notebooks/40_cagi6_mapk1_submission.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "26c81e96-6b41-41fa-80a6-6360c482eb58",
   "metadata": {},
   "source": [
    "## Summary\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ec276dec-6afa-4b57-bbfc-a11e04a8767e",
   "metadata": {},
   "source": [
    "## Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "58a8599f-8b1e-4448-9c07-754ea7c57c0e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "env: CUDA_VISIBLE_DEVICES=0\n",
      "env: TF_FORCE_UNIFIED_MEMORY=1\n",
      "env: XLA_PYTHON_CLIENT_MEM_FRACTION=4.0\n"
     ]
    }
   ],
   "source": [
    "%env CUDA_VISIBLE_DEVICES=0\n",
    "%env TF_FORCE_UNIFIED_MEMORY=1\n",
    "%env XLA_PYTHON_CLIENT_MEM_FRACTION=4.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "539f4edf-ca04-4bef-b5c9-a6d402f7570b",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Warning: importing 'simtk.openmm' is deprecated.  Import 'openmm' instead.\n"
     ]
    }
   ],
   "source": [
    "import concurrent.futures\n",
    "import contextlib\n",
    "import itertools\n",
    "import json\n",
    "import os\n",
    "import sys\n",
    "import tempfile\n",
    "import urllib.request\n",
    "from datetime import datetime\n",
    "from pathlib import Path\n",
    "\n",
    "import dotenv\n",
    "import elaspic2 as el2\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import pyarrow as pa\n",
    "import pyarrow.parquet as pq\n",
    "import seaborn as sns\n",
    "import torch\n",
    "from elaspic2.plugins.msa import analyze_msa\n",
    "from elaspic2.plugins.rosetta_ddg import RosettaDDG\n",
    "from kmbio import PDB\n",
    "from kmtools import structure_tools\n",
    "from tqdm.auto import tqdm"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5be0ca05-c038-414f-ba84-b2ae8013088b",
   "metadata": {},
   "source": [
    "## Parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "10e1faee-9f6a-40d8-8819-272b1c9837bb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "PosixPath('/home/kimlab5/strokach/workspace/elaspic/elaspic2-cagi6/notebooks/40_cagi6_mapk1_submission')"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "NOTEBOOK_DIR = Path(\"40_cagi6_mapk1_submission\").resolve()\n",
    "NOTEBOOK_DIR.mkdir(exist_ok=True)\n",
    "\n",
    "NOTEBOOK_DIR"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "2b7e8ea5-8f73-4c64-9f72-3fbefc9ddf1d",
   "metadata": {},
   "outputs": [],
   "source": [
    "src_dir = str(NOTEBOOK_DIR.parents[1].joinpath(\"src\"))\n",
    "if src_dir not in sys.path:\n",
    "    sys.path.insert(0, src_dir)\n",
    "\n",
    "import helpers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "4ffc9d11-d255-47dc-85dd-550d53787ba4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'P28482'"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "UNIPROT_ID = \"P28482\"\n",
    "\n",
    "UNIPROT_ID"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "1a161a46-46da-410c-833f-cc487c097541",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "HCyTzII-HD1t",
    "outputId": "92f63199-1727-48d3-cb4b-6ff798fa75b6"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "device(type='cpu')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "\n",
    "device"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "92bd5cc7-d64c-4f0d-bdfc-3d44ede20480",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "i3LyUBcjSUzb",
    "outputId": "054b8acc-0717-48cf-83d9-ca0b886f7b10"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'2021-10-01T21'"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "version = datetime.now().isoformat(timespec=\"hours\")\n",
    "\n",
    "version"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ba95e079-5284-4944-a22e-32c22becc8dd",
   "metadata": {},
   "source": [
    "## Download data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "9688e150-c466-4320-b201-3ebe4b32edb0",
   "metadata": {},
   "outputs": [],
   "source": [
    "def download(url, filename):\n",
    "    urllib.request.urlretrieve(url, filename)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "de7bafde-0507-4f52-9fb1-12c2f6f47b1b",
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_sequence(sequence_file):\n",
    "    with sequence_file.open(\"rt\") as fin:\n",
    "        data = fin.read()\n",
    "    chunks = []\n",
    "    for line in data.split(\"\\n\"):\n",
    "        if line.startswith(\">\"):\n",
    "            continue\n",
    "        chunks.append(line.strip())\n",
    "    return \"\".join(chunks)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "ce54f842-cfd5-467c-960e-e4706d6fc8f2",
   "metadata": {},
   "outputs": [],
   "source": [
    "if not NOTEBOOK_DIR.joinpath(\"validation.py\").is_file():\n",
    "    import synapseclient\n",
    "    import synapseutils\n",
    "\n",
    "    dotenv.load_dotenv(\"../.env\")\n",
    "    syn = synapseclient.Synapse()\n",
    "    syn.login(os.environ[\"SYNAPSE_USERNAME\"], os.environ[\"SYNAPSE_PASSWORD\"])\n",
    "    _ = synapseutils.syncFromSynapse(syn, \"syn25955327\", path=NOTEBOOK_DIR)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "e62c0d9c-f89a-427e-9200-e00939cdf402",
   "metadata": {},
   "outputs": [],
   "source": [
    "sequence_file = NOTEBOOK_DIR.joinpath(f\"{UNIPROT_ID}.fasta\")\n",
    "\n",
    "if not sequence_file.is_file():\n",
    "    download(f\"https://www.uniprot.org/uniprot/{UNIPROT_ID}.fasta\", sequence_file)\n",
    "\n",
    "sequence = load_sequence(sequence_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "bc4e24de-1b1a-4fe0-92eb-4dd5910d5289",
   "metadata": {},
   "outputs": [],
   "source": [
    "structure_file = NOTEBOOK_DIR.joinpath(f\"{UNIPROT_ID}.pdb\")\n",
    "\n",
    "if not structure_file.is_file():\n",
    "    download(f\"https://alphafold.ebi.ac.uk/files/AF-{UNIPROT_ID}-F1-model_v1.pdb\", structure_file)\n",
    "\n",
    "with structure_file.open(\"r\") as fin:\n",
    "    structure_blob = fin.read()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "6a937350-73cb-4515-8205-870357bcc175",
   "metadata": {},
   "outputs": [],
   "source": [
    "alignment_file = NOTEBOOK_DIR.joinpath(f\"{UNIPROT_ID}.a3m.gz\")\n",
    "\n",
    "if not alignment_file.is_file():\n",
    "    from elaspic2.plugins.alphafold import mmseqs2\n",
    "\n",
    "    dotenv.load_dotenv(\"../.env\")\n",
    "    with mmseqs2.api_gateway(mmseqs2.MMSEQS2_HOST_URL) as gateway:\n",
    "        alignment = mmseqs2.run_mmseqs2(sequence, gateway=gateway)\n",
    "        assert alignment[1] == f\"{sequence}\\n\"\n",
    "        alignment_df = pd.DataFrame({\"alignment\": alignment})\n",
    "        pq.write_table(pa.Table.from_pandas(alignment_df, preserve_index=False), alignment_file)\n",
    "\n",
    "alignment = pq.read_table(alignment_file).to_pandas()[\"alignment\"].values.tolist()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "af4fdd39-dd7b-49f3-b37a-6747aef9c571",
   "metadata": {},
   "source": [
    "## Load data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "d8c86eef-3016-4305-b977-1037dd6d97c6",
   "metadata": {},
   "outputs": [],
   "source": [
    "def mutation_matches_sequence(mutation, sequence):\n",
    "    wt, pos, mut = mutation[0], mutation[1:-1], mutation[-1]\n",
    "    pos = int(pos)\n",
    "    return sequence[pos - 1] == wt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "9b3b0ba8-47d5-447c-b309-646698238fb2",
   "metadata": {
    "id": "gJG9gYqPHD15"
   },
   "outputs": [],
   "source": [
    "def sequence_matches_structure(sequence, structure_blob):\n",
    "    with tempfile.NamedTemporaryFile(suffix=\".pdb\") as tmp_file:\n",
    "        with open(tmp_file.name, \"wt\") as fout:\n",
    "            fout.write(structure_blob)\n",
    "        structure = PDB.load(tmp_file.name)\n",
    "\n",
    "    chain_sequence = structure_tools.get_chain_sequence(\n",
    "        structure[0][\"A\"], if_unknown=\"replace\", unknown_residue_marker=\"\"\n",
    "    )\n",
    "    return sequence == chain_sequence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "8e273fa4-c6d1-4b41-9edc-74a5d86d9eab",
   "metadata": {},
   "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>MAPK1-variant</th>\n",
       "      <th>DDG-NotPO4</th>\n",
       "      <th>Standard-Deviation-DDG-NotPO4</th>\n",
       "      <th>DDG-PO4</th>\n",
       "      <th>Standard-Deviation-DDG-PO4</th>\n",
       "      <th>Ratio-kcat/Km</th>\n",
       "      <th>Standard-Deviation-Ratio-kcat/Km</th>\n",
       "      <th>Comment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>p.E33Q</td>\n",
       "      <td>*</td>\n",
       "      <td>*</td>\n",
       "      <td>*</td>\n",
       "      <td>*</td>\n",
       "      <td>*</td>\n",
       "      <td>*</td>\n",
       "      <td>*</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>p.E81K</td>\n",
       "      <td>*</td>\n",
       "      <td>*</td>\n",
       "      <td>*</td>\n",
       "      <td>*</td>\n",
       "      <td>*</td>\n",
       "      <td>*</td>\n",
       "      <td>*</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  MAPK1-variant DDG-NotPO4 Standard-Deviation-DDG-NotPO4 DDG-PO4  \\\n",
       "0        p.E33Q          *                             *       *   \n",
       "1        p.E81K          *                             *       *   \n",
       "\n",
       "   Standard-Deviation-DDG-PO4 Ratio-kcat/Km Standard-Deviation-Ratio-kcat/Km  \\\n",
       "0                           *             *                                *   \n",
       "1                           *             *                                *   \n",
       "\n",
       "  Comment  \n",
       "0       *  \n",
       "1       *  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result_template_df = pd.read_csv(NOTEBOOK_DIR.joinpath(\"mapk1_predictions.tsv\"), sep=\"\\t\")\n",
    "\n",
    "result_template_df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "5bfc1d8d-c316-4b98-9de4-4e7282d632b7",
   "metadata": {},
   "outputs": [],
   "source": [
    "result_df = result_template_df.copy()\n",
    "result_df[\"mut\"] = result_df[\"MAPK1-variant\"].str[2:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "0a500b9f-bd8b-47c9-be38-2182a73b2632",
   "metadata": {},
   "outputs": [],
   "source": [
    "assert all([mutation_matches_sequence(mut, sequence) for mut in result_df[\"mut\"]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "7251f319-8562-4309-aa9c-ade5de2d1288",
   "metadata": {},
   "outputs": [],
   "source": [
    "assert sequence_matches_structure(sequence, structure_blob)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "e85e5625-2b0b-429d-9b43-80d3566415c0",
   "metadata": {},
   "outputs": [],
   "source": [
    "assert alignment[1] == f\"{sequence}\\n\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4fa0b6a3-2e3a-4155-bb87-538568d08ca0",
   "metadata": {},
   "source": [
    "## Run Rosetta"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "2073be3f-3e84-4851-a738-b904cab52e7e",
   "metadata": {},
   "outputs": [],
   "source": [
    "rosetta_ddg_data = RosettaDDG.build(\n",
    "    structure_file,\n",
    "    protocol=\"cartesian_ddg\",\n",
    "    energy_function=\"beta_nov16_cart\",\n",
    "    interface=0,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "38e42bcb-6ce2-4440-8c91-1bf8a6880fd2",
   "metadata": {},
   "outputs": [],
   "source": [
    "def rosetta_ddg_worker(mut, data):\n",
    "    results = RosettaDDG.analyze_mutation(f\"A_{mut}\", data)\n",
    "    results = {\"mut\": mut} | {f\"rosetta_{key}\": value for key, value in results.items()}\n",
    "    return results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "192db9b4-9f12-4362-bc92-b5f3787e5b39",
   "metadata": {},
   "outputs": [],
   "source": [
    "rosetta_results_file = NOTEBOOK_DIR.joinpath(\"rosetta-results.parquet\")\n",
    "\n",
    "if rosetta_results_file.is_file():\n",
    "    rosetta_results_df = pq.read_table(rosetta_results_file).to_pandas()\n",
    "else:\n",
    "    with concurrent.futures.ThreadPoolExecutor(len(result_df)) as pool:\n",
    "        rosetta_results = list(\n",
    "            tqdm(\n",
    "                pool.map(\n",
    "                    rosetta_ddg_worker,\n",
    "                    result_df[\"mut\"].values.tolist(),\n",
    "                    itertools.repeat(rosetta_ddg_data),\n",
    "                ),\n",
    "                total=len(result_df),\n",
    "            )\n",
    "        )\n",
    "    rosetta_results_df = pd.DataFrame(rosetta_results)\n",
    "    pq.write_table(\n",
    "        pa.Table.from_pandas(rosetta_results_df, preserve_index=False), rosetta_results_file\n",
    "    )"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3aa40f55-6795-4a8a-a00e-22116cdc6d74",
   "metadata": {},
   "source": [
    "## Run AlphaFold"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "9117368f-887a-4156-890b-449ca33dc775",
   "metadata": {},
   "outputs": [],
   "source": [
    "def worker(mutation, data):\n",
    "    try:\n",
    "        results = AlphaFold.analyze_mutation(f\"A_{mutation}\", data)\n",
    "    except Exception as error:\n",
    "        print(f\"{error!r}\")\n",
    "        return None\n",
    "    else:\n",
    "        results = {f\"alphafold_core_{key}\": value for key, value in results.items()}\n",
    "        return results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "f36721a1-d8e9-4b3e-aaea-3a7f711ee11c",
   "metadata": {},
   "outputs": [],
   "source": [
    "alphafold_results_file = NOTEBOOK_DIR.joinpath(\"alphafold-results.parquet\")\n",
    "\n",
    "if alphafold_results_file.is_file():\n",
    "    alphafold_results_df = pq.read_table(alphafold_results_file).to_pandas()\n",
    "else:\n",
    "    from elaspic2.plugins.alphafold import AlphaFold, AlphaFoldAnalyzeError, AlphaFoldBuildError\n",
    "    from jax.lib import xla_bridge\n",
    "\n",
    "    jax_device = xla_bridge.get_backend().platform\n",
    "    print(f\"Device: {jax_device}\")\n",
    "\n",
    "    AlphaFold.load_model(device=jax_device)\n",
    "    data = AlphaFold.build(sequence, ligand_sequence=None, msa=alignment)\n",
    "\n",
    "    result_list = []\n",
    "    for mut in tqdm(result_df[\"mut\"], leave=False):\n",
    "        result = worker(mut, data)\n",
    "        result_list.append({\"mut\": mut} | result)\n",
    "    alphafold_results_df = pd.DataFrame(result_list)\n",
    "\n",
    "    pq.write_table(\n",
    "        pa.Table.from_pandas(alphafold_results_df, preserve_index=False), alphafold_results_file\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "010f06f4-f320-491e-bc9a-ca066db6a5d1",
   "metadata": {},
   "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>mut</th>\n",
       "      <th>alphafold_core_scores_residue_plddt_wt</th>\n",
       "      <th>alphafold_core_scores_protein_plddt_wt</th>\n",
       "      <th>alphafold_core_scores_protein_max_predicted_aligned_error_wt</th>\n",
       "      <th>alphafold_core_scores_proten_ptm_wt</th>\n",
       "      <th>alphafold_core_features_residue_experimentally_resolved_wt</th>\n",
       "      <th>alphafold_core_features_residue_predicted_lddt_wt</th>\n",
       "      <th>alphafold_core_features_residue_msa_first_row_wt</th>\n",
       "      <th>alphafold_core_features_residue_single_wt</th>\n",
       "      <th>alphafold_core_features_residue_structure_module_wt</th>\n",
       "      <th>...</th>\n",
       "      <th>alphafold_core_features_residue_experimentally_resolved_mut</th>\n",
       "      <th>alphafold_core_features_residue_predicted_lddt_mut</th>\n",
       "      <th>alphafold_core_features_residue_msa_first_row_mut</th>\n",
       "      <th>alphafold_core_features_residue_single_mut</th>\n",
       "      <th>alphafold_core_features_residue_structure_module_mut</th>\n",
       "      <th>alphafold_core_features_protein_experimentally_resolved_mut</th>\n",
       "      <th>alphafold_core_features_protein_predicted_lddt_mut</th>\n",
       "      <th>alphafold_core_features_protein_msa_first_row_mut</th>\n",
       "      <th>alphafold_core_features_protein_single_mut</th>\n",
       "      <th>alphafold_core_features_protein_structure_module_mut</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>E33Q</td>\n",
       "      <td>84.271057</td>\n",
       "      <td>87.747954</td>\n",
       "      <td>31.75</td>\n",
       "      <td>0.887063</td>\n",
       "      <td>[2.8770702, 2.9461424, 3.2922769, 2.7614207, 3...</td>\n",
       "      <td>[-8.669692, -9.221002, -8.246704, -7.5423164, ...</td>\n",
       "      <td>[18.622025, 11.388911, 4.587965, 14.636523, -1...</td>\n",
       "      <td>[9.641211, 44.703594, 14.128213, 2.6826615, 3....</td>\n",
       "      <td>[0.0021715537, 0.011204451, -0.006013179, 0.00...</td>\n",
       "      <td>...</td>\n",
       "      <td>[2.6185238, 2.6907444, 3.006825, 2.516821, 3.1...</td>\n",
       "      <td>[-8.608849, -9.146437, -8.178978, -7.4733496, ...</td>\n",
       "      <td>[18.210424, 10.107504, 3.882144, 15.387014, -1...</td>\n",
       "      <td>[9.321604, 51.92668, 20.040691, 2.2599945, 5.8...</td>\n",
       "      <td>[0.0020749941, 0.0112785995, -0.006013153, 0.0...</td>\n",
       "      <td>[2.9144516, 3.056479, 3.4794421, 2.8902307, 3....</td>\n",
       "      <td>[-8.733058, -9.778259, -8.551963, -7.715663, -...</td>\n",
       "      <td>[2.2348666, 5.301391, 0.3843421, 1.4361696, -5...</td>\n",
       "      <td>[27.884449, 9.957067, 9.515727, 4.9400554, 4.9...</td>\n",
       "      <td>[0.0006174341, 0.010583821, -0.0059063067, 0.0...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>E81K</td>\n",
       "      <td>95.295345</td>\n",
       "      <td>87.747954</td>\n",
       "      <td>31.75</td>\n",
       "      <td>0.887063</td>\n",
       "      <td>[3.2879066, 3.3683913, 3.9344985, 3.2070408, 4...</td>\n",
       "      <td>[-11.546194, -12.453913, -10.83298, -9.735328,...</td>\n",
       "      <td>[13.546079, 6.353491, -4.5482187, 1.0821763, -...</td>\n",
       "      <td>[46.411034, 37.813553, 2.90744, 16.750113, 3.5...</td>\n",
       "      <td>[6.083399e-05, 0.006591566, -0.005965011, 0.00...</td>\n",
       "      <td>...</td>\n",
       "      <td>[2.9272072, 3.0048463, 3.4825618, 2.8184836, 3...</td>\n",
       "      <td>[-11.615765, -12.529614, -10.889041, -9.788271...</td>\n",
       "      <td>[13.030119, 5.61815, -3.9496727, 1.913593, -7....</td>\n",
       "      <td>[37.746056, 44.183926, 5.7669396, 7.617279, 0....</td>\n",
       "      <td>[0.00011404604, 0.0066877753, -0.005954543, 0....</td>\n",
       "      <td>[2.92088, 3.0634732, 3.4871912, 2.8968136, 3.4...</td>\n",
       "      <td>[-8.746692, -9.791034, -8.560947, -7.7228723, ...</td>\n",
       "      <td>[2.2382922, 5.3104186, 0.37440258, 1.425295, -...</td>\n",
       "      <td>[27.87276, 10.037536, 9.543559, 4.8875713, 4.9...</td>\n",
       "      <td>[0.0006678155, 0.01063951, -0.005907728, 0.003...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 29 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    mut  alphafold_core_scores_residue_plddt_wt  \\\n",
       "0  E33Q                               84.271057   \n",
       "1  E81K                               95.295345   \n",
       "\n",
       "   alphafold_core_scores_protein_plddt_wt  \\\n",
       "0                               87.747954   \n",
       "1                               87.747954   \n",
       "\n",
       "   alphafold_core_scores_protein_max_predicted_aligned_error_wt  \\\n",
       "0                                              31.75              \n",
       "1                                              31.75              \n",
       "\n",
       "   alphafold_core_scores_proten_ptm_wt  \\\n",
       "0                             0.887063   \n",
       "1                             0.887063   \n",
       "\n",
       "  alphafold_core_features_residue_experimentally_resolved_wt  \\\n",
       "0  [2.8770702, 2.9461424, 3.2922769, 2.7614207, 3...           \n",
       "1  [3.2879066, 3.3683913, 3.9344985, 3.2070408, 4...           \n",
       "\n",
       "   alphafold_core_features_residue_predicted_lddt_wt  \\\n",
       "0  [-8.669692, -9.221002, -8.246704, -7.5423164, ...   \n",
       "1  [-11.546194, -12.453913, -10.83298, -9.735328,...   \n",
       "\n",
       "    alphafold_core_features_residue_msa_first_row_wt  \\\n",
       "0  [18.622025, 11.388911, 4.587965, 14.636523, -1...   \n",
       "1  [13.546079, 6.353491, -4.5482187, 1.0821763, -...   \n",
       "\n",
       "           alphafold_core_features_residue_single_wt  \\\n",
       "0  [9.641211, 44.703594, 14.128213, 2.6826615, 3....   \n",
       "1  [46.411034, 37.813553, 2.90744, 16.750113, 3.5...   \n",
       "\n",
       "  alphafold_core_features_residue_structure_module_wt  ...  \\\n",
       "0  [0.0021715537, 0.011204451, -0.006013179, 0.00...   ...   \n",
       "1  [6.083399e-05, 0.006591566, -0.005965011, 0.00...   ...   \n",
       "\n",
       "  alphafold_core_features_residue_experimentally_resolved_mut  \\\n",
       "0  [2.6185238, 2.6907444, 3.006825, 2.516821, 3.1...            \n",
       "1  [2.9272072, 3.0048463, 3.4825618, 2.8184836, 3...            \n",
       "\n",
       "  alphafold_core_features_residue_predicted_lddt_mut  \\\n",
       "0  [-8.608849, -9.146437, -8.178978, -7.4733496, ...   \n",
       "1  [-11.615765, -12.529614, -10.889041, -9.788271...   \n",
       "\n",
       "   alphafold_core_features_residue_msa_first_row_mut  \\\n",
       "0  [18.210424, 10.107504, 3.882144, 15.387014, -1...   \n",
       "1  [13.030119, 5.61815, -3.9496727, 1.913593, -7....   \n",
       "\n",
       "          alphafold_core_features_residue_single_mut  \\\n",
       "0  [9.321604, 51.92668, 20.040691, 2.2599945, 5.8...   \n",
       "1  [37.746056, 44.183926, 5.7669396, 7.617279, 0....   \n",
       "\n",
       "  alphafold_core_features_residue_structure_module_mut  \\\n",
       "0  [0.0020749941, 0.0112785995, -0.006013153, 0.0...     \n",
       "1  [0.00011404604, 0.0066877753, -0.005954543, 0....     \n",
       "\n",
       "   alphafold_core_features_protein_experimentally_resolved_mut  \\\n",
       "0  [2.9144516, 3.056479, 3.4794421, 2.8902307, 3....             \n",
       "1  [2.92088, 3.0634732, 3.4871912, 2.8968136, 3.4...             \n",
       "\n",
       "   alphafold_core_features_protein_predicted_lddt_mut  \\\n",
       "0  [-8.733058, -9.778259, -8.551963, -7.715663, -...    \n",
       "1  [-8.746692, -9.791034, -8.560947, -7.7228723, ...    \n",
       "\n",
       "   alphafold_core_features_protein_msa_first_row_mut  \\\n",
       "0  [2.2348666, 5.301391, 0.3843421, 1.4361696, -5...   \n",
