSkMHz6cVq1a2VSOmJLsf4UhQ4aQnZ3N3LlzGT9+vCljckgPTiqHhoY+9pNvmJsb/3ygu4oSWBEe/sjRHdzbPb2TlxfuSiXuSiWdvLzo5OVltMuZubm55ObmsnPnTuC/C24E49FqtaSkpJCWloa/v7/ZNvY1FpEj5nHkyBGSk5NRqVRm6Yt5Q6PhbFFRtf9u3F9jYE1kj/BKSkpo1KgRkiSxf/9+IiIiTBmXQ7l58yZxcXEGTSpPDQ5m/pUrZGs0bGnXjoEWnr/x8/MjLCyMtm3b6gv2Rx99ZNGY7ElZWRnJyclcuHCBVq1aMXz48CrzugDbt29n1qxZLF68WL8oxJqIHDG933//nZ07d+Lr60tMTIzJC94NjYaXTpzgTkUFFZJEXkUFvs7OOCsU+Dk780v79jSwog4uskd4qamp+tsShg0bZsqYHEpmZiarV6+mqKiIESNGGLSCysvJiWZqtcWLHdzbfLZLly5oNBrq1atHUVERW7dutXRYdkGj0RAfH8+FCxfo2LFjtUVMACtXriQyMpITJ04QGRnJypUrLRTto4kcMa1du3axc+dOAgMDmTJlillGd3kVFdypqMBVocAZKNHpcAZcFQruVFSQV1Fh8hgMIXuE5+fnx/79+3G9P+cTGhpqsqAciYeHB2q1muHDh+sn8p+kcsVlie7eLJ4hvTQLtFqyNRq25+aapFCWlpbStGlTmjZtyt69e43++o7IxcUFDw8PevTowYsvvljtcndGRgaxsbFIkgTcu1Q4bdo0Bg4cSFhYmCVCfiyRI6bh6+tLWFgY48ePN/uONqr7c4QKwFWpRKVUUqbVmjUGOWQXvKtXr9KtWzecnJzEyiojKCoqwsPDAx8fH6ZPn27QpPLDKy4j0tKAJ6+4NMe9ez4+Pnz66adIkoSfn59RX9vRFBUVoVarUSqVjBo16pE5cv78eXS6qut2tVotFy5csMqCJ3LEeLRaLeXl5bi5uREREUHHjh3FApXHMGiVZllZmd1vH2EOR48eZdu2bYwfP56nnnrK4AStXHH5sMetuKzNvXuGys3NZciQIZSUlFBRUWHX+2qZWk5ODnFxcXTo0IF+/fo9Nkcqd6d+sOg5OTnJvmJgTiJHjKesrIwNGzZQWlrKxIkTcXV1FcXuCWQXvM6dO3Pnzh3u3LkDIJYS19Iff/zBr7/+iq+vb61/2YPv99I0RG3u3TPUnj17yMrKqvLYjBkzjPLajiQjI4OEhATKyspkdU4JCwvjn//8J7Gxseh0OpRKJStWrLDK0Z3IEeMoKSkhMTGRjIwMOnXqVG1O19w0Oh1lOh0SUKZ7+J3Gesj+OODi4sKZM2dwdXXF19fX4BPdvn2b0aNHA7B3714++OAD5s6di9YKr/OagiRJ7Nq1i19//dWsk8qVKu/de9CT7t0z1MNvZLXh6Hly8eJFfvjhB7RaLePHj6fN/fsln2Tq1KmkpqbyzDPPkJqaytSpU00cae2IHKm7goIC1qxZQ0ZGBj179uTll1+22MjO19kZP2dnyiSJCsBdqaQCKJMk/Jyd8bVwIX6Y7GiOHj2Kh4cHL730Eu+//z59+vQx6ESrVq0iJCQEgC1btvDll1+yadMmDhw4QM+ePas9Pycnh5s3b1Z7/OLFiwad1xIebuMlSRK/7dtHxvHjtLLQpHLlvXv/3/3LmnLu3TPUsGHDWL58Oa+++ipKpZI1a9YY/BqG5Ikt50hNzpw5w8aNG3F1dWX8+PE0bNjQoOMHDhxolbcjPMjcOQL2lSd5eXmsXbuWvLw8XnzxxRq/X3NqoFLxS/v2Na7G9HV2tqpbEsCAggdw584dDh48SFkNPRkftmbNGvbv3w9Au3bt6N+/P0lJSQD6Ti2BgYHcunWrxuPXr1/PsmXLqj1eWlpqSMgWUWMbL7WaV1q35tPISP1KV3Mz9b17oaGhVFRUEBAQgFarNXme2HKO1MTT01PfZcde+9WaO0fAvvLEzc0Nd3d3evfuTadOnSwdDnCv6FlbYXsUhVS5lvkJTpw4wcmTJ9HpdAwePNigy3Hvvfce7u7uHDx4kJkzZ7J3714WLlzIpk2bqF+/Pj169Kh2zKM+lZ0+fZrRo0eTlpZmNT/wh1WO8Eb8+ScAP7ZtiyRJNHBxIdTMI7uHtTl0CIBTXbua5PUvX77Mxo0bUSgUjBkzxqBRiqF5Yss58qDKFbuAfg7OnpkzR8A8eXJDo6nVKEfucY6WI6Yie4S3ceNGZs2ahaenp8En+eKLLwCYPXs2AwcORKVS8eGHH6LVavn4449rPKZ+/frUr1+/2uMaK2xX87BglQpfnY6yggIUCgXPqNUOs9niqVOnKC8vp0OHDly8eNGgNzND88ScOWKKLiaSJLF06VIOHjzIyJEjGTZsmEO8kZkzR8D0eVLbbiNyj7tw4QLJycm8/PLLtGvXDqVSWesC6+hkFzylUsnChQv1X9dmP7zPPvsMgD59+hg8B2hLCgoKiI+Pp7xRI/19VJZWl5vVDVHXuV6wvjxZuXIlr732GpIkERkZyXfffVfnRSE6nY63336bb7/9FkmSSEpKMsrrGoOpW5TZW4482G0ESaJEpyOAqt1GaipCco7LvXCBlJQUXF1d9fcs2lo7L2siu+A988wz+j+LG88f7c6dO8TFxXHnzh08WrVCrVZbxb9XbW9Wrw1D5nqtnSm6mFRUVPD999/ri52xXtcYTFHca2JPOVKptt1GHnXcqVOn+HPbNry8vIiOjtaPUuUWWDEKrE52watsJVZYWMjOnTt55ZVXTBWTzaq8WbioqIghQ4awwYr6yNXmZvXaGDp0KCdPnuTChQt20RjY2F1MysvLWbduHQcOHODh6XNTdEcxZLRmrhZl9pYjpqDRaPj3oUOE16tHdHR0jR1pHldgxSiwZrILXmXCV1RUkJKSYrKAbJm7uztubm5ERkbSunVruL9AxBrU5mb12qjLXK81MnYXEycnJ9RqNX369CEuLs6k3VEMHa2Zq0WZveWIKSiVSgIDA3l11Kha/TvV9jKrvZNd8CqLnLOzs1XMM1iT4uJi1Go1Xl5evP766/oFKqZu1GyNjDHXa02M1cWkuLgYd3d3lEolw4YNQ6FQoNVqTdYdpTajNXO1KLO3HKlU224jDx6nuT9Cc3FxYeTIkXX+UGArTZ3NRfZqilatWjF69GgyMzO5fPmyKWMymyyNhiMFBdX+yzJg9dapU6f4+uuvuXDhAoC+2FU2ai7jXqPmlUboMGELpkyZgr+/P/Xr1+fNN9+0dDhGUdcuJrdv3+a7775jx44dAPqd7E3ZHeVxo7VHqSzulYusTNWizN5ypLbdRh48rlyS8Cwvp7CoCI1Oh5+zM/5PmG54uMBqrLill7WQPcLbu3cvGo2G7t27k5aWxvDhw00Zl1nUeIM48hdypKWl8a9//QtPT88qfTHN0ajZWq1bt47p06ej0Wj4xz/+waeffmrpkIyitl1MsrOziYuLo7i4uMal8abqjlLb0drUqVMJCwvj/fff5/PPPzdJbPaWI7XtNlJ5XE5JCampqWRkZNCieXP6de+Ov6vrI4+rLJR3KiqqFFgeKLDWtg+dtZBd8EpLS9m0aRPvvPOO/pOqratcyDHy1CkANt7vW/ikhRySJPH777+za9cu/Pz8iImJqTKpbI5GzdZKq9WSl5eHTqdDkiQyMzP1baAczbVr10hMTKSiooJRo0bRqlUrs527LpdiTd2izB5zpLbdRry1WrYmJ6PJzCSyc2ciIyOfeBuTnAJb+Xe20tTZXGQXvIULF1JRUYFOp3vkDZ62pnIhh/v9BOvk5SXruJ07d/LHH38QFBREdHR0tevslY2aH0wvYzdqtlZeXl5s2rQJuLeyNyUlxSG74V+8eJF169ahUCgYP348TZo0MXsM5hit1YbIkXsKCgqIi4vj5s2b9O7dmxdeeEH2LUxPKrByRoGOSPZ3PXfuXM6fP0+DBg24ceMGP/74oynjsmoeHh40bNiQcePG1dgE2hyNmq2VI75x1cTd3R1PT09Gjhypv6XHEqyxobTIkXtcXV1xcXFhwIABdO/e3aivbWtNnc1FdsHz9PQkNjaW4OBgUlNTTRmTVaqoqECr1aJSqejRowfPPvvsY9uFmbpRs2CdKlfshoSEMGPGDIdpKSfIV5kjKpWKKVOmmCxHbKmps7nIXqXp5uaGv78/y5cv56+//jJlTFZHo9GQmJjIunXrqLj/iUlOkno5OdFMrXaYYvf999/r/3znzh3efvttC0ZjXpIksXfvXr799lt9135R7Kpz5BwBuHr1KkuWLOHo0aOAyBFzkz3C+5//+R8uXbrERx99VONd//aquLiYhIQEMjMz6dy5s0jQx/D09GTNmjW0bt2a77//nlmzZlk6JLOQJIkdO3Zw4MABGjRogJsDXLquLUfNEYBz586RnJyMUqms1SbaQt3JHuF98sknJCQkcO3aNebOnWvKmKzG3bt3WbNmDZmZmfTu3ZtBgwZZRV9MazV27FjUajXffvst//u//2v0G5atkU6nY8uWLRw4cIBGjRoxadIk0UHkMRwxRwBOnjzJ+vXrcXFxYdKkSTxt5P61gjwGzeGpVCq6dOnCli1bTBmTVbh9+zY//PAD+fn5JplUtkdz5swB7i3HnzVrFiqVym66aNREq9WyceNGzp49S/PmzRk1apR+Q1KhZo6WIwCHDx8mNTUVb29voqOjCQwMtHRIDkt2wWvVqhUbNmzgjTfesPi28uagUqlwcXFh6NChdOjQwdLh2IRFixZRXl6Oi4sLBQUFeMm8zcNWKZVKVCoV7dq1Y+jQoeJytwyOliNwb8VuQEAA0dHR+Pj4WDochya74PXp04ewsDB0Op1dX9arXEHl4eFRpS+mIcy195y1+b//+z/CwsIYMmQISUlJVFRUMH36dEuHZXSlpaW4urqiVCqJiopCoVDY9e+EMTlKjkiSRGlpKe7u7rRt25ZWrVqJD0RWQPYc3rx587h69SrXr1/n+vXrpozJ7Aq0Wi4UF/P98eN88803pN+/f662CboiM5OItDQul5ZyubSUiLQ0ItLSWJGZacywrU56ejpDhgwBIDY21m56rj4oPz+flStXkpqaiiRJ+r6YgjyOkCNarZZNmzaxevVqSkpKALEa01rIHuGFh4cTERGBq5H3T7O0yibPAK/dvs2IoCAm1nHRgbn2nrM2D+dG5S+7vcjNzSUuLo78/HwiIiJEoasFe8+R8vJyNm7cyLlz52jRogXODtrRxFrJ/mmcOHGCsrIy/ScVe+iW8HCTZxQKNjVpwtd1vG/OXHvPWZuXX36ZN998k0aNGvHXX3/Z1SbBWVlZxMfHU1JSwiuvvMIzzzxj6ZBskj3niEajISkpiatXr9K+fXuioqLEyM7KGDTCi42NrbGVlq1y5CbPpvDcc8/Rs2dP8vPz8fX1Zc+ePZYOySiuXr1KYmIiWq2WMWPGEB4ebumQbJa95khxcTHx8fFkZWXRtWtXXnrpJXEFwArJnsPLy8tjwYIFzJkzR7+02NZVNnl+kKM0eTaV3NxcEhMTiYqKeuzea7bEzc0Nd3d3oqOjRbEzAnvMERcXF5ydnXn++edFsbNiskd48+bNM2EY5qXT6SgvL3foJs+mMHv2bDw9PenXrx95eXnExsZaOqQ6KSkpwd3dnaCgIN58801xecoI7DVHKm8oFzli3WSP8OxFRUUFGzZsICEhgfLycqYGB9NQpcIVSG3XjqnBwZYO0Wa1a9eO3Nxc9u3bR35+vqXDqZODBw+ydOlSsrOzAbHKzljsKUcyMzNZtmwZhw8fBkSO2ALZBe/DDz9k27ZtlJeXmzIek9JoNCQkJJCeno6fn59+o0VHa/JsKhMmTGDevHlERUXh6+vL66+/bumQDCZJEnv27OGXX35BrVaLvphGZg85AnDlyhXWrl1LWVmZuJnchsi+pDl//nz279/Pm2++SWhoKBMnTqRx48amjM2oioqKSEhIEJPKJjRnzhzc3d0JCQnhgw8+sLkPR5Ik8csvv3Do0CGCg4OZMGECHh4elg7Lrth6jsC9ewmTk5NxdnYmOjrapt4HHZ3sEd6aNWvYunUrzz//PK+99hqff/65KeMyqrt377J69WqysrLEpLIJSZLEhx9+yKVLlwBsqq+kTqdj06ZNHDp0iMaNGzNp0iRR7EzAlnME4Pjx46xfvx6VSsWkSZNEsbMxskd43bp149VXXwWgsLDQpgqei4sLTk5OvPTSSzz77LOWDsdu6e63UpMkycKRGE6hUODi4kJ4eDgjRoywuTdiW2HLOQL3euz6+voyYcIE/MUUiM2RXfDWrl3L+fPnCQ4OJisrix9//NGUcRlF5Qoqd3d3XnvtNdH1wMScnJz0t6xU/t/aO+FrNBpcXFxQKpUMGjQIQD+3KxifLebIg30xW7ZsSbNmzcR7iY0yaHug2NhYgoOD2bZtmyljMoorV66wbt06hgwZQps2bUSCmkHXrl2Be6MlLy8vevfubeGIHq+oqIj4+HiCgoIYOnSoKHRmYGs5IkkS27Zt49KlS7z66qt4eHiI9xIbJvs33M3NDX9/f5YvX861a9dMGVOdpaenEx8fDyA24zSj0NBQQkNDCQkJobi42KobFOTl5bF69Wpu3LiBn5+fpcNxGLaUI1qtlp9++onDhw/j7e0tCp0dkP0TbNKkCZ07d6ZNmzZWfb/J8ePH2bx5s74zRrC4r85sKj+9V/r3v/9toUge7+bNm8THx3P37l0iIyOrxS2Yjq3kSHl5OcnJyZw/f56WLVsyYsQIUfDsgOyfYGpqKkVFRfoNG62x6evBgwf55Zdf8PHxISYmRkwqm1nmA9sfZWdnU1xcbMFoapaZmUl8fDwajYbhw4fTrl07S4fkUGwhRzQaDYmJiVy7do0OHTowZMgQcbnbTsgueP379+fpp582ZSx15uLiYtDOwo66UauppKSkAPfmZ7y9vfnss88sHFF1Li4uuLi4MGzYMJo3b27pcByOLeSIk5MTSqWSbt26MWDAAHELkx2RXfBOnTrF4MGDrW5OTJIkysrKUKlUdOrUifbt28u+9LAiM5P5V6/qv45ISwPgo8aNmWflxd0ahYaG6v+sUCjYs2eP1VwJKC0txc3NjcDAQN58801xecpCbCFHnJ2dmTBhAk5OTqLY2RnZv/VKpZKFCxfqv7aGpcRarZbNmzeTm5vLxIkTUalUBr2ROepGraZS+Wam1WpZtWoVAQEBVvFmduzYMbZv305MTAwhISGi2FmQtebIzZs3iYuLo3v37nTv3l3kiJ2S/VONiYnh5MmTSJJEYWGhKWOSpaKigvXr13P+/HnCw8NrtZDGUTdqNZWuXbty7tw5lixZwsSJE61iyfn+/fvZsWMHPj4+qMTP2uKsMUdycnLYuXMnGo1Gv0ZBsE+yC95XX31FcHAwbdq04cSJEwad5OLFi6xevRqlUklMTAxZWVls374dhULB/Pnza1WsUlNTUSgUYlLZiixbtozTp0/zzjvv4OnpSWZmJiEhIbKPN3aeHDp0iOzsbAIDA4mOjsbb29vQb0kwMmvLEYCtW7cSEhLC2LFjxbyunZNdJfz8/Ojbty+enp4GdwffuHEjQUFBKBQKQkND2bJlCwsXLqRz584cOHDA4KABsrKy6NatG1FRUaLYWZHWrVuzY8cOUlJS9AsU5DJ2nhw9epTQ0FAmT54sip0VsaYcAfTFUxQ7+yd7hDdw4ECCgoJISEggMDDwic9fs2YN+/fvB+A///kPe/bs4fLly2zdulXfpzAwMJBbt27VeHxOTg43b96s9vjp06cBqF+/PgEBARw9elTutyCYWI8ePao9duTIkccek5qaypUrV3BycjI4T56UIwqFgrZt23L27NnafDuCCZg7R+DJedKyZUtu3br1yOMF29CyZUvUavVjnyO74PXq1QuAv//977KeP3nyZCZPngzAe++9h1qtxtvbG51Opx+R3bx5k/r169d4/Pr161m2bFm1x+/evQvAJ598wieffCI3fMGKpaWl0alTJ4Pz5Ek5smLFClasWGHa4AWzqG2OwJPzZMaMGaYLXDCbyhx5HIVkhrblJ0+eZOXKlXh6evLuu