       "1  [2.2382922, 5.3104186, 0.37440258, 1.425295, -...   \n",
       "\n",
       "          alphafold_core_features_protein_single_mut  \\\n",
       "0  [27.884449, 9.957067, 9.515727, 4.9400554, 4.9...   \n",
       "1  [27.87276, 10.037536, 9.543559, 4.8875713, 4.9...   \n",
       "\n",
       "  alphafold_core_features_protein_structure_module_mut  \n",
       "0  [0.0006174341, 0.010583821, -0.0059063067, 0.0...    \n",
       "1  [0.0006678155, 0.01063951, -0.005907728, 0.003...    \n",
       "\n",
       "[2 rows x 29 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "11\n"
     ]
    }
   ],
   "source": [
    "display(alphafold_results_df.head(2))\n",
    "print(len(alphafold_results_df))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e1bbd9f2-f7f1-4ef3-b110-bc924c1b08df",
   "metadata": {},
   "source": [
    "## Run ProteinSolver"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "7444eab9-13dd-4609-9937-97e874204e5b",
   "metadata": {},
   "outputs": [],
   "source": [
    "def worker(mutation, data):\n",
    "    try:\n",
    "        results = ProteinSolver.analyze_mutation(f\"A_{mutation}\", data)\n",
    "    except Exception as error:\n",
    "        print(f\"{error!r}\")\n",
    "        return None\n",
    "    else:\n",
    "        results = {f\"proteinsolver_core_{key}\": value for key, value in results.items()}\n",
    "        return results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "585b17c1-e812-49e3-945b-c0b6d76f7579",
   "metadata": {},
   "outputs": [],
   "source": [
    "proteinsolver_results_file = NOTEBOOK_DIR.joinpath(\"proteinsolver-results.parquet\")\n",
    "\n",
    "if proteinsolver_results_file.is_file():\n",
    "    proteinsolver_results_df = pq.read_table(proteinsolver_results_file).to_pandas()\n",
    "else:\n",
    "    from elaspic2.plugins.proteinsolver import ProteinSolver\n",
    "\n",
    "    ProteinSolver.load_model(device=device)\n",
    "\n",
    "    data = ProteinSolver.build(structure_file, sequence, None, remove_hetatms=False)\n",
    "\n",
    "    result_list = []\n",
    "    for mut in tqdm(result_df[\"mut\"], leave=False):\n",
    "        result = worker(mut, data)\n",
    "        result_list.append({\"mut\": mut} | result)\n",
    "    proteinsolver_results_df = pd.DataFrame(result_list)\n",
    "\n",
    "    pq.write_table(\n",
    "        pa.Table.from_pandas(proteinsolver_results_df, preserve_index=False),\n",
    "        proteinsolver_results_file,\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "64bb9f9e-47c9-452f-b2cc-17b1a3e2d3f5",
   "metadata": {},
   "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>mut</th>\n",
       "      <th>proteinsolver_core_score_wt</th>\n",
       "      <th>proteinsolver_core_score_mut</th>\n",
       "      <th>proteinsolver_core_features_residue_wt</th>\n",
       "      <th>proteinsolver_core_features_protein_wt</th>\n",
       "      <th>proteinsolver_core_features_residue_mut</th>\n",
       "      <th>proteinsolver_core_features_protein_mut</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>E33Q</td>\n",
       "      <td>0.138828</td>\n",
       "      <td>0.072097</td>\n",
       "      <td>[2.932053804397583, 1.4533919095993042, -0.066...</td>\n",
       "      <td>[-1.43948233127594, 0.8285224437713623, 0.0743...</td>\n",
       "      <td>[-0.5347995162010193, 0.6960828304290771, -0.5...</td>\n",
       "      <td>[-1.4501885175704956, 0.8256354928016663, 0.07...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>E81K</td>\n",
       "      <td>0.175108</td>\n",
       "      <td>0.015804</td>\n",
       "      <td>[3.316300630569458, 1.5210775136947632, -0.088...</td>\n",
       "      <td>[-1.43948233127594, 0.8285225033760071, 0.0743...</td>\n",
       "      <td>[-1.2187674045562744, 2.1976358890533447, 0.00...</td>\n",
       "      <td>[-1.455361008644104, 0.8385410904884338, 0.074...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    mut  proteinsolver_core_score_wt  proteinsolver_core_score_mut  \\\n",
       "0  E33Q                     0.138828                      0.072097   \n",
       "1  E81K                     0.175108                      0.015804   \n",
       "\n",
       "              proteinsolver_core_features_residue_wt  \\\n",
       "0  [2.932053804397583, 1.4533919095993042, -0.066...   \n",
       "1  [3.316300630569458, 1.5210775136947632, -0.088...   \n",
       "\n",
       "              proteinsolver_core_features_protein_wt  \\\n",
       "0  [-1.43948233127594, 0.8285224437713623, 0.0743...   \n",
       "1  [-1.43948233127594, 0.8285225033760071, 0.0743...   \n",
       "\n",
       "             proteinsolver_core_features_residue_mut  \\\n",
       "0  [-0.5347995162010193, 0.6960828304290771, -0.5...   \n",
       "1  [-1.2187674045562744, 2.1976358890533447, 0.00...   \n",
       "\n",
       "             proteinsolver_core_features_protein_mut  \n",
       "0  [-1.4501885175704956, 0.8256354928016663, 0.07...  \n",
       "1  [-1.455361008644104, 0.8385410904884338, 0.074...  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "11\n"
     ]
    }
   ],
   "source": [
    "display(proteinsolver_results_df.head(2))\n",
    "print(len(proteinsolver_results_df))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "03d5b438-42fb-4a7e-9fa6-574a071b57c1",
   "metadata": {},
   "source": [
    "## Run ProtBert"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "a8e263f3-8f66-420b-b746-06c1a6b11c58",
   "metadata": {},
   "outputs": [],
   "source": [
    "def worker(mutation, data):\n",
    "    try:\n",
    "        results = ProtBert.analyze_mutation(f\"A_{mutation}\", data)\n",
    "    except Exception as error:\n",
    "        print(f\"{error!r}\")\n",
    "        return None\n",
    "    else:\n",
    "        results = {f\"protbert_core_{key}\": value for key, value in results.items()}\n",
    "        return results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "060e37b8-379a-404f-9464-4631a91a7e92",
   "metadata": {},
   "outputs": [],
   "source": [
    "protbert_results_file = NOTEBOOK_DIR.joinpath(\"protbert-results.parquet\")\n",
    "\n",
    "if protbert_results_file.is_file():\n",
    "    protbert_results_df = pq.read_table(protbert_results_file).to_pandas()\n",
    "else:\n",
    "    from elaspic2.plugins.protbert import ProtBert\n",
    "\n",
    "    ProtBert.load_model(device=device)\n",
    "    data = ProtBert.build(sequence, ligand_sequence=None)\n",
    "\n",
    "    result_list = []\n",
    "    for mut in tqdm(result_df[\"mut\"], leave=False):\n",
    "        result = worker(mut, data)\n",
    "        result_list.append({\"mut\": mut} | result)\n",
    "    protbert_results_df = pd.DataFrame(result_list)\n",
    "\n",
    "    pq.write_table(\n",
    "        pa.Table.from_pandas(protbert_results_df, preserve_index=False), protbert_results_file\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "065e2410-bf71-48c4-bff0-4c1a5f75099d",
   "metadata": {},
   "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>mut</th>\n",
       "      <th>protbert_core_score_wt</th>\n",
       "      <th>protbert_core_score_mut</th>\n",
       "      <th>protbert_core_features_residue_wt</th>\n",
       "      <th>protbert_core_features_protein_wt</th>\n",
       "      <th>protbert_core_features_residue_mut</th>\n",
       "      <th>protbert_core_features_protein_mut</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>E33Q</td>\n",
       "      <td>0.998497</td>\n",
       "      <td>0.000154</td>\n",
       "      <td>[-0.09428805112838745, -0.06660676002502441, 0...</td>\n",
       "      <td>[0.07330284267663956, -0.02223372645676136, -0...</td>\n",
       "      <td>[-0.08288051187992096, -0.06935720145702362, 0...</td>\n",
       "      <td>[0.07342074066400528, -0.020175088196992874, -...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>E81K</td>\n",
       "      <td>0.994313</td>\n",
       "      <td>0.000130</td>\n",
       "      <td>[0.09070268273353577, -0.028208790346980095, -...</td>\n",
       "      <td>[0.07330284267663956, -0.02223372645676136, -0...</td>\n",
       "      <td>[0.11418430507183075, -0.009324148297309875, -...</td>\n",
       "      <td>[0.07354287058115005, -0.021542491391301155, -...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    mut  protbert_core_score_wt  protbert_core_score_mut  \\\n",
       "0  E33Q                0.998497                 0.000154   \n",
       "1  E81K                0.994313                 0.000130   \n",
       "\n",
       "                   protbert_core_features_residue_wt  \\\n",
       "0  [-0.09428805112838745, -0.06660676002502441, 0...   \n",
       "1  [0.09070268273353577, -0.028208790346980095, -...   \n",
       "\n",
       "                   protbert_core_features_protein_wt  \\\n",
       "0  [0.07330284267663956, -0.02223372645676136, -0...   \n",
       "1  [0.07330284267663956, -0.02223372645676136, -0...   \n",
       "\n",
       "                  protbert_core_features_residue_mut  \\\n",
       "0  [-0.08288051187992096, -0.06935720145702362, 0...   \n",
       "1  [0.11418430507183075, -0.009324148297309875, -...   \n",
       "\n",
       "                  protbert_core_features_protein_mut  \n",
       "0  [0.07342074066400528, -0.020175088196992874, -...  \n",
       "1  [0.07354287058115005, -0.021542491391301155, -...  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "11\n"
     ]
    }
   ],
   "source": [
    "display(protbert_results_df.head(2))\n",
    "print(len(protbert_results_df))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3c0b4064-1f4d-4d9a-9dfe-e75d60af2bbf",
   "metadata": {},
   "source": [
    "## Run MSA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "bf8e6561-0b38-4c1c-ac38-79c9e4d0d64c",
   "metadata": {},
   "outputs": [],
   "source": [
    "msa_results_file = NOTEBOOK_DIR.joinpath(\"msa-results.parquet\")\n",
    "\n",
    "if msa_results_file.is_file():\n",
    "    msa_results_df = pq.read_table(msa_results_file).to_pandas()\n",
    "else:\n",
    "    msa_sequences = analyze_msa.alignment_to_sequences(alignment)\n",
    "    assert sequence == msa_sequences[0]\n",
    "    assert all([len(sequence) == len(seq) for seq in msa_sequences])\n",
    "\n",
    "    msa_counts_mat = analyze_msa.sequences_to_counts(msa_sequences)\n",
    "    msa_probas_mat = analyze_msa.counts_to_probas(msa_counts_mat)\n",
    "    msa_length = len(msa_sequences)\n",
    "    msa_proba_mean = np.mean(\n",
    "        [msa_probas_mat[i, analyze_msa.AMINO_ACIDS.index(aa)] for i, aa in enumerate(sequence)]\n",
    "    )\n",
    "    msa_conservation_df = analyze_msa.run_convervation_script(msa_sequences)\n",
    "\n",
    "    msa_results = []\n",
    "    for mutation in result_df[\"mut\"]:\n",
    "        msa_result = analyze_msa.featurize_mutation(\n",
    "            mutation,\n",
    "            sequence=sequence,\n",
    "            msa_counts_mat=msa_counts_mat,\n",
    "            msa_probas_mat=msa_probas_mat,\n",
    "            msa_conservation_df=msa_conservation_df,\n",
    "            msa_length=msa_length,\n",
    "            msa_proba_mean=msa_proba_mean,\n",
    "        )\n",
    "        msa_results.append(msa_result)\n",
    "    msa_results_df = pd.DataFrame(msa_results)\n",
    "    pq.write_table(pa.Table.from_pandas(msa_results_df, preserve_index=False), msa_results_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "dda25ee5-b48f-4081-bc98-ff0cbdeefc24",
   "metadata": {},
   "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>mut</th>\n",
       "      <th>msa_count_wt</th>\n",
       "      <th>msa_count_mut</th>\n",
       "      <th>msa_count_total</th>\n",
       "      <th>msa_proba_wt</th>\n",
       "      <th>msa_proba_mut</th>\n",
       "      <th>msa_proba_total</th>\n",
       "      <th>msa_length</th>\n",
       "      <th>msa_proba</th>\n",
       "      <th>msa_H</th>\n",
       "      <th>msa_KL</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>E33Q</td>\n",
       "      <td>3459.0</td>\n",
       "      <td>1070.0</td>\n",
       "      <td>11827.0</td>\n",
       "      <td>-1.230806</td>\n",
       "      <td>-2.403482</td>\n",
       "      <td>-80.568258</td>\n",
       "      <td>18675</td>\n",
       "      <td>-1.239175</td>\n",
       "      <td>2.191395</td>\n",
       "      <td>0.777022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>E81K</td>\n",
       "      <td>4750.0</td>\n",
       "      <td>875.0</td>\n",
       "      <td>13804.0</td>\n",
       "      <td>-1.068051</td>\n",
       "      <td>-2.758795</td>\n",
       "      <td>-93.442122</td>\n",
       "      <td>18675</td>\n",
       "      <td>-1.239175</td>\n",
       "      <td>1.782619</td>\n",
       "      <td>1.294940</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    mut  msa_count_wt  msa_count_mut  msa_count_total  msa_proba_wt  \\\n",
       "0  E33Q        3459.0         1070.0          11827.0     -1.230806   \n",
       "1  E81K        4750.0          875.0          13804.0     -1.068051   \n",
       "\n",
       "   msa_proba_mut  msa_proba_total  msa_length  msa_proba     msa_H    msa_KL  \n",
       "0      -2.403482       -80.568258       18675  -1.239175  2.191395  0.777022  \n",
       "1      -2.758795       -93.442122       18675  -1.239175  1.782619  1.294940  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "11\n"
     ]
    }
   ],
   "source": [
    "display(msa_results_df.head(2))\n",
    "print(len(msa_results_df))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "116596ce-a27a-43da-8497-3cda1bef8653",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Run `ELASPIC2`"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bb43e36d-46de-4989-ab9c-6d4216afeeda",
   "metadata": {
    "id": "4jTx3zGeHD2e"
   },
   "source": [
    "### Initialize the `ELASPIC2` model\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "39bf1842-3592-4c78-a2fe-c779fb0b9061",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/strokach/envs/default-v46/lib/python3.9/site-packages/sklearn/base.py:310: UserWarning: Trying to unpickle estimator PCA from version 0.23.2 when using version 0.24.2. This might lead to breaking code or invalid results. Use at your own risk.\n",
      "  warnings.warn(\n",
      "Some weights of the model checkpoint at /home/kimlab5/strokach/workspace/elaspic/elaspic2/src/elaspic2/plugins/protbert/data/prot_bert_bfd were not used when initializing BertModel: ['cls.predictions.transform.LayerNorm.weight', 'cls.predictions.bias', 'cls.predictions.transform.dense.bias', 'cls.seq_relationship.bias', 'cls.predictions.decoder.bias', 'cls.predictions.transform.dense.weight', 'cls.seq_relationship.weight', 'cls.predictions.decoder.weight', 'cls.predictions.transform.LayerNorm.bias']\n",
      "- This IS expected if you are initializing BertModel from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n",
      "- This IS NOT expected if you are initializing BertModel from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n"
     ]
    }
   ],
   "source": [
    "model = el2.ELASPIC2(device=device)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b7a8816c-def8-4c35-8071-681af15d82e4",
   "metadata": {
    "id": "i8OkFJbj8sYO"
   },
   "source": [
    "### Make predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "54b21cac-0df6-4d73-b525-5644e4aadf03",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "ac6mTuNYHD2i",
    "outputId": "b2427457-6fb6-4e2c-dac0-d037fafc8282"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "ELASPIC2Data(is_interface=False, protbert_data=ProtBertData(sequence='MAAAAAAGAGPEMVRGQVFDVGPRYTNLSYIGEGAYGMVCSAYDNVNKVRVAIKKISPFEHQTYCQRTLREIKILLRFRHENIIGINDIIRAPTIEQMKDVYIVQDLMETDLYKLLKTQHLSNDHICYFLYQILRGLKYIHSANVLHRDLKPSNLLLNTTCDLKICDFGLARVADPDHDHTGFLTEYVATRWYRAPEIMLNSKGYTKSIDIWSVGCILAEMLSNRPIFPGKHYLDQLNHILGILGSPSQEDLNCIINLKARNYLLSLPHKNKVPWNRLFPNADSKALDLLDKMLTFNPHKRIEVEQALAHPYLEQYYDPSDEPIAEAPFKFDMELDDLPKEKLKELIFEETARFQPGYRS'), proteinsolver_data=Data(edge_attr=[18572, 2], edge_index=[2, 18572], x=[360]))"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "protein_features = model.build(\n",
    "    structure_file=structure_file,\n",
    "    protein_sequence=sequence,\n",
    "    ligand_sequence=None,\n",
    "    remove_hetatms=True,\n",
    ")\n",
    "\n",
    "protein_features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "3e360633-ea88-43e3-9963-5c34380085de",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2b6658cc94324f33bebf9b61029b7b92",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/11 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "mutation_features = list(\n",
    "    tqdm(\n",
    "        (model.analyze_mutation(mut, protein_features) for mut in result_df[\"mut\"]),\n",
    "        total=len(result_df),\n",
    "    )\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "7c9c2807-c269-4b54-af24-d531321dc2d6",
   "metadata": {},
   "outputs": [],
   "source": [
    "mutation_features = [\n",
    "    {k: v for k, v in features.items() if \"proteinsolver_core_features\" not in k}\n",
    "    for features in mutation_features\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "5af30b7a-4101-48f6-aace-5ee124a889b3",
   "metadata": {
    "id": "XJPiK7eVRBqi"
   },
   "outputs": [],
   "source": [
    "# In all cases, higher scores means less stable (same as ΔΔG)\n",
    "result_df[\"protbert_score\"] = [\n",
    "    f[\"protbert_core_score_wt\"] - f[\"protbert_core_score_mut\"] for f in mutation_features\n",
    "]\n",
    "result_df[\"proteinsolver_score\"] = [\n",
    "    f[\"proteinsolver_core_score_wt\"] - f[\"proteinsolver_core_score_mut\"] for f in mutation_features\n",
    "]\n",
    "result_df[\"el2_score\"] = model.predict_mutation_effect(mutation_features).tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "c1714a6e-2152-4c26-92db-db009acb59a5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# In all cases, higher scores means less stable (same as ΔΔG)\n",
    "result_df[\"protbert_ratio\"] = [\n",
    "    f[\"protbert_core_score_mut\"] / f[\"protbert_core_score_wt\"] for f in mutation_features\n",
    "]\n",
    "result_df[\"proteinsolver_ratio\"] = [\n",
    "    f[\"proteinsolver_core_score_mut\"] / f[\"proteinsolver_core_score_wt\"] for f in mutation_features\n",
    "]\n",
    "result_df[\"el2_ratio\"] = (0.5 / model.predict_mutation_effect(mutation_features)).tolist()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "52815912-8fd8-4218-b038-50b4b624d539",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Run ELASPIC2 + AlphaFold"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "d0356e9d-3ebd-498c-9a01-6f1986cf51d1",
   "metadata": {},
   "outputs": [],
   "source": [
    "prediction_df = (\n",
    "    proteinsolver_results_df\n",
    "    #\n",
    "    .merge(protbert_results_df, on=[\"mut\"])\n",
    "    .merge(rosetta_results_df, on=[\"mut\"])\n",
    "    .merge(msa_results_df, on=[\"mut\"])\n",
    "    .merge(alphafold_results_df, on=[\"mut\"])\n",
    "    .merge(result_df[[\"mut\", \"el2_score\"]], on=[\"mut\"])\n",
    ")\n",
    "\n",
    "prediction_df[\"aa_wt_onehot\"] = (\n",
    "    prediction_df[\"mut\"].str[0].map(analyze_msa.RESIDUE_ENCODING_WT).apply(list)\n",
    ")\n",
    "prediction_df[\"aa_mut_onehot\"] = (\n",
    "    prediction_df[\"mut\"].str[-1].map(analyze_msa.RESIDUE_ENCODING_MUT).apply(list)\n",
    ")\n",
    "\n",
    "for column in list(prediction_df):\n",
    "    if not column.endswith(\"_mut\"):\n",
    "        continue\n",
    "\n",
    "    column_wt = column.removesuffix(\"_mut\") + \"_wt\"\n",
    "    if column_wt not in prediction_df:\n",
    "        print(column_wt)\n",
    "        continue\n",
    "\n",
    "    column_change = column.removesuffix(\"_mut\") + \"_change\"\n",
    "    prediction_df[column_change] = prediction_df[column_wt] - prediction_df[column]\n",
    "    del prediction_df[column]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "6d39fda1-a17f-4f5d-866c-d2a1dcb896a2",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "model_infos = [\n",
    "    (\"7f9826be\", \"optimized\"),\n",
    "    (\"900500fe\", \"optimized\"),\n",
    "]\n",
    "\n",
    "for unique_id, model_type in model_infos:\n",
    "    scalar_features, vector_features = helpers.load_features(NOTEBOOK_DIR.parent, unique_id)\n",
    "    features_to_exclude = helpers.load_features_to_exclude(\n",
    "        NOTEBOOK_DIR.parent, unique_id, model_type\n",
    "    )\n",
    "    best_parameters = helpers.load_best_parameters(NOTEBOOK_DIR.parent, unique_id, model_type)\n",
    "    models = helpers.load_best_models(NOTEBOOK_DIR.parent, unique_id, model_type)\n",
    "\n",
    "    mask = ~prediction_df[scalar_features + vector_features].isnull().any(axis=1)\n",
    "\n",
    "    X_ref = np.c_[\n",
    "        prediction_df.loc[mask, scalar_features].values,\n",
    "        np.hstack([np.vstack(prediction_df.loc[mask, col].values) for col in vector_features]),\n",
    "    ]\n",
    "\n",
    "    if features_to_exclude is not None:\n",
    "        feature_mask = np.ones(X_ref.shape[1], dtype=bool)\n",
    "        feature_mask[np.array(features_to_exclude, dtype=int)] = False\n",
    "        X = X_ref[:, feature_mask]\n",
    "        assert len(features_to_exclude) == X_ref.shape[1] - X.shape[1]\n",
    "    else:\n",
    "        X = X_ref\n",
    "\n",
    "    for model_idx, model in enumerate(models):\n",
    "        prediction_df.loc[mask, f\"pred_{unique_id}_{model_type}_{model_idx}\"] = model.predict(X)\n",
    "\n",
    "    pred_column = f\"pred_{unique_id}_{model_type}\"\n",
    "    prediction_df[pred_column] = prediction_df.loc[\n",
    "        mask, [f\"pred_{unique_id}_{model_type}_{model_idx}\" for model_idx in range(len(models))]\n",
    "    ].mean(axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "81c29bc9-e114-4bcb-af22-9587a969b48f",
   "metadata": {},
   "source": [
    "## Combine"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "74757401-a202-4435-ac7b-858f6e65ea65",
   "metadata": {},
   "outputs": [],
   "source": [
    "rosetta_results_df[\"rosetta_dg_mut\"] = (\n",
    "    rosetta_results_df[\"rosetta_dg_change\"] + rosetta_results_df[\"rosetta_dg_wt\"]\n",
    ")\n",
    "assert np.allclose(\n",
    "    rosetta_results_df[\"rosetta_dg_change\"],\n",
    "    rosetta_results_df[\"rosetta_dg_mut\"] - rosetta_results_df[\"rosetta_dg_wt\"],\n",
    ")\n",
    "\n",
    "rosetta_results_df[\"rosetta_dg_score\"] = rosetta_results_df[\"rosetta_dg_change\"]\n",
    "rosetta_results_df[\"rosetta_dg_ratio\"] = (\n",
    "    rosetta_results_df[\"rosetta_dg_wt\"] / rosetta_results_df[\"rosetta_dg_mut\"]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "ff9fc37a-43b9-427d-b559-d3db6f40324f",
   "metadata": {},
   "outputs": [],
   "source": [
    "for column in [\"pred_7f9826be_optimized\", \"pred_900500fe_optimized\"]:\n",
    "    prediction_df[f\"{column}_score\"] = prediction_df[column]\n",
    "    prediction_df[f\"{column}_ratio\"] = 1 - prediction_df[column]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "b23fea27-6c4e-4394-ac27-628695b814d4",
   "metadata": {},
   "outputs": [],
   "source": [
    "try:\n",
    "    del result_df[\"rosetta_dg_score\"], result_df[\"rosetta_dg_ratio\"]\n",
    "except KeyError:\n",
    "    pass\n",
    "\n",
    "final_result_df = (\n",
    "    result_df\n",
    "    #\n",
    "    .merge(\n",
    "        rosetta_results_df[[\"mut\", \"rosetta_dg_score\", \"rosetta_dg_ratio\"]], on=\"mut\", how=\"left\"\n",
    "    ).merge(\n",
    "        prediction_df[\n",
    "            [\"mut\"]\n",
    "            + [\n",
    "                c\n",
    "                for c in prediction_df\n",
    "                if c.startswith(\"pred_\") and (c.endswith(\"_score\") or c.endswith(\"_ratio\"))\n",
    "            ]\n",
    "        ]\n",
    "    )\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b4ce2d7f-12f9-45f6-9625-e2e30406bd3e",
   "metadata": {
    "tags": []
   },
   "source": [
    "### Visualize output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "d8365929-af43-4d33-9dc7-2518aa77ce95",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABDoAAAQkCAYAAACc8X5SAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAAAsTAAALEwEAmpwYAAD+F0lEQVR4nOzde3wU9b3/8ffuJoQYWHAJSTYERVCU2lBF1HJsFQUEFRuxohVsqVasIOIPq5Wi5eIdtOWoiFRaEYRiS62lBEQRtF4O2lqwotZLbdQCu4QQcSGEkOzu7w9O9hBzm032Mt/J6/l4+DAMs7ufmXnPdzYf5uKKRqNRAQAAAAAAOIA73QUAAAAAAAAkCo0OAAAAAADgGDQ6AAAAAACAY9DoAAAAAAAAjuGYRkddXZ22b9+uurq6dJcCtIiswgTkFKYgqzAFWYUJyCmcwjGNjmAwqGHDhikYDKa7FKBFZBUmIKcwBVmFKcgqTEBO4RSOaXQAAAAAAADQ6AAAAAAAAI5BowMAAAAAADgGjQ4AAAAAAOAYGekuAAASKRKJKlBRpcpQtXzebPlzc+R2u9JdFuBo7HctY/0AiAdjBkxh56zS6ADgGJFIVJu3BTR/5RbV1IaVlenRtCsHaUix3zaDLuA07HctY/0AiAdjBkxh96xy6QoAxwhUVMUGW0mqqQ1r/sotClRUpbkywLnY71rG+gEQD8YMmMLuWaXRAcAxKkPVscG2Xk1tWJX7qtNUEeB87HctY/0AiAdjBkxh96zS6ADgGD5vtrIyPQ2mZWV65OuanaaKAOdjv2sZ6wdAPBgzYAq7Z5VGBwDH8OfmaNqVg2KDbv21gv7cnDRXBjgX+13LWD8A4sGYAVPYPavcjBSAY7jdLg0p9quPf6gq91XL19Ved38GnIj9rmWsHwDxYMyAKeyeVRodABzF7XapV14X9crrku5SgA6D/a5lrB8A8WDMgCnsnFUuXQEAAAAAAI5BowMAAAAAADgGjQ4AAAAAAOAYNDoAAAAAAIBjcDNSoAOLRKIKVFSpMlQtn9ded0oGYA7GEudhmwLpw/4HU9g5qzQ6gA4qEolq87aA5q/copracOzZ10OK/bYZoADYH2OJ87BNgfRh/4Mp7J5VLl0BOqhARVVsYJKkmtqw5q/cokBFVZorA2ASxhLnYZsC6cP+B1PYPas0OoAOqjJUHRuY6tXUhlW5rzpNFQEwEWOJ87BNgfRh/4Mp7J5VGh1AB+XzZisr09NgWlamR76u2WmqCICJGEuch20KpA/7H0xh96zS6AA6KH9ujqZdOSg2QNVfV+fPzUlzZQBMwljiPGxTIH3Y/2AKu2eVm5ECHZTb7dKQYr/6+Ieqcl+1fF3tdadkAGZgLHEetimQPux/MIXds0qjA+jA3G6XeuV1Ua+8LukuBYDBGEuch20KpA/7H0xh56xy6QoAAAAAAHAMGh0AAAAAAMAxaHQAAAAAAADHoNEBAAAAAAAcg0YHAAAAAABwDBodAAAAAADAMWh0AAAAAAAAx6DRAQAAAAAAHINGBwAAAAAAcAwaHQAAAAAAwDFodAAAAAAAAMeg0QEAAAAAAByDRgcAAAAAAHCMjFR8yBdffKGf/vSn+vzzz9WpUycde+yxuvPOO+Xz+RrMFw6Hdffdd+vVV1+Vy+XSddddp7Fjx6aiRAAAAAAA4AApOaPD5XLp2muv1fPPP681a9aod+/eevDBBxvNt2bNGn3++ed64YUX9Lvf/U6PPPKItm/fnooSAQAAAACAA6TkjI7u3bvrzDPPjP35lFNO0cqVKxvNt27dOo0dO1Zut1s+n0/Dhw/X+vXrde211zaYLxQKKRQKNZgWDAaTUzzQDmQVJiCnMAVZhSnIKkxATuFkKWl0HCkSiWjlypU677zzGv1dIBBQYWFh7M9+v7/JnW3p0qVasGBBUusEEoGswgTkFKYgqzAFWYUJyCmcLOWNjrvuuktHHXWUrrrqqja/x4QJEzRmzJgG04LBoMaPH9/e8oCEIqswATmFKcgqTEFWYQJyCidLaaNj7ty5+uyzz7Ro0SK53Y1vD+L3+7Vz504NHDhQUuMzPOp5vV55vd6k1wu0F1mFCcgpTEFWYQqyChOQUzhZyh4vO3/+fL377rt69NFH1alTpybnGTVqlFatWqVIJKLKykq9+OKLGjlyZKpKBAAAAAAAhktJo+Pjjz/WokWLVF5eru9973sqKSnRDTfcIEmaOHGitm3bJkkqKSlRUVGRzj//fF1++eW64YYb1Lt371SUCAAAAAAAHCAll66ccMIJ+vDDD5v8u8WLF8d+9ng8mjNnTipKAgAAAAAADpSyS1cAAAAAAACSjUYHAAAAAABwDBodAAAAAADAMWh0AAAAAAAAx6DRAQAAAAAAHINGBwAAAAAAcAwaHQAAAAAAwDFodAAAAAAAAMeIq9Fx6NAhzZ8/X8OGDdNpp50mSXrttde0fPnypBQHAAAAAAAQj7gaHffee68++ugjPfjgg3K5XJKkE044QStXrkxKcQAAAAAAAPHIiGfmF198US+88IKOOuooud2HeyT5+fnatWtXUooDAAAAAACIR1xndGRmZiocDjeYVllZqe7duyeyJgAAAAAAgDaJq9ExatQo3XbbbfrPf/4jSSovL9edd96piy66KCnFAQAAAAAAxCOuRse0adPUq1cvfec731EoFNLIkSOVl5enG264IVn1AQAAAAAAWGb5Hh3hcFiPPfaYbr31Vt1+++2qrKzU0UcfHbspKQAAAAAAQLpZPqPD4/Hot7/9rTIyDvdGfD4fTQ4AAAAAAGArcV26cskll/AoWQAAAAAAYFtxPV72nXfe0fLly/Wb3/xGBQUFDc7oWLFiRcKLAwAAAAAAiEdcjY7LL79cl19+ebJqAQAAAAAAaJe4Gh1jxoxJVh0AAAAAAADtFlejQ5KeeeYZrV69Wrt27VJ+fr5KSkr03e9+Nxm1AQAAAAAAxCWuRsdjjz2mP/3pT7rmmmtUWFionTt36te//rXKy8s1adKkZNUIAAkViUQVqKhSZahaPm+2/Lk5crt5ihTQFuxP9sb2AczDfgtT2DmrcTU6Vq1apaeeekq9evWKTfvWt76lq666ikYHkER2HkRME4lEtXlbQPNXblFNbVhZmR5Nu3KQhhT7WadAnCKRqN765y59/J+9ikSjcrtcOqF3dw0ekM/+dIR0jeGMd4B5GFdhCrtnNa5GR3V1tXw+X4Np3bt318GDBxNaFID/wxfVxApUVMXWpSTV1IY1f+UW9fEPVa+8LmmuDjBLcE+VPg+G9OzL/4qNT98b0V9FeV1U2JP9SUrvGM54B5iHcRWmsHtW3fHM/O1vf1u33HKL/v3vf+vgwYP65JNPNH36dH3rW99KVn1Ah9fcF9VARVVCPycSiWpH+X5t+9du7Sjfr0gkmtD3t4vKUHVsXdarqQ2rcl91mioCzFUZOqinN3zUYHx6esNHqgzxDyD1UjWGN6W58W5PiPEOsCvGVZjC7lmN64yOmTNn6s4771RJSYnq6uqUkZGhCy64QHfccUey6gM6vJZ+MU/Uv8h1pLNGfN5sZWV6GqzTrEyPfF2z01gVYKbqmromx6fqmro0VWQ/qRjDm9PceBcORxWJRB03vgNOwLgKU9g9q3Gd0dGlSxfNmzdP//jHP/Taa6/pH//4h+bNmyev15us+oAOr/6L6pES/Yt5Ov/FMdX8uTmaduWg2Dqtb+r4c3PSXBlgnu5ds5ocn7p3yUpTRfaTijG8Of7cHN14+SkNxrsrRvTXY8/8w5HjO+AEjKswhd2zGlej409/+pM++OADud1u9ejRQ263Wx988IH+9Kc/Jak8AKn4xbwjXc7hdrs0pNivh24eqnsn/5ceunmoI89cAVKhti6scSNPbDA+jRt5omoj4VZe2XGks7nqdrvU8+jOKjm7ny4f3l8lZ/fTuv8pU2DPAUeO74ATMK7CFHbPalyXrjz00EONmhoFBQWaNGmSLrnkkgSWBaBe/S/mffxDVbmvWr6uib9jf0e7nMPtdqlXXhduxge0U7eczlq/+VOVnN1PckmKSus3f6ozT/anuzTbSMUY3pJuOZ21+pVPOsz4DpiOcRWmsHtW42p07N+/X126NPzFoGvXrgqFQgktCkBDVn4xb8/jC+v/xfGr9+jgcg4ALfHn5mjCRSc3GDtuvPwU5fuOSndptnosdzqbq4zvgFmaGlfZZ2FHds9qXI2Ofv366fnnn9eFF14Ym7Zhwwb169cv4YUBsK69NxNN9784AjCT2+3SmScX6Parz9D7ZXsUiUgr1v9TGR53Wi8J60g3WG4N4ztgnswMt8YMPV6RaFRul0uZGXHdbQBIGTtnNa5Gxy233KLrrrtOzz33nHr37q3PP/9cmzdv1uOPP56s+gBY0NzNRPv4h1r+F0Qu5wDQFrsqD+ieJX9tcGlEvONPoiViTHQSxnfAHIGKKs176q1Gl5s9dHPHHL9gX3bPalwtl8GDB6u0tFTFxcWqrq7WwIEDVVpaqtNOOy1Z9QGwoCPdTBSAvdhx/LFjTQBgBeMXTGH3rMZ1RockFRYW6rrrrpMkHTx4UG63fU5PATqqjnYzUQD2Ycfxx441AYAVjF8whd2zGleXYu7cuXrnnXckSS+//LLOOOMMnX766dq0aVNSigNgTTofXwigY7Pj+GPHmgDACsYvmMLuWY3rjI41a9Zo6tSpkqRHH31UDzzwgLp27ar77rtP5513XlIKBNA6bjYHIF3sOP7YsSYAsILxC6awe1bjanRUV1crOztbX3zxhf7zn/9o5MiRkqQdO3a0+Lq5c+fq+eef144dO7RmzRr179+/0TyPPPKIfvvb3yovL0+SNGjQIM2aNSue8mBRbV0kKXfETdb7whpuNgcgXew4/tixJgCwgvELprBzVuNqdPTp00d//vOf9fnnn+uss86SJFVWVqpz584tvm7YsGH6wQ9+oPHjx7c43yWXXKLbbrstnpLQBpkZbs1Y+HrC3/feyWcl/D0BAAAAAIhHXI2OWbNm6d5771VmZqbuueceSdJrr70Wa3o0Z/DgwW2vsAmhUEihUKjBtGAwmNDPABKBrMIE5BSmIKswBVmFCcgpnCyuRsfAgQP19NNPN5j2ne98R9/5zndif3788cdjT2WJ19q1a/Xaa6+pZ8+euvHGG3Xqqac2Od/SpUu1YMGCNn0GkEpkFSYgpzAFWYUpyCpMQE7hZK5oNBpN5BsOGjRIW7ZsafLvzjvvPC1atKjJe3Ts3r1b3bt3V2Zmpl5//XXdcsstWrdunY4++uhG8zbXfRw/frw2btyooqKixCyMg3HpSmqQVZiAnMIUZBWmIKswATmFk8V1RocVbe2b9OzZM/bzWWedJb/fr48//lhnnHFGo3m9Xq+8Xm+bawRShazCBOQUpiCrMAVZhQnIKZws4Y/IcLna9jiZXbt2xX7+5z//qR07dui4445LVFkAAAAAAKADSPgZHU25++679cILL6iiokJXX321unfvrrVr12rixImaOnWqiouL9ctf/lLvvfee3G63MjMzNW/evAZneQAAAAAAALQmJZeu3HHHHbrjjjsaTV+8eHHs57lz5ya6FAAAAAAA0MEkvNGR6EfJAh1NJBJVoKJKlaFq+bzZ8ufmyO1u2yVhAJAKjFswBVmFCcgpTGHnrMbV6Pjyyy/VrVu3RtODwaAKCgokNTxLA0B8IpGoNm8LaP7KLaqpDSsr06NpVw7SkGK/bQYNJI+dDxZAcxi3zNaRxh2yChPU1UX0+js79cjv3yansDW7Z9XSzUjLysp0wQUX6Mwzz9S3v/1trVu3rsHfX3jhhUkpDuhoAhVVsS9gklRTG9b8lVsUqKhKc2VItvov4Df98mXNeOx/dNMvX9bmbQFFIgl9AjiQcDt2729y3Nqxe3+aK0NrOtq4wzEWdheJRLXtk4rYL44SOYU9mZBVS42Oe+65R6NGjdKbb76pWbNm6f7779fjjz8e+/u2PlIWQEOVoerYYFGvpjasyn3VaaoIqcIXcJhq156qJsetXXvIrt11tHGHYyzsLlBRpffL9pBT2J4JWbXU6Ni2bZumTJmibt26afjw4fr973+v1atXa/78+cmuD+hQfN5sZWV6GkzLyvTI1zU7TRUhVfgCDlN1zspoctzqnJWSB7uhHTrauMMxFnZXGapWJCpyCtszIauWGh1ut1tVVf/X3S8oKNBTTz2lV155RXfffXfSigM6Gn9ujqZdOSg2aNRf6+bPzUlzZUg2voDDVD5vZ31vRP8G49b3RvSXz9s5zZWhNR1t3OEYC7vzebP16tbtuuIrY+qNl59CTmErJmTV0j+3nHrqqdqwYYO++93vxqb5fD4tXbpU11xzjQ4ePJi0AoGOxO12aUixX338Q1W5r1q+rs6+MRz+T/0X8K/eJM8uBwugOQU9cnRMgVdjhh6vSDQqt8ulYwq8KuhBdu2uo407HGNhd/7cHE246GQtXfueSs7uJ7db+tpxPVTcL5ecwlZMyKqlRsdPf/pThUKhRtO9Xq+efPJJvfjiiwkvDOio3G6XeuV1Ua+8LukuBSnEF3CYyu12afCAfPXq2YXsGqYjjjscY2Fn/7dPejvMPgkzmZBVS42OPn36NPt3Xbp00SWXXJKgcgCg4+ILOExFds3FtgPshX0SprB7VlttdNx6661yuVrvzMybNy8hBQEAAAAAALRVq42OY489NhV1AAAAAAAAtFurjY4pU6Y0+PPrr7+u0tJSVVZW6le/+pW2bdum/fv3J61AoCOIRKIKVFSpMlQtn9d+17gBQHMYv9qH9QfgSIwJMImd8xrXQ+6feuopLVu2TGPHjtULL7wgSercubPuueceDRkyJCkFAk4XiUS1eVug0V3vhxT7bTNQAEBTGL/ah/UH4EiMCTCJ3fPqjmfmpUuXasmSJbruuuvkdh9+ad++fVVWVpaU4oCOIFBRFRsgJKmmNqz5K7coUFGV5soAoGWMX+3D+gNwJMYEmMTueY2r0VFVVSW/3y9JsRuU1tXVKTMzM/GVAR1EZag6NkDUq6kNq3JfdZoqAgBrGL/ah/UH4EiMCTCJ3fMaV6Pj9NNP1+OPP95g2rJly3TmmWcmtCigI/F5s5WV6WkwLSvTI1/X7DRVBADWMH61D+sPwJEYE2ASu+c1rkbHHXfcoQ0bNui8885TVVWVRo4cqfXr12v69OnJqg9wPH9ujqZdOSg2UNRf3+bPzUlzZQDQMsav9mH9ATgSYwJMYve8xnUz0ry8PD3zzDPatm2bduzYIb/fr4EDB8bu1wEgfm63S0OK/erjH6rKfdXydbXXHYsBoDmMX+3D+gNwJMYEmMTueY2r0SEdvjfHwIEDNXDgwGTUA3RIbrdLvfK6qFdel3SXAgBxYfxqH9YfgCMxJsAkds4rp2IAAAAAAADHoNEBAAAAAAAcg0YHAAAAAABwDBodAAAAAADAMWh0AAAAAAAAx6DRAQAAAAAAHINGBwAAAAAAcAwaHQAAAAAAwDFodAAAAAAAAMeg0QEAAAAAAByjwzQ6ausivK+hkrkuWM8AAAAA4CwZ6S4gVTIz3Jqx8PWEv++9k89K+HtKyatXSl7NycK6AAAAAABY1WHO6AAAAAAAAM5HowMAAAAAADgGjQ4AAAAAAOAYNDoAAAAAAIBj0OgAAAAAAACOkZJGx9y5c3XeeefpxBNP1EcffdTkPOFwWHPmzNHw4cM1YsQIrVq1KhWlAQAAAAAAB0lJo2PYsGFasWKFevXq1ew8a9as0eeff64XXnhBv/vd7/TII49o+/btqSgPAAAAAAA4REYqPmTw4MGtzrNu3TqNHTtWbrdbPp9Pw4cP1/r163Xttdc2mjcUCikUCjWYFgwGE1YvkChkFSYgpzAFWYUpyCpMQE7hZClpdFgRCARUWFgY+7Pf7292R1u6dKkWLFiQqtKANiOrMAE5hSnIKkxBVmECcgons02jIx4TJkzQmDFjGkwLBoMaP358mioCmkZWYQJyClOQVZiCrMIE5BROZptGh9/v186dOzVw4EBJjc/wOJLX65XX601leUCbkFWYgJzCFGQVpiCrMAE5hZPZ5vGyo0aN0qpVqxSJRFRZWakXX3xRI0eOTHdZAAAAAADAIClpdNx99906++yzFQwGdfXVV+uiiy6SJE2cOFHbtm2TJJWUlKioqEjnn3++Lr/8ct1www3q3bt3KsoDAAAAAAAOkZJLV+644w7dcccdjaYvXrw49rPH49GcOXNSUQ4AAAAAAHAo21y6AgAAAAAA0F40OgAAAAAAgGPQ6AAAAAAAAI5Bo6Odausi6S4BAAAAAAD8r5TcjNTJMjPcmrHw9YS/772Tz0r4ewJIvUgkqkBFlSpD1fJ5s+XPzZHb7Up3WW3mtOUBUoH9Bui42P/hZHbON40OAEiSSCSqzdsCmr9yi2pqw8rK9GjalYM0pNhvm4NAPJy2PEAqsN8AHRf7P5zM7vnm0hUASJJARVVs8Jekmtqw5q/cokBFVZoraxunLQ+QCuw3QMfF/g8ns3u+aXQAQJJUhqpjg3+9mtqwKvdVp6mi9nHa8gCpwH4DdFzs/3Ayu+ebRgcAJInPm62sTE+DaVmZHvm6ZqepovZx2vIAqcB+A3Rc7P9wMrvnm0YHACSJPzdH064cFDsI1F+76M/NSXNlbeO05QFSgf0G6LjY/+Fkds83NyMFgCRxu10aUuxXH/9QVe6rlq+rve5GHS+nLQ+QCuw3QMfF/g8ns3u+aXQAQBK53S71yuuiXnld0l1KQjhteYBUYL8BOi72fziZnfNNowMAbMrOzyaH/ZEfNIds2AfbAqYgqzANjQ4AsCG7P5sc9kZ+0ByyYR9sC5iCrMJE3IwUAGzI7s8mh72RHzSHbNgH2wKmIKswEY0OALAhuz+bHPZGftAcsmEfbAuYgqzCRDQ6AMCG7P5sctgb+UFzyIZ9sC1gCrIKE9HoAAAbsvuzyWFv5AfNIRv2wbaAKcgqTMTNSAHAhuz+bHLYG/lBc8iGfbAtYAqyChPR6AAAm7Lzs8lhf+QHzSEb9sG2gCnIKkzDpSsAAAAAAMAxHHNGRzh8+E7AwWCwyb8vKipSTc3BpHy2ae+bzPfevn17wt8zmdtOar7mgoICZWQkfhdpLatAvJKRVXKKZCCrMAHHf5iCMRWmSNa42hJXNBqNpvQTk+Stt97S+PHj010GHGTjxo0qKipK+PuSVSRaMrJKTpEMZBUm4PgPUzCmwhTJGldb4phGx8GDB/Xuu++qZ8+e8ng8rb+gjYLBoMaPH68VK1aooKAgaZ/TXibUafcak9V5jCerdl9HTaHm1Diy5lNOOSXhWU3VmNoaE7dNa5y4TJK15UrGuGqXrDbHqdu7Pey+Tuxw/G+O3dedRI2JwphqDyZkJd3SldXWOObSlc6dO2vw4MEp+7yCgoKUd6XawoQ6TagxkdqSVRPXETWnRrIOHKkeU1tj4rZpjROXSUr9ctktq81x6vZuj462ThKZVRPWHTUmBmOqPZiQlXSz2zriZqQAAAAAAMAxaHQAAAAAAADHoNEBAAAAAAAcg0ZHnLxer6ZMmSKv15vuUlpkQp0m1JhuJq4jak4NE2tuCycupxOXSXLucrUX66Ux1knbmbDuqDExTKixI2A7tM6u68gxT10BAAAAAADgjA4AAAAAAOAYNDoAAAAAAIBj0OgAAAAAAACO4ZhGR11dnbZv3666urp0lwK0iKzCBOQUpiCrMAVZhQnIKZzCMY2OYDCoYcOGKRgMprsUoEVkFSYgpzAFWYUpyCpMQE7hFI5pdAAAAAAAANDoAAAAAAAAjkGjAwAAAAAAOEZGugtItkgkqkBFlSpD1fJ5s+XPzZHb7Up3WQAAJBXHPzSHbNgH2wKmIKswjaMbHZFIVJu3BTR/5RbV1IaVlenRtCsHaUixnx0TAOBYHP/QHLJhH2wLmIKswkSOvnQlUFEV2yElqaY2rPkrtyhQUZXmygAASB6Of2gO2bAPtgVMQVZhIkc3OipD1bEdsl5NbViV+6rTVBEAAMnH8Q/NIRv2wbaAKcgqTOToRofPm62sTE+DaVmZHvm6ZqepIgAAko/jH5pDNuyDbQFTkFWYyNGNDn9ujqZdOSi2Y9ZfT+bPzUlzZW0XiUS1o3y/tv1rt3aU71ckEk13SQAAm0n18Y9jkzmaysaNl5+iL6sOsu1SzInfU9F+dhxPySqaYsesHsnRNyN1u10aUuxXH/9QVe6rlq+r2XcI5kZAAAArUnn849hkliOzsSdUrXA4qsee+YcCew6w7VLMad9T0X52HU/JKr7Krlk9kqPP6JAO75i98rqouF9P9crrYpsV3xbcCAgAYFWqjn8cm8xTn40e3mzds+SvCuw5IIltlw5O+p6K9rPzeEpWcSQ7Z7We4xsdTsKNgAAAdsOxyVxsO8Be2CdhChOySqPDINwICABgNxybzMW2A+yFfRKmMCGrNDoMwo2AAAB2w7HJXGw7wF7YJ2EKE7Lq6JuROg03AgIA2A3HJnOx7QB7YZ+EKUzIKo0Ow9TfCKhXXpd0lwIAgCSOTSZj2wH2wj4JU9g9q1y6AgAAAAAAHINGBwAAAAAAcAwaHQAAAAAAwDFodAAAAAAAAMfo0DcjjUSiClRUqTJULZ/XfneKBQC7YdwE4sd+A3Rc7P9wMjvnu8M2OiKRqDZvC2j+yi2qqQ3Hnv07pNhvm40DAHbCuAnEj/0G6LjY/+Fkds93h710JVBRFdsoklRTG9b8lVsUqKhKc2UAYE+Mm0D82G+Ajov9H05m93xbbnREo1H9/ve/1w9+8ANdfPHFkqS//e1vWrduXdKKS6bKUHVso9SrqQ2rcl91mioCAHtj3ATix34DdFzs/3Ayu+fbcqPjoYce0h/+8AddccUVCgQCkqSCggL9+te/TlpxyeTzZisr09NgWlamR76u2WmqCADsjXETiB/7DdBxsf/Dyeyeb8uNjmeffVaLFi3SRRddJJfr8DU3RUVF+s9//pO04pLJn5ujaVcOim2c+muK/Lk5aa4MAOyJcROIH/sN0HGx/8PJ7J5vyzcjDYfDysk5XHR9o6OqqkpHHXWUpdeXlZVp+vTp2rt3r7p37665c+eqT58+jeZbt26dHnvsMUWjUblcLi1ZskS5ublWy7TM7XZpSLFfffxDVbmvWr6u9rpLLADYDeMmED/2G6DjYv+Hk9k935YbHWeffbbuu+8+zZgxQ9Lhe3Y89NBDOvfccy29ftasWRo3bpxKSkq0evVqzZw5U8uWLWswz7Zt27RgwQItXbpUPXv21L59+9SpU6c4Fic+brdLvfK6qFdel6R9BgA4CeMmED/2G6DjYv+Hk9k535YbHTNmzNBtt92m0047TXV1dTr11FN11llnae7cua2+ds+ePXr//fe1ZMkSSdLo0aN11113qbKyUj6fLzbfk08+qWuuuUY9e/aUJHXt2rXJ9wuFQgqFQg2mBYNBq4sCpAxZhQnIKUxBVmEKsgoTkFM4maVGRzgc1vr16/XLX/5S+/fv144dO+T3+2MNidYEAgHl5+fL4zl8/Y7H41FeXp4CgUCDRscnn3yioqIijR8/XgcOHNCIESM0adKk2KUy9ZYuXaoFCxZYXUYgbcgqTEBOYQqyClOQVZiAnMLJLDU6PB6P7r//fl122WXKyspSjx49klJMOBzWhx9+qCVLlujQoUO69tprVVhYqEsuuaTBfBMmTNCYMWMaTAsGgxo/fnxS6gLaiqzCBOQUpiCrMAVZhQnIKZzM8qUr5557rjZt2qTzzjsv7g/x+/3atWuXwuGwPB6PwuGwysvL5ff7G8xXWFioUaNGqVOnTurUqZOGDRumd955p1Gjw+v1yuv1xl0HkGpkFSYgpzAFWYUpyCpMQE7hZJYbHTU1NZo6dapOPfVUFRQUNLicZN68eS2+tkePHhowYIBKS0tVUlKi0tJSDRgwoMFlK9Lhe3f85S9/UUlJierq6vTGG29o5MiRcS4SAAAAAADoqCw3Ovr376/+/fu3+YNmz56t6dOna+HChfJ6vbGbmE6cOFFTp05VcXGxLrroIr377ru68MIL5Xa79a1vfUuXXXZZmz8TAAAAAAB0LJYbHVOmTGnXB/Xr10+rVq1qNH3x4sWxn91ut372s5/pZz/7Wbs+CwAAAAAAdEyWGx2S9MYbb2j16tUqLy9XXl6evvOd72jIkCHJqg0AAAAAACAubqszrlq1StOmTVPPnj01YsQI5eXl6ZZbbtHvf//7ZNYHAAAAAABgmeUzOn79619ryZIlOumkk2LTLrjgAk2dOlWXX355UooDAAAAAACIh+UzOvbu3at+/fo1mNa3b199+eWXCS8KAAAAAACgLSw3OgYNGqT7779f1dXVkqQDBw5o3rx5OvXUU5NWHAAAAAAAQDwsX7oyZ84c3XzzzRo8eLC6deumL7/8Uqeeeqp+8YtfJLM+AAAAAAAAyyw3OvLy8rR8+XIFg8HYU1cKCgqSWRsAAAAAAEBcLDc6XnvtNfXq1UvHHXdcrMHx73//W4FAQGeddVbSCgQAAAAAALDK8j067rzzTuXk5DSYlpOTozvvvDPhRQEAAAAAALSF5UbHnj17lJeX12BaXl6edu/enfCiAAAAAAAA2sJyo6N3797avHlzg2lvvvmmioqKEl4UAAAAAABAW1i+R8eUKVN044036rLLLlPv3r31n//8R3/84x917733JrM+AAAAAAAAyyyf0TF8+HA98cQTOnDggP7yl7/owIED+vWvf63hw4cnsz4AAAAAAADLLJ/RIUkDBw7UwIEDk1ULAAAAAABAu1g+o2PJkiX65z//KUn6xz/+oaFDh2rYsGHaunVr0opD6kQiUe0o369t/9qtHeX7FYlE010SAMAAHD/sje0DmIf9Fqawc1Ytn9Hx5JNP6rLLLpMkPfjgg/rhD3+onJwc3XvvvVq1alXSCkTyRSJRbd4W0PyVW1RTG1ZWpkfTrhykIcV+ud2udJcHALApjh/2xvYBzMN+C1PYPauWz+jYt2+funbtqv379+vDDz/U97//fY0dO1ZlZWXJrA8pEKioigVUkmpqw5q/cosCFVVprgwAYGccP+yN7QOYh/0WprB7Vi03Ovx+v7Zs2aJ169Zp8ODB8ng82r9/vzweTzLrQwpUhqpjAa1XUxtW5b7qNFUEADABxw97Y/sA5mG/hSnsnlXLl6789Kc/1dSpU9WpUyc9/PDDkqSXXnpJxcXFSSsOqeHzZisr09MgqFmZHvm6ZqexKgCA3XH8sDe2D2Ae9luYwu5ZtXxGxznnnKPXXntNmzZt0te//nVJ0qhRo/TYY4/F5iktLU18hUg6f26Opl05SFmZh8/Oqb++yp+bk+bKAAB2xvHD3tg+gHnYb2EKu2c1rsfLflVmZmaDP8+cOVOjR49uV0FIPbfbpSHFfvXxD1Xlvmr5umbLn5tji5vIAADsi+OHvbF9APOw38IUds9quxodXxWN2udxMoiP2+1Sr7wu6pXXJd2lAAAMwvHD3tg+gHnYb2EKO2fV8qUrVrhc9ujeAAAAAACAjimhZ3QApohEogpUVKkyVC2f116nWQGAaRhTW8b6ARAPxgyYws5ZpdGBDicSiWrztkDsuc/1N84ZUuy3zY4JAKZgTG0Z6wdAPBgzYAq7Z9XSpSuRSESbN2/WoUOHWpyvsLAwIUUByRSoqIrtkNLh5z3PX7lFgYqqNFcGAOZhTG0Z6wdAPBgzYAq7Z9VSo8Ptdmvy5Mnq1KlTi/PxeFmYoDJU3eB5z9LhHbNyX3WaKgIAczGmtoz1AyAejBkwhd2zavlmpKeffrrefvvtJJYCpIbPmx173nO9rEyPfF2z01QRAJiLMbVlrB8A8WDMgCnsnlXLjY7CwkJNnDhR06dP13//93/roYceiv0HmMSfm6NpVw6K7Zj115P5c3PSXBkAmIcxtWWsHwDxYMyAKeyeVcs3I62pqdHw4cMlSbt27Yr7g8rKyjR9+nTt3btX3bt319y5c9WnT58m5/33v/+tMWPGaNy4cbrtttvi/iygJW63S0OK/erjH6rKfdXydbXXHYIBwCSMqS1j/QCIB2MGTGH3rFpudNx3333t+qBZs2Zp3LhxKikp0erVqzVz5kwtW7as0XzhcFizZs2KNVWAZHC7XeqV10W98rqkuxQAMB5jastYPwDiwZgBU9g5q5YvXZGkTz75RI8++qjuvPNOSYfPvPjggw9afd2ePXv0/vvva/To0ZKk0aNH6/3331dlZWWjeR9//HENHTq02bM9AAAAAAAAmmP5jI7nnntOc+bM0fnnn6/S0lLNnDlTVVVV+sUvfqEnn3yyxdcGAgHl5+fL4zl8/Y7H41FeXp4CgYB8Pl9svg8++ECvvfaali1bpoULFzb7fqFQSKFQqMG0YDBodVGAlCGrMAE5hSnIKkxBVmECcgons9zoePjhh7VkyRINGDBAzz33nCTppJNOsnRGhxW1tbX6+c9/rvvuuy/WEGnO0qVLtWDBgoR8LpBMZBUmIKcwBVmFKcgqTEBO4WSWGx2VlZU66aSTJEkulyv2//qfW+L3+7Vr1y6Fw2F5PB6Fw2GVl5fL7/fH5tm9e7c+//xzXXfddZIOdxij0aj279+vu+66q8H7TZgwQWPGjGkwLRgMavz48VYXB0gJsgoTkFOYgqzCFGQVJiCncDLLjY6TTz5Zq1ev1iWXXBKbtnbtWg0cOLDV1/bo0UMDBgxQaWmpSkpKVFpaqgEDBjS4bKWwsFBvvvlm7M+PPPKIDhw40ORTV7xer7xer9XSgbQhqzABOYUpyCpMQVZhAnIKJ7Pc6Lj99tv1ox/9SH/4wx904MAB/ehHP1JZWZmeeOIJS6+fPXu2pk+froULF8rr9Wru3LmSpIkTJ2rq1KkqLi5u2xIAAAAAAAD8L8uNjn79+um5557TSy+9pKFDh8rv92vo0KHKycmx/PpVq1Y1mr548eIm57/xxhutlgYAAAAAACApjkbHiy++qKFDh+rCCy9MZj0AAAAAAABt5rY64yOPPKL/+q//0u23397gXhoAAAAAAAB2YfmMjtWrV+tf//qX1qxZozvuuEM1NTW68MILNXr0aH39619PZo0AAAAAAACWWD6jQ5KOP/54TZs2TRs2bNDDDz+sjz76SGPHjk1WbQAAAAAAAHGxfEZHvUAgoLVr16q0tFQ7d+7UpZdemoy6AAAAAAAA4ma50bFixQqVlpbqgw8+0DnnnKMpU6bo7LPPVqdOnZJZHwAAAAAAgGWWGx0vvfSSrrjiCo0YMcLyI2UBAAAAAABSyXKj49e//nUy6wAAAAAAAGi3Fhsdt956q1wuV6tvMm/evIQVBAAAAAAA0FYtNjqOPfbYVNUBQFJtXUSZGXE9DMmIzwIAAACAVGmx0TFlypRU1QFAUmaGWzMWvp6Sz7p38lkp+RwAAAAASKW4Hi/7xhtvaPXq1SovL1deXp6+853vaMiQIcmqDQAAAAAAIC6Wz1tftWqVpk2bpp49e2rEiBHKy8vTLbfcot///vfJrA8AAAAAAMCyuJ66smTJEp100kmxaRdccIGmTp2qyy+/PCnFAQAAAAAAxMPyGR179+5Vv379Gkzr27evvvzyy4QXBQAAAAAA0BaWGx2DBg3S/fffr+rqaknSgQMHNG/ePJ166qlJKw4AAAAAACAeli9dmTNnjm6++WYNHjxY3bp105dffqlTTz1Vv/jFL5JZHwAAAAAAgGWWGx15eXlavny5gsFg7KkrBQUFyawNAAAAAAAgLpYbHZWVlcrKylJBQYF69uypP/3pT/J4PPrOd74jt9vyFTAAAAAAAABJY7lD8eMf/1ifffaZJGn+/Pl64okntGTJEt1///1JKw4AAAAAACAelhsdn376qQYMGCBJ+vOf/6zFixdr6dKlWrduXdKKAwAAAAAAiIflS1fcbrdqa2tVVlamrl27qrCwUJFIRFVVVcmsDwAAAAAAwDLLjY6zzz5bN910k/bu3asLL7xQkvSvf/1L+fn5SSsOAAAAAAAgHpYbHffcc4+effZZZWRkqKSkRJL0xRdf6MYbb0xacQAAAAAAAPGw3Ojo1KmTrrjiigbTzjzzzIQXBAAAAAAA0FaWGx179+7VE088oX/+8586cOBAg79bsWJFwgsDAAAAAACIl+VGx09+8hMdOnRIF1xwgbKzs5NZEwAAAAAAQJtYbnRs3bpVb7zxhjp16pTMegDAGLV1EWVmWH5Kt+0/BwAAAHACy42OE088UcFgUMccc0wy6wEAY2RmuDVj4etJ/5x7J5+V9M8AAAAAnMJyo+Ob3/ymrr32Wl166aXKzc1t8HeXXXZZwgsDAAAAAACIl+VGx1tvvaX8/Hy9/nrDf710uVyWGh1lZWWaPn269u7dq+7du2vu3Lnq06dPg3keffRRrVu3Th6PRxkZGZo2bZq+/e1vWy0RAAAAAAB0cJYbHU899VS7PmjWrFkaN26cSkpKtHr1as2cOVPLli1rMM/AgQN1zTXXKDs7Wx988IGuuuoqvfbaa+rcuXO7PhsAAAAAAHQMLTY6otGoXC6XJCkSiTQ7n9vd8k3y9uzZo/fff19LliyRJI0ePVp33XWXKisr5fP5YvMdefbGiSeeqGg0qr1796qgoKDB+4VCIYVCoQbTgsFgizUA6UBWYQJyClOQVZiCrMIE5BRO1mKj47TTTtOWLVskSV/72tdiTY969Y2Qf/7zny1+SCAQUH5+vjwejyTJ4/EoLy9PgUCgQaPjSH/60590zDHHNGpySNLSpUu1YMGCFj8TsAOyChOQU5iCrMIUZBUmIKdwshYbHWvXro39vHHjxqQXU++vf/2rHnroIT3xxBNN/v2ECRM0ZsyYBtOCwaDGjx+fivIAy8gqTEBOYQqyClOQVZiAnMLJWmx0+P3+2M+9evWSdPgSloqKCuXl5Vn+EL/fr127dikcDsvj8SgcDqu8vLzB+9fbunWrbr31Vi1cuFB9+/Zt8v28Xq+8Xq/lzwfShazCBOQUpiCrMAVZhQnIKZys5ZtrHCEUCuknP/mJBg4cqPPPP1/S4bM85s+f3+pre/TooQEDBqi0tFSSVFpaqgEDBjS6bOWdd97RtGnT9PDDD+vkk0+OZzkAAAAAAACsNzpmzZqlLl26aNOmTcrMzJQknXrqqXruuecsvX727Nlavny5Ro4cqeXLl2vOnDmSpIkTJ2rbtm2SpDlz5ujgwYOaOXOmSkpKVFJSog8//DDeZQIAAAAAAB2U5cfLbt68Wa+++qoyMzNjNyX1+Xzas2ePpdf369dPq1atajR98eLFsZ+feeYZq+UgQSKRqAIVVaoMVcvnzZY/N0dut6v1FwIAII4jJmAbAWZhn4Up7JxVy42Orl276osvvmhwb46dO3eqZ8+eSSkMyReJRLV5W0DzV25RTW1YWZkeTbtykIYU+20TUACAfXEcsT+2EWAW9lmYwu5ZtXzpytixYzV16lS98cYbikQi2rp1q2677TZ973vfS2Z9SKJARVUsmJJUUxvW/JVbFKioSnNlAAATcByxP7YRYBb2WZjC7lm13OiYOHGiRo0apTvvvFN1dXWaMWOGhg0bpgkTJiSzPiRRZag6Fsx6NbVhVe6rTlNFAACTcByxP7YRYBb2WZjC7lm1fOlKRUWFfvjDH+qHP/xhg+m7d+/m8hVD+bzZysr0NAhoVqZHvq7Zaayq7ex8jRgAOFH9caRrTqbOO+0YySW5XS718Jp5HHGith7rOaYC6eHzZsvf4yh9+5Qi6X93uVe3bjf2+zmcy+5ZtdzoGDlypLZs2dJo+kUXXaS//vWvCS0KqeHPzdG0Kwc1uq7Kn5uT7tLiZvdrxADAify5Ofrp9wfr82BIT2/4KDb+9vF7VdCDX4ztoC3Heo6pQPrk+47S5cNP1KI/vhPb/66/dKDyfUeluzSgAbtn1XKjIxqNNpq2f//+2BNYYB6326UhxX718Q9V5b5q+bqa+y82zV0j1sc/VL3yuqS5OgBwJrfbJX9ujuY99Rbjr0215VjPMRVIn12VB2K/OEqH979Ff3xHA/r42P9gK3bPaquNjnPOOUcul0s1NTUaOnRog7/bu3evLrroomTVhhRwu13qldfFFmFsj5auETN92QDAzvbuO8j4a3PxHus5pgLpw/4HU9g9q602Oh544AFFo1Fdd911mjdvXmy6y+VSjx491Ldv36QWCFjhtPuNpEJtXUSZGZbvR8xnAWgS46/zsE2B9GH/gynsntVWGx1nnHGGJOmNN95QdrY9iga+ykn3G0mVzAy3Zix8PSWfde/ks1LyOZJzlwuwK8Zf52GbAunD/gdT2D2rlu/RkZGRoYcfflirV69WeXm58vLyVFJSouuvv16dOnVKZo1Aq5x0vxEAMAnjr/OwTYH0Yf+DKeyeVcuNjgceeEDvvPOO5syZo8LCQu3cuVMLFy7U/v37NWPGjGTWCFjilPuNAIBpGH+dh20KpA/7H0xh56xabnSsX79eq1ev1tFHHy1J6tu3r772ta+ppKSERgcAAAAAALAFy3fsa+rxsi1NBwAAAAAASDXLjY5Ro0Zp0qRJevXVV/XJJ5/olVde0Q033KALLrggmfUBAAAAAABYZvnSlVtvvVWPPfaY7rzzTpWXlys/P18XXnihJk+enMz6AAAAAAAALLPU6AiHw/r5z3+uu+66SzfddFOyawIAAAAAAGgTS5eueDwevf7663K57PGoGAAAAAAAgKZYvkfHhAkT9Mgjj+jQoUPJrAdJFolEtaN8v7b9a7d2lO9XJMLNZAEA7WPHY4sdawIAKxi/YAo7Z9XyPTqWL1+uiooKLVmyRD6fTy6XS9FoVC6XSy+//HISS0SiRCJRbd4W0PyVW1RTG1ZWpkfTrhykIcV+ud2crQMAiJ8djy12rAkArGD8ginsnlXLjY4HHnggmXUgBQIVVbEgSlJNbVjzV25RH/9Q9crrkubqAAAmsuOxxY41AYAVjF8whd2zarnRccopp+ixxx7T2rVrVV5erry8PF144YWaNGlSMutDAlWGqmNBrFdTG1blvmpbhBEAYB47HlvsWBMAWMH4BVPYPauWGx2zZ89WWVmZbr/9dvXq1Us7duzQ448/rl27dum+++5LZo1IEJ83W1mZngaBzMr0yNc1O41VAQBMZsdjix1rAgArGL9gCrtn1fLNSDdu3Khf/epXOuecc3T88cfrnHPO0aOPPqqNGzcmsz4kkD83R9OuHKSsTI8kxa6j8ufmpLkyAICp7HhssWNNAGAF4xdMYfesWj6jIzc3V9XV1fJ6vbFpNTU16tmzZ1IKQ+K53S4NKfarj3+oKvdVy9c1W/7cHFvcLAYAYCY7HlvsWBMAWMH4BVPYPauWGx0lJSW69tpr9f3vf1/5+fkKBoNasWKFSkpKtHnz5th8Q4YMSUqhSAy326VeeV1scd0UAMAZ7HhssWNNAGAF4xdMYeesWm50PP3005KkRYsWNZpe/3cul4tLWQAAAAAAQNpYbnRs2rQpmXUAAAAAAAC0m+VGB5wtEokqUFGlylC1fF57XV8FALAvjh+tYx0BiAdjBkxh56zS6IAikag2bwto/sotqqkNx+6YO6TYb5ugAgDsh+NH61hHAOLBmAFT2D2rlh8vC+cKVFTFAipJNbVhzV+5RYGKqjRXBgCwM44frWMdAYgHYwZMYfespqzRUVZWpiuuuEIjR47UFVdcoU8//bTRPOFwWHPmzNHw4cM1YsQIrVq1KlXldWiVoepYQOvV1IZVua86TRUBAEzA8aN1rCMA8WDMgCnsntWUNTpmzZqlcePG6fnnn9e4ceM0c+bMRvOsWbNGn3/+uV544QX97ne/0yOPPKLt27enqsQOy+fNVlamp8G0rEyPfF2z01QRAMAEHD9axzoCEA/GDJjC7llNSaNjz549ev/99zV69GhJ0ujRo/X++++rsrKywXzr1q3T2LFj5Xa75fP5NHz4cK1fv77R+4VCIW3fvr3Bf8FgMBWLknaRSFQ7yvdr2792a0f5fkUi0Xa/pz83R9OuHBQLav31Vf7cnHa/d0fXkbMKc5BTtFVTx48bLz9F+b6jkvJ5JmaVY6y9JON7VFNMzCrsIZVjBjlFe9j9+JaSm5EGAgHl5+fL4zm8Ejwej/Ly8hQIBOTz+RrMV1hYGPuz3+9vcmdbunSpFixYkPzCbSZZN3xxu10aUuxXH/9QVe6rlq+rve6Ya7KOmlWYhZyirdxul848uUC3X32G3i/bo0hEWrH+n8rwuJNyMzITs8ox1j5SeeM8E7MK+8jMcGvM0OMViUbldrmUmZGcf5smp2ivVGW1LYx86sqECRM0ZsyYBtOCwaDGjx+fpopSo7kbvvTxD1WvvC7tem+326VeeV3a/T5oqKNmFWYhp2iPXZUHdM+Svza4TjdRx6avMjWrHGPtIZnfo77K1Kwi/QIVVZr31FsNxtSsTI8eupmcwl5SmdW2SEmjw+/3a9euXQqHw/J4PAqHwyovL5ff7280386dOzVw4EBJjc/wqOf1euX1elNRespYeQZxSzd8sUOY0JgTswrnIadojy+rDqrk7H7S/x6yNr31uSr2HkzKscmJWbVy/EdipPJ7lBOzitRgTIUpKkPV6pqTqZLTGubVLr+bpqTR0aNHDw0YMEClpaUqKSlRaWmpBgwY0OCyFUkaNWqUVq1apfPPP1979+7Viy++qBUrVqSixLSyeipl/Q1fvto1s8sNXwAAHUskEtXuLw5q9SufxI5fV4zorw1vfsaxyYJUXkoBvkfB/hhTYZIe3bJ10X8dp6c3fBTL6/dG9FcPrz2ymrKLaGbPnq3ly5dr5MiRWr58uebMmSNJmjhxorZt2yZJKikpUVFRkc4//3xdfvnluuGGG9S7d+9UlZg2Vp9BbPcbvgAAOpZARZUe+f3bDY5fv9vwkSZ99xscmyywevxHYvA9CnbHmAqThCPRWJNDOpzXpzd8pHCSbvIcr5Tdo6Nfv35atWpVo+mLFy+O/ezxeGINkI7E6qmU3NAMAGAnzR2/PB4XxyYLuCQ1tfgeBbtjTIVJ9u472GRe9+4/qN75XdNU1f8x8makThPPqZTc0AwAYBfNHb/sctqq3XEpRerxPQp2xpgKk9j9GGaf5790YJxKCQAwEcev9mH9ATgSYwJMYve8ckaHDXAqJQDARBy/2of1B+BIjAkwid3zSqPDJjiVEgBgIo5f7cP6A3AkxgSYxM555dIVAAAAAADgGI45oyMcPnwTlGAwmOZK4BQFBQXKyEj8LtJSVouKilRTczDhn9mcVH7W9