+9y7Ngxdu/ejVar5eOPP67xkuSjPpUdPXqU9957jy+++ILWrVubOnSjuXjxIu+++y5ffvklTZs2tXQ4spkj7spPZobmicgR62DNOQIiT6yBOXPksSQb8+eff0otWrSQ/vzzT0uHYhARt/nYYsySJOI2NxG3+VhLzGK1hyAIguAQRMETBEEQHIIoeIIgCIJDEAVPEARBcAg2V/ACAwOZMWOGrHsBrYmI23xsMWYQcZubiNt8rCVms9yWIAiCIAiWZnMjPEEQBEGoDVHwBEEQBIdgU5s+maJTujncvn2b119/nQ0bNrB3716biPnw4cOsW7eOsrIy3nrrLZv5t7bVHAGRJ+YicsS8rClHbGqEZ4pO6eawatUq/RYothJzYWEhixcvZvr06fz+++82E7et5giIPDEXkSPmZU05YvUjvLruumAJD8bcrl07+vfvT1JSEoDVxgxV4+7cuTPdu3cnPj6emTNnkp6eDlh/3LaSIyDyxJxEjpiXteaI1Re8uu66YAkPx5yTk8PRo0fZvn271cYMVePOz89n4cKFzJgxA39/f5uJ21ZyBESemJPIEfOy1hyxqdsSatMp3VrMnj2bzz77jD179thEzHPmzCEnJwc/Pz9efPFF3NzcbCJuW84REHliDiJHzMuacsSmCp4gCIIg1JZ1fiQQBEEQBCMTBU8QBEFwCKLgCYIgCA5BFDxBEATBIYiCJwiCIDgEUfBMKDs729IhCFZO5Iggh8gT43D4grd06VIyMjIAOHjwIMnJyUyfPl3/90lJSQa3v0lJSeHgwYN8/fXXACxfvtzguDIyMli6dGm1+FJSUvTPuXnzJsnJyVWe8ySzZ88G4MSJE/znP/+RdUzl97NmzRpKSkoM/E5sn8iRJ3P0HAGRJ3JYOk+svtOKuTk5OdG8eXMuXbpEkyZNOHr0KGPHjgXuJc6cOXMIDw+nqKiI4OBgVCoVo0ePZunSpQB07NgRgNLSUk6dOsXOnTu5cuUKGRkZ/P3vf6d9+/Z4e3szcuRIPvnkE7y9vcnJydEff+nSJb777juUSiUNGjR4bKzx8fHExsayatUqCgsLee+995g7dy7Lly+nvLychg0b0qRJE7Zv387t27eZOXOm/tiDBw/St29fYmJiavx+Fi5ciFqtplevXvpjXnjhBTZu3EhMTIxR/81tjcgRkSNyiDyxvjxx+BGeQqGo9vWIESP4+eefuXLlCs2aNavynLZt2zJ79mw0Gg1vvfUWly5d4ueff0aj0aBWqzlx4gQAbm5utG7dmhdffFF/bKtWrfh//+//cebMGVJTU5k0aRJz5sxBq9Xqn5OcnMz777/PlClT9I8tXryY2bNn889//rNKrDdu3MDDwwO413botdde07frmTt3Lj179iQkJITIyEhCQkI4efKk/tjS0lJUKtUjv5+JEyeyYMECduzYoT+mcePGVV7DUYgcETkih8gT688Thx/heXl5kZOTQ1hYGNnZ2Xh5edGoUSOys7P517/+xejRo6s8X61W4+zsjKurK3AvqXU6HYMGDaJbt25s376d0tJS/d89yN3dHQBnZ2cqKiqQJKnGX5IH+/sBzJo1i7CwMA4ePMj169f1jz+4rUZsbCwbNmxg5MiRVDbPycjIYO/evfTo0YO2bdui0+mAe1uM1KtX75Hfz4NxPdyIx9nZ8VJG5IjIETlEnlh/njhmZj5gyJAh/O1vf9Mn0KJFiwD0W1fIaW46ePBg5s+fT0pKCoMHD9YnacOGDUlISHjkMYsWLcLX17fK42PGjOGLL77AxcWFwMDAx543MDCQwsJC4N7lj/T0dMrKysjMzOSjjz6iRYsWBAQE8Mcff5Cfn0/79u0B2LdvX5XLCzXF9vnnn+Ph4cHgwYPJy8sD7l0iadu27RP/PeyNyJGaYxM5UpXIk5pjs6Y8Eb00LeTSpUskJyej1Wrp2rVrlcsVcuXl5fHrr78yatQoE0RY3dq1axk/frx+WxLBtESOCHKIPJFPFDxBEATBITj8ohVBEATBMYiCJwiCIDgEUfAEQRAEhyAKniAIguAQRMETBEEQHIIoeIIgCIJDEAVPEARBcAii4AmCIAgOQRQ8QRAEwSH8//F+56NN49EWAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 501.969x137.795 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 501.969x137.795 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# for export\n",
    "series = 'All'\n",
    "plot_query = 'n_curve > 0'\n",
    "df_q = agg_df.dropna(subset=['Tm', 'dG_37'])#.query(plot_query)\n",
    "df_q[['dG_37_uv_std', 'Tm_uv_std', 'dH_uv_std']] = 0\n",
    "\n",
    "def mae_norm(x, xerr, y):\n",
    "    epsilon = 1e-2\n",
    "    aerr = np.abs(x - y)\n",
    "    return np.mean(aerr / (xerr + epsilon))\n",
    "\n",
    "def plot_uv_comparison(df_q, param, lim, array_suffix=''):\n",
    "    unit_dict = dict(dH='kcal/mol', dS='kcal/mol', dG_37='kcal/mol', Tm='°C')\n",
    "    color_dict = dict(dH='c', Tm='cornflowerblue', dG_37='teal', dS='steelblue')\n",
    "    param_name_dict = dict(dH=r'$dH°$', dS=r'$dS°$', dG_37=r'$dG°_{37}$', Tm=r'$T_m$')\n",
    "    unit = unit_dict[param]\n",
    "    param_name = param_name_dict[param]\n",
    "    \n",
    "    kwargs = dict(fmt = '.', alpha=1, marker='o', markersize=3, capsize=2)\n",
    "    \n",
    "    fig, ax = plt.subplots(1, 3, figsize=(3*4.25*cm, 3.5*cm))\n",
    "    x = df_q[param + '_uv']\n",
    "    y = df_q[param+array_suffix]\n",
    "    xerr = df_q[param+'_uv_std'].values\n",
    "    ax[0].plot(lim, lim, '--', c='gray')\n",
    "    ax[0].errorbar(x, y,\n",
    "                   xerr=xerr, yerr=df_q[param+'_se'],\n",
    "                   color=color_dict[param], **kwargs)\n",
    "    ax[0].set_xlim(lim)\n",
    "    ax[0].set_ylim(lim)\n",
    "    ax[0].set_xlabel(f'UV melting {param} ({unit})', fontsize=tick_font_size)\n",
    "    ax[0].set_ylabel(f'array measurement {param} ({unit})', fontsize=tick_font_size)\n",
    "    # ax[0].set_title('Array Melt Measurement\\ncorr = %.4f, MAE = %.4f' %\n",
    "    #                 (pearsonr(x, y)[0], mae_norm(x, xerr, y)))\n",