u3bU/I5Tl2HLa2/ZGSVMRXJQFZhgnQc/4G2YEyFKZI1rrbEFY1G7fGg23Z66623NH78+HSXAQfZuHGjioqKEv6+ZBWJloysklMkA1mFCTj+wxSMqTBFssbVljim0XHw4EG9++676tmzpzweT9I+JxgMavz48VqxYoUKCgqS9jntZUKddq8xWZ3HeLJq93XUFGpOjSNrPuWUUxKe1VSNqa0xcdu0xonLJFlbrmSMq3bJanOcur3bw+7rxA7H/+bYfd1J1JgojKn2YEJW0i1dWW2NYy5d6dy5swYPHpyyzysoKEh5V6otTKjThBoTqS1ZNXEdUXNqJOvAkeoxtTUmbpvWOHGZpNQvl92y2hynbu/26GjrJJFZNWHdUWNiMKbagwlZSTe7rSNuRgoAAAAAAByDRgcAAAAAAHAMGh0AAAAAAMAxaHTEyev1asqUKfJ6vekupUUm1GlCjelm4jqi5tQwsea2cOJyOnGZJOcuV3uxXhpjnbSdCeuOGhPDhBo7ArZD6+y6jhzz1BUAAAAAAADO6AAAAAAAAI5BowMAAAAAADgGjQ4AAAAAAOAYjml01NXVafv27aqrq0t3KUCLyCpMQE5hCrIKU5BVmICcwikc0+gIBoMaNmyYgsFguksBWkRWYQJyClOQVZiCrMIE5BRO4ZhGBwAAAAAAAI0OAAAAAADgGDQ6AAAAAACAY2Sk4kPmzp2r559/Xjt27NCaNWvUv3//RvOEw2HdfffdevXVV+VyuXTddddp7NixqSjPKHV1Ef1755eq+LJaud2y1bewmzIyrPerIpGoAhVVqgxVy+fNlj83R263K4kVp45Tlq2ty9GW1zllnTkF2wNIHPYna5pbT05df05dLjhHJBJVYM9+VX5Zo737a5TbvbP6FXZXp06edJcGNGD3rKak0TFs2DD94Ac/0Pjx45udZ82aNfr888/1wgsvaO/evbrkkks0ZMgQFRUVpaLEtLNy4K2ri+jlLdu16I/vqKY2rKxMj66/dKCGDiqy1OyIRKLavC2g+Su3xF4/7cpBGlLsVyQSbVcDJd1aWramvsDY7YtOfT17vqxWOBLVY8/8Q4E9B2LLcebJBdpVeaDZeuNd/ra+xgRWtq3dtn99TU7cHkA62G1/aukfKZI9HrX0/s2tpzNPLtCb7wVts/4SxW65AL4qEonq3X/vVjQS1aG6qKLRqL4I1ejv+4I67cQC2/wCCUQiUf3j43LV1YXl8XgUDkdUXVOndz4p18B+ebbIakoaHYMHD251nnXr1mns2LFyu93y+XwaPny41q9fr2uvvbbRvKFQSKFQqME0k+8MbPXA+++dX8aaHJJUUxvWoj++o2MKuqr/MUe3+jmBiqrYZ9S/funa9+TveZQ+2f6lfvXHbW1qoNhBU8s2f+UW9fEPVa+8Lg3mTeUXHStZbaqeK0b017r/KVPF3oNauvY91YUjeuT3bzdbbzzLX6+513Tr8k0d3bWzLX75j5eVbWvXL7pt2YaJ4rQxFc5lNavp3J++qqV/pHC7XUkdj5oa7268/BSdNbBQGRnuZtfTvZPPss36S6RU5oJxFW0R3FOlL/cf0q49B/T0ho9i++11l3xdn+0K6YTerX/fjwc5RVsFKqrk8bhUsbdWv3r2/44xPx5TrP/s3qd+vbqnu8TUNDqsCAQCKiwsjP3Z7/c3u6MtXbpUCxYsSFVpSWf1wFvxZXVsnno1tWHt+bJaUusDX2Wo4etzu3fWiDOP1Y5dVbEmR/17xtNAsYOvLpt0eDkq91U3+vKSyi86VrLaVD2/2/CRSs7up99v/EjfPqUo1uRort54lr9ec6/Z+tFurf7LJ7b45T9eVratnX4BOlJbtmGiOG1MhXNZzWo696evaukfKXI6ZyZ1PGpqvHvk92/Lm9NJ3zihZ7PrqbnvG+lYf4mUylwwrqItKkMH9Xlwv559+V8N9tvH//SuZl57ZsI/j5yirSpD1XLJrV892/B3yF89u02zJn4zzdUdZptGRzwmTJigMWPGNJgWDAZbvDTGzqweeHO7ZSsr09Ng3qxMj3p0y7b0OT5vw9efd9ox+t2Gj3TdJcXtaqDYwVeXTTq8bnxdG6+bVH7RsZLV5urR//YX3G61Wm88y1+vudcoap9f/uNlZdva6RegI7VlGyaK08ZUOJfVrKZzf/qqlv6RouZQXVLHo+bGu/fL9ijv6KOaXU/Nfd9Ix/pLpFTmgnEVbXGwpk6RaLTJ/Ta0/1DCP4+coq183mx9smNvk1ndu+9gmqpqyDbXJfj9fu3cuTP250AgoIKCgibn9Xq9KioqavBfc/OaoP7Ae6SmDrx9C7vp+ksHxuatP/21b2E3S5/jz83RtCsHxV5f/wt0dmdPk59vtYFiB19dtvrTf/25OY3mtbq+E8FKVpurR9HD///acT1arTee5W/pNVeM6K9Nf/9c0v992TaJlW2byu0fj7Zsw0Rx2pgK57Ka1XTuT19V3zQ4Uv0xNtnjUXPvH4lIlfuqm11PfQu72Wb9JVIqc8G4irYo6JEjt8vV5H57tDcr4Z9HTtFW/twc+bp1bjqrXTunqaqGbHNGx6hRo7Rq1Sqdf/752rt3r1588UWtWLEi3WWlRP2B96vX6H71wJuR4dbQQUU6pqCr9nxZrR5x3jTU7XZpSLFfffxDVbmvWlmZGXr25U/07Mv/0rUlX9evV7/b4Pphqw0UO/jqsvm6Nn9DN6vrO1WaqufGy09RT19nnTe4t/J9R7VabzzL39RrAhX79eHne2P3BZHs8ct/vKxsW7tt/3pt2YYAmman/an+Hym+eo+OvoXd5Ha7kjoe+XNzdOPlpzS4x9MVI/prw5uf6bzBvVtcT3ZZf4nk1OWCcxT27KK+hV59b0T/BvfouP7SYvXoZo9fHgHp8Hiae3Rn/XhMcezylfp7dOQebY+spqTRcffdd+uFF15QRUWFrr76anXv3l1r167VxIkTNXXqVBUXF6ukpET/+Mc/dP7550uSbrjhBvXu3TsV5aVdPAfejAz3/943o22XlLjdLvXK66JeeV0UiURjX7BeePNT3XTFqZIrqvyjj1LfXt2NuRFpvSOXrbX57PRFx0o9Vuq1uvxNvcafm6Oa2oj2VdVKMvdf76ysS7tt/yO1ZRsCaJpd9qfW/pEimeOR2+3SWQML5c3ppPfL9igSkTa8+ZkmXHRybHxvbj3ZZf0lmlOXC87gdrt0+tcK5P1PpWb88HRV19QpJztTbreUd7RZ38ngfHndclSYV62ZPzpTe/fXqHuXLHkypLzu9shqShodd9xxh+64445G0xcvXhz72ePxaM6cOakox5bSceC18y98yWa3Lzqt1ZPsep2UBSvrym7bH4CztfSPFMkejzIy3PrGCT2Vd/RRqtxXrfMG9zZ2fAc6gowMtwYc2+PwY6EN/04GZ3O7XRpwbK4CFVVyuWW7rNrm0hWkB7/woR5ZAABnYnwHzMI+C1PYOatmXZsAAAAAAADQAhodAAAAAADAMWh0AAAAAAAAx6DRAQAAAAAAHINGBwAAAAAAcAwaHQAAAAAAwDFodAAAAAAAAMeg0QEAAAAAAByDRgcAAAAAAHAMGh0AAAAAAMAxaHQAAAAAAADHoNEBAAAAAAAcg0YHAAAAAABwDBodAAAAAADAMWh0AAAAAAAAx6DRAQAAAAAAHINGBwAAAAAAcAwaHQAAAAAAwDFodAAAAAAAAMfISNUHlZWVafr06dq7d6+6d++uuXPnqk+fPg3m2b17t2bOnKnt27errq5O119/vUpKStr1uZFIVIGKKlWGquXzZsufmyO329Wu9wSAVGD8ApAMjC32wbaAKcgqTJOyRsesWbM0btw4lZSUaPXq1Zo5c6aWLVvWYJ77779fX//61/XYY4+psrJSl156qc444wz5/f42fWYkEtXmbQHNX7lFNbVhZWV6NO3KQRpS7GfHBGBrjF8AkoGxxT7YFjAFWYWJUtLo2LNnj95//30tWbJEkjR69GjdddddqqyslM/ni833wQcfaMKECZIkn8+nk046Sc8995yuueaaBu8XCoUUCoUaTAsGg40+N1BRFdshJammNqz5K7eoj3+oeuV1SegyAk2xmlXgq1I5fpFTmIKsth/fjVLDSlbZFkg3fqeCk6Wk0REIBJSfny+PxyNJ8ng8ysvLUyAQaNDoOPnkk7Vu3ToVFxdr+/bt2rp1q4qKihq939KlS7VgwYJWP7cyVB3bIevV1IZVua+anRIpYTWrwFelcvwipzAFWW0/vhulhpWssi2QbvxOBSdL2aUrVkyfPl333nuvSkpKVFhYqG9+85vKyGhc4oQJEzRmzJgG04LBoMaPH99gms+braxMT4MdMyvTI1/X7OQsAPAVVrMKfFUqxy9yClOQ1fbju1FqWMkq2wLpxu9UcLKUNDr8fr927dqlcDgsj8ejcDis8vLyRvfe8Pl8evDBB2N/njhxovr169fo/bxer7xeb+ufm5ujaVcOanQ9mT83p/0LBVhgNavAV6Vy/CKnMAVZbT++G6WGlayyLZBu/E4FJ0tJo6NHjx4aMGCASktLVVJSotLSUg0YMKDBZSuS9MUXX6hr167KyMjQ5s2b9dFHH+nhhx9u8+e63S4NKfarj3+oKvdVy9eVOwQDMAPjF4BkYGyxD7YFTEFWYaK4Gh2vv/661q5dq8rKSi1atEjbtm3T/v37NWTIkFZfO3v2bE2fPl0LFy6U1+vV3LlzJR0+a2Pq1KkqLi7WO++8o3vuuUdut1tHH320Fi1apOzs9p0S5Xa71CuvC9ePATAO4xeAZGBssQ+2BUxBVmEay42Op556SsuWLdPYsWP1/PPPS5I6d+6se+65x1Kjo1+/flq1alWj6YsXL479fM455+icc86xWhIAAAAAAEADbqszLl26VEuWLNF1110nt/vwy/r27auysrKkFQcAAAAAABAPy42Oqqqq2M1DXa7D12PV1dUpMzMzOZUBAAAAAADEyXKj4/TTT9fjjz/eYNqyZct05plnJrwoAAAAAACAtrDc6Ljjjju0YcMGnXfeeaqqqtLIkSO1fv16TZ8+PZn1AQAAAzzw1FvpLgEAAECSxZuRRiIRffLJJ/rtb3+rjz76SDt27JDf79fAgQNj9+sAAAAd16G6SLpLAAAAkGSx0eF2uzV58mRt3bpVAwcO1MCBA5NdFwAAAAAAQNziukfH22+/ncRSAAAAAAAA2sfSGR2SVFhYqIkTJ2rYsGEqKCiIPXlFkm666aakFAcAAAAAABAPy42OmpoaDR8+XJK0a9eupBUEAAAAAADQVpYbHffdd18y6wAAAAAAAGg3y40OSfr0009VWlqq8vJy5eXlafTo0erTp0+SSgMAAAAAAIiP5ZuRbtq0SZdeeqnKysrUrVs3lZWV6bvf/a42btyYzPoAAAAAAAAss3xGx/z587Vw4UJ985vfjE178803ddddd2nYsGFJKQ4AAAAAACAels/oCAaDGjx4cINpp512moLBYMKLAgAAAAAAaAvLjY6TTjpJTzzxRINpS5Ys0YABAxJeFAAAAAAAQFtYvnRl9uzZmjRpkpYtWya/369AIKCjjjpKjz32WDLrAwAAAAAAsMxyo6Nfv35at26d3n777dhTV77xjW8oMzMzmfUBAAAAAABYZrnR8c9//lPdu3dvcJ+OQCCgL7/8UieddFJSigMAAAAAAIiH5Xt03Hrrraqrq2swrba2VrfeemvCiwIAAAAAAGgLy42OnTt3qnfv3g2mHXPMMdqxY0fCiwIAAAAAAGgLy42OgoICvffeew2mvffee8rLy7P0+rKyMl1xxRUaOXKkrrjiCn366aeN5tmzZ4+uu+46XXzxxRo1apRmz57d6CwSAAAAAACA5lhudPzwhz/U5MmT9dRTT+kvf/mLnnrqKU2ZMkVXX321pdfPmjVL48aN0/PPP69x48Zp5syZjeZZtGiR+vXrpzVr1mjNmjV677339MILL1hfGgAAAAAA0KFZvhnp5Zdfrq5du+oPf/iDgsGgCgoKdNttt2nUqFGtvnbPnj16//33tWTJEknS6NGjddddd6myslI+ny82n8vlUlVVlSKRiA4dOqTa2lrl5+c3er9QKKRQKNRgWjAYtLooQMqQVZiAnMIUZBWmIKswATmFk1ludEjSBRdcoAsuuCDuDwkEAsrPz5fH45EkeTwe5eXlKRAINGh0TJ48WTfeeKO+9a1vqbq6WuPHj9dpp53W6P2WLl2qBQsWxF0HkGpkFSYgpzAFWYUpyCpMQE7hZJYbHaWlpRowYID69eunsrIy/fznP5fb7dasWbPUr1+/hBSzfv16nXjiiVq6dKmqqqo0ceJErV+/vtFZIxMmTNCYMWMaTAsGgxo/fnxC6gAShazCBOQUpiCrMAVZhQnIKZzMcqPjv//7v/X0009LkubOnavi4mIdddRRmjNnjpYtW9bia/1+v3bt2qVwOCyPx6NwOKzy8nL5/f4G8y1fvlz33nuv3G63unbtqvPOO09vvvlmo0aH1+uV1+u1WjqQNmQVJiCnMAVZhSnIKkxATuFklm9GWllZqdzcXNXU1Ojvf/+7pk2bphtuuEEffPBBq6/t0aOHBgwYoNLSUkn/d3bIkZetSFJRUZFeeeUVSdKhQ4e0efNmnXDCCfEsDwAAAAAA6MAsNzp8Pp8+++wzvfLKKyouLlanTp1UU1OjaDRq6fWzZ8/W8uXLNXLkSC1fvlxz5syRJE2cOFHbtm2TJM2YMUN///vfdfHFF+uSSy5Rnz59dPnll7dhsQAAAAAAQEdk+dKVyZMn69JLL5XH49H8+fMlSZs3b9ZJJ51k6fX9+vXTqlWrGk1fvHhx7Odjjjkm9mQWAAAAAACAeFludFx66aWxJ65kZ2dLkr7xjW/ol7/8ZWyev//9700+JQUAAAAAACAVLF+6Ih1ucNQ3OaTD997o2bNn7M8TJ05MXGUAAAAAAABxiqvR0Rqr9+sAAAAAAABIhoQ2OlwuVyLfDgAAAAAAIC4JbXQAAAAAAACkE40OAAAAAADgGNyjAwAAAAAAOIblRscll1zS5PRLL7009vPWrVvbXRAAAAAAAEBbWW50fPbZZ42mRaNRbd++PaEFAQAAAAAAtFVGazP89Kc/lSTV1tbGfq63Y8cOHX/88cmpDAAAAAAAIE6tNjqOOeaYJn+WpEGDBmnUqFGJrwoAAAAAAKANWm10TJkyReFwWEcffbTGjh2rTp06paIuAAAAAACAuFm6R4fH49H8+fNpcgAAAAAAAFuzfDPSc889V5s2bUpmLQAAAAAAAO3S6qUr9WpqajR16lSdeuqpKigokMvliv3dvHnzklIcAAAAAABAPCw3Ovr376/+/fsnsxYAAAAAAIB2sdzomDJlSjLrAAAAAAAAaDfLjQ5Jev3117V27VpVVlZq0aJF2rZtm/bv368hQ4Ykqz4AAAAAAADLLN+M9KmnntLs2bPVp08f/e1vf5Mkde7cWQ899FDSigMAAAAAAIiH5TM6li5dqieffFJFRUVavHixJKlv374qKyuz9PqysjJNnz5de/fuVffu3TV37lz16dOnwTw//elP9eGHH8b+/OGHH+rRRx/VsGHDrJYJAAAAAAA6MMuNjqqqKvn9fkmKPXGlrq5OmZmZll4/a9YsjRs3TiUlJVq9erVmzpypZcuWNZjnyKe3fPDBB5owYYK+/e1vWy0RAAAAAAB0cJYvXTn99NP1+OOPN5i2bNkynXnmma2+ds+ePXr//fc1evRoSdLo0aP1/vvvq7KystnX/OEPf9DFF1+sTp06WS0RAAAAAAB0cJbP6Ljjjjt0/fXXa9WqVaqqqtLIkSPVpUsXLVq0qNXXBgIB5efny+PxSJI8Ho/y8vIUCATk8/kazX/o0CGtWbNGTz75ZJPvFwqFFAqFGkwLBoNWFwVIGbIKE5BTmIKswhRkFSYgp3Ayy42OvLw8PfPMM9q2bZt27Nghv9+vgQMHyu22fFKIZS+++KIKCws1YMCAJv9+6dKlWrBgQcI/F0g0sgoTkFOYgqzCFGQVJiCncLK4Hi8biURUW1srSQqHw4pGo5Ze5/f7tWvXLoXDYXk8HoXDYZWXl8fu+fFVzzzzjL773e82+34TJkzQmDFjGkwLBoMaP368xSUBUoOswgTkFKYgqzAFWYUJyCmczHKj44MPPtANN9ygQ4cOKT8/X8FgUFlZWXr00Ud10kkntfjaHj16aMCAASotLVVJSYlKS0s1YMCAJi9bCQaD+vvf/65f/OIXzb6f1+uV1+u1WjqQNmQVJiCnMAVZhSnIKkxATuFklq87mTFjhsaPH69XXnlFf/jDH/Tqq6/qqquu0owZMyy9fvbs2Vq+fLlGjhyp5cuXa86cOZKkiRMnatu2bbH5nn32WZ177rnq3r17fEsCAAAAAAA6PMtndHz66aeaMGFC7NGyLpdLP/jBD/TII49Yen2/fv20atWqRtMXL17c4M+TJk2yWhIAAAAAAEADls/oOOecc7Rp06YG01566SUNHTo00TUBAAAAAAC0SYtndNx6662xMzjC4bCmTZumr3/96yooKFAwGNS7776rYcOGpaRQAAAAAACA1rTY6Dj22GMb/Ll///6xn48//nh961vfSk5VAAAAAAAAbdBio2PKlCmxn3fv3q2ePXs2mmf37t2JrwoAABjjv1duSXcJAAAAMZbv0TFy5Mgmp1900UUJKwYAAJjnUF0k3SUAAADEWG50RKPRRtP2798fu4cHAAAAAABAurX6eNlzzjlHLpdLNTU1jZ6wsnfvXs7oAAAAAAAAttFqo+OBBx5QNBrVddddp3nz5sWmu1wu9ejRQ3379k1qgQAAAAAAAFa12ug444wzJElvvPGGsrOzk14QAAAAAABAW1m+R0dGRoYefvhhDRs2TMXFxRo2bJgefvhhHTp0KJn1AQAAAAAAWNbqGR31HnjgAb3zzjuaM2eOCgsLtXPnTi1cuFD79+/XjBkzklkjAAAAAACAJZYbHevXr9fq1at19NFHS5L69u2rr33tayopKaHRAQAAAAAAbKFdj5dtaToAAAAAAECqWW50jBo1SpMmTdKrr76qTz75RK+88opuuOEGXXDBBcmsDwAAAAAAwDLLl67ceuuteuyxx3TnnXeqvLxc+fn5uvDCCzV58uRk1gcAAAAAAGCZpUZHOBzWz3/+c91111266aabkl0TAAAAAABAm1i6dMXj8ej111+Xy+VKdj0AAAAAAABtZvkeHRMmTNAjjzyiQ4cOJbMeAAAAAACANrN8j47ly5eroqJCS5Yskc/nk8vlUjQalcvl0ssvv5zEEgEAAAAAAKyx3Oh44IEHklkHAAAAAABAu1ludJxyyil67LHHtHbtWpWXlysvL08XXnihJk2aZOn1ZWVlmj59uvbu3avu3btr7ty56tOnT6P51q1bp8ceeyx2tsiSJUuUm5treYEAAAAAAEDHZbnRMXv2bJWVlen2229Xr169tGPHDj3++OPatWuX7rvvvlZfP2vWLI0bN04lJSVavXq1Zs6cqWXLljWYZ9u2bVqwYIGWLl2qnj17at++ferUqVP8SwUAAAAAADoky42OjRs3asOGDfJ6vZKk448/Xt/4xjd0/vnnt/raPXv26P3339eSJUskSaNHj9Zdd92lyspK+Xy+2HxPPvmkrrnmGvXs2VOS1LVr1ybfLxQKKRQKNZgWDAatLgqQMmQVJiCnMAVZhSnIKkxATuFklhsdubm5qq6ujjU6JKmmpibWlGhJIBBQfn6+PB6PpMOPq83Ly1MgEGjQ6Pjkk09UVFSk8ePH68CBAxoxYoQmTZrU6LG2S5cu1YIFC6yWDqQNWYUJyClMQVZhCrIKE5BTOJnlRkdJSYmuvfZaff/731d+fr6CwaBWrFihkpISbd68OTbfkCFD2lxMOBzWhx9+qCVLlujQoUO69tprVVhYqEsuuaTBfBMmTNCYMWMaTAsGgxo/fnybPxtIBrIKE5BTmIKswhRkFSYgp3Ayy42Op59+WpK0aNGiRtPr/87lcmnjxo2NXuv3+7Vr1y6Fw2F5PB6Fw2GVl5fL7/c3mK+wsFCjRo1Sp06d1KlTJw0bNkzvvPNOo0aH1+ttcGYJYFdkFSYgpzAFWYUpyCpMQE7hZJYbHZs2bWrzh/To0UMDBgxQaWmpSkpKVFpaqgEDBjS4bEU6fO+Ov/zlLyopKVFdXZ3eeOMNjRw5ss2fCwAAAAAAOhZ3qj5o9uzZWr58uUaOHKnly5drzpw5kqSJEydq27ZtkqSLLrpIPXr00IUXXqhLLrlExx9/vC677LJUlQgAAAAAAAxn+YyO9urXr59WrVrVaPrixYtjP7vdbv3sZz/Tz372s1SVBQAAAAAAHCRlZ3QAAAAAAAAkG40OAAAAAADgGDQ6AAAAAACAY9DoAAAAAAAAjkGjAwAAAAAAOAaNDgAAAAAA4Bg0OgAAAAAAgGPQ6AAAAAAAAI5BowMAAAAAADgGjQ4AAAAAAOAYNDoAAAAAAIBj0OgAAAAAAACOQaMDABC3B556K90lwAK2EwAA6IhodAAA4naoLpLuEmAB2wkAAHRENDoAAAAAAIBj0OgAAAAAAACOQaMDAAAAAAA4Bo0OAAAAAADgGDQ6AAAAAACAY9DoAAAAAAAAjpGRqg8qKyvT9OnTtXfvXnXv3l1z585Vnz59GszzyCOP6Le//a3y8vIkSYMGDdKsWbOSVlMkElWgokqVoWr5vNny5+bI7XYl7fMAdDyMMwAYB4COi/0fTmbnfKes0TFr1iyNGzdOJSUlWr16tWbOnKlly5Y1mu+SSy7RbbfdlvR6IpGoNm8LaP7KLaqpDSsr06NpVw7SkGK/bTYOALMxzgBgHAA6LvZ/OJnd852SRseePXv0/vvva8mSJZKk0aNH66677lJlZaV8Pl/c7xcKhRQKhRpMCwaDcb1HoKIqtlEkqaY2rPkrt6iPf6h65XWJuyagKYnIKsxlyjhDTmEKE7NqyjiAxDIxq0g8u+//5BTtYfd8p6TREQgElJ+fL4/HI0nyeDzKy8tTIBBo1OhYu3atXnvtNfXs2VM33nijTj311Ebvt3TpUi1YsKBdNVWGqmMbpV5NbViV+6ptsWHgDInIKsxlyjhDTmEKE7NqyjiAxDIxq0g8u+//5BTtYfd8p+zSFSu+973v6frrr1dmZqZef/11TZ48WevWrdPRRx/dYL4JEyZozJgxDaYFg0GNHz/e8mf5vNnKyvQ02DhZmR75uma3byGAIyQiqzCXKeMMOYUpTMyqKeMAEsvErCLx7L7/k1O0h93znZJGh9/v165duxQOh+XxeBQOh1VeXi6/399gvp49e8Z+Puuss+T3+/Xxxx/rjDPOaDCf1+uV1+ttX025OZp25aBG1xT5c3Pa9b7AkRKRVZjLlHGGnMIUJmbVlHEAiWViVpF4dt//ySnaw+75Tkmjo0ePHhowYIBKS0tVUlKi0tJSDRgwoNFlK7t27VJ+fr4k6Z///Kd27Nih4447Lik1ud0uDSn2q49/qCr3VcvX1V53iQVgPsYZAIwDQMfF/g8ns3u+U3bpyuzZszV9+nQtXLhQXq9Xc+fOlSRNnDhRU6dOVXFxsX75y1/qvffek9vtVmZmpubNm9fgLI9Ec7td6pXXxRbXEAFwJsYZAIwDQMfF/g8ns3O+U9bo6Nevn1atWtVo+uLFi2M/1zc/AAAAAAAA2sJWNyNtj3D48E1QeCQSEqWgoEAZGYnfRcgqEi0ZWW0ppy6XS4rUaufOnYpGown9XCROqrZT/ecoEtWOHTta/KxUZxVoC47/MAVjKkyRrHG1Ja6oQ76lvvXWW9whGAm1ceNGFRUVJfx9ySoSLRlZJadIBrIKE3D8hykYU2GKZI2rLXFMo+PgwYN699131bNnT3k8nqR9Tv0jl1asWKGCgoKkfU57mVCn3WtMVucxnqzafR01hZpT48iaTznllIRnNVVjamtM3DatceIySdaWKxnjql2y2hynbu/2sPs6scPxvzl2X3cSNSYKY6o9mJCVdEtXVlvjmEtXOnfurMGDB6fs8woKClLelWoLE+o0ocZEaktWTVxH1JwayTpwpHpMbY2J26Y1TlwmKfXLZbesNsep27s9Oto6SWRWTVh31JgYjKn2YEJW0s1u68id7gIAAAAAAAAShUYHAAAAAABwDBodAAAAAADAMWh0xMnr9WrKlCnyer3pLqVFJtRpQo3pZuI6oubUMLHmtnDicjpxmSTnLld7sV4aY520nQnrjhoTw4QaOwK2Q+vsuo4c89QVAAAAAAAAzugAAAAAAACOQaMDAAAAAAA4Bo0OAAAAAADgGI5pdNTV1Wn79u2qq6tLdylAi8gqTEBOYQqyClOQVZiAnMIpHNPoCAaDGjZsmILBYLpLAVpEVmECcgpTkFWYgqzCBOQUTuGYRgcAAAAAAACNDgAAAAAA4Bg0OgAAAAAAgGPQ6AAAAAAAAI6Rke4CEJ9IJKpARZUqQ9XyebPlz82R2+1Kd1lIILYxANMwbgFA4jCmwhR2ziqNDoNEIlFt3hbQ/JVbVFMbVlamR9OuHKQhxX7bBArtwzYGYBrGLQBIHMZUmMLuWeXSFYMEKqpiQZKkmtqw5q/cokBFVZorQ6KwjQGYhnELABKHMRWmsHtWaXQYpDJUHQtSvZrasCr3VaepIiQa2xiAaRi3ACBxGFNhCrtnlUaHQXzebGVlehpMy8r0yNc1O00VIdHYxgBMw7gFAInDmApT2D2rNDoM4s/N0bQrB8UCVX8dlD83J82VIVHYxgBMw7gFAInDmApT2D2r3IzUIG63S0OK/erjH6rKfdXydbXXnW3RfmxjAKZh3AKAxGFMhSnsnlUaHYZxu13qlddFvfK6pLsUJAnbGIBpGLcAIHEYU2EKO2eVS1cAAAAAAIBj0OgAAAAAAACOQaMDAAAAAAA4Bo0OAAAAAADgGDQ6AAAAAACAY9DoAAAAAAAAjkGjAwAAAAAAOAaNDgAAAAAA4Bg0OgAAAAAAgGPQ6AAAAAAAAI5BowMAAAAAADhGRroL+Kqamhrde++92rx5s7KysnTKKaforrvuSndZAAAAAADAALZrdDzwwAPKysrS888/L5fLpYqKinSXBAAAAAAADGGrRkdVVZX+9Kc/6S9/+YtcLpckKTc3t9F8oVBIoVCowbRgMJiSGoF4kFWYgJzCFGQVpiCrMAE5hZPZqtHxn//8R927d9eCBQv05ptvKicnRzfddJMGDx7cYL6lS5dqwYIFaaoSsI6swgTkFKYgqzAFWYUJyCmczBWNRqPpLqLeu+++q+9+97t68MEHdfHFF+sf//iHrr/+em3YsEFdunSJzddc93H8+PHauHGjioqKUl060CSyChOQU5iCrMIUZBUmIKdwMlud0VFYWKiMjAyNHj1akvSNb3xDRx99tMrKylRcXBybz+v1yuv1pqtMwDKyChOQU5iCrMIUZBUmIKdwMls9Xtbn8+nMM8/U66+/LkkqKyvTnj17dOyxx6a5MgAAAAAAYAJbndEhSXPmzNGMGTM0d+5cZWRkaN68eXQaAQAAAACAJXE3Oj755BOtX79eFRUVmjVrlj755BPV1tbqpJNOSkhBvXv31lNPPZWQ9wIAAAAAAB1LXJeuPPfcc7rqqqu0a9curV69WpJ04MAB3X///UkpDgAAAAAAIB5xndHx8MMP64knntCAAQP03HPPSZJOOukkffDBB0kpDgAAAAAAIB5xndFRWVkZu0TF5XLF/l//MwAAAAAAQDrF1eg4+eSTY5es1Fu7dq0GDhyY0KIAAAAAAADaIq5LV26//Xb96Ec/0h/+8AcdOHBAP/rRj1RWVqYnnngiWfUBAAAAAABYZrnREY1G1alTJ5WWluqVV17R0KFD5ff7NXToUOXk5CSzRgAAAAAAAEssNzpcLpcuvvhibdmyRRdeeGEyawIAAAAAAGiTuO7RMWDAAJWVlSWrFgAAAAAAgHaJ6x4dZ5xxhiZOnKgxY8aooKCgwdNWLrvssoQXBwAAAAAAEI+4Gh1btmxRr1699Ne//rXBdJfLRaMDAAAAAACkXVyNjqeeeipZdQAAAAAAALRbXI0OSfryyy/10ksvadeuXcrPz9e5556rbt26JaM2AAAAAACAuMR1M9KtW7dqxIgRevrpp/Xhhx/q6aef1ogRI7R169Zk1QcAAAAAAGBZXGd03HvvvZo1a5Yuuuii2LR169bp7rvv1jPPPJPw4gAAAAAAAOIR1xkdn376qS644IIG00aOHKnPP/88oUUBAAAAAAC0RVyNjmOPPVZr165tMG39+vXq3bt3QosCAAAAAABoi7guXZkxY4auv/56PfXUUyosLNSOHTv02WefadGiRcmqDwAAAAAAwLK4Gh2DBg3Shg0b9PLLL6u8vFznnnuuzjnnHHXv3j1J5QEAAAAAAFgXV6Nj165d6ty5s0pKSmLTvvzyy9ijZgEAAAAAANIprnt0TJ48WcFgsMG0YDCoKVOmJLQoAAAAAACAtoj7qSsnnnhig2knnnii/v3vfye0KElasGCBTjzxRH300UcJf28AAAAAAOBMcTU6fD6fPvvsswbTPvvss4Tfo+O9997T22+/rcLCwoS+LwAAAAAAcLa47tHx3e9+VzfeeKOmTZum3r176/PPP9dDDz2ksWPHJqygQ4cO6c4779SDDz6oCRMmNDlPKBRSKBRqMO2rl9QAdkBWYQJyClOQVZiCrMIE5BROFlej47rrrlNGRobmzp2rYDCogoICjR07VldffXXCCnrooYf0ne98R7179252nqVLl2rBggUJ+0wgWcgqTEBOYQqyClOQVZiAnMLJXNFoNJruIupt3bpV8+fP19KlS+VyuXTeeedp0aJF6t+/f4P5mus+jh8/Xhs3blRRUVEqywaaRVZhAnIKU5BVmIKswgTkFE4W1xkdb7zxhnr16qXevXtr9+7devDBB+V2u3XzzTerZ8+e7S7mb3/7m/79739r2LBhkg7vaD/60Y9033336Vvf+lZsPq/XK6/X2+7PA5KNrMIE5BSmIKswBVmFCcgpnCyum5HOmTNHHo9HknT//ferrq5OLpdLP//5zxNSzHXXXafXXntNmzZt0qZNm1RQUKDf/OY3DZocAAAAAAAAzYnrjI5du3apsLBQdXV1sYZEZmamvv3tbyerPgAAAAAAAMvianR06dJFFRUV+vjjj9WvXz/l5OTo0KFDqqurS0pxmzZtSsr7AgAAAAAAZ4qr0XHVVVfpsssuU21trWbMmCFJ2rJli/r27ZuU4oB4RCJRBSqqVBmqls+bLX9ujtxuV7rLsjXWGYBEYCwBgMRhTIUp7JzVuB8vO2LECHk8Hh1zzDGSpPz8fN19992xeeofOwukUiQS1eZtAc1fuUU1tWFlZXo07cpBGlLst83OZjesMwCJwFgCAInDmApT2D2rcd2MVJKOO+64WJOj/s8nnnhi7M8XXnhhYioD4hCoqIrtZJJUUxvW/JVbFKioSnNl9sU6A5AIjCWpV1sXceRnAWBMhTnsntW4zuiwIhqNJvotgVZVhqpjO1m9mtqwKvdVq1delzRVZW+sMwCJwFiSepkZbs1Y+HpKPuveyWel5HMAHMaYClPYPatxn9HRGpcr/aepoOPxebOVlelpMC0r0yNf1+w0VWR/rDMAicBYAgCJw5gKU9g9qwlvdADp4M/N0bQrB8V2tvprxPy5OWmuzL5YZwASgbEEABKHMRWmsHtWE37pCpLLzne2TSe326UhxX718Q9V5b5q+bqyblrDOrMf9m+YyO126cyTC3Tv5LO058tq9eiWrb6F3cguALTRcYVezZ74TR2sqVN+jxz16tmFMRW2ZOesco8Og9j9zrbp5na71Cuviy2uCTMF68w+2L9hqkgkqjffC5JdAGin5r4L9OrJ9zTYiwlZTfilK+vWrUv0W+J/2f3OtgDajv0bpiK7AJAYjKcwhQlZjeuMjnPOOafJm4126tRJ+fn5Ov/883XllVcmrDg0ZPc72wJoO/ZvmIrsAkBiMJ7CFCZkNa5Gx/e//339+c9/1ve//335/X4FAgGtWLFCo0aNUrdu3bRkyRIFAgH99Kc/TVa9cXPSNe/1d7Y9MlR2urMtgLZrav/29zhKWZkZ2vav3caPX0i9VB3/ODYBbeek76loPzuPp2QVR7JzVuvF1eh49tln9cQTTyg/Pz827eyzz9Y111yjtWvX6swzz9TVV19tm0aH0655r7+z7VeXxy53tgXQdl/dv/09jtLlw0/UjIWvO2L8Qmql8vjHsQloG6d9T0X72XU8Jav4Krtm9UhxNTp2796tnJyGxWdnZ6u8vFySdNxxxykUCiWuunZq7tqhPv6htjmlJh48JQNwrq/u31mZGbEmh2T++IXUSuXxj2MT0DZO+56K9rPreEpW8VV2zeqR4mp0nHvuuZo0aZImTZqk/Px87dq1S7/61a907rnnSpK2bt2qoqKipBTaFiZcOxQvnpIBONeR+/e2f+123PiF1En18Y9jExA/J35PRfvZcTwlq2iKHbN6pLgaHXfeeaceeeQRzZw5U+Xl5erZs6cuuOAC3XDDDZKk3r1761e/+lVSCm0LE64dSjeut0M9smAvjF9oj1Tmh7EDaBvGeTTHbuMqWUVz7JbVI8XV6MjKytItt9yiW265pcm/79mzZ0KKShQTrh1KpyOvt+uak6nhpx+rYwq66ji/V4U9u9gmpEi+jnbtpZ0H5XqMX2iPVOUnEonqrX/u0sf/2atINCq3y6UTenfX4AH5ttunALthnEdT7DiuklU0xY5ZPVKrjY7NmzdbeqMhQ4a0u5hEM+HaoXSqv96ua06mLvyv4/S7DR91iF9y0VhHuvbSlKYO4xfaI1X5Ce6p0ufBkJ59+V+x/el7I/qrKK+LCns6a+wAEo1xHk2x47hKVtEUO2b1SK02Om6//fYGf66/8Wj37t21d+9eSVJ+fr42btyY+OoSwO7XDqVT/fV2Jaf1izU5JGf/koumdaRrL01q6jB+oT1SkZ/K0EE9/ZXjx9MbPtJJfXy2+JID2B3jPL7KruMqWcVX2TWr9VptdGzatCn286JFi7R3717ddNNNys7OVnV1tR5++GF17949mTUiSeqvt5NLHeaXXDStI1172ZGaOkCyHaypa3J/OlhTl6aKAMBsjKswhd2z6o5n5ieffFI/+clPlJ19+Jef7Oxs3XzzzVqyZElSikNy1V9v53a5Djc8juDUX3LRtPos1OfAyddexhp8RyDvQNsU9Mhpcn/K7+G8sQMAUoFxFaawe1bjanQcddRReueddxpM27ZtW6zxAbPUX2937mlFmnzZwA7xSy6aVp+Fh24eqnsn/5ceunmo7e5ZkSgdqakDJFthzy5N7k+9bHDKKgCYiHEVprB7VuN66srUqVN17bXX6rzzzlNBQYGCwaBeeuklzZw5MyHFfPHFF/rpT3+qzz//XJ06ddKxxx6rO++8Uz6fLyHvj8bcbpcKe3ZRQY8cnXiMjxsMdWAd5dpLbqgFJA770/+prYsoMyOufz8CgEYYV2EKu2c1rkbHJZdcoq9//et6/vnnVV5eruOOO06TJk3S8ccfn5BiXC6Xrr32Wp155pmSpLlz5+rBBx/Uvffem5D3R/M6yi+5gETegURifzosM8OtGQtfT8ln3Tv5rJR8DoD0YFyFKeyc1bgaHZJ0/PHHt9jYuPjii7VmzZo2FdO9e/dYk0OSTjnlFK1cubLRfKFQSKFQqMG0YDDYps8EkomswgTkFKYgqzAFWYUJyCmcLO5GR2u2b9+ekPeJRCJauXKlzjvvvEZ/t3TpUi1YsCAhnwMkE1mFCcgpTEFWYQqyChOQUzhZwhsdLldirsm56667dNRRR+mqq65q9HcTJkzQmDFjGkwLBoMaP358Qj4bSBSyChOQU5iCrMIUZBUmIKdwsoQ3OhJh7ty5+uyzz7Ro0SK53Y1v7OX1euX1etNQGRAfsgoTkFOYgqzCFGQVJiCncDLbNTrmz5+vd999V48//rg6deqU7nIAAAAAAIBBEt7oiEajbX7txx9/rEWLFqlPnz763ve+J0kqKirSo48+mqjyAAAAAACAgyW80XHnnXe2+bUnnHCCPvzwwwRWAwAAAAAAOpK4Gx0bN27U3/72N33xxRcNzt6YN2+epMOPlwUAAAAAAEiHxnf6bMGCBQs0a9YsRSIRrV+/Xt27d9drr73GTWwAAAAAAIAtxNXoeOaZZ/TEE09oxowZyszM1IwZM7Ro0SJt3749WfUBAAAAAABYFlejIxQKqX///pKkzMxM1dbWauDAgfrb3/6WlOIAAAAAAADiEdc9Oo455hh9/PHHOuGEE3TCCSdo5cqV8nq96tatW7LqAwAAAAAAsCyuRsf/+3//T3v37pUk3XLLLfrJT36iAwcOaObMmcmoDQAAAAAAIC5xNTrOOeec2M8DBw7Uhg0bEl4QAAAAAABAW8V1j44zzjijyelDhgxJSDEAAAAAAADtEVejo7a2tslpkUgkYQUBAAAAAAC0laVLV8aNGyeXy6VDhw5p/PjxDf4uGAzq1FNPTUpxAAAAAAAA8bDU6Bg7dqyi0ai2bdumyy67LDbd5XKpR48e+uY3v5m0AgEAAAAAAKyy1OgYM2aMJKlv37465ZRTGv39O++8o4EDBya0sI4mEokqUFGlylC1fN5s+XNz5Ha70l0WUoTtD8BUjF8AkDiMqTCJnfMa11NXrrnmGm3ZsqXR9GuvvVZ//etfE1ZURxOJRLV5W0DzV25RTW1YWZkeTbtykIYU+20TFCQP2x+AqRi/ACBxGFNhErvn1dLNSCORiMLhsKLRqKLRqCKRSOy/Tz/9VB6PJ9l1OlqgoioWEEmqqQ1r/sotClRUpbkypALbH4CpGL8AIHEYU2ESu+fV0hkdX/va1+RyuWI/H8ntduv6669PfGUdSGWoOhaQejW1YVXuq1avvC5pqgqpwvYHYCrGLwBIHMZUmMTuebXU6Ni4caOi0ai+//3va/ny5YpGo3K5XHK5XPL5fOrcuXOy63Q0nzdbWZmeBkHJyvTI1zU7jVUhVdj+AEzF+AUAicOYCpPYPa+WLl3p1auXioqK9NJLL6lXr14qLCxUZmamCgsLaXIkgD83R9OuHKSszMOXANVf3+TPzUlzZUgFtj8AUzF+AUDiMKbCJHbPa1w3Iw2FQpozZ46ef/55ZWRk6O2339bGjRv1zjvvaNq0acmq0fHcbpeGFPvVxz9Ulfuq5etqrzvWIrnY/gBMxfgFAInDmAqT2D2vls7oqDdr1ix16dJFmzZtUmZmpiTp1FNP1XPPPZeU4pItEolqR/l+bfvXbu0o369IJJq2Wtxul3rldVFxv57qldcl5QGx07pIBbstb2vb32712pnd1pXd6mkvpy0PEiMSjaqmJqwv9h1UcE8VuQCA/xXvcdPtdinfd5SyMjP02a6Q/rV9r+rqIimqFohPXV1Eoaoa7d13SKGqGltlNa4zOjZv3qxXX31VmZmZsZuT+nw+7dmzJynFJZPdH4eTSh1tXZi2vKbVm052W1d2q6e9nLY8aL+mMvG9Ef11TIFXgwfkkwsAHVpbjpt1dRG9vGW7Fv3xndhrrr90oIYOKlJGRlz/Rg0k1aFDYf1l63b96tltsaz+eEyxzjm1SJ06pf+prHHtLV27dtUXX3zRYNrOnTvVs2fPhBVUVlamK664QiNHjtQVV1yhTz/9NGHvfSS7Pw4nlTraujBteU2rN53stq7sVk97OW150H5NZeLpDR/p4//sJRcAOry2HDf/vfPLWJOj/jWL/viO/r3zy5TUDFj1yY69sSaHdDirv3p2mz7ZsTe9hf2vuBodY8eO1dSpU/XGG28oEolo69atuu222/S9730vYQXNmjVL48aN0/PPP69x48Zp5syZCXvvI7X0OJyOpqOtC9OW17R608lu68pu9bSX05YH7ddcJiLRKLkA0OG15bhZ8WXTr9nzJWMq7KXiy4PNZPVgmipqKK5LVyZOnKhOnTrpzjvvVF1dnWbMmKErrrhCEyZMSEgxe/bs0fvvv68lS5ZIkkaPHq277rpLlZWV8vl8sflCoZBCoVCD1waDwbg+y+6Pw0mljrYuUrm8ZDW17Lau7FZPc6zm1JTlQeo0lwm3y2XbMRVIBbIKqW3HzdxuTb+mRzfGVNhLbrfOzWTVHk9ljavR4XK59MMf/lA//OEPk1JMIBBQfn6+PJ7D1/R4PB7l5eUpEAg0aHQsXbpUCxYsaNdn1T8O56vXzNnlcTip1NHWRSqXl6ymlt3Wld3qaY7VnJqyPEidpjJRf48Ou46pQCqQVUhtO272Leym6y8d2OgeHX0LuyW8PnKK9ujXq7t+PKa40T06+vXqnu7SJEmuaDRq+dbob7zxhnr16qXevXtr9+7devDBB+V2u3XzzTcn5D4d7777rm677TatXbs2Nu3CCy/UAw88oJNPPjk2rbnu4/jx47Vx40YVFRVZ+rxIJKpARZUtH4eTah1tXaRqeclq6tltXdmtnqbEk1MTlgepFYlEtWP3fu3aU6XOWRnyeTuroIe9x9RkmbHw9ZR8zr2Tz0rpZyF+ds8qUqctx826uoj+vfNL7fmyWj26ZatvYbek3IiUnKK9Dh0K65Mde7Xny4Pq0a2z+vXqbosbkUpxntExZ84c/eY3v5Ek3X///ZKkrKws/fznP9eiRYvaXYzf79euXbsUDofl8XgUDodVXl4uv9/fYD6v1yuv19vuz6t/pGevvC7tfi/TdbR1karlJaupZ7d1Zbd6mhJPTk1YHqSW2+1S7/yu6p3fNemflagxFUg2sop6bTluZmS41f+YoyUdnbzCRE7Rfp06eTTguB7pLqNJcTU6du3apcLCQtXV1em1117Tpk2blJmZqW9/+9sJKaZHjx4aMGCASktLVVJSotLSUg0YMKDBZSsAAAAAAADNiavR0aVLF1VUVOjjjz9Wv379lJOTo0OHDqmuri5hBc2ePVvTp0/XwoUL5fV6NXfu3IS9NwAAAAAAcLa4Gh1XXXWVLrvsMtXW1mrGjBmSpC1btqhv374JK6hfv35atWpVwt4PAAAAAAB0HHE1Oq677jqNGDFCHo9HxxxzjCQpPz9fd999d1KKAwAAAAAAiEdcjQ5J6t27t7Zu3ap33nlH+fn5OvXUU5WREffbAAAAAAAAJFxcHYpPPvlEkyZN0sGDB+X3+xUIBJSVlaVFixapX79+yaoRAAAAAADAkrgfL3v55ZfrRz/6kVyuw89//s1vfqPZs2frqaeeSkqBAAAAAAAAVrnjmfmDDz7Q1VdfHWtySNKECRP0wQcfJLwwAAAAAACAeMXV6MjLy9Nf//rXBtPeeust5eXlJbQoAAAAAACAtojr0pWbb75ZkydP1tChQ1VYWKidO3fq5Zdf1gMPPJCs+gAAAAAAACyzfEZHOBzW//t//08rV67UCSecoKqqKp1wwgn64x//qOHDhyezRgAAAAAAAEssn9Hh8XjUp08fHX300Zo8eXIyawIAAAAAAGiTuC5dufjii3X99dfrBz/4gQoKChr83ZAhQxJaGAAAAAAAQLzianSsXLlSkvTII480mO5yubRx48bEVQUAAAAAANAGcTU6Nm3alKw6AAAAAAAA2i2ux8sCAAAAAADYGY0OAAAAAADgGDQ6AAAAAACAY9DoAAAAAAAAjkGjAwAAAAAAOAaNDgAAAAAA4Bg0OgAAAAAAgGPQ6AAAAI5VWxdJdwmwKJXbilwAgLNlpLsAAACAZMnMcGvGwteT/jn3Tj4r6Z/hdKnaVhLbCwCczjaNjjlz5mjz5s3q1KmTjjrqKN1+++0qLi5Od1kAAAAAAMAgtrl05eyzz9aaNWv05z//WT/+8Y81bdq0dJcEAAAAAAAMY5szOs4999zYz6eccoqCwaAikYjc7sa9mFAopFAo1GBaMBhMeo1AvMgqTEBOYQqyClOQVZiAnMLJbNPoONKKFSs0dOjQJpsckrR06VItWLAgxVUB8SOrMAE5hSnIKkxBVmECcgonS1mjY8yYMdq5c2eTf/c///M/8ng8kqS1a9dqzZo1WrFiRbPvNWHCBI0ZM6bBtGAwqPHjxyeuYCAByCpMQE5hCrIKU5BVmICcwslS1uh49tlnW51nw4YNmj9/vp588knl5uY2O5/X65XX601keUBSkFWYgJzCFGQVpiCrMAE5hZPZ5tKVl156Sffdd5+WLFmioqKidJfjWJFIVIGKKlWGquXzZsufmyO325XuspAAbFsA6cL4AwCJw5gKU9g5q7ZpdPzsZz9TZmampk6dGpv25JNP6uijj05jVc4SiUS1eVtA81duUU1tWFmZHk27cpCGFPttE0i0DdsWQLow/gBA4jCmwhR2z6ptGh1vvPFGuktwvEBFVSyIklRTG9b8lVvUxz9UvfK6pLk6tAfbFkC6MP4AQOIwpsIUds9q0481gSNVhqpjQaxXUxtW5b7qNFWERGHbAkgXxh8ASBzGVJjC7lml0dGB+LzZysr0NJiWlemRr2t2mipCorBtAaQL4w8AJA5jKkxh96zS6OhA/Lk5mnbloFgg66+j8ufmpLkytBfbFkC6MP4AQOIwpsIUds+qbe7RgeRzu10aUuxXH/9QVe6rlq+rve6Mi7Zj2wJIF8YfAEgcxlSYwu5ZpdHRwbjdLvXK62KLG8Qgsdi2ANKF8QcAEocxFaawc1a5dAUAAAAAADhGXGd0vP7661q7dq0qKyu1aNEibdu2Tfv379eQIUOSVR8AAAAAAIBllhsdTz31lJYtW6axY8fq+eeflyR17txZ99xzjy0aHeHw4UfbBIPBNFcCpygoKFBGRuKv7iKrSLRkZJWcIhnSkdWioiLV1BxM6Gc2J1Wfk+rP2r59e0o+J5XbSmp+uTj+wxQc/2GKZI2rLXFFo9GolRmHDx+uJ598UkVFRTr99NP1t7/9TeFwWP/1X/+lN998M9l1tuqtt97S+PHj010GHGTjxo0qKipK+PuSVSRaMrJKTpEMZBUm4PgPUzCmwhTJGldbYrnRMWTIEL322mvyeDw644wz9Ne//lU1NTUaNmyYXnvttWTX2aqDBw/q3XffVc+ePeXxeFp/QRsFg0GNHz9eK1asUEFBQdI+p71MqNPuNSar8xhPVu2+jppCzalxZM2nnHJKwrOaqjG1NSZum9Y4cZkka8uVjHHVLlltjlO3d3vYfZ3Y4fjfHLuvO4kaE4Ux1R5MyEq6pSurrbH8aaeffroef/xxTZo0KTZt2bJlOvPMM5NSWLw6d+6swYMHp+zzCgoKUt6VagsT6jShxkRqS1ZNXEfUnBrJOnCkekxtjYnbpjVOXCYp9ctlt6w2x6nbuz062jpJZFZNWHfUmBiMqfZgQlbSzW7ryPK34zvuuEPXX3+9Vq1apaqqKo0cOVJdunTRokWLklkfAAAAAACAZZYaHZFIRJ988ol++9vf6qOPPtKOHTvk9/s1cOBAud08oRYAAAAAANiDpUaH2+3W5MmTtXXrVg0cOFADBw5Mdl0AAAAAAABxs3w6xumnn6633347iaWYwev1asqUKfJ6vekupUUm1GlCjelm4jqi5tQwsea2cOJyOnGZJOcuV3uxXhpjnbSdCeuOGhPDhBo7ArZD6+y6jiw/dWX27Nlau3athg0bpoKCArlcrtjf3XTTTUkrEAAAAAAAwCrLNyOtqanR8OHDJUm7du1KWkEAAAAAAABtZfmMDgAAAAAAALuzfEaHJH366acqLS1VeXm58vLyNHr0aPXp0ydJpQEAAAAAAMTH8s1IN23apEsvvVRlZWXq1q2bysrK9N3vflcbN25MZn2W1dXVafv27aqrq0t3KUCLyCpMQE5hCrIKU5BVmICcwiksNzrmz5+vhQsX6he/+IV+8pOf6Be/+IUWLlyo+fPnJ7M+y4LBoIYNG6ZgMJjuUoAWkVWYgJzCFGQVpiCrMAE5hVNYbnQEg0ENHjy4wbTTTjuNnQAAAAAAANiG5UbHSSedpCeeeKLBtCVLlmjAgAEJLwoAAAAAAKAtLN+MdPbs2Zo0aZKWLVsmv9+vQCCgo446So899lgy6wMAAAAAALDMcqOjX79+Wrdund5+++3YU1e+8Y1vKDMzs9XXzp07V88//7x27NihNWvWqH///o3mCYfDuvvuu/Xqq6/K5XLpuuuu09ixY+Nbmg4gEokqUFGlylC1fN5s+XNz5Ha72vz6fN9R2lV5QPuraxSJSF/sq1Fu92z1LeymjAzLJ/zYQjzrpr3rMZn+f3t3HhZV3f4P/D0zrAqIIKtLKm7kkiiKpobigguGWGqKZpq4a/mouYtLZmRp7iQVmZaluYXiFj5uieb25F5qmAsDKIgguzPn94c/ztcRmDkDMzCD79d1eTlz5iz3+Zz7fM7MzVmKiw2AznhLs07PT1Pd3gZyuQypj02vTfQhpR1MefsTmZv8fBUSlOlIzciDjaUCVWwtUNetGmxs9Hqwm8GZ0n7+9Kka/yQ+xsPHOahRrXyPsWq1gPsPniA5NQs21hZwcrCBu7N+x4eKbj9DqqzrRURUEbKy8pCQlIm0jFw4V7NBXTd7VK1qXdFhAdCj0HHt2jU4Ojpq3KdDqVTi8ePHaNKkidZpu3btinfffRehoaEljhMTE4M7d+7g4MGDSE9PR79+/dC+fXvUqlVLaohmTeqPs/hLSqzYch55BSpYWyowZXArtG/uIekg/eL0Hs5VMLBbYxw8lYCANq/g692XxfmO7d8CnVvVMptihz5tU9Z2NFb8yodZSH2cA5VawPrtf0KZmg1rSwU+GuaLgqdqrfGWZp2Km+ad7o2w92QCMrMKKrxNSkNKO5ji9n8+fn4BJ3OSn6/CiYv3se6Xi+L+NCSwCZJTs+H3qkeFFTtMaT9/+lSNI+fvIXLHxWKPscbc70vq5+u4O8DX2w1yuazE4rqptJ8hmVJeEBGZuydPcnHqajIid1x67vjWHO2buptEsUPyN5Dp06cXuUyloKAA06dPR0xMjNZpX7yJaXFiY2MxYMAAyOVyODk5oVu3bti/fz9GjRpVZNyMjAxkZGRoDDPnm6JKPfDef/BEHAcA8gpUWLHlPOq4+6O2m73O5SgfZmlM36llLUTuuIjJg1pi1c//05hv5I6LqONuj0Z1qhthjQ3vxXUrbJu6Hp1R09Wu1OOWlZRcLW77D+reCLEnE/AwPRc37qZj55GbWuMtzToVN81Ph/5G8Bte2Br3t9HaxJiktEN5bn99VOQX8MrWp1L5uZWYLhY5gGf7048HrmNQ90a4lZiOpvVrGHR5UnPVlPbzfxIfi0WOwlgKj7ENajkadb8vqZ8P6dwANV3s4FGjarHLr+NubzLtZ0imdvwnqmjMUyottVrAvylPxCIHUHh8uwSPGlXRtL4ZFToSExNRu3ZtjWF16tTB/fv3DRKIUqmEp6en+N7Dw6PEHW3jxo1Ys2aNQZZrCqQeeJNTs8RxCuUVqJCcmiWp0JGWkaM5vezZ9Ll5qmLnm/o4B4B5FDqKrBuerUNaZk6RLy/6jFtWUnK1uO3/83MFB7Ug6Iy3NOtU0jSQSZveFElph/Lc/vqoyB9mla1PpfKTlpFX7P5Urao10jJyDb48qblqSvv5w8fFx5L6OAdVbSyNut+X1A5qQUBaZg4AFLv8me/6mkz7GZKpHf+JtFm26SymD9P9x+KyYJ5SaSkfZiE9s/jvAOlP8isoKk2SCx3u7u64cuUKmjZtKg67cuUKXF1djRKYNsOHD0dISIjGsKSkJK2XxpgyqQdeG2sLWFsqNMa1tlTAxlraZnRysC12elsbRbHDnavZlnaVyl1J6+ZkX3Qd9Bm3rKTkqq6Cg1wm0xlvadappGkgSJveFElph/Lc/vqoyB9mla1PpfJT1ab445KttQLV7W0MvjypuWpK+3mNasXH4lzN1uj7fUntIJfJ4GRf8vJtStiuFd1PlpWpHf+JtMl/qjb6MpinVFppGTmoYmtZbJ9apYLv0VVI8g0Y3nvvPYwfPx6bNm3C0aNHsWnTJkycOBEjRowwSCAeHh5ITEwU3yuVSri7uxc7roODA2rVqqXxr6RxzUHhgfd5xR14nRxs8E73RuK4hdfaOjlI+zLpUaMqpgxuJU5//MI9jO3fAjHHbmFUcDON+Y7t3wL1PauVddXKzYvrVnj6beG1xqUdt6yk5GpJ2x/Cs/8b1nbUGW9p1qm4ad7p3giHz90xapsYk5R2KM/trw+p/YAxVLY+lcpPDUfbYo9L9lUs4VTN8KetSs1VU9rP63tWw9j+LYo9xhp7vy+pn29Y2xEeNapqWb6NybSfIZna8Z+oojFPqbScHGyR+SSv2O8Azg4Vf9kKoMcZHQMHDoS9vT1++eUXJCUlwd3dHTNmzEDPnj0NEkjPnj2xbds29OjRA+np6fjtt9/www8/GGTepq7wwPviNbIvHnjdnauijrsDQjo3gFoQIJfJUMfdAe7O0g7QcrkM7Zt7oK5HZ6Rl5sDJ/tlTV7zrOuFJbh4Wj2mPR0/yyv2O8IZQ3LqVdEM3fcYtD8Vt/0kDW8LFyQYBvrXFPNAWb2nW6cVpHO1soJDL0Lhu9Qpvk9KS0g6mtv0LSe0HiEyJZw071Haz1zguudeoAksrOdydKu4SB1Pazy0s5OjcqhbquNsj9XEOnJ87xhp7vy9shzru/sU+daWk5bs7V4W7c1WTaD9DMqW8ICIyZx41quL+gydwtFdpfAeo7WaPWq4OFR0eAD0KHQDQq1cv9OrVS++FfPzxxzh48CAePnyIESNGwNHREXv37kVYWBgmT56M5s2bIzg4GH/++Sd69OgBAJgwYUKRe4JUVlIPvHK5DL7ebqjpYlfqA7RcLkNNVzuNU2KfvTbfa24LFbduhhjX2KRuf13xlmadipvG06Xi26QspLSDKW3/QvwCTuZILpehzavu8HSx0/vxpeURm6ns5xYW8v9/c2/N+16Vx34vlz/74lncvbx0Ld9U2s+QTCkviIjMVeHv0qTULLg75yI37yncnKuipotdhR//C0kudOzZswfe3t7w8vJCQkIC5s2bB7lcjvDwcHh5eWmddu7cuZg7d26R4VFRUeJrhUKBhQsX6hF65SL1wMsDdOXE7UoA84DMk7Yf0qRbRe/3Fb18IiIyT3K5DJ4udib7R1LJ1yZ8+eWXqFbt2T0bIiIi0Lx5c7Rp0+alLk4QERERERERkWmRfEZHWloaatSogby8PJw7dw6rVq2ChYUF2rVrZ8z4iIiIiIiIiIgkk1zocHJywr///ou///4bzZs3h5WVFXJyciAIgjHjIyIiIiIiIiKSTHKhY/z48ejfvz8UCgVWrFgBAIiPj0eTJk2MFhwRERERERERkT4kFzr69+8vPnHF1vbZ891fe+01LF++XBzn3LlzaN26tYFDJCIiIiIiIiKSRq/HyxYWOAo5OztrvA8LC8P58+fLHlU5UasFKB9mIS0jB04OfJQjERleZetnKtv6EJUH7jdERFQZmfLxTa9Chy7mdL8OtVpA/CUlVmw5j7wCFawtFZgyuBXaN/cwmY1DROatsvUzlW19iMoD9xsiIqqMTP34JvnxslLIZBW/QlIpH2aJGwUA8gpUWLHlPJQPsyo4MiKqLCpbP1PZ1oeoPHC/ISKiysjUj28GLXSYk7SMHHGjFMorUCEtM6eCIiKiyqay9TOVbX2IygP3GyIiqoxM/fj20hY6nBxsYW2p0BhmbamAk71tCVMQEemnsvUzlW19iMoD9xsiIqqMTP34ZtBChzndo8OjRlVMGdxK3DiF1xR51KhawZERUWVR2fqZyrY+ROWB+w0REVVGpn5803ozUrVaLWkmcvmzesmFCxfKHlE5kctlaN/cA3U9OiMtMwdO9qZ1l1giMn+VrZ+pbOtDVB643xARUWVk6sc3rYWOV199VdINRq9du2awgMqTXC5DTVc71HS1q+hQiKiSqmz9TGVbH6LywP2GiIgqI1M+vmktdMTFxYmvjxw5ggMHDmDMmDHw9PREYmIioqKi0KNHD6MHSUREREREREQkhdZCR82aNcXX3333HbZv3w4HBwcAQL169dCsWTO89dZbGDJkiHGjJCIiIiIiIiKSQPLNSDMzM5GTo/momNzcXGRmZho8KCIiIiIiIiKi0tB6RsfzQkJCMGLECAwfPhzu7u5ISkrCpk2bEBISYsz4iIiIiIiIiIgkk1zomD59OurUqYPY2FikpKTAxcUFoaGhGDhwoDHjIyIiIiIiIiKSTHKhQy6XY/DgwRg8eLAx4yEiIiIiIiIiKjXJ9+gQBAFbt27F8OHD0bdvXwDAmTNnEBsba7TgiIiIiIiIiIj0IbnQsXLlSvzyyy8YOHAglEolAMDd3R1ff/210YIjIiIiIiIiItKH5ELHzp07ERkZiT59+kAmkwEAatWqhbt37xotOCIiIiIiIiIifUgudKhUKlStWhUAxEJHVlYWqlSpYpzIiIiIiIiIiIj0JLnQ4e/vj6VLlyI/Px/As3t2rFy5El26dDFacERERERERERE+pBc6Jg1axZSUlLQunVrZGZmwsfHB4mJiZg2bZox4yMiIiIiIiIikkzy42Xt7Oywbt06pKam4v79+/Dw8ICLi4sxYyMiIiIiIiIi0oteZ3QcP34czs7OaNGihVjkWLBggaTpExISMGjQIAQGBmLQoEG4fft2kXFSU1MxevRo9O3bFz179sSCBQvw9OlTqSESERERERER0UtOcqEjJiYGc+bMwTfffKMx/Ndff5U0fXh4OIYMGYIDBw5gyJAhmD9/fpFxIiMj4eXlhZiYGMTExODKlSs4ePCg1BCJiIiIiIiI6CUn+dIVKysrbN26FePHj8f169exZMkSWFlZQRAEndOmpqbi6tWriI6OBgAEBQVh8eLFSEtLg5OTkzieTCZDVlYW1Go18vPzUVBQADc3tyLzy8jIQEZGhsawpKQkqatCVG6Yq2QOmKdkLpirZC6Yq2QOmKdUmUkudACAu7s7fvzxR8yZMweDBw/GunXrxEfNaqNUKuHm5gaFQgEAUCgUcHV1hVKp1Ch0jB8/HpMmTULHjh2Rk5OD0NBQtG7dusj8Nm7ciDVr1ugTOlGFYK6SOWCekrlgrpK5YK6SOWCeUmUmudBReOaGjY0NvvjiC2zYsAFvv/22+LhZQ9i/fz8aN26MjRs3IisrC2FhYdi/fz969uypMd7w4cMREhKiMSwpKQmhoaEGi4XIEJirZA6Yp2QumKtkLpirZA6Yp1SZSS50TJgwQeP96NGj0bhxY+zfv1/ntB4eHkhOToZKpYJCoYBKpUJKSgo8PDw0xtu8eTM++eQTyOVy2NvbIyAgAKdPny5S6HBwcICDg4PU0IkqDHOVzAHzlMwFc5XMBXOVzAHzlCozyTcjHTVqVJFh/v7+WLp0qc5pnZ2d4e3tjT179gAA9uzZA29vb43LVgCgVq1aOHbsGAAgPz8f8fHxaNiwodQQiYiIiIiIiOglp/WMjvfff198ysqQIUNKvB/HDz/8oHNBCxYswMyZM7Fu3To4ODggIiICABAWFobJkyejefPmmD17NsLDw9G3b1+oVCr4+flh4MCB+q4TEREREREREb2ktBY6+vXrJ74eMGBAmRbk5eWFbdu2FRkeFRUlvq5Tp474ZBYiIiIiIiIiIn1pLXT07dtXfP3ijWqIiIiIiIiIiEyNXo+XPXv2LK5evYrs7GyN4WPHjjVoUEREREREREREpSG50LF48WLs27cPvr6+sLa2FoeXdN8OIiIiIiIiIqLyJrnQERMTg5iYGLi5uRkzHiIiIiIiIiKiUpP8eFl3d3dYWVkZMxYiIiIiIiIiojKRfEbHkiVLMG/ePPTp0wc1atTQ+KxNmzYGD4yIiIiIiIiISF+SCx1XrlzBsWPHcObMGdjY2IjDZTIZjhw5YozYiIiIiIiIiIj0IrnQsWLFCkRGRuL11183ZjxERERERERERKUm+R4dtra28PX1NWYsRERERERERERlIrnQMXnyZHzyySd48OAB1Gq1xj8iIiIiIiIiIlMg+dKV2bNnAwB+/vlncZggCJDJZLh27ZrhIyMiIiIiIiIi0pPkQkdcXJwx4yAiIiIiIiIiKjPJhY6aNWsaMw4iIiIiIiIiojLTWuiYN28eFi9eDACYPn06ZDJZseN99tlnho+MiIiIiIiIiEhPWgsdtWrVEl+/8sorRg+GiIiIiIiIiKgstBY6xowZI74eNGgQXFxciozz4MEDw0dFRERERERERFQKkh8vGxgYWOzwPn36GCwYIiIiIiIiIqKykFzoEAShyLAnT56UeN8OIiIiIiIiIqLypvOpK/7+/pDJZMjLy0Pnzp01PktPT+cZHURERERERERkMnQWOpYtWwZBEDB69GiNp6vIZDI4Ozujfv36Rg2QiIiIiIiIiEgqnYWOtm3bAgBOnToFW1tbowdERERERERERFRaOgsdhRQKBVauXIm9e/ciJSUFrq6u6N27N8aNGwdra2tjxkhEREREREREJInkQseCBQuQkJCAOXPmoGbNmrh//z42bNiA5ORkLF261JgxEhERERERERFJIrnQERcXh0OHDsHBwQEA0KBBA7z22mvo0aOH0YIjIiIiIiIiItKH5MfL1qhRAzk5ORrD8vLy4OLiYvCgiIiIiIiIiIhKQ/IZHcHBwRg1ahSGDRsGNzc3JCUl4YcffkBwcDDi4+PF8dq3b2+UQImIiIiIiIiIdJFc6Pjpp58AAJGRkUWGF34mk8kQFxdX7PQJCQmYOXMm0tPT4ejoiIiICNStW7fIeLGxsVi/fj0EQYBMJkN0dDRq1KghNUwiIiIiIiIieolJLnQcPny4TAsKDw/HkCFDEBwcjN27d2P+/Pn4/vvvNca5dOkS1qxZg40bN8LFxQWZmZmwsrIq03KJiIiIiIiI6OUhudABAE+fPsWFCxeQnJwMd3d3tGzZEhYWumeRmpqKq1evIjo6GgAQFBSExYsXIy0tDU5OTuJ43333HUaOHCne98Pe3r7Y+WVkZCAjI0NjWFJSkj6rQlQumKtkDpinZC6Yq2QumKtkDpinVJlJLnTcunUL48aNQ25uLjw8PKBUKmFtbY3IyEh4eXlpnVapVMLNzQ0KhQIAoFAo4OrqCqVSqVHouHXrFmrVqoXQ0FBkZ2eje/fuGDduHGQymcb8Nm7ciDVr1uiznkQVgrlK5oB5SuaCuUrmgrlK5oB5SpWZ5ELHwoULMXDgQLz//vti4eGbb77BggULsGnTJoMEo1Kp8NdffyE6Ohr5+fkYNWoUPD090a9fP43xhg8fjpCQEI1hSUlJCA0NNUgcRIbCXCVzwDwlc8FcJXPBXCVzwDylykxyoeP69euIjo7WOLti+PDhRW5OWhwPDw8kJydDpVJBoVBApVIhJSUFHh4eGuN5enqiZ8+esLKygpWVFbp27YqLFy8WKXQ4ODjAwcFBauhEFYa5SuaAeUrmgrlK5oK5SuaAeUqVmVzqiK6urvjjjz80hp09exaurq46p3V2doa3tzf27NkDANizZw+8vb01LlsBnt2748SJExAEAQUFBTh16hSaNGkiNUQiIiIiIiIieslJPqNjypQpGD9+PDp37gxPT08kJibiyJEjWLZsmaTpFyxYgJkzZ2LdunVwcHBAREQEACAsLAyTJ09G8+bN0adPH1y+fBm9e/eGXC5Hx44d8fbbb5duzYiIiIiIiIjopSO50NG1a1fs2LED+/btQ0pKCho2bIjJkyejXr16kqb38vLCtm3bigyPiooSX8vlcsyaNQuzZs2SGhYRERERERERkUivx8vWq1cP48ePL/Hzvn37IiYmpsxBERERERERERGVhuR7dEhx7949Q86OiIiIiIiIiEgvBi10PP9EFiIiIiIiIiKi8mbQQgcRERERERERUUVioYOIiIiIiIiIKg2DFjoEQTDk7IiIiIiIiIiI9KJ3oUOtViMlJaXYzxYtWlTmgIiIiIiIiIiISktyoSMjIwNTp05FixYt0KNHDwBAXFwcVqxYIY7Tt29fw0dIRERERERERCSR5EJHeHg47OzscPjwYVhaWgIAfHx8sG/fPqMFR0RERERERESkDwupI8bHx+P48eOwtLQUHyPr5OSE1NRUowVHRERERERERKQPyWd02Nvb49GjRxrDEhMT4eLiYvCgiIiIiIiIiIhKQ3KhY8CAAZg8eTJOnToFtVqNCxcuYMaMGXjnnXeMGR8RERERERERkWSSL10JCwuDlZUVFi1ahKdPn2L27NkYNGgQhg8fbsz4iIiIiIiIiIgkk1zokMlkeO+99/Dee+8ZMRwiIiIiIiIiotKTXOgAnt2QdO/evUhJSYGrqyv69OmD9u3bGys2IiIiIiIiIiK9SL5HR3R0NP7zn/+gWrVq8Pf3h6OjI6ZOnYpvv/3WmPEREREREREREUkm+YyOb7/9Fhs3bkSjRo3EYcHBwRgxYgRGjhxplOCIiIiIiIiIiPQh+YwOAHjllVc03teuXRsymcygARERERERERERlZbWQodarRb/TZo0CbNnz8bt27eRm5uLhIQEzJs3D5MnTy6vWImIiIiIiIiItNJ66cqrr74qnrEhCAIAYO/evRrD9uzZgwEDBhg5TCIiIiIiIiIi3bQWOuLi4sorDiIiIiIiIiKiMtNa6KhZs2Z5xUFEREREREREVGY6b0a6f/9+3L17FwDw+PFjTJs2DX5+fvDz88PMmTORmZlp9CCJiIiIiIiIiKTQWeiIiIiAg4MDAGDJkiXIz8/Ht99+i2+++Qa5ublYsmSJ0YMkIiIiIiIiIpJC66UrAJCeng57e3sAQHx8PA4ePAhbW1sAwNKlS9G1a1fjRkhEREREREREJJHOMzrq1auHkydPAgAcHR3x4MED8bOHDx+KT2MhIiIiIiIiIqpoOs/omDZtGqZNm4Z33nkH3bp1w4gRI/DWW28BAHbs2IHQ0FBJC0pISMDMmTORnp4OR0dHREREoG7dusWO+88//yAkJARDhgzBjBkzpK8NEREREREREb3UdJ7R8frrr+P777/Hv//+i5iYGCQnJ+Pbb7/F77//jnHjxmHixImSFhQeHo4hQ4bgwIEDGDJkCObPn1/seCqVCuHh4ejWrZt+a0JERERERERELz2dZ3QAQIMGDfDFF1+UeiGpqam4evUqoqOjAQBBQUFYvHgx0tLS4OTkpDHuhg0b0LlzZ2RnZyM7O7vY+WVkZCAjI0NjWFJSUqnjIzIW5iqZA+YpmQvmKpkL5iqZA+YpVWaSCh1lpVQq4ebmBoVCAQBQKBRwdXWFUqnUKHRcv34dJ06cwPfff49169aVOL+NGzdizZo1Ro+bqKyYq2QOmKdkLpirZC6Yq2QOmKdUmeksdGRlZWHZsmU4f/48GjZsiA8++AB16tQRP+/bty9iYmLKHEhBQQHmzZuHpUuXigWRkgwfPhwhISEaw5KSkiTfL4SovDBXyRwwT8lcMFfJXDBXyRwwT6ky01noiIiIgFKpxKRJk/DHH39gwIABWLVqFfz8/AAA9+7d07kQDw8PJCcnQ6VSQaFQQKVSISUlBR4eHuI4Dx48wJ07dzB69GgAz06lEgQBT548weLFizXm5+DgAAcHB71WlKgiMFfJHDBPyVwwV8lcMFfJHDBPqTLTWeg4fPgw9u3bB3t7e3Tv3h2dO3fGhx9+iE8//RT+/v6SFuLs7Axvb2/s2bMHwcHB2LNnD7y9vTUuW/H09MTp06fF96tXr0Z2djafukJEREREREREkul86kpeXh6srKzE9x06dMD69esxe/Zs7Nu3DzKZTNKCFixYgM2bNyMwMBCbN2/GwoULAQBhYWG4dOlSKcMnIiIiIiIiIvo/Os/oaNiwIc6cOYOOHTuKw1q2bImoqCiEhYUhNzdX0oK8vLywbdu2IsOjoqKKHX/SpEmS5ktEREREREREVEjnGR1jxoxBenp6keGvvvoqoqOjERwcbIy4iIiIiIiIiIj0pvOMDm334WjUqBGWLl1q0ICIiIiIiIiIiEpLZ6EDAP773//ixo0b6NChA5o2bYotW7bg6NGj8Pb2xtixY2FtbW3sOImIiIiIiIiIdNJZ6Fi7di22bNmC1q1b48cff8Tbb7+N48ePo3fv3ti3bx9SU1OxaNGi8oiViIiIiIiIiEgrnYWObdu2YcuWLahduzb++ecf9OnTB0eOHIGbmxt69+6NkJAQFjqIiIiIiIiIyCTovBnpkydPULt2bQBAnTp1IJfL4eLiAgCoUaMG8vLyjBshEREREREREZFEOgsdjRs3xsqVK3Hr1i2sXLkSNWvWxJ49ewAAsbGxeOWVV4weJBERERERERGRFDoLHfPmzcPBgwfx9ttvQy6XIyIiAgsXLoSfnx8WLlyI6dOnl0ecREREREREREQ66bxHR5MmTbB3716NYYcPH8bdu3dRt25d2NnZGS04IiIiIiIiIiJ9SHq87IuqVauGatWqGToWIiIiIiIiIqIykVTo2LlzJ7Zv344bN24gJycH7u7uaN68OcaPHw8vLy9jx0hEREREREREJInOe3SsWbMG69atQ6dOnfDee+/Bzc0Nffr0Qa1atTB06FCcPHmyPOIkIiIiIiIiItJJ5xkdW7ZswdatW1GzZk0AQO/evTFx4kTExMSgVatWWLZsGXbu3Gn0QImIiIiIiIiIdNF5RocgCHBxcRHf16hRA48ePQIAdOrUCbdv3zZacERERERERERE+tB5Rscbb7yBjz76CGFhYVCr1YiKikK7du0AABkZGXzqChERERERERGZDJ1ndMyZMwe2trZ4//33ERYWBmtra8yfPx8AkJmZidmzZxs9SCIiIiIiIiIiKbSe0bF582YMHToUS5cuxe3bt1G3bl2Nz2vXro3atWsbMz4iIiIiIiIiIsm0ntGxYsUK8XX//v2NHgwRERERERERUVloPaOjTp06+PTTT9GgQQM8ffoUv/zyS7Hjvf3220YJjoiIiIiIiIhIH1oLHcuXL8fXX3+NvXv34unTp9i9e3eRcWQyGQsdRERERERERGQStBY66tWrhyVLlgAAhg8fjo0bN5ZLUEREREREREREpaHzqSuFWOQgIiIiIjK8ZZvOVnQIZGTcxkTlS2eho3Xr1pgzZw7OnuXOSURERERkaPlP1RUdAhkZtzFR+dJZ6MjPz4darcbo0aPRvXt3rFmzBvfu3SuP2IiIiIiIiIiI9KKz0GFpaYmlS5fixIkTmDBhAs6ePYvAwEAMGzYMO3fuRHZ2dnnESURERERERESkk+R7dFSpUgX9+vXDd999h99++w0dOnTAV199hQ4dOhgzPiIiIiIiIiIiybQ+dQUABEEoMszDwwNjx47F2LFj8b///U/SghISEjBz5kykp6fD0dERERERqFu3rsY4a9euRWxsLBQKBSwsLDBlyhR06tRJ0vyJiIiIiIiIiHQWOt58802tn7ds2VLSgsLDwzFkyBAEBwdj9+7dmD9/Pr7//nuNcVq0aIGRI0fC1tYW169fx9ChQ3HixAnY2NhIWgYRERERERERvdx0XrqycOFCyTPbsGFDscNTU1Nx9epVBAUFAQCCgoJw9epVpKWlaYzXqVMn2NraAgAaN24MQRCQnp4ueflERERERERE9HLTeUaHPiIjIzF69Ogiw5VKJdzc3KBQKAAACoUCrq6uUCqVcHJyKnZeu3btQp06deDu7l7ks4yMDGRkZGgMS0pKMsAaEBkWc5XMAfOUzAVzlcwFc5XMAfOUKjODFjqKu59Hafzxxx9YuXIlvv3222I/37hxI9asWSNpXmq1AOXDLKRl5MDJwRYeNapCLpcZJE4iXfTJVaIXlVf/xTytnCrj8Y+5ahiVMTdMDXOVzAF/U1FlZtBCh0xWfLJ7eHggOTkZKpUKCoUCKpUKKSkp8PDwKDLuhQsXMH36dKxbtw7169cvdn7Dhw9HSEiIxrCkpCSEhoZqDFOrBcRfUmLFlvPIK1DB2lKBKYNboX1zD+6YVC6k5irRi8qz/2KeVj6V9fjHXC27ypobpoa5SuaAv6moMjNooaMkzs7O8Pb2xp49exAcHIw9e/bA29u7yGUrFy9exJQpU7Bq1So0bdq0xPk5ODjAwcFB53KVD7PEHRIA8gpUWLHlPOp6dEZNV7uyrRSRBFJzlehF5dl/MU8rn8p6/GOull1lzQ1Tw1wlc8DfVFSZ6bwZqT60XbqyYMECbN68GYGBgdi8ebN4k9OwsDBcunQJwLMbn+bm5mL+/PkIDg5GcHAw/vrrr1LHk5aRI+6QhfIKVEjLzCn1PImIygP7LyoL5g+VhLlBRPpiv0HmyKBndPj6+pb4mZeXF7Zt21ZkeFRUlPh6+/bthgwHTg62sLZUaOyY1pYKONnbGnQ5RESGxv6LyoL5QyVhbhCRvthvkDnSWuiIj4+XNJP27dsD0CxamAKPGlUxZXCrIteTedSoWtGhERFpxf6LyoL5QyVhbhCRvthvkDnSWuiYM2eOxvuUlBQAgKOjI9LT0wEAbm5uiIuLM050ZSSXy9C+uQfqenRGWmYOnOx5h2AiMg/sv6gsmD9UEuYGEemL/QaZI62FjsOHD4uvIyMjkZ6ejg8++AC2trbIycnBqlWr4OjoaOwYy0Qul6Gmqx1vlENEZof9F5UF84dKwtwgIn2x3yBzI/lmpN999x3at2+PxYsXY+zYsbC1tUWPHj2wYcMGY8ZHRERERERERCSZ5JuRqtVqzJs3D0OHDsWBAwcAAHfu3EFeXp7RgtOHSvXs5jhJSUkVHAlVFu7u7rCwMPwTmJmrZGjGyFXmKRkDc5XMQXkf/2UyGaAuQGJiotYnGJL50ncby+VyQP0U9+/f1zo++1QyF8bqV7WRvDSFQoHHjx/jr7/+Qn5+PqZOnYrDhw/D0tLSmPFJ9uDBAwBAaGhoBUdClUVcXBxq1apl8PkyV8nQjJGrzFMyBuYqmYOKOv5//5nBF0kmRt9trGt89qlkLozVr2ojEySWjtu3b4/vvvsOv/32GyIjIzFmzBgEBARg9OjROHHihLHj1Ck3NxeXL1+Gi4sLFAqF0ZaTlJSE0NBQ/PDDD3B3dzfacsrKHOI09RiNVXnUJ1dNvY2Kw5jLx/Mxt2zZ0uC5Wl59qi7muG10qYzrBEhbL2P0q6aSqyWprNu7LEy9TUzh+F8SU287gDEaCvtU02AOuVLRKipXdZG8tDZt2uDw4cMYN24coqOjMXHiRERFRcHPz8+Y8UlmY2MDX1/fclueu7t7uVelSsMc4jSHGA2pNLlqjm3EmMuHsQ4c5d2n6mKO20aXyrhOQPmvl6nlakkq6/Yui5etTQyZq+bQdozRMNinmgZzyJWKZmptJPlmpB9++CGio6Ph7e2NzMxMBAYGYuvWrXBycjJmfEREREREREREkkkqdKjVaoSHh6N79+746quvYGtri4iICGzZsgVHjx41doxERERERERERJJIOt9ZLpfjjz/+wLfffgtLS0tYWVmhZcuWAIDU1FRjxkdEREREREREJJnkS1dsbGyK3HQ0MTERLi4uBg/KlDk4OGDixIlwcHCo6FC0Moc4zSHGimaObcSYy4c5xlwalXE9K+M6AZV3vcqK7VIU26T0zKHtGKNhmEOMLwNuB91MtY0kP3Vl4MCBuHz5Mtq3b48//vgDQUFBOHXqFOrVq4dvv/3W2HESEREREREREekk+Vb99evXh1qtxp9//gmVSoWDBw+iXr16cHNzM2Z8RERERERERESSSSp0qFQqCIKAH3/8EVZWVsaOiYiIiIiIiIioVCQVOhQKBX7//XfcuXMH+/fvR0pKClxdXREUFIS6desaOUQiIiIiIiIiImkk34y0Y8eO6NevH27evIlq1aohISEBb731FuLi4owZn2RPnz7FvXv38PTp04oOhUgr5iqZA+YpmQvmKpkL5iqZA+YpVRaSCx0xMTFQqVSIi4vDr7/+inPnzsHS0hKTJ082ZnySJSUloWvXrkhKSqroUIi0Yq6SOWCekrlgrpK5YK6SOWCeUmUh+WakVlZWWLduHRQKhThMpVJh/PjxRgmMiIiIiIiIiEhfkgsdTZs2RWRkJJRKpXiPDg8PD7z66qvGjI+IiIiIiIiISDLJl644OjriwoULSE9PR+3atZGeno4LFy7A0dHRiOEREREREREREUkn+YyOM2fO4MiRI/jnn3/EMzrq1q2L3r1765w2IiICBw4cwP379xETE4NGjRoVGUelUuHjjz/G8ePHIZPJMHr0aAwYMEC/tdGTWi1A+TALaRk5cHKwhUeNqpDLZUZdZlljKWvML07v5lQFyWnZeJKTB7UaeJSZhxqOtqjvWQ0WFpLrYCZBn7YxpW3/YjzV7W0gl8uQ+vj/YgOgM97SrJOu5VZkm5SWlHYoz+1varlWVpVtfajscnOf4h9lOlIf56FaVSvY2ChQx8UBNjaSv2IYhSnlan6+Crfup+Ph41zUqGYDr5qOsLJSlEucarWA+w+eIDk1CzbWFnBysIG7s37Hh4puP0OqrOtFpkvfnFOrBdxLyUBqRh6ycwpQ3cEar7jao2pV63KMmsj8Sf4WYm9vj3v37sHX11ccdvHiRTg4OOictmvXrnj33XcRGhpa4jgxMTG4c+cODh48iPT0dPTr1w/t27dHrVq1pIaoF7VaQPwlJVZsOY+8AhWsLRWYMrgV2jf3KPcDntRYyhrzi9N7OFfBwG6NcfBUAgLavIKvd18W5zu2fwt0blXLbIod+rSNKW37kuJ5p3sj7D2ZgMysAnw0zBcFT9Va4y3NOulabkW2SWlJaYfy3P76LsvUv4Cb2r5DFS839ymOX7yPr3ZcEnMiLLgZUh/lwKeRW4UVO0wpV/PzVTh64R6+2vl/bTQmpDn8fZ4dY40ZZ0n9fB13B/h6u0EulxXb7wAwmfYzJFPKC3o5lOZ7wJmrSbibnImfDv393Pfy5mjf1J3FDiI9SP4Vm52djenTp2Pr1q04evQofv75Z3z00UfIyclBfHy8+K84vr6+8PDw0Dr/2NhYDBgwAHK5HE5OTujWrRv2799f7LgZGRm4d++exj997wysfJgldjoAkFegwoot56F8mKXXfAxBaiz3Hzwpdrz7D56UajmdWtZC5I6LCOrkJRY5CucbueMi/kl8bKhVNDp9tmd5bnspuVpcPD8d+hsBresgr0CFG3fTdcZbmnXStdyK2h/KQko7lOf212dZhV+GPlh+BLPXn8QHy48g/pISarVg8LheJLVPNaV+k0zDrcR0scgBPMuJqN2X8VT17DNDM8dcvXU/XSxyFMby1c5LuHU/3ehxltTP37j7bNkl9Tslfd8w933d1I7/VPnpm3P3HzzBzXuPxSJH4TSROy4hISnT4PExT6kyk/ynlrS0NNjY2CAyMlJjeGpqKubMmQMAkMlkiIuLK1UgSqUSnp6e4nsPD48Sd7SNGzdizZo1pVpOobSMHLEDKZRXoEJaZg5qutqVad7GiiU5NavY8ZJTs1DbzV7/5cieTZ+bpyp2vqmPcwBU13+FKoA+27M8t72UXC0pHvz/Qr9aEHTGW5p10rXcitofykJKO5Tn9tdnWSV9Garr0dno20Bqn2pK/SaZhtTHucXmRE7+U6RlGL5IZ465+rCENkp9nIunKrVR4yypHdSCgLTMHAAott+Z+a6vybSfIZna8Z8qP31zLjk1q8TvfY8ycw0eH/OUKjPJhY769etj2bJlaNq0qTjsypUrmDp1aolnXhjL8OHDERISojEsKSlJ66UxL3JysIW1pUKjI7G2VMDJ3tZgcRo6Fhtri2LHs7GWthlLWo6tjaLY4c7Vyr8tSkuf7Vme215KrpYUD/7/bwS5TKYz3tKsk67lVtT+UBZS2qE8t78+y6rIH2ZS+1RT6jfJNDhXsyn+uGJlger2hj/F2hxztUYJbeRczQYOVa2NGmdJ7SCXyeBkb1tiv2NjU/z3DXPf103t+E+Vn745Z2NtUeL3vur2NgaPj3lKlZnkS1fee+89jB8/Hps2bcLRo0exadMmTJw4ESNGjDBIIB4eHkhMTBTfK5VKuLu7Fzuug4MDatWqpfGvpHFLXF6NqpgyuNWzH3aAeM1c4bWp5UlqLE4ONnineyON8d7p3ghODtI6vheXc/zCPYzt3wIxx25hVHAzjfmO7d8C9T2rGWoVjU6f7Vme215KrhYXzzvdG+HwuTuwtlSgYW1HnfGWZp10Lbei9oeykNIO5bn99VlW4Zeh55XXDwupfaop9ZtkGrw8HTGmf3ONnAgLbgYLBeBUzfCFDnPMVa+ajhgTotlGY0Kaw6umo9HjLKmfb1j72bJL7ndsTKb9DMnUjv9U+embc04ONnBzKvp9f2z/5qjnrvvsbX0xT6kykwmCIPnc0n379uGXX35BUlIS3N3dMWDAAPTs2VPywgICAhAZGVnsU1d27NiBvXv3IioqSrwZ6Q8//IDatWtLmve9e/fQtWtXxMXFSb6BqXgDrswcONmbyFNXtMSiVgs4ey0ZN+6mQy0IkMtkaFjbUbyhWGmWIz51JTcPahXw6EkealQz86euSNieFbnti8vV5+NxtLOBQi5Dasb/xQZAUn7ou066lmuON2eTui+V1/aXuixTu0leSX2qKfWbZBpycgpwK/ExHmU+e+qKwkIGlUqNZvVdyiU3zCFXC5+6kvo4F84lPXXFSHFqe+qKtn4H0H3cMUemdvynyk/f76dnryUjNy8f9lVtkJ33FI52VnjF1Q52doY/o6M4zFOqLPQqdJTWxx9/jIMHD+Lhw4eoXr06HB0dsXfvXoSFhWHy5Mlo3rw5VCoVFi1ahN9//x0AEBYWhkGDBklexsuyU5rSFzcqnZclV0l/prR/M09JH/zxaL5Mqd+p7JirJEVF75PMU6osJN+jY8+ePfD29oaXlxcSEhIwb948yOVyhIeHw8vLS+u0c+fOxdy5c4sMj4qKEl8rFAosXLhQj9BfTnK5DDVd7cz6ZmBEVDzu32SumLvmi9uOyLRwnyQyDMnXJnz55ZeoVu3ZPRsiIiLQvHlztGnThsUJIiIiIiIiIjIZej1etkaNGsjLy8O5c+ewatUqWFhYoF27dsaMj4iIiIiIiIhIMsmFDicnJ/z777/4+++/0bx5c1hZWSEnJwflcIsPIiIiIiIiIiJJJBc6xo8fj/79+0OhUGDFihUAgPj4eDRp0sRowRERERERERER6UNyoaN///7o1asXAMDW1hYA8Nprr2H58uXiOOfOnUPr1q0NHCIRERERERERkTSSb0YKPCtwFBY5AMDZ2RkuLi7i+7CwMMNFRkRERERERESkJ70KHbrwfh1EREREREREVJEMWuiQyWSGnB0RERERERERkV4MWuggIiIiIiIiIqpILHQQERERERERUaXBe3QQERERERERUaWh9fGyarVa0kzk8mf1kgsXLpQ9IiIiIiIiIiKiUtJa6Hj11Vcl3WD02rVrBguIiIiIiIiIiKi0tBY64uLixNdHjhzBgQMHMGbMGHh6eiIxMRFRUVHo0aOH0YMkIiIiIiIiIpJCa6GjZs2a4uvvvvsO27dvh4ODAwCgXr16aNasGd566y0MGTLEuFESEREREREREUkg+WakmZmZyMnJ0RiWm5uLzMxMgwdFRERERERERFQaWs/oeF5ISAhGjBiB4cOHw93dHUlJSdi0aRNCQkKMGR8RERERERERkWSSCx3Tp09HnTp1EBsbi5SUFLi4uCA0NBQDBw40ZnxERERERERERJJJLnTI5XIMHjwYgwcPNmY8RERERERERESlJvkeHYIgYOvWrRg+fDj69u0LADhz5gxiY2ONFhwRERERERERkT4kFzpWrlyJX375BQMHDoRSqQQAuLu74+uvvzZacERERERERERE+pBc6Ni5cyciIyPRp08fyGQyAECtWrVw9+5dowVHRERERERERKQPyffoUKlUqFq1KgCIhY6srCxUqVJF0vQJCQmYOXMm0tPT4ejoiIiICNStW1djnNTUVMyaNQtKpRIFBQVo164d5s6dCwsLyWESERERERER0UtM8hkd/v7+WLp0KfLz8wE8u2fHypUr0aVLF0nTh4eHY8iQIThw4ACGDBmC+fPnFxknMjISXl5eiImJQUxMDK5cuYKDBw9KDZGIiIiIiIiIXnKSCx2zZs1CSkoKWrdujczMTPj4+CAxMRHTpk3TOW1qaiquXr2KoKAgAEBQUBCuXr2KtLQ0jfFkMhmysrKgVquRn5+PgoICuLm56blKRERERERERPSyknxNiJ2dHdatW4fU1FTcv38fHh4ecHFxkTStUqmEm5sbFAoFAEChUMDV1RVKpRJOTk7ieOPHj8ekSZPQsWNH5OTkIDQ0FK1bty4yv4yMDGRkZGgMS0pKkroqROWGuUrmgHlK5oK5SuaCuUrmgHlKlZnkQsesWbPQu3dvdOrUCc7OzuLwBQsWYMGCBQYJZv/+/WjcuDE2btyIrKwshIWFYf/+/ejZs6fGeBs3bsSaNWsMskwiY2KukjlgnpK5YK6SuWCukjlgnlJlJhMEQZAyYrNmzeDk5IThw4fj/fffF4e3atUK58+f1zptamoqAgMDcfr0aSgUCqhUKvj5+eHgwYMaZ3QEBQXhk08+QYsWLQAAGzZsgFKpRHh4uMb8Sqo+hoaGIi4uDrVq1ZKySkRGx1wlc8A8JXPBXCVzwVwlc8A8pcpM8hkdVlZW2Lp1K8aPH4/r169jyZIlsLKygpQ6ibOzM7y9vbFnzx4EBwdjz5498Pb21ihyAM8eV3vs2DG0aNEC+fn5iI+PR/fu3YvMz8HBAQ4ODlJDJ6owzFUyB8xTMhfMVTIXzFUyB8xTqswk34wUANzd3fHjjz9CrVZj8ODBSE5OFh81q8uCBQuwefNmBAYGYvPmzVi4cCEAICwsDJcuXQIAzJ49G+fOnUPfvn3Rr18/1K1bFwMHDtRzlYiIiIiIiIjoZSX5jI7CMzdsbGzwxRdfYMOGDXj77bfFx83q4uXlhW3bthUZHhUVJb6uU6cOoqOjpYZERERERERERKRBcqFjwoQJGu9Hjx6Nxo0bY//+/QYPioiIiIiIiIioNCQXOkaNGlVkmL+/P/z9/Q0aEBERERERERFRaWktdLz//vv45ptvAABDhgwp8X4cP/zwg+EjIyIiIiIiIiLSk9ZCR79+/cTXAwYMMHYsRERERERERERlorXQ0bdvX/F1SEiI0YMhIiIiIiIiIioLyffoAICzZ8/i6tWryM7O1hg+duxYgwZFRERERERERFQakgsdixcvxr59++Dr6wtra2txeEn37SAiIiIiIiIiKm+SCx0xMTGIiYmBm5ubMeMhIiIiIiIiIio1udQR3d3dYWVlZcxYiIiIiIiIiIjKRPIZHUuWLMG8efPQp08f1KhRQ+OzNm3aGDwwIiIiIiIiIiJ9SS50XLlyBceOHcOZM2dgY2MjDpfJZDhy5IgxYiMiIiIiIiIi0ovkQseKFSsQGRmJ119/3ZjxEBERERERERGVmuR7dNja2sLX19eYsRARERERERERlYnkQsfkyZPxySef4MGDB1Cr1Rr/iIiIiIiIiIhMgeRLV2bPng0A+Pnnn8VhgiBAJpPh2rVrho+MiIiIiIiIiEhPkgsdcXFxxoyDiIiIiIiIiKjMJBc6atasacw4iIiIiIiIiIjKTGuhY968eVi8eDEAYPr06ZDJZMWO99lnnxk+MiIiIiIiIiIiPWktdNSqVUt8/corrxg9GCIiIiIiIiKistBa6BgzZoz4etCgQXBxcSkyzoMHDwwfFRERERERERFRKUh+vGxgYGCxw/v06WOwYIiIiIiIiIiIykJyoUMQhCLDnjx5UuJ9O4iIiIiIiIiIypvOp674+/tDJpMhLy8PnTt31vgsPT2dZ3QQERERERERkcnQWehYtmwZBEHA6NGjNZ6uIpPJ4OzsjPr160taUEJCAmbOnIn09HQ4OjoiIiICdevWLTJebGws1q9fD0EQIJPJEB0djRo1akhfIyIiIiIiIiJ6aeksdLRt2xYAcOrUKdja2pZ6QeHh4RgyZAiCg4Oxe/duzJ8/H99//73GOJcuXcKaNWuwceNGuLi4IDMzE1ZWVqVeJhERERERERG9XHQWOgopFAqsXLkSe/fuRUpKClxdXdG7d2+MGzcO1tbWWqdNTU3F1atXER0dDQAICgrC4sWLkZaWBicnJ3G87777DiNHjhSf7mJvb1/s/DIyMpCRkaExLCkpSeqqEJUb5iqZA+YpmQvmKpkL5iqZA+YpVWaSCx0LFixAQkIC5syZg5o1a+L+/fvYsGEDkpOTsXTpUq3TKpVKuLm5QaFQAHhWNHF1dYVSqdQodNy6dQu1atVCaGgosrOz0b17d4wbN67IDU83btyINWvW6LOeRBWCuUrmgHlK5oK5SuaCuUrmgHlKlZnkQkdcXBwOHToEBwcHAECDBg3w2muvoUePHgYLRqVS4a+//kJ0dDTy8/MxatQoeHp6ol+/fhrjDR8+HCEhIRrDkpKSEBoaarBYiAyBuUrmgHlK5oK5SuaCuUrmgHlKlZnkQkeNGjWQk5MjFjoAIC8vT7zMRBsPDw8kJydDpVJBoVBApVIhJSUFHh4eGuN5enqiZ8+esLKygpWVFbp27YqLFy8WKXQ4ODhoxEFkqpirZA6Yp2QumKtkLpirZA6Yp1SZyaWOGBwcjFGjRmHr1q04evQofv75Z4SFhSE4OBjx8fHiv+I4OzvD29sbe/bsAQDs2bMH3t7eGpetAM/u3XHixAkIgoCCggKcOnUKTZo0KcPqEREREREREdHLRPIZHT/99BMAIDIyssjwws9kMhni4uKKnX7BggWYOXMm1q1bBwcHB0RERAAAwsLCMHnyZDRv3hx9+vTB5cuX0bt3b8jlcnTs2BFvv/12qVaMiIiIiIiIiF4+kgsdhw8fLtOCvLy8sG3btiLDo6KixNdyuRyzZs3CrFmzyrQsIiIiIiIiIno5SS50AMDTp09x4cIFJCcnw93dHS1btoSFhV6zICIiIiIiIiIyGslVilu3bmHcuHHIzc2Fh4cHlEolrK2tERkZCS8vL2PGSEREREREREQkieRCx8KFCzFw4EC8//77kMlkAIBvvvkGCxYswKZNm4wWIBERERERERGRVJKfunL9+nWMGDFCLHIAz569fP36daMERkRERERERESkL8mFDldXV/zxxx8aw86ePQtXV1eDB0VEREREREREVBqSL12ZMmUKxo8fj86dO8PT0xOJiYk4cuQIli1bZsz4iIiIiIiIiIgkk3xGR9euXbFjxw40bNgQWVlZaNiwIXbs2IFu3boZMz4iIiIiIiIiIsn0ejZsvXr1MH78+BI/79u3L2JiYsocFBERERERERFRaUg+o0OKe/fuGXJ2RERERERERER6MWih4/knshARERERERERlTeDFjqIiIiIiIiIiCoSCx1EREREREREVGkYtNAhCIIhZ0dEREREREREpBe9Cx1qtRopKSnFfrZo0aIyB0REREREREREVFqSCx0ZGRmYOnUqWrRogR49egAA4uLisGLFCnGcvn37Gj5CIiIiIiIiIiKJJBc6wsPDYWdnh8OHD8PS0hIA4OPjg3379hktOCIiIiIiIiIifVhIHTE+Ph7Hjx+HpaWl+BhZJycnpKamGi04IiIiIiIiIiJ9SD6jw97eHo8ePdIYlpiYCBcXF4MHRURERERERERUGpILHQMGDMDkyZNx6tQpqNVqXLhwATNmzMA777xjzPiIiIiIiIiIiCSTfOlKWFgYrKyssGjRIjx9+hSzZ8/GoEGDMHz4cGPGR0REREREREQkmdZCR0REBGbMmAEAOHXqFN577z2899575REXEREREREREZHetF66snXrVvH1hAkTjB4MEREREREREVFZaD2jo0mTJpg8eTK8vLyQn5+PlStXFjveBx98oHNBCQkJmDlzJtLT0+Ho6IiIiAjUrVu32HH/+ecfhISEYMiQIeIZJUREREREREREumg9o2PVqlVo0qQJHjx4AABISkoq9p8U4eHhGDJkCA4cOIAhQ4Zg/vz5xY6nUqkQHh6Obt266bkqRERERERERPSy03pGh7OzM8aPHw/gWQFi6dKlpVpIamoqrl69iujoaABAUFAQFi9ejLS0NDg5OWmMu2HDBnTu3BnZ2dnIzs4u1fKIiIiIiIiI6OUk+akr8+bNw9WrV5GVlYWqVauibt26qFKliqRplUol3NzcoFAoAAAKhQKurq5QKpUahY7r16/jxIkT+P7777Fu3boS55eRkYGMjAyNYVLPLCEqT8xVMgfMUzIXzFUyF8xVMgfMU6rMdBY6MjIysGDBAhw6dAgWFhawt7fHkydPUFBQgB49eiA8PBwODg5lDqSgoADz5s3D0qVLxYJISTZu3Ig1a9aUeZlExsZcJXPAPCVzwVwlc8FcJXPAPKXKTCYIgqBthIkTJ8La2hoffvghateuLQ6/e/cuVq1ahZycHJ07SGpqKgIDA3H69GkoFAqoVCr4+fnh4MGD4hkdiYmJCAkJQdWqVQE8K7AIgoDevXtj8eLFGvMrqfoYGhqKuLg41KpVS3oLEBkRc5XMAfOUzAVzlcwFc5XMAfOUKjOdZ3T8/vvvOHnyJGxtbTWG165dGwsXLkSHDh10LsTZ2Rne3t7Ys2cPgoODsWfPHnh7e2tctuLp6YnTp0+L71evXo3s7Oxin7ri4OBgkLNIiIyNuUrmgHlK5oK5SuaCuUrmgHlKlZnWp64AQPXq1XH16tViP7t27RocHR0lLWjBggXYvHkzAgMDsXnzZixcuBAAEBYWhkuXLkmPmIiIiIiIiIioBDrP6JgyZQrCwsIQEBCAJk2aiPfouH79Ov773/+KBQtdvLy8sG3btiLDo6Kiih1/0qRJkuZLRERERERERFRIZ6Gjb9++aNKkCfbs2YPz588jOzsbVapUQcOGDfHTTz+hQYMG5REnEREREREREZFOkh4v27BhQ0yZMsXYsRARERERERERlYmkQsejR49w8OBB3LhxA1lZWahatSoaNmyIHj16oHr16saOkYiIiIiIiIhIEp03I42Pj0ePHj3w66+/QhAEuLq6AgBiYmIQGBiIU6dOGT1IIiIiIiIiIiIpdJ7RsXjxYixZsgQ9evQo8tmhQ4ewcOFC7Nu3zyjBERERERERERHpQ+cZHYmJiejcuXOxn/n7+yMxMdHQMRERERERERERlYrOQkeLFi2wYsUKZGdnawzPzs7Gl19+iRYtWhgtOCIiIiIiIiIifei8dGXp0qWYOnUq2rVrh9q1a8Pe3h5PnjzB3bt34e3tjeXLl5dHnEREREREREREOuksdNSsWRM//fQTbt++jZs3b4pPXWnQoAHq1q1bDiESEREREREREUkj6fGyAFC3bl0WNoiIiIiIiIjIpOksdNy7dw+1atUS38fGxmL//v0AgK5duyI4ONh40RERERERERER6UHnzUjffPNN8fWWLVvwySefoFmzZmjevDm++OIL/PDDD0YNkIiIiIiIiIhIKp1ndAiCIL7+8ccfsXr1avj4+AAA/Pz8MHPmTISGhhovQiIiIiIiIiIiiXSe0SGTycTXKSkpaNmypfi+RYsWSEpKMkpgRERERERERET60nlGR35+Pj766CMAgFqtxsOHD+Hi4gIAyMjIgKWlpXEjJCIiIiIiIiKSSGehY+zYseLr4cOHIyMjQyx0nDlzBh07djRedEREREREREREetBZ6Jg4cWKJn3Xt2hVdu3Y1aEBERERERERERKWls9BRKCEhATdv3kRWVhaqVq2KBg0aoF69esaMjYiIiIiIiIhILzoLHYmJiZgyZQquX7+OOnXqwM7ODk+ePMHdu3fRpEkTLF++HJ6enuURKxERERERERGRVjoLHbNmzULr1q3x3XffwdbWVhyenZ2NtWvXYubMmfj++++NGiQRERERERERkRQ6Cx0XL15EVFQUrKysNIZXqVIFH3zwAdq2bWu04IiIiIiIiIiI9CHXNYK7uzuOHDlS7GdHjx6Fh4eHoWMiIiIiIiIiIioVnWd0zJ8/H5MmTUJ0dDSaNGkCe3t7PHnyBNeuXcPNmzexatWq8oiTiIiIiIiIiEgnnYWO9u3b49ChQzh06BBu3LiB1NRUVKlSBSEhIejWrRucnJwkLSghIQEzZ85Eeno6HB0dERERgbp162qMs3btWsTGxkKhUMDCwgJTpkxBp06dSrViRERERERERPTykfR42erVq2PgwIFlWlB4eDiGDBmC4OBg7N69G/Pnzy9yE9MWLVpg5MiRsLW1xfXr1zF06FCcOHECNjY2ZVo2EREREREREb0cJBU6bt26hd27d+PGjRvIyspC1apV0bBhQwQHB8PLy0vn9Kmpqbh69Sqio6MBAEFBQVi8eDHS0tI0zgh5/uyNxo0bQxAEpKenw93dXWN+GRkZyMjI0BiWlJQkZVWIyhVzlcwB85TMBXOVzAVzlcwB85QqM52Fjj179mDBggUICAhAmzZtxHt0XL9+He+88w4WLlyI3r17a52HUqmEm5sbFAoFAEChUMDV1RVKpbLES1927dqFOnXqFClyAMDGjRuxZs0aKetHVKGYq2QOmKdkLpirZC6Yq2QOmKdUmeksdCxfvhxfffUVWrduXeSzc+fOYfr06ToLHfr6448/sHLlSnz77bfFfj58+HCEhIRoDEtKSkJoaKhB4yAqK+YqmQPmKZkL5iqZC+YqmQPmKVVmOgsdjx49QtOmTYv97NVXX8WjR490LsTDwwPJyclQqVRQKBRQqVRISUkp9tG0Fy5cwPTp07Fu3TrUr1+/2Pk5ODjAwcFB53KJKhpzlcwB85TMBXOVzAVzlcwB85QqM7muEV5//XXMnj0bd+7c0Rh+584dzJ07F6+//rrOhTg7O8Pb2xt79uwB8OxyGG9v7yKXrVy8eBFTpkzBqlWrSiyuEBERERERERGVRGeh45NPPgEA9O7dGy1btkTHjh3h4+ODPn36aHyuy4IFC7B582YEBgZi8+bNWLhwIQAgLCwMly5dAgAsXLgQubm5mD9/PoKDgxEcHIy//vqrVCtGRERERERERC8fnZeuVKtWDcuXL0dOTg5u374tPnWlbt26sLW1lbwgLy8vbNu2rcjwqKgo8fX27dslz4+IiIiIiIiI6EWSHi8LALa2tqhduzZyc3Ph7OwMmUxmzLiIiIiIiIiIiPSms9ChUqmwfv16bN++XXyusoWFBZo2bYoPPvgA7du3N3qQRERERERERERS6Cx0LFmyBFeuXMHUqVOhVquxefNmdOvWDdWqVcOsWbMwffp08X4dREREREREREQVSWehY+/evdi/fz+qV68OAGjfvj2GDBmCQ4cOoXHjxpg1axYLHURERERERERkEnQ+dcXKygoWFv9XD1EoFMjLywMAtGzZUrychYiIiIiIiIioouk8o6NXr14YN24chg0bBkEQ8P3336NHjx4AgJSUFPFMDyIiIiIiIiKiiqaz0DFt2jRs2LABGzZsgCAIeOONNzBu3DgAgCAI+Oyzz4weJBERERERERGRFDoLHVZWVpg4cSImTpxY5DM3Nze4ubmJ7zds2IDRo0cbNkIiIiIiIiIiIol03qNDH5GRkYacHRERERERERGRXnSe0aEPQRAMOTuDUKsFKB9mIS0jB04OtvCoURVyuayiwyIi0on9F5UF84dKwtwwHdwWRETGYdBCh0xmWh2zWi0g/pISK7acR16BCtaWCkwZ3Artm3vwIEJEJo39F5UF84dKwtwwHdwWRETGY9BLV0yN8mGWePAAgLwCFVZsOQ/lw6wKjoyISDv2X1QWzB8qCXPDdHBbEBEZj0ELHaZ26UpaRo548CiUV6BCWmZOBUVERCQN+y8qC+YPlYS5YTq4LYiIjMeghQ5fX19Dzq7MnBxsYW2p0BhmbamAk71tBUVERCQN+y8qC+YPlYS5YTq4LYiIjEfrPTri4+MlzaR9+/YAgKioqLJHZEAeNapiyuBWRa599KhRtaJDIyLSiv0XlQXzh0rC3DAd3BZERMajtdAxZ84cjfcpKSkAAEdHR6SnpwMA3NzcEBcXZ5zoykgul6F9cw/U9eiMtMwcONnzbtZEZB7Yf1FZMH+oJMwN08FtQURkPFoLHYcPHxZfR0ZGIj09HR988AFsbW2Rk5ODVatWwdHR0dgxlolcLkNNVzvUdLWr6FCIiPTC/ovKgvlDJWFumA5uCyIi45B8j47vvvsOU6dOha3ts+sGbW1t8Z///AfR0dFGC46IiIiIiIiISB9az+h4XpUqVXDx4kW0bt1aHHbp0iWx8FHRVKpnd61OSkqq4EiosnB3d4eFheRdRDLmKhmaMXKVeUrGwFwlc8DjP5kL9qlkLozVr2ojeWmTJ0/GqFGjEBAQAHd3dyQlJeG///0v5s+fb8z4JHvw4AEAIDQ0tIIjocoiLi4OtWrVMvh8matkaMbIVeYpGQNzlcwBj/9kLtinkrkwVr+qjUwQBEHqyDdv3sSBAweQkpICFxcX9OzZEw0aNDBmfJLl5ubi8uXLcHFxgUKh0D1BKSUlJSE0NBQ//PAD3N3djbacsjKHOE09RmNVHvXJVVNvo+Iw5vLxfMwtW7Y0eK6WV5+qizluG10q4zoB0tbLGP2qqeRqSSrr9i4LU28TUzj+l8TU2w5gjIbCPtU0mEOuVLSKylVd9FpagwYNUL9+fTx8+BCurq7GiqlUbGxs4OvrW27Lc3d3L/eqVGmYQ5zmEKMhlSZXzbGNGHP5MNaBo7z7VF3McdvoUhnXCSj/9TK1XC1JZd3eZfGytYkhc9Uc2o4xGgb7VNNgDrlS0UytjSTfjDQjIwNTp05FixYt0KNHDwDPTkFZsWKF0YIjIiIiIiIiItKH5EJHeHg47OzscPjwYVhaWgIAfHx8sG/fPqMFR0RERERERESkD8nnO8fHx+P48eOwtLSETCYDADg5OSE1NdVowRERERERERER6UPyGR329vZ49OiRxrDExES4uLgYPChT5uDggIkTJ8LBwaGiQ9HKHOI0hxgrmjm2EWMuH+YYc2lUxvWsjOsEVN71Kiu2S1Fsk9Izh7ZjjIZhDjG+DLgddDPVNpL81JUNGzbg8OHD+PDDDzFx4kRERUVh+fLl6Nq1K9577z0jh0lEREREREREpJvkQocgCNi4cSO2bt2KxMREeHh4YNCgQRg+fLh4KQsRERERERERUUWSVOhQqVSYPXs2Fi9eDCsrq/KIi4iIiIiIiIhIb5Lu0aFQKPD777/zzA0iIiIiIiIiMmmSL12JiopCZmYmJk6c+FKc1REREYEDBw7g/v37iImJQaNGjYqMo1Kp8PHHH+P48eOQyWQYPXo0BgwYYFIxrl69Gj/++CNcXV0BAK1atUJ4eHi5xfjo0SN89NFHuHPnDqysrPDKK69g0aJFcHJy0hivotuyvCUkJGDmzJlIT0+Ho6MjIiIiULduXY1xtLVJRbSXlJjXrl2L2NhYKBQKWFhYYMqUKejUqROAistFKXFri81U2/qjjz7CX3/9Jb7/66+/sHbtWnTt2rXC9/vSkLLOABAbG4v169dDEATIZDJER0ejRo0a5R+wRFLWKzU1FbNmzYJSqURBQQHatWuHuXPnwsJC8oPRypU5HB9NidTcflkEBATAysoK1tbWAIBp06aJx4mXnTkcZ83hmGrqx9Cy9qHsX42jrLld2ZnlsV+Q6I033hBeffVVoVmzZsIbb7wh+Pv7i/9XRmfOnBESExOFLl26CH/99Vex4+zcuVMYOXKkoFKphNTUVKFTp07C3bt3TSrGVatWCZ9++mm5xfSiR48eCadOnRLff/rpp8KsWbOKjFfRbVnehg0bJuzatUsQBEHYtWuXMGzYsCLjaGuTimgvKTEfO3ZMyM7OFgRBEK5duya0bt1ayMnJEQSh4nJRStzaYjPVtn7etWvXhLZt2wp5eXmCIFT8fl8aUtb54sWLQq9evYSUlBRBEAQhIyNDyM3NLdc49SVlvT7++GNxe+Xn5wtvv/22sHfv3nKNUx/mcHw0Jfruz5Wdtrx52ZnDcdYcjqmmfgwtax/K/tU4yprblZ05HvslP1522bJliI6OxjfffINly5bhs88+E/+vjHx9feHh4aF1nNjYWAwYMAByuRxOTk7o1q0b9u/fX04RSouxojk6OsLPz09837JlSyQmJhYZr6Lbsjylpqbi6tWrCAoKAgAEBQXh6tWrSEtL0xhPW5uUd3tJjblTp06wtbUFADRu3BiCICA9Pd1ocekiNW5tTLWtn/fLL7+gb9++Znu2ndR1/u677zBy5Ejxseb29vbiX4VNkdT1kslkyMrKglqtRn5+PgoKCuDm5lYRIUtiDsdHU2GIPoheDuZwnDWHY6o5HEPL2oeyfzU89tW6meOxX3Kho2XLloiPj8fcuXMxevRozJ07FydPnsRrr71mzPhMmlKphKenp/jew8MDSUlJFRhR8fbu3Yu+ffti5MiRuHDhQoXFoVarsWXLFgQEBBT5zFza0hCUSiXc3NygUCgAPLsHjqurK5RKZZHxSmqT8m4vqTE/b9euXahTpw7c3d3FYeWdi/rEXVJspt7W+fn5iImJwVtvvaUx3FT2eymkrvOtW7dw9+5dhIaGIiQkBOvWrYMg7erLCiF1vcaPH4+EhAR07NhR/Ne6deuKCNlgXqY+XZvS9J0vg2nTpqFv375YsGABMjIyKjock2AOx1lzOKZWlmOoKX0HfBkYIrfJ9HJTcqFjwYIFOHXqFObMmYNffvkFc+bMwZkzZ7BgwQIjhkdl9c477yAuLg4xMTF4//33MX78eDx69KhCYlm8eDGqVKmCoUOHVsjyqfz88ccfWLlyJb744gtxmCnl4otMOTZdfvvtN3h6esLb21scZs7ro41KpcJff/2F6OhobNq0CceOHcPu3bsrOqwy279/Pxo3bowTJ07g2LFjOHv2LP86R5XWDz/8gF9//RXbt2+HIAhYtGhRRYdklkz5OGsqcUjxMh1DqeyYG+ZFcqEjLi4OX331Ffz9/dGgQQP4+/tj7dq1iIuLM2Z8Js3Dw0PjMgylUqlRVTcFLi4usLS0BAB06NABHh4euHHjRrnHERERgX///Rdffvkl5PKiaWcObWkoHh4eSE5OhkqlAvDsx1tKSkqR08G0tUl5t5fUmAHgwoULmD59OtauXYv69euLwysiF6XGrS02U25rANi+fXuRv0SZyn4vldR19vT0RM+ePWFlZQU7Ozt07doVFy9erIiQJZG6Xps3b8abb74JuVwOe3t7BAQE4PTp0xURssG8TH26Nvruzy+DwnW3srLCkCFDcP78+QqOyDSYw3HWHI6pleUYakrfAV8GhshtMr3clFzoqFGjBnJycjSG5eXliddKv4x69uyJbdu2Qa1WIy0tDb/99hsCAwMrOiwNycnJ4utr167h/v37qFevXrnGsGLFCly+fBlr164t8fpHc2hLQ3F2doa3tzf27NkDANizZw+8vb2LPIlGW5uUd3tJjfnixYuYMmUKVq1ahaZNm2p8VhG5KDVubbGZalsDQFJSEs6dOydeUyplfUyR1HUOCgrCiRMnIAgCCgoKcOrUKTRp0qQiQpZE6nrVqlULx44dA/DsNOr4+Hg0bNiw3OM1pJepT9dGn/35ZZCdnY3MzEwAgCAIiI2N1fhL+svMHI6z5nBMrSzHUFP6DvgyMERuk+nlpuTHy27YsAExMTEYNmwY3NzckJSUhB9++AFBQUFo3ry5OF779u2NFmx5+vjjj3Hw4EE8fPgQ1atXh6OjI/bu3YuwsDBMnjwZzZs3h0qlwqJFi/D7778DAMLCwjBo0CCTinHGjBm4cuUK5HI5LC0tMXnyZPj7+5dbjDdu3EBQUBDq1q0LGxsbAM++1K9du9ak2rK83bp1CzNnzkRGRgYcHBwQERGB+vXrS26TimgvKTG/9dZbuH//vsaNFD/77DM0bty4wnJRStzaYjPVtgaA9evX4++//8aKFSs0pq/o/b40pKyzWq1GREQEjh07Brlcjo4dO2LGjBnFniVmKqSs1507dxAeHo6HDx9CpVLBz88Pc+bMMdnHy5rD8dGUlJQDL6O7d+9i0qRJUKlUUKvV8PLywty5c8VHNb7szOE4aw7HVFM/hpa1D2X/ahxlze3KzhyP/ZILHcXdQLLIzGSyl/pSFiIiIiIiIiKqWJILHUREREREREREps50z/clIiIiIiIiItITCx1EREREREREVGmw0EFERERERERElQYLHURERERERERUabDQQUSSJSYmwsfHByqVymjLWL16NaZNm2a0+RMxx+hlFxAQgJMnT1Z0GC8lQ7T9P//8g379+sHHxwfff/+9gSLT371799C4cWM8ffq0wmLQJTIyEnPmzCnVtL/++itGjhxp0HjMoc0KVeZ+4uHDhwgNDYWPjw8+/fTTig4HAHPVGFjoMHH8Qk7GdPr0abzxxhuSx/f09MSFCxegUCiMGBWR4eib46XBfprKqrgvlDt27MDgwYMrMCqqKPPnz4ePj4/4r1mzZvDx8RE///rrr9G2bVtcuHAB7777LjIyMjBjxgy0b98e7du3x+rVqzXmd+3aNQwZMgStW7fGG2+8gTVr1oifHTlyBIMHD4avry86dOiAuXPn4smTJxrTnzx5EiEhIWjZsiX8/f0RGxtr3AYopeL6+7Fjx2LJkiWlmt+bb76Jb7/91hChVVr5+fn45JNP0LFjR7Rp0wYLFixAQUGB+Hl6ejomTJiAli1bokuXLoiJidGYPj4+Hj179sRrr72GYcOG4f79++Jnq1evRtOmTTX2hbt374qf37t3D8OGDcNrr72Gnj17FinKxMTEoEuXLmjZsiXGjx+P9PR08bOff/4Z1atXx/nz5zFz5kwDt4puzNXywUKHGUlISMC4cePQrl07tG3bFu+//z7++eefig6LTJg5V2ENzZhnoejC7WA8Fd22Fb38F1Vknr8sTG2bV0aCIECtVld0GDoZKxcWLVqECxcuiP+CgoLQs2dP8fPExEQ0bNhQfL906VLk5OTg8OHD2LZtG3bv3o3t27eLn0+dOhVt2rTBH3/8gc2bN+Onn35CXFwcACAzMxPjxo3D8ePHERsbi6SkJHz22WfitDdv3sTUqVPx4Ycf4uzZs9i1axeaNWtmlPUm43kxVw11rNiwYQMuX76MPXv24MCBA7h69SrWr18vfr5o0SJYWlri999/x7Jly7BgwQLcuHEDAJCWloaJEyfigw8+wB9//IFmzZphypQpGvPv1auXxr5Qu3Zt8bOpU6fi1VdfxenTpzFlyhRMnjwZaWlpAIAbN25g/vz5+Oyzz/D777/D1tYWCxcuFKdNTEyEl5cXZDKZQdqBjKcsucpChxnJzMxEQEAA9u/fj99//x3NmzfH+PHjKzosfumrAAEBAfjqq6/Qu3dvtGnTBrNmzUJeXp5YId6wYQM6dOiAWbNmIT8/H0uWLEHHjh3RsWNHLFmyBPn5+cjOzkZYWBhSUlLESnlycjLUajU2bNiAbt26wc/PDx988IFYBX/xr47Dhg3Dl19+iXfeeQc+Pj4YOXKkeJDJy8vDtGnT4OfnB19fX7z11lt4+PAhACA5ORljx45F27Zt0b17d2zdurXY9Xz//fexefNmjWFvvvkmDh48CAC4desWRowYgbZt2yIwMFDjr0wzZ85EeHg4wsLC0LJlS5w+fbrE9jx69Ch69+4NHx8fdOrUCd9884342W+//Ybg4GC0atUK3bp1w7Fjx3Suw+rVqzF58mRMmzYNrVq1ws6dO5GZmYnZs2ejY8eO6NSpE1asWMEfpVoYM8eBZ3+F+vDDD+Hj44OQkBBcv35dXHZycjImTZqEdu3aISAgQOPU8Be37U8//YSvvvoK+/btg4+PD958802t67Vjxw507doVPj4+CAgIwK+//ip+tnXrVvTq1Qs+Pj7o3bs3rly5AuBZng8bNgy+vr7o06eP+AMFKD7PtcVPpRMQEIANGzagb9++aNmyJeLi4tCnTx/4+vpi2LBhuHXrljjuhg0b0KlTJ/j4+CAwMBDx8fEAoLVvHTp0KACgTZs28PHxwYULFxAeHo7//e9/8PHxga+vL4Bnf33v168fWrVqBX9//yJ/uS/Jrl270KVLF/j5+Wn8CAGA3NxczJgxA23atEGvXr0QFRUl6SyoktZTpVIhMjIS3bp1g4+PD/r37w+lUgkAOH/+PN566y20bt0ab731Fs6fPy/Ob9iwYVixYgXeeecdvPbaa7h7967WPt6Y9Ol/tG1XQHvbS5GdnY0DBw4gJCQEAPDuu+/i9OnTWLRoEXx8fJCQkIDDhw9j1KhRsLW1Ra1atfD2229rFDru37+Pvn37QqFQoE6dOmjVqhVu3rwJAOjbty/eeOMN2Nraolq1ahg4cCAuXLggTrt+/XoMGjQI/v7+sLCwQPXq1VGnTh2NGLdv3y72v8//VVlX25RE2/5V0rYpqb9//oy7wu8w27dvh7+/P9q0aYMtW7bg4sWL6Nu3L3x9fbFo0SJxWc+fVRUVFaVxZkHTpk3FMwG0Hd9VKhUiIiLg5+eHrl274ujRo7o3OqQfK5o2bYpFixahd+/eaN26NTp06IDWrVuLOVeYqz179sT8+fMRGBiIJk2aoE2bNhg7dizGjh0rHiumT5+ud64ePnwYw4YNg6OjI5ycnDBs2DAx97Kzs3Hw4EF88MEHqFq1Knx9fREQEIDdu3cDAA4dOoSGDRuiV69esLa2xqRJk3D9+nWN7V2ShIQEXLlyBZMmTYKNjQ0CAwPRqFEjHDhwAMCzszkCAgLQpk0bVK1aFR988AEOHTqEJ0+eYObMmdi1axe++eYb+Pj44OTJk8xVGD9XK+R7jUAmISkpSZg4caLg5+cndOnSRdi4caMgCIKwatUqYerUqcVO8+jRI6FRo0ZCWlqa1nn/+eefQkhIiODj4yO0b99e+OSTT8TPzpw5IwwaNEho3bq18MYbbwjbt28XBEEQMjIyhOnTpwt+fn5C586dhbVr1woqlUoQBEHYvn27MGjQIGHJkiVCmzZthOXLlwt5eXnCp59+Kvj7+wvt27cX5s2bJ+Tk5BiiaagYXbp0Efr06SMkJiYKjx49EgYNGiQsX75cOHXqlODt7S189tlnQl5enpCTkyN8+eWXwoABA4SHDx8KqampwqBBg4QVK1YIgiAIp06dEjp16qQx7+joaGHAgAGCUqkU8vLyhHnz5glTpkwRBEEQ7t69KzRq1EgoKCgQBEEQhg4dKnTt2lX4559/hJycHGHo0KHCsmXLBEEQhC1btghjxowRsrOzhadPnwqXLl0SMjMzBUEQhNDQUCE8PFzIzc0Vrl69Kvj5+QknT54UBEEz53fu3CkMGjRIjO3GjRtC69athby8PCErK0t44403hF9++UUoKCgQLl++LLRt21b4+++/BUEQhBkzZgitWrUSzp49K6hUKiE3N7fE9uzQoYNw5swZQRAEIT09Xbh8+bIgCM/2nVatWgknTpwQVCqVkJSUJNy8eVPSOrz66qvCoUOHBJVKJeTk5Ajjxo0T5s2bJ2RlZQkPHz4U3nrrLWHLli2l2fwvBWPmeOH22bdvn5Cfny98/fXXQpcuXYT8/HxBpVIJISEhwurVq4W8vDzhzp07QkBAgHDs2DGNaZ/fttr66edlZWUJPj4+wq1btwRBEITk5GQxX2NjY4WOHTsKf/75p6BWq4Xbt28L9+7dE/Lz84Vu3boJ69evF/Ly8oSTJ08KLVu2FOfxYp5nZ2drjZ9Kp0uXLsKbb74pJCYmCteuXRNee+014cSJE0J+fr6wYcMGoVu3bkJeXp5w69Yt4Y033hCSkpIEQXjWZ/7777+CIOjXtwrCs2PtO++8oxHHqVOnhOvXrwsqlUq4du2a0L59e+HQoUNaY79x44bQsmVL4Y8//hDy8vKETz75RPD29hZ+//13QRAEYdmyZUJoaKiQnp4uKJVKISgoqMg+8yJt6xkVFSUEBQUJt27dEtRqtXDt2jUhLS1NePTokeDr6yvs3LlTKCgoEGJiYgRfX1/xO8zQoUMFf39/4e+//xYKCgqEjIwMrX28MenT/2jbrrraXoqdO3cKAQEBglqtFocNHTpU2Lp1q/i+bdu2wp9//im+X7duneDr6yu+/+KLL4Rly5YJ+fn5wq1bt4ROnTppjP+8jz/+WPjwww/F9wEBAcKKFSuEoKAgoUOHDsLUqVOFR48eCYLwf3k7ZcoUISsrS7h+/brg5+cnrp+2tinJP//8U+L+JQglbxtBKLm/L+yfC+OdN2+ekJubKxw/flxo1qyZMG7cOOHhw4dCUlKS0K5dO+H06dOCIBS/DwqCICQmJgodOnQQjhw5IgiCoPX4/uOPPwqBgYFivEOHDi2yr79In2NFp06dhB49egj//vuv0KVLF8Hf319YtmyZ8M033wiNGjUSZs+eLeTl5QnTpk0TmjVrJvTu3Vu4f/++kJ6eLvj5+Ql9+vQR8vLyhOPHjwuNGzcWvv76a71yNSQkRNi7d6/4fvfu3UKjRo2EjIwM4cqVK0Lz5s01xv/666+FMWPGCIIgCIsXLxbmz5+v8XmfPn2E/fv3i9uuVatWQps2bYTevXsLP/zwgzjewYMHhZ49e2pMu3DhQmHRokWCIAjC2LFjha+++krj85YtWwqXLl0SBOHZsbMwbwSBuSoIxs/VivhewzM6TIBarca4cePQuHFjHDt2DBs3bsTGjRtx/PhxrdOdPXsWLi4uqF69utbxlixZgnfffRfnz5/HoUOH0KtXLwDPTtsKCwvD0KFDER8fj127dsHb2xsAsHjxYmRmZuK3337Dpk2bipwGefHiRdSuXRsnT57EuHHjsGzZMiQkJGDXrl04ePAgUlJSsHbt2jK2DGkTGhoKDw8PODo6Yty4cdi7dy8AQC6XY/LkybCysoKNjQ1iYmIwYcIEODs7w8nJCRMmTNCotr7o559/xpQpU+Du7g4rKytMnDgRBw4cKPHMnf79+6NevXqwsbFBz549ce3aNQCAhYUF0tPT8e+//0KhUKBZs2aws7ODUqnEuXPnMG3aNFhbW8Pb2xsDBgwQK/zP69atG65fvy5esxkTE4Pu3bvDysoKR44cQc2aNfHWW2/BwsICTZs2RWBgoFjNB4CuXbuidevWkMvlsLa2LnGdLSwscPPmTTx58gTVqlVD06ZNAQC//PIL3nrrLXTo0AFyuRxubm7w8vKStA4tW7ZEt27dIJfL8eTJExw7dgyzZ89GlSpV4OzsjPfee0/cZlQ8Y+U4ADRt2hQ9e/aEpaUlRowYgfz8fPz555+4dOmSeDqtlZUVateujYEDB2r8Jfn5bWtjY6PXOsnlcty4cQO5ublwdXUVTz//5ZdfMGrUKLRo0QIymQyvvPIKatasiT///BPZ2dkYPXo0rKys0L59e3Tp0kUjd57P87///ltn/FQ6w4YNg4eHB+Li4uDv748OHTrA0tIS77//PnJzc8X7F+Xn5+PWrVsoKChArVq1xL9+69u3FsfPzw+NGzeGXC5HkyZN0KdPH/zxxx9ap9m/fz86d+6MNm3awMrKCh988AHk8v/7+rdv3z6MGTMG1apVg7u7O959912dcWhbz23btuGDDz5A/fr1IZPJ0KRJE1SvXh1HjhzBK6+8gn79+sHCwgJBQUGoX78+/vvf/4rzDQkJQcOGDWFhYYHjx4/r7OONSWr/o2276mp7KXbu3Il+/fppPcW+U6dO2LBhA548eYJ///0X27dvR05Ojvh5586dceDAAbz22mvo1asX3n77bbRo0aLIfH7//Xfs2rULkydPFoclJyfj119/xapVq3DgwAHk5eVh8eLFGtNNmDABVapUQePGjdG/f3/s2bMHQOlyPjY2tsT9q1BJ20aqCRMmwNraGh07dkSVKlUQFBQEZ2dnuLm5wdfXF1evXi1x2tzcXEyYMAHvvvsu/P398fDhQ63H93379mH48OFivGPGjJEUo9RjhYWFBd577z2kpKQgLy8P4eHh2L9/P5o2bQqZTIYaNWrAysoKCoUCVlZWmDt3Ljw9PXH79m1YWloiISEBcrkc//vf/9CoUSPcvHlTr1zt1KkTvv/+e6SlpeHBgwfYtGkTACAnJwfZ2dmwt7fXGN/e3h5ZWVkAUOzndnZ24ue9evVCbGws4uPjsXjxYqxbt07MraysrDLN+0XMVePnakV8r7GQtAZkVM9/sQagsQE9PT2LnSYpKQkLFy6UdAMdCwsL3LlzB2lpaXByckLLli0BPPvR+PrrryMoKAgAUL16dVSvXh0qlQqxsbHYtWsX7OzsYGdnhxEjRuDXX3/FgAEDAACurq4YNmwYgGdfeLZt24Zff/0Vjo6OAIAxY8Zg6tSpmDp1almahrTw8PAQX3t6eiIlJQXAs+34/I/6lJQUjTx6ftziJCYmYsKECRoHOLlcjtTU1GLHd3FxEV/b2toiOzsbABAcHIykpCT85z//QUZGBt58801MmTIFKSkpqFatGuzs7DRiunz5cpF529nZwd/fH3v37sXo0aOxd+9e8QvW/fv3cfHiRfGUbuDZaXfPXzrwfBtps2rVKqxfvx5ffPEFGjdujKlTp8LHxwdKpRL+/v5FxpeyDu7u7uLrxMREPH36FB07dhSHqdVqyfG9rIyV44Dm9iksYhVOk5KSUiSvnn///LT6qFKlClasWIFvv/0Wc+bMQatWrTBjxgyxePbi6eCFsbi7u2vsj56enuIlOIBmO92/f19n/FQ6he38Yr7J5XJ4eHggOTkZfn5+mD17NlavXo2bN2+iY8eOmDlzJtzc3PTuW4vz559/4vPPP8eNGzdQUFCA/Px8jXs3FKcwhwpVqVJFPFYXfv58DknJ71deeaXE9UxKSioxl1/8TqMrl3X18cYktf/Rtl11tb0uSqUSZ86cwccff6x1vLlz52Lx4sUIDAyEo6Mj+vTpI/5oSE9Px6hRozB//nwEBQXh4cOHmDx5MpydnREaGirO43//+x+mTp2KVatWoV69euJwa2tr8Q8awLPvdyNGjNBY/vNtVbNmTfz9998628bNza3YddG2fxW3PCn9/YucnZ011u/F94XfY4ozZ84c1KtXD6NHjwag+/j+4v5V0vf65+l7rPDw8BBzrVatWmJ72NraavQvhT985XI5CgoKkJubCwBo27Yt8vPzIQiCuF2k5uq4ceOQmZmJ4OBgWFlZYeDAgbh27RqcnZ3x8OHDIje2ffLkCapWrSou48XPs7KyxM8bNGggDm/VqhXeffddHDhwAEFBQahatare837+8xcxV8snVwuXU17fa1joMAHaNmBxSZaWloaRI0diyJAhYpFCmyVLlmDVqlXo1asXatWqhYkTJ6JLly4lJuCjR49QUFBQ5IfD8wn4/IE7LS0NOTk56N+/vzhMMJMbiZmzwmuegWedl6urKwAU+auPq6urxo3LlEplieMCz7btJ598gtatWxf57N69e5Ljs7S0xMSJEzFx4kTcu3cPo0ePRr169dCxY0c8fvwYT548EQsFSqWyxANJUFAQ1qxZgzZt2iA3Nxd+fn4AnnWCbdq0QXR0tOSYStKiRQusX78eBQUF+OGHH/Dhhx/i6NGj8PDwwJ07d4qM7+rqqnMdnm/bwr8QnDp1ChYW7HalMlaOA8+KxYXUajWSk5Ph6uoKhUKBWrVqifeBKc6L89TnZmadOnVCp06dkJubiy+//BLz5s3Djz/+qDXXkpKSoFarxS8FSqUSdevWLXb+Hh4eOuOn0inczq6uruIPOeDZ8e75/b9v377o27cvnjx5gvnz5+Pzzz/HsmXLtPatzz9p4MXlPW/q1KkYOnQovv76a1hbW2PJkiV49OiR1rhdXV01rhvPycnRuPbcxcUFSUlJ4o+K5/cNbbSt5507d9CoUaMicSQmJmoMUyqV6NSpU7HrbMg+vjSk9j/atquuttdl165d8PHx0bgBY3EcHR3xxRdfiO+XL18unrFx9+5dKBQK9OvXT4y3d+/eOHbsmFjouHr1KsaNG4dPPvkE7du315h348aNdcapVCrh5eUFQLOttLVNSXTtX4XLK6Rt2xjahg0bkJCQgB9//FEcpuv47uLiohHv86+10edYoVQq0bBhQyQlJeH+/ftie6jVao12q1q1KtavX4/WrVvjwoULmDFjhnisWLNmDW7duoUVK1YAkJ6rNjY2mD9/PubPnw/g2ZkRTZs2hUKhQN26daFSqXD79m3xmHX9+nWxr2nYsCF27twpzis7Oxt37tzRKHC8SBAEAM+KIHfv3tX4Hnb9+nXxd1HDhg017r119+5dFBQUlHjsZK6WT64C5fu9hpeumIDCDXj27Fnx34ULFxAVFVVk3MePH2PkyJEICAjAuHHjJM2/bt26WL58OeLj4xEWFobJkycjOzu7xASsXr06LC0tNb6QaPsRV716ddjY2GDv3r1i/OfOndM4dYsM78cff0RSUhLS09PFmx0Vp0+fPli/fj3S0tKQlpaGtWvXom/fvgCeVYrT09ORmZkpjj948GB8+eWX4hfvtLQ0/Pbbb3rHd+rUKfz1119QqVSws7ODhYUFFAoFPDw84OPjg+XLlyMvLw/Xr1/HL7/8Isb0In9/fyQmJmLVqlXo3bu32Cl27twZt2/fxq5du1BQUICCggJcvHhR0k2snpefn49ff/0VmZmZsLS0RNWqVcXH57799tvYsWMH4uPjxR/Dt27d0nsdXF1d0aFDB3z66ad48uQJ1Go17ty5o/OU85edsXIcAK5cuYKDBw/i6dOn2LhxI6ysrPDaa6+hRYsWsLOzw4YNG5CbmwuVSoW///4bFy9eLDFOZ2dn3L9/X2dx9+HDh4iLi0N2djasrKxQpUoVjVz79ttvcfnyZQiCgH///Rf3799HixYtYGtri6+//hoFBQU4ffo0Dh8+XGJblCZ+0k+vXr1w9OhRxMfHo6CgAN9++y2srKzg4+ODf/75B/Hx8cjPz4eVlRWsra3Fbaytb3VycoJcLtd4dKKzszOSk5ORn58vDsvKykK1atVgbW2NixcviqdxaxMYGIgjR47g7NmzyM/Px6pVqzRytVevXvjqq6/w+PFjJCcnF7kBdHG0reeAAQOwcuVK3L59G4Ig4Pr163j06BH8/f1x+/ZtxMTE4OnTp4iNjcXNmzfRuXPnYpdhqD6+tKT2P9q2q66212XXrl3iTUi1uXPnDh49egSVSoWjR4/i559/Fr8j1qtXD4IgICYmBmq1Gg8ePMC+ffvQpEkTAMDff/+NUaNGYd68eQgICCgy7/79+2PHjh24e/cucnJyEBUVVWSbrVu3Djk5Obhx4wZ27NghtlVpvk9o278KlbRtSurvDeHo0aP4/vvvsXbtWo1LFnUd33v16oVNmzYhKSkJjx8/xoYNG3QuS59jReExzNXVFVZWVli0aBECAwNx7do15OXlaeTtq6++Km6PFi1awMbGBh999BFyc3PRrVs3HD58GD///LNeuZqcnIzk5GQIgoD//e9/WLduHSZNmgTg2V/7u3fvjlWrViE7Oxvnzp1DXFwcgoODAQDdu3fHjRs3xEui1q5di8aNG4tFs99++w2PHz+GIAi4ePEiNm3ahK5duwJ4ltfe3t5Yu3Yt8vLycOjQIfz1118IDAwE8KwQ+9///hdnz55FdnY2Vq5cie7du2ucifs85qrxc7Uivtew0GECpG7AJ0+e4P3330erVq3Eu/JKsXv3bqSlpUEul8PBwQHAs8tN+vbti5MnTyI2NhZPnz7Fo0ePcO3aNSgUCvTs2RMrVqzAkydPcP/+fURHR5d4uqhcLseAAQPwySefiKfIJScn67zHCJVNUFAQRo4ciW7duqF27dolFr7Gjx+PZs2a4c0338Sbb76Jpk2bik/r8fLyQp8+fdCtWzf4+voiOTkZ7777LgICAjBy5Ej4+Phg4MCBpfqRVHh6bOvWrdG7d2+0bdtWzKHly5fj/v376NSpEyZOnIhJkyahQ4cOxc7HysoK3bt3x8mTJzXOYLKzs8M333yD2NhYdOrUCR07dsTnn3+u8aNAqt27dyMgIEB8ikbho/VatGiBpUuXilX+oUOHigVAfdYBAD777DMUFBSId9+ePHkyHjx4oHesLxNj5Tjw7PrP2NhYtGnTBrt378bq1athaWkJhUKB9evX4/r16+jatSvatWuHuXPnFjkF9nmFlw74+flp/VGiVqsRHR2NTp06oW3btjhz5gzCw8MBPPuCMXbsWEydOhWtWrXChAkT8PjxY1hZWWH9+vU4duwY2rVrh4ULF+Kzzz4Tvwi+qDTxk37q16+PZcuWYfHixWjXrh3++9//IjIyElZWVsjPz8cXX3wBPz8/dOzYEWlpaeLjErX1rba2thg7diwGDx4MX19f/O9//0O7du3QoEEDdOzYUTyTLTw8HKtWrYKPjw/Wrl0r3nNLm4YNG2L+/PmYNm0aOnXqBAcHB42zMidMmAB3d3d07doV7733HgIDA2FlZaV1ntrWc8SIEejVqxdGjhyJVq1aYc6cOcjLy0P16tURGRmJ6Oho+Pn54euvv0ZkZCScnJyKXYYh+/jSkNr/aNuuutpemwsXLiA5OVnnpUkAcPnyZfTt2xetWrXC8uXL8fnnn4tnuNnZ2WH16tX47rvv0KZNG/Tr1w8NGzbE2LFjAQDR0dFIS0vDnDlzxKc09OnTR5z322+/jX79+mHAgAHo0qWLeJ+H5xU+fey9997DyJEjxVPjS/N9Qtv+VaikbVNSf28I+/btw6NHj8QntPn4+IhnMWg7vg8cOBAdO3ZEcHAwQkJC0KNHD53L0udYkZqaig4dOmDs2LF4/PgxcnNzsWXLFmzcuBHVqlXTOFa0aNFC3B6+vr7IzMzE5cuX0bVrVwwbNgwuLi748ssv9crVO3fuYPDgwWjZsiVmzJiBqVOnalwaER4ejtzcXLz++uuYOnUqFixYIOamk5MTVq9ejRUrVqBNmza4ePEili9fLk4bGxuLHj16oFWrVvjoo48QFhamcYxdvnw5Ll++jDZt2uDzzz/HqlWrxP6kYcOGWLhwIaZNm4bXX38dWVlZYhsWh7lq/FytiO81MqHwHCCqUMnJyYiIiMDp06eRn5+PevXq4cMPP8S5c+fw77//4vPPP8fOnTsxc+ZM2NraapxRsXfvXq3XUU2bNg2///47cnNz4enpiSlTpqBbt24Ant3QNCIiArdu3YK9vT0+/PBDhISE4PHjx1i8eDFOnDgBa2trDBgwAOPHj4dcLseOHTuwbds2bNmyRVxGYSV27969ePToEdzc3DB48GBJNzUj/QUEBODjjz/G66+/XtGhEBkFc5yoYvz444+IjY2VdGZHZcX+x3Rx22hie5gubpuKx4vFTYSbm5tGFbPQ8ztHSEiIpFMYX/T555+X+Jmvry+2bdtWZHi1atVKnK5///4a9+MAnt0Q5z//+Q/+85//6B0fERERVYyUlBTcvXsXPj4+uH37NqKjozVuUklERGSOWOggopdCnz59itwIDwAWLlxYbnfxp5fD89fnPi8qKopPPyGj+PXXX4s9LdvT01Pn4wwLCgoQHh6Oe/fuwd7eHn369MGQIUOQmJiocQnD83SdSUq6jRo1CufOnSsyfMyYMeJlJZVVWfK1MjGXYwVzlblqLrn6Il66Ukm8zJ0QERERERERUSEWOoiIiIiIiIio0uBTV4iIiIiIiIio0mChg4iIiIiIiIgqDRY6iIiIiIiIiKjSYKGDiIiIiIiIiCqN/wfglAPghLftzAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 1080x1080 with 42 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set(style=\"ticks\", color_codes=True)\n",
    "g = sns.pairplot(\n",
    "    final_result_df[\n",
    "        [\n",
    "            \"el2_score\",\n",
    "            \"proteinsolver_score\",\n",
    "            \"protbert_score\",\n",
    "            \"rosetta_dg_score\",\n",
    "            \"pred_7f9826be_optimized_score\",\n",
    "            \"pred_900500fe_optimized_score\",\n",
    "        ]\n",
    "    ]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "5e483f61-0149-43d1-bbcc-c836d7609de2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x1080 with 42 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set(style=\"ticks\", color_codes=True)\n",
    "g = sns.pairplot(\n",
    "    final_result_df[\n",
    "        [\n",
    "            \"el2_ratio\",\n",
    "            \"proteinsolver_ratio\",\n",
    "            \"protbert_ratio\",\n",
    "            \"rosetta_dg_ratio\",\n",
    "            \"pred_7f9826be_optimized_ratio\",\n",
    "            \"pred_900500fe_optimized_ratio\",\n",
    "        ]\n",
    "    ]\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fc28cbfa-441b-4755-92b9-15237c94a4ba",
   "metadata": {},
   "source": [
    "## Write results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "031f73ce-59c4-4d8b-931f-758a7569822f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "PosixPath('/home/kimlab5/strokach/workspace/elaspic/elaspic2-cagi6/notebooks/40_cagi6_mapk1_submission/submission')"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "output_dir = NOTEBOOK_DIR.joinpath(\"submission\")\n",
    "output_dir.mkdir(exist_ok=True)\n",
    "\n",
    "output_dir"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "0a143cac-53e1-4cac-a9a0-6a6eca5fab13",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Overwriting /home/kimlab5/strokach/workspace/elaspic/elaspic2-cagi6/notebooks/40_cagi6_mapk1_submission/submission/strokach_desc.md\n"
     ]
    }
   ],
   "source": [
    "%%file {output_dir}/strokach_desc.md\n",
    "# Submission for CAGI6—MAPK1 challenge\n",
    "\n",
    "## Overview\n",
    "\n",
    "- `strokach_modelnumber_1.tsv` → Predictions made using ELASPIC2 [1].\n",
    "- `strokach_modelnumber_2.tsv` → Predictions made using ProteinSolver [2].\n",
    "- `strokach_modelnumber_3.tsv` → Predictions made using ProtBert [3].\n",
    "- `strokach_modelnumber_4.tsv` → Predictions made using Rosetta's cartesian_ddg protocol [4].\n",
    "- `strokach_modelnumber_5.tsv` → Predictions made using ELASPIC2 with AlphaFold [5] features for wildtype protein.\n",
    "- `strokach_modelnumber_6.tsv` → Predictions made using ELASPIC2 with AlphaFold [5] features for wildtype and mutant proteins.\n",
    "\n",
    "## References\n",
    "\n",
    "- [1] Strokach et al. (2021). _ELASPIC2 (EL2): Combining Contextualized Language Models and Graph Neural Networks to Predict Effects of Mutations._ https://doi.org/10.1016/j.jmb.2021.166810\n",
    "- [2] Strokach et al. (2020). _Fast and Flexible Protein Design Using Deep Graph Neural Networks._ https://doi.org/10.1016/j.cels.2020.08.016\n",
    "- [3] Elnaggar et al. (2020). _ProtTrans: Towards Cracking the Language of Life’s Code Through Self-Supervised Deep Learning and High Performance Computing._ https://doi.org/10.1101/2020.07.12.199554\n",
    "- [4] Park et al. (2016). _Simultaneous Optimization of Biomolecular Energy Functions on Features from Small Molecules and Macromolecules._ https://doi.org/10.1021/acs.jctc.6b00819\n",
    "- [5] Jumper et al. (2021). _Highly accurate protein structure prediction with AlphaFold._ https://doi.org/10.1038/s41586-021-03819-2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "30ef90b3-aefa-4cf1-951c-378075ef4553",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "el2\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>MAPK1-variant</th>\n",
       "      <th>DDG-NotPO4</th>\n",
       "      <th>Standard-Deviation-DDG-NotPO4</th>\n",
       "      <th>DDG-PO4</th>\n",
       "      <th>Standard-Deviation-DDG-PO4</th>\n",
       "      <th>Ratio-kcat/Km</th>\n",
       "      <th>Standard-Deviation-Ratio-kcat/Km</th>\n",
       "      <th>Comment</th>\n",
       "      <th>mut</th>\n",
       "      <th>protbert_score</th>\n",
       "      <th>...</th>\n",
       "      <th>el2_score</th>\n",
       "      <th>protbert_ratio</th>\n",
       "      <th>proteinsolver_ratio</th>\n",
       "      <th>el2_ratio</th>\n",
       "      <th>rosetta_dg_score</th>\n",
       "      <th>rosetta_dg_ratio</th>\n",
       "      <th>pred_7f9826be_optimized_score</th>\n",
       "      <th>pred_7f9826be_optimized_ratio</th>\n",
       "      <th>pred_900500fe_optimized_score</th>\n",
       "      <th>pred_900500fe_optimized_ratio</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>p.E33Q</td>\n",
       "      <td>-1.582201</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-1.582201</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.316016</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>E33Q</td>\n",
       "      <td>0.998343</td>\n",
       "      <td>...</td>\n",
       "      <td>1.582201</td>\n",
       "      <td>0.000155</td>\n",
       "      <td>0.519324</td>\n",
       "      <td>0.316016</td>\n",
       "      <td>-2.457000</td>\n",
       "      <td>0.996143</td>\n",
       "      <td>0.999970</td>\n",
       "      <td>0.000030</td>\n",
       "      <td>0.999987</td>\n",
       "      <td>0.000013</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>p.E81K</td>\n",
       "      <td>-1.630727</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-1.630727</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.306612</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>E81K</td>\n",
       "      <td>0.994183</td>\n",
       "      <td>...</td>\n",
       "      <td>1.630727</td>\n",
       "      <td>0.000131</td>\n",
       "      <td>0.090255</td>\n",
       "      <td>0.306612</td>\n",
       "      <td>2.456000</td>\n",
       "      <td>1.003807</td>\n",
       "      <td>0.999963</td>\n",
       "      <td>0.000037</td>\n",
       "      <td>0.999974</td>\n",
       "      <td>0.000026</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>p.L121I</td>\n",
       "      <td>-1.390088</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-1.390088</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.359689</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>L121I</td>\n",
       "      <td>0.999190</td>\n",
       "      <td>...</td>\n",
       "      <td>1.390088</td>\n",
       "      <td>0.000108</td>\n",
       "      <td>0.157240</td>\n",
       "      <td>0.359689</td>\n",
       "      <td>2.965667</td>\n",
       "      <td>1.004512</td>\n",
       "      <td>0.994864</td>\n",
       "      <td>0.005136</td>\n",
       "      <td>0.992272</td>\n",
       "      <td>0.007728</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>p.R135K</td>\n",
       "      <td>-1.664755</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-1.664755</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.300345</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>R135K</td>\n",
       "      <td>0.997200</td>\n",
       "      <td>...</td>\n",
       "      <td>1.664755</td>\n",
       "      <td>0.000537</td>\n",
       "      <td>0.310411</td>\n",
       "      <td>0.300345</td>\n",
       "      <td>2.395333</td>\n",
       "      <td>1.003528</td>\n",
       "      <td>0.999991</td>\n",
       "      <td>0.000009</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.000007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>p.D162G</td>\n",
       "      <td>-2.243989</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-2.243989</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.222818</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>D162G</td>\n",
       "      <td>0.998402</td>\n",
       "      <td>...</td>\n",
       "      <td>2.243989</td>\n",
       "      <td>0.000142</td>\n",
       "      <td>0.053119</td>\n",
       "      <td>0.222818</td>\n",
       "      <td>3.375667</td>\n",
       "      <td>1.005196</td>\n",
       "      <td>0.999963</td>\n",
       "      <td>0.000037</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.000007</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  MAPK1-variant  DDG-NotPO4  Standard-Deviation-DDG-NotPO4   DDG-PO4  \\\n",
       "0        p.E33Q   -1.582201                            1.0 -1.582201   \n",
       "1        p.E81K   -1.630727                            1.0 -1.630727   \n",
       "2       p.L121I   -1.390088                            1.0 -1.390088   \n",
       "3       p.R135K   -1.664755                            1.0 -1.664755   \n",
       "4       p.D162G   -2.243989                            1.0 -2.243989   \n",
       "\n",
       "    Standard-Deviation-DDG-PO4  Ratio-kcat/Km  \\\n",
       "0                          1.0       0.316016   \n",
       "1                          1.0       0.306612   \n",
       "2                          1.0       0.359689   \n",
       "3                          1.0       0.300345   \n",
       "4                          1.0       0.222818   \n",
       "\n",
       "   Standard-Deviation-Ratio-kcat/Km Comment    mut  protbert_score  ...  \\\n",
       "0                               1.0       *   E33Q        0.998343  ...   \n",
       "1                               1.0       *   E81K        0.994183  ...   \n",
       "2                               1.0       *  L121I        0.999190  ...   \n",
       "3                               1.0       *  R135K        0.997200  ...   \n",
       "4                               1.0       *  D162G        0.998402  ...   \n",
       "\n",
       "   el2_score  protbert_ratio  proteinsolver_ratio  el2_ratio  \\\n",
       "0   1.582201        0.000155             0.519324   0.316016   \n",
       "1   1.630727        0.000131             0.090255   0.306612   \n",
       "2   1.390088        0.000108             0.157240   0.359689   \n",
       "3   1.664755        0.000537             0.310411   0.300345   \n",
       "4   2.243989        0.000142             0.053119   0.222818   \n",
       "\n",
       "   rosetta_dg_score  rosetta_dg_ratio  pred_7f9826be_optimized_score  \\\n",
       "0         -2.457000          0.996143                       0.999970   \n",
       "1          2.456000          1.003807                       0.999963   \n",
       "2          2.965667          1.004512                       0.994864   \n",
       "3          2.395333          1.003528                       0.999991   \n",
       "4          3.375667          1.005196                       0.999963   \n",
       "\n",
       "   pred_7f9826be_optimized_ratio  pred_900500fe_optimized_score  \\\n",
       "0                       0.000030                       0.999987   \n",
       "1                       0.000037                       0.999974   \n",
       "2                       0.005136                       0.992272   \n",
       "3                       0.000009                       0.999993   \n",
       "4                       0.000037                       0.999993   \n",
       "\n",
       "   pred_900500fe_optimized_ratio  \n",
       "0                       0.000013  \n",
       "1                       0.000026  \n",
       "2                       0.007728  \n",
       "3                       0.000007  \n",
       "4                       0.000007  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The file's format is valid! You are good to submit now!\n",
      "proteinsolver\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>MAPK1-variant</th>\n",
       "      <th>DDG-NotPO4</th>\n",
       "      <th>Standard-Deviation-DDG-NotPO4</th>\n",
       "      <th>DDG-PO4</th>\n",
       "      <th>Standard-Deviation-DDG-PO4</th>\n",
       "      <th>Ratio-kcat/Km</th>\n",
       "      <th>Standard-Deviation-Ratio-kcat/Km</th>\n",
       "      <th>Comment</th>\n",
       "      <th>mut</th>\n",
       "      <th>protbert_score</th>\n",
       "      <th>...</th>\n",
       "      <th>el2_score</th>\n",
       "      <th>protbert_ratio</th>\n",
       "      <th>proteinsolver_ratio</th>\n",
       "      <th>el2_ratio</th>\n",
       "      <th>rosetta_dg_score</th>\n",
       "      <th>rosetta_dg_ratio</th>\n",
       "      <th>pred_7f9826be_optimized_score</th>\n",
       "      <th>pred_7f9826be_optimized_ratio</th>\n",
       "      <th>pred_900500fe_optimized_score</th>\n",
       "      <th>pred_900500fe_optimized_ratio</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>p.E33Q</td>\n",
       "      <td>-0.066731</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.066731</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.519324</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>E33Q</td>\n",
       "      <td>0.998343</td>\n",
       "      <td>...</td>\n",
       "      <td>1.582201</td>\n",
       "      <td>0.000155</td>\n",
       "      <td>0.519324</td>\n",
       "      <td>0.316016</td>\n",
       "      <td>-2.457000</td>\n",
       "      <td>0.996143</td>\n",
       "      <td>0.999970</td>\n",
       "      <td>0.000030</td>\n",
       "      <td>0.999987</td>\n",
       "      <td>0.000013</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>p.E81K</td>\n",
       "      <td>-0.159303</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.159303</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.090255</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>E81K</td>\n",
       "      <td>0.994183</td>\n",
       "      <td>...</td>\n",
       "      <td>1.630727</td>\n",
       "      <td>0.000131</td>\n",
       "      <td>0.090255</td>\n",
       "      <td>0.306612</td>\n",
       "      <td>2.456000</td>\n",
       "      <td>1.003807</td>\n",
       "      <td>0.999963</td>\n",
       "      <td>0.000037</td>\n",
       "      <td>0.999974</td>\n",
       "      <td>0.000026</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>p.L121I</td>\n",
       "      <td>-0.692270</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.692270</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.157240</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>L121I</td>\n",
       "      <td>0.999190</td>\n",
       "      <td>...</td>\n",
       "      <td>1.390088</td>\n",
       "      <td>0.000108</td>\n",
       "      <td>0.157240</td>\n",
       "      <td>0.359689</td>\n",
       "      <td>2.965667</td>\n",
       "      <td>1.004512</td>\n",
       "      <td>0.994864</td>\n",
       "      <td>0.005136</td>\n",
       "      <td>0.992272</td>\n",
       "      <td>0.007728</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>p.R135K</td>\n",
       "      <td>-0.481067</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.481067</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.310411</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>R135K</td>\n",
       "      <td>0.997200</td>\n",
       "      <td>...</td>\n",
       "      <td>1.664755</td>\n",
       "      <td>0.000537</td>\n",
       "      <td>0.310411</td>\n",
       "      <td>0.300345</td>\n",
       "      <td>2.395333</td>\n",
       "      <td>1.003528</td>\n",
       "      <td>0.999991</td>\n",
       "      <td>0.000009</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.000007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>p.D162G</td>\n",
       "      <td>-0.140639</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.140639</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.053119</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>D162G</td>\n",
       "      <td>0.998402</td>\n",
       "      <td>...</td>\n",
       "      <td>2.243989</td>\n",
       "      <td>0.000142</td>\n",
       "      <td>0.053119</td>\n",
       "      <td>0.222818</td>\n",
       "      <td>3.375667</td>\n",
       "      <td>1.005196</td>\n",
       "      <td>0.999963</td>\n",
       "      <td>0.000037</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.000007</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  MAPK1-variant  DDG-NotPO4  Standard-Deviation-DDG-NotPO4   DDG-PO4  \\\n",
       "0        p.E33Q   -0.066731                            1.0 -0.066731   \n",
       "1        p.E81K   -0.159303                            1.0 -0.159303   \n",
       "2       p.L121I   -0.692270                            1.0 -0.692270   \n",
       "3       p.R135K   -0.481067                            1.0 -0.481067   \n",
       "4       p.D162G   -0.140639                            1.0 -0.140639   \n",
       "\n",
       "    Standard-Deviation-DDG-PO4  Ratio-kcat/Km  \\\n",
       "0                          1.0       0.519324   \n",
       "1                          1.0       0.090255   \n",
       "2                          1.0       0.157240   \n",
       "3                          1.0       0.310411   \n",
       "4                          1.0       0.053119   \n",
       "\n",
       "   Standard-Deviation-Ratio-kcat/Km Comment    mut  protbert_score  ...  \\\n",
       "0                               1.0       *   E33Q        0.998343  ...   \n",
       "1                               1.0       *   E81K        0.994183  ...   \n",
       "2                               1.0       *  L121I        0.999190  ...   \n",
       "3                               1.0       *  R135K        0.997200  ...   \n",
       "4                               1.0       *  D162G        0.998402  ...   \n",
       "\n",
       "   el2_score  protbert_ratio  proteinsolver_ratio  el2_ratio  \\\n",
       "0   1.582201        0.000155             0.519324   0.316016   \n",
       "1   1.630727        0.000131             0.090255   0.306612   \n",
       "2   1.390088        0.000108             0.157240   0.359689   \n",
       "3   1.664755        0.000537             0.310411   0.300345   \n",
       "4   2.243989        0.000142             0.053119   0.222818   \n",
       "\n",
       "   rosetta_dg_score  rosetta_dg_ratio  pred_7f9826be_optimized_score  \\\n",
       "0         -2.457000          0.996143                       0.999970   \n",
       "1          2.456000          1.003807                       0.999963   \n",
       "2          2.965667          1.004512                       0.994864   \n",
       "3          2.395333          1.003528                       0.999991   \n",
       "4          3.375667          1.005196                       0.999963   \n",
       "\n",
       "   pred_7f9826be_optimized_ratio  pred_900500fe_optimized_score  \\\n",
       "0                       0.000030                       0.999987   \n",
       "1                       0.000037                       0.999974   \n",
       "2                       0.005136                       0.992272   \n",
       "3                       0.000009                       0.999993   \n",
       "4                       0.000037                       0.999993   \n",
       "\n",
       "   pred_900500fe_optimized_ratio  \n",
       "0                       0.000013  \n",
       "1                       0.000026  \n",
       "2                       0.007728  \n",
       "3                       0.000007  \n",
       "4                       0.000007  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The file's format is valid! You are good to submit now!\n",
      "protbert\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>MAPK1-variant</th>\n",
       "      <th>DDG-NotPO4</th>\n",
       "      <th>Standard-Deviation-DDG-NotPO4</th>\n",
       "      <th>DDG-PO4</th>\n",
       "      <th>Standard-Deviation-DDG-PO4</th>\n",
       "      <th>Ratio-kcat/Km</th>\n",
       "      <th>Standard-Deviation-Ratio-kcat/Km</th>\n",
       "      <th>Comment</th>\n",
       "      <th>mut</th>\n",
       "      <th>protbert_score</th>\n",
       "      <th>...</th>\n",
       "      <th>el2_score</th>\n",
       "      <th>protbert_ratio</th>\n",
       "      <th>proteinsolver_ratio</th>\n",
       "      <th>el2_ratio</th>\n",
       "      <th>rosetta_dg_score</th>\n",
       "      <th>rosetta_dg_ratio</th>\n",
       "      <th>pred_7f9826be_optimized_score</th>\n",
       "      <th>pred_7f9826be_optimized_ratio</th>\n",
       "      <th>pred_900500fe_optimized_score</th>\n",
       "      <th>pred_900500fe_optimized_ratio</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>p.E33Q</td>\n",
       "      <td>-0.998343</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.998343</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000155</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>E33Q</td>\n",
       "      <td>0.998343</td>\n",
       "      <td>...</td>\n",
       "      <td>1.582201</td>\n",
       "      <td>0.000155</td>\n",
       "      <td>0.519324</td>\n",
       "      <td>0.316016</td>\n",
       "      <td>-2.457000</td>\n",
       "      <td>0.996143</td>\n",
       "      <td>0.999970</td>\n",
       "      <td>0.000030</td>\n",
       "      <td>0.999987</td>\n",
       "      <td>0.000013</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>p.E81K</td>\n",
       "      <td>-0.994183</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.994183</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000131</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>E81K</td>\n",
       "      <td>0.994183</td>\n",
       "      <td>...</td>\n",
       "      <td>1.630727</td>\n",
       "      <td>0.000131</td>\n",
       "      <td>0.090255</td>\n",
       "      <td>0.306612</td>\n",
       "      <td>2.456000</td>\n",
       "      <td>1.003807</td>\n",
       "      <td>0.999963</td>\n",
       "      <td>0.000037</td>\n",
       "      <td>0.999974</td>\n",
       "      <td>0.000026</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>p.L121I</td>\n",
       "      <td>-0.999190</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.999190</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000108</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>L121I</td>\n",
       "      <td>0.999190</td>\n",
       "      <td>...</td>\n",
       "      <td>1.390088</td>\n",
       "      <td>0.000108</td>\n",
       "      <td>0.157240</td>\n",
       "      <td>0.359689</td>\n",
       "      <td>2.965667</td>\n",
       "      <td>1.004512</td>\n",
       "      <td>0.994864</td>\n",
       "      <td>0.005136</td>\n",
       "      <td>0.992272</td>\n",
       "      <td>0.007728</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>p.R135K</td>\n",
       "      <td>-0.997200</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.997200</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000537</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>R135K</td>\n",
       "      <td>0.997200</td>\n",
       "      <td>...</td>\n",
       "      <td>1.664755</td>\n",
       "      <td>0.000537</td>\n",
       "      <td>0.310411</td>\n",
       "      <td>0.300345</td>\n",
       "      <td>2.395333</td>\n",
       "      <td>1.003528</td>\n",
       "      <td>0.999991</td>\n",
       "      <td>0.000009</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.000007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>p.D162G</td>\n",
       "      <td>-0.998402</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.998402</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000142</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>D162G</td>\n",
       "      <td>0.998402</td>\n",
       "      <td>...</td>\n",
       "      <td>2.243989</td>\n",
       "      <td>0.000142</td>\n",
       "      <td>0.053119</td>\n",
       "      <td>0.222818</td>\n",
       "      <td>3.375667</td>\n",
       "      <td>1.005196</td>\n",
       "      <td>0.999963</td>\n",
       "      <td>0.000037</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.000007</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  MAPK1-variant  DDG-NotPO4  Standard-Deviation-DDG-NotPO4   DDG-PO4  \\\n",
       "0        p.E33Q   -0.998343                            1.0 -0.998343   \n",
       "1        p.E81K   -0.994183                            1.0 -0.994183   \n",
       "2       p.L121I   -0.999190                            1.0 -0.999190   \n",
       "3       p.R135K   -0.997200                            1.0 -0.997200   \n",
       "4       p.D162G   -0.998402                            1.0 -0.998402   \n",
       "\n",
       "    Standard-Deviation-DDG-PO4  Ratio-kcat/Km  \\\n",
       "0                          1.0       0.000155   \n",
       "1                          1.0       0.000131   \n",
       "2                          1.0       0.000108   \n",
       "3                          1.0       0.000537   \n",
       "4                          1.0       0.000142   \n",
       "\n",
       "   Standard-Deviation-Ratio-kcat/Km Comment    mut  protbert_score  ...  \\\n",
       "0                               1.0       *   E33Q        0.998343  ...   \n",
       "1                               1.0       *   E81K        0.994183  ...   \n",
       "2                               1.0       *  L121I        0.999190  ...   \n",
       "3                               1.0       *  R135K        0.997200  ...   \n",