    "    \n",
    "    x = df_q[param+'_uv'].values\n",
    "    y = df_q[param+'_NUPACK_salt_corrected'].values\n",
    "    ax[1].errorbar(x, y,\n",
    "                   xerr=df_q[param+'_uv_std'],\n",
    "                   color='k', **kwargs)\n",
    "    ax[1].plot(lim, lim, '--', c='gray')\n",
    "    ax[1].set_xlim(lim)\n",
    "    ax[1].set_ylim(lim)\n",
    "    ax[1].set_xlabel(f'UV melting {param} ({unit})', fontsize=tick_font_size)\n",
    "    ax[1].set_ylabel(f'NUPACK prediction {param} ({unit})', fontsize=tick_font_size)\n",
    "    # ax[1].set_title('NUPACK prediction\\ncorr = %.4f, MAE = %4f' %\n",
    "    #                 (pearsonr(x, y)[0], mae_norm(x, xerr, y)))\n",
    "\n",
    "    x = df_q[param+array_suffix].values\n",
    "    y = df_q[param+'_NUPACK_salt_corrected'].values\n",
    "    kwargs['marker'] = 's'\n",
    "    kwargs['markersize'] = 5\n",
    "    kwargs['alpha'] = 0.8\n",
    "    ax[2].errorbar(x, y,\n",
    "               xerr=df_q[param+'_uv_std'],\n",
    "               color=color_dict[param], **kwargs)\n",
    "    ax[2].plot(lim, lim, '--', c='gray')\n",
    "    ax[2].set_xlim(lim)\n",
    "    ax[2].set_ylim(lim)\n",
    "    ax[2].set_xlabel(f'UV melting {param} ({unit})', fontsize=tick_font_size)\n",
    "    ax[2].set_ylabel(f'NUPACK prediction {param} ({unit})', fontsize=tick_font_size)\n",
    "    # ax[2].set_title('Array Melt vs NUPACK prediction\\ncorr = %.4f, MAE = %.4f' %\n",
    "    #                 (pearsonr(x, y)[0], mae(x, y)))\n",
    "    for i in range(3):\n",
    "        ax[i].set_yticklabels(ax[i].get_yticklabels(), fontsize=tick_font_size)\n",
    "        ax[i].set_xticklabels(ax[i].get_xticklabels(), fontsize=tick_font_size)\n",
    "\n",
    "    plt.suptitle(param)\n",
    "    sns.despine()\n",
    "    # plt.show()\n",
    "    \n",
    "plot_uv_comparison(df_q, 'dG_37', lim = (-5.5, 1))\n",
    "# util.save_fig(f'./fig/Fig2/Fig2b/{series}_dG_37.pdf')\n",
    "plot_uv_comparison(df_q, 'dH', lim = (-60, -10))\n",
    "# util.save_fig(f'./fig/Fig2/Fig2b/{series}_dH.pdf')\n",
    "plot_uv_comparison(df_q, 'Tm', lim = (30, 90))\n",
    "# util.save_fig(f'./fig/Fig2/Fig2b/{series}_Tm.pdf')\n",
    "# plot_uv_comparison(df_q, 'dS', lim = (-.18, -.01))\n",
    "# util.save_fig(f'./fig/Fig2/Fig2b/{series}_dS.pdf')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7a62de96-65b8-47fa-b0a9-cad5efcca476",
   "metadata": {},
   "source": [
    "## Fig. S2b: WC68 over concentration range"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "368e8aba-64a5-4c16-860d-30980c8b5af1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 167.323x137.795 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "result_df = pd.read_csv('./data/uv_melt/uvmelt_230502_2.csv', index_col=0).query('pass_qc')\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(4.25*cm, 3.5*cm))\n",
    "df = result_df.query('SEQID == \"WC68\" & (curve_name == \"CoolingCurve\" | curve_name == \"SecondaryCoolingCurve\")')\n",
    "sns.swarmplot(data=df, y='Tm', x='conc_uM', #hue='curve_name', \n",
    "              color='cornflowerblue', alpha=.7, marker='D', size=4, ax=ax)\n",
    "# ax.get_legend().remove()\n",
    "sns.boxplot(data=df, y='Tm', x='conc_uM', color='white', linewidth=.5, ax=ax)\n",
    "\n",
    "ax.set_xlabel('DNA concentration ($\\mu M$)', fontsize=label_font_size)\n",
    "ax.set_ylabel('UV melting $Tm$ (°C)', fontsize=label_font_size)\n",
    "plt.ylim([50,90])\n",
    "# ax.set_xticklabels(ax.get_xticklabels(), fontsize=tick_font_size)\n",
    "ax.yaxis.set_major_locator(MultipleLocator(10))\n",
    "ax.yaxis.set_minor_locator(MultipleLocator(2.5))\n",
    "ax.tick_params(axis='both', which='major', labelsize=tick_font_size)\n",
    "\n",
    "sns.despine()\n",
    "util.save_fig('./fig/Fig2/FigS2b_WC68_conc.pdf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9ef1aa48-a856-4e0e-9200-57a3d0601525",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.8.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}