       "4                               1.0       *  D162G        0.998402  ...   \n",
       "\n",
       "   el2_score  protbert_ratio  proteinsolver_ratio  el2_ratio  \\\n",
       "0   1.582201        0.000155             0.519324   0.316016   \n",
       "1   1.630727        0.000131             0.090255   0.306612   \n",
       "2   1.390088        0.000108             0.157240   0.359689   \n",
       "3   1.664755        0.000537             0.310411   0.300345   \n",
       "4   2.243989        0.000142             0.053119   0.222818   \n",
       "\n",
       "   rosetta_dg_score  rosetta_dg_ratio  pred_7f9826be_optimized_score  \\\n",
       "0         -2.457000          0.996143                       0.999970   \n",
       "1          2.456000          1.003807                       0.999963   \n",
       "2          2.965667          1.004512                       0.994864   \n",
       "3          2.395333          1.003528                       0.999991   \n",
       "4          3.375667          1.005196                       0.999963   \n",
       "\n",
       "   pred_7f9826be_optimized_ratio  pred_900500fe_optimized_score  \\\n",
       "0                       0.000030                       0.999987   \n",
       "1                       0.000037                       0.999974   \n",
       "2                       0.005136                       0.992272   \n",
       "3                       0.000009                       0.999993   \n",
       "4                       0.000037                       0.999993   \n",
       "\n",
       "   pred_900500fe_optimized_ratio  \n",
       "0                       0.000013  \n",
       "1                       0.000026  \n",
       "2                       0.007728  \n",
       "3                       0.000007  \n",
       "4                       0.000007  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The file's format is valid! You are good to submit now!\n",
      "rosetta_dg\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>MAPK1-variant</th>\n",
       "      <th>DDG-NotPO4</th>\n",
       "      <th>Standard-Deviation-DDG-NotPO4</th>\n",
       "      <th>DDG-PO4</th>\n",
       "      <th>Standard-Deviation-DDG-PO4</th>\n",
       "      <th>Ratio-kcat/Km</th>\n",
       "      <th>Standard-Deviation-Ratio-kcat/Km</th>\n",
       "      <th>Comment</th>\n",
       "      <th>mut</th>\n",
       "      <th>protbert_score</th>\n",
       "      <th>...</th>\n",
       "      <th>el2_score</th>\n",
       "      <th>protbert_ratio</th>\n",
       "      <th>proteinsolver_ratio</th>\n",
       "      <th>el2_ratio</th>\n",
       "      <th>rosetta_dg_score</th>\n",
       "      <th>rosetta_dg_ratio</th>\n",
       "      <th>pred_7f9826be_optimized_score</th>\n",
       "      <th>pred_7f9826be_optimized_ratio</th>\n",
       "      <th>pred_900500fe_optimized_score</th>\n",
       "      <th>pred_900500fe_optimized_ratio</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>p.E33Q</td>\n",
       "      <td>2.457000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.457000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.996143</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>E33Q</td>\n",
       "      <td>0.998343</td>\n",
       "      <td>...</td>\n",
       "      <td>1.582201</td>\n",
       "      <td>0.000155</td>\n",
       "      <td>0.519324</td>\n",
       "      <td>0.316016</td>\n",
       "      <td>-2.457000</td>\n",
       "      <td>0.996143</td>\n",
       "      <td>0.999970</td>\n",
       "      <td>0.000030</td>\n",
       "      <td>0.999987</td>\n",
       "      <td>0.000013</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>p.E81K</td>\n",
       "      <td>-2.456000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-2.456000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.003807</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>E81K</td>\n",
       "      <td>0.994183</td>\n",
       "      <td>...</td>\n",
       "      <td>1.630727</td>\n",
       "      <td>0.000131</td>\n",
       "      <td>0.090255</td>\n",
       "      <td>0.306612</td>\n",
       "      <td>2.456000</td>\n",
       "      <td>1.003807</td>\n",
       "      <td>0.999963</td>\n",
       "      <td>0.000037</td>\n",
       "      <td>0.999974</td>\n",
       "      <td>0.000026</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>p.L121I</td>\n",
       "      <td>-2.965667</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-2.965667</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.004512</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>L121I</td>\n",
       "      <td>0.999190</td>\n",
       "      <td>...</td>\n",
       "      <td>1.390088</td>\n",
       "      <td>0.000108</td>\n",
       "      <td>0.157240</td>\n",
       "      <td>0.359689</td>\n",
       "      <td>2.965667</td>\n",
       "      <td>1.004512</td>\n",
       "      <td>0.994864</td>\n",
       "      <td>0.005136</td>\n",
       "      <td>0.992272</td>\n",
       "      <td>0.007728</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>p.R135K</td>\n",
       "      <td>-2.395333</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-2.395333</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.003528</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>R135K</td>\n",
       "      <td>0.997200</td>\n",
       "      <td>...</td>\n",
       "      <td>1.664755</td>\n",
       "      <td>0.000537</td>\n",
       "      <td>0.310411</td>\n",
       "      <td>0.300345</td>\n",
       "      <td>2.395333</td>\n",
       "      <td>1.003528</td>\n",
       "      <td>0.999991</td>\n",
       "      <td>0.000009</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.000007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>p.D162G</td>\n",
       "      <td>-3.375667</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-3.375667</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.005196</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>D162G</td>\n",
       "      <td>0.998402</td>\n",
       "      <td>...</td>\n",
       "      <td>2.243989</td>\n",
       "      <td>0.000142</td>\n",
       "      <td>0.053119</td>\n",
       "      <td>0.222818</td>\n",
       "      <td>3.375667</td>\n",
       "      <td>1.005196</td>\n",
       "      <td>0.999963</td>\n",
       "      <td>0.000037</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.000007</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  MAPK1-variant  DDG-NotPO4  Standard-Deviation-DDG-NotPO4   DDG-PO4  \\\n",
       "0        p.E33Q    2.457000                            1.0  2.457000   \n",
       "1        p.E81K   -2.456000                            1.0 -2.456000   \n",
       "2       p.L121I   -2.965667                            1.0 -2.965667   \n",
       "3       p.R135K   -2.395333                            1.0 -2.395333   \n",
       "4       p.D162G   -3.375667                            1.0 -3.375667   \n",
       "\n",
       "    Standard-Deviation-DDG-PO4  Ratio-kcat/Km  \\\n",
       "0                          1.0       0.996143   \n",
       "1                          1.0       1.003807   \n",
       "2                          1.0       1.004512   \n",
       "3                          1.0       1.003528   \n",
       "4                          1.0       1.005196   \n",
       "\n",
       "   Standard-Deviation-Ratio-kcat/Km Comment    mut  protbert_score  ...  \\\n",
       "0                               1.0       *   E33Q        0.998343  ...   \n",
       "1                               1.0       *   E81K        0.994183  ...   \n",
       "2                               1.0       *  L121I        0.999190  ...   \n",
       "3                               1.0       *  R135K        0.997200  ...   \n",
       "4                               1.0       *  D162G        0.998402  ...   \n",
       "\n",
       "   el2_score  protbert_ratio  proteinsolver_ratio  el2_ratio  \\\n",
       "0   1.582201        0.000155             0.519324   0.316016   \n",
       "1   1.630727        0.000131             0.090255   0.306612   \n",
       "2   1.390088        0.000108             0.157240   0.359689   \n",
       "3   1.664755        0.000537             0.310411   0.300345   \n",
       "4   2.243989        0.000142             0.053119   0.222818   \n",
       "\n",
       "   rosetta_dg_score  rosetta_dg_ratio  pred_7f9826be_optimized_score  \\\n",
       "0         -2.457000          0.996143                       0.999970   \n",
       "1          2.456000          1.003807                       0.999963   \n",
       "2          2.965667          1.004512                       0.994864   \n",
       "3          2.395333          1.003528                       0.999991   \n",
       "4          3.375667          1.005196                       0.999963   \n",
       "\n",
       "   pred_7f9826be_optimized_ratio  pred_900500fe_optimized_score  \\\n",
       "0                       0.000030                       0.999987   \n",
       "1                       0.000037                       0.999974   \n",
       "2                       0.005136                       0.992272   \n",
       "3                       0.000009                       0.999993   \n",
       "4                       0.000037                       0.999993   \n",
       "\n",
       "   pred_900500fe_optimized_ratio  \n",
       "0                       0.000013  \n",
       "1                       0.000026  \n",
       "2                       0.007728  \n",
       "3                       0.000007  \n",
       "4                       0.000007  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The file's format is valid! You are good to submit now!\n",
      "pred_7f9826be_optimized\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>MAPK1-variant</th>\n",
       "      <th>DDG-NotPO4</th>\n",
       "      <th>Standard-Deviation-DDG-NotPO4</th>\n",
       "      <th>DDG-PO4</th>\n",
       "      <th>Standard-Deviation-DDG-PO4</th>\n",
       "      <th>Ratio-kcat/Km</th>\n",
       "      <th>Standard-Deviation-Ratio-kcat/Km</th>\n",
       "      <th>Comment</th>\n",
       "      <th>mut</th>\n",
       "      <th>protbert_score</th>\n",
       "      <th>...</th>\n",
       "      <th>el2_score</th>\n",
       "      <th>protbert_ratio</th>\n",
       "      <th>proteinsolver_ratio</th>\n",
       "      <th>el2_ratio</th>\n",
       "      <th>rosetta_dg_score</th>\n",
       "      <th>rosetta_dg_ratio</th>\n",
       "      <th>pred_7f9826be_optimized_score</th>\n",
       "      <th>pred_7f9826be_optimized_ratio</th>\n",
       "      <th>pred_900500fe_optimized_score</th>\n",
       "      <th>pred_900500fe_optimized_ratio</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>p.E33Q</td>\n",
       "      <td>-0.999970</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.999970</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000030</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>E33Q</td>\n",
       "      <td>0.998343</td>\n",
       "      <td>...</td>\n",
       "      <td>1.582201</td>\n",
       "      <td>0.000155</td>\n",
       "      <td>0.519324</td>\n",
       "      <td>0.316016</td>\n",
       "      <td>-2.457000</td>\n",
       "      <td>0.996143</td>\n",
       "      <td>0.999970</td>\n",
       "      <td>0.000030</td>\n",
       "      <td>0.999987</td>\n",
       "      <td>0.000013</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>p.E81K</td>\n",
       "      <td>-0.999963</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.999963</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000037</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>E81K</td>\n",
       "      <td>0.994183</td>\n",
       "      <td>...</td>\n",
       "      <td>1.630727</td>\n",
       "      <td>0.000131</td>\n",
       "      <td>0.090255</td>\n",
       "      <td>0.306612</td>\n",
       "      <td>2.456000</td>\n",
       "      <td>1.003807</td>\n",
       "      <td>0.999963</td>\n",
       "      <td>0.000037</td>\n",
       "      <td>0.999974</td>\n",
       "      <td>0.000026</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>p.L121I</td>\n",
       "      <td>-0.994864</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.994864</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.005136</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>L121I</td>\n",
       "      <td>0.999190</td>\n",
       "      <td>...</td>\n",
       "      <td>1.390088</td>\n",
       "      <td>0.000108</td>\n",
       "      <td>0.157240</td>\n",
       "      <td>0.359689</td>\n",
       "      <td>2.965667</td>\n",
       "      <td>1.004512</td>\n",
       "      <td>0.994864</td>\n",
       "      <td>0.005136</td>\n",
       "      <td>0.992272</td>\n",
       "      <td>0.007728</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>p.R135K</td>\n",
       "      <td>-0.999991</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.999991</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000009</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>R135K</td>\n",
       "      <td>0.997200</td>\n",
       "      <td>...</td>\n",
       "      <td>1.664755</td>\n",
       "      <td>0.000537</td>\n",
       "      <td>0.310411</td>\n",
       "      <td>0.300345</td>\n",
       "      <td>2.395333</td>\n",
       "      <td>1.003528</td>\n",
       "      <td>0.999991</td>\n",
       "      <td>0.000009</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.000007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>p.D162G</td>\n",
       "      <td>-0.999963</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.999963</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000037</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>D162G</td>\n",
       "      <td>0.998402</td>\n",
       "      <td>...</td>\n",
       "      <td>2.243989</td>\n",
       "      <td>0.000142</td>\n",
       "      <td>0.053119</td>\n",
       "      <td>0.222818</td>\n",
       "      <td>3.375667</td>\n",
       "      <td>1.005196</td>\n",
       "      <td>0.999963</td>\n",
       "      <td>0.000037</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.000007</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  MAPK1-variant  DDG-NotPO4  Standard-Deviation-DDG-NotPO4   DDG-PO4  \\\n",
       "0        p.E33Q   -0.999970                            1.0 -0.999970   \n",
       "1        p.E81K   -0.999963                            1.0 -0.999963   \n",
       "2       p.L121I   -0.994864                            1.0 -0.994864   \n",
       "3       p.R135K   -0.999991                            1.0 -0.999991   \n",
       "4       p.D162G   -0.999963                            1.0 -0.999963   \n",
       "\n",
       "    Standard-Deviation-DDG-PO4  Ratio-kcat/Km  \\\n",
       "0                          1.0       0.000030   \n",
       "1                          1.0       0.000037   \n",
       "2                          1.0       0.005136   \n",
       "3                          1.0       0.000009   \n",
       "4                          1.0       0.000037   \n",
       "\n",
       "   Standard-Deviation-Ratio-kcat/Km Comment    mut  protbert_score  ...  \\\n",
       "0                               1.0       *   E33Q        0.998343  ...   \n",
       "1                               1.0       *   E81K        0.994183  ...   \n",
       "2                               1.0       *  L121I        0.999190  ...   \n",
       "3                               1.0       *  R135K        0.997200  ...   \n",
       "4                               1.0       *  D162G        0.998402  ...   \n",
       "\n",
       "   el2_score  protbert_ratio  proteinsolver_ratio  el2_ratio  \\\n",
       "0   1.582201        0.000155             0.519324   0.316016   \n",
       "1   1.630727        0.000131             0.090255   0.306612   \n",
       "2   1.390088        0.000108             0.157240   0.359689   \n",
       "3   1.664755        0.000537             0.310411   0.300345   \n",
       "4   2.243989        0.000142             0.053119   0.222818   \n",
       "\n",
       "   rosetta_dg_score  rosetta_dg_ratio  pred_7f9826be_optimized_score  \\\n",
       "0         -2.457000          0.996143                       0.999970   \n",
       "1          2.456000          1.003807                       0.999963   \n",
       "2          2.965667          1.004512                       0.994864   \n",
       "3          2.395333          1.003528                       0.999991   \n",
       "4          3.375667          1.005196                       0.999963   \n",
       "\n",
       "   pred_7f9826be_optimized_ratio  pred_900500fe_optimized_score  \\\n",
       "0                       0.000030                       0.999987   \n",
       "1                       0.000037                       0.999974   \n",
       "2                       0.005136                       0.992272   \n",
       "3                       0.000009                       0.999993   \n",
       "4                       0.000037                       0.999993   \n",
       "\n",
       "   pred_900500fe_optimized_ratio  \n",
       "0                       0.000013  \n",
       "1                       0.000026  \n",
       "2                       0.007728  \n",
       "3                       0.000007  \n",
       "4                       0.000007  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The file's format is valid! You are good to submit now!\n",
      "pred_900500fe_optimized\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>MAPK1-variant</th>\n",
       "      <th>DDG-NotPO4</th>\n",
       "      <th>Standard-Deviation-DDG-NotPO4</th>\n",
       "      <th>DDG-PO4</th>\n",
       "      <th>Standard-Deviation-DDG-PO4</th>\n",
       "      <th>Ratio-kcat/Km</th>\n",
       "      <th>Standard-Deviation-Ratio-kcat/Km</th>\n",
       "      <th>Comment</th>\n",
       "      <th>mut</th>\n",
       "      <th>protbert_score</th>\n",
       "      <th>...</th>\n",
       "      <th>el2_score</th>\n",
       "      <th>protbert_ratio</th>\n",
       "      <th>proteinsolver_ratio</th>\n",
       "      <th>el2_ratio</th>\n",
       "      <th>rosetta_dg_score</th>\n",
       "      <th>rosetta_dg_ratio</th>\n",
       "      <th>pred_7f9826be_optimized_score</th>\n",
       "      <th>pred_7f9826be_optimized_ratio</th>\n",
       "      <th>pred_900500fe_optimized_score</th>\n",
       "      <th>pred_900500fe_optimized_ratio</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>p.E33Q</td>\n",
       "      <td>-0.999987</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.999987</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000013</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>E33Q</td>\n",
       "      <td>0.998343</td>\n",
       "      <td>...</td>\n",
       "      <td>1.582201</td>\n",
       "      <td>0.000155</td>\n",
       "      <td>0.519324</td>\n",
       "      <td>0.316016</td>\n",
       "      <td>-2.457000</td>\n",
       "      <td>0.996143</td>\n",
       "      <td>0.999970</td>\n",
       "      <td>0.000030</td>\n",
       "      <td>0.999987</td>\n",
       "      <td>0.000013</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>p.E81K</td>\n",
       "      <td>-0.999974</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.999974</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000026</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>E81K</td>\n",
       "      <td>0.994183</td>\n",
       "      <td>...</td>\n",
       "      <td>1.630727</td>\n",
       "      <td>0.000131</td>\n",
       "      <td>0.090255</td>\n",
       "      <td>0.306612</td>\n",
       "      <td>2.456000</td>\n",
       "      <td>1.003807</td>\n",
       "      <td>0.999963</td>\n",
       "      <td>0.000037</td>\n",
       "      <td>0.999974</td>\n",
       "      <td>0.000026</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>p.L121I</td>\n",
       "      <td>-0.992272</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.992272</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.007728</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>L121I</td>\n",
       "      <td>0.999190</td>\n",
       "      <td>...</td>\n",
       "      <td>1.390088</td>\n",
       "      <td>0.000108</td>\n",
       "      <td>0.157240</td>\n",
       "      <td>0.359689</td>\n",
       "      <td>2.965667</td>\n",
       "      <td>1.004512</td>\n",
       "      <td>0.994864</td>\n",
       "      <td>0.005136</td>\n",
       "      <td>0.992272</td>\n",
       "      <td>0.007728</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>p.R135K</td>\n",
       "      <td>-0.999993</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.999993</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000007</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>R135K</td>\n",
       "      <td>0.997200</td>\n",
       "      <td>...</td>\n",
       "      <td>1.664755</td>\n",
       "      <td>0.000537</td>\n",
       "      <td>0.310411</td>\n",
       "      <td>0.300345</td>\n",
       "      <td>2.395333</td>\n",
       "      <td>1.003528</td>\n",
       "      <td>0.999991</td>\n",
       "      <td>0.000009</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.000007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>p.D162G</td>\n",
       "      <td>-0.999993</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.999993</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000007</td>\n",
       "      <td>1.0</td>\n",
       "      <td>*</td>\n",
       "      <td>D162G</td>\n",
       "      <td>0.998402</td>\n",
       "      <td>...</td>\n",
       "      <td>2.243989</td>\n",
       "      <td>0.000142</td>\n",
       "      <td>0.053119</td>\n",
       "      <td>0.222818</td>\n",
       "      <td>3.375667</td>\n",
       "      <td>1.005196</td>\n",
       "      <td>0.999963</td>\n",
       "      <td>0.000037</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.000007</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  MAPK1-variant  DDG-NotPO4  Standard-Deviation-DDG-NotPO4   DDG-PO4  \\\n",
       "0        p.E33Q   -0.999987                            1.0 -0.999987   \n",
       "1        p.E81K   -0.999974                            1.0 -0.999974   \n",
       "2       p.L121I   -0.992272                            1.0 -0.992272   \n",
       "3       p.R135K   -0.999993                            1.0 -0.999993   \n",
       "4       p.D162G   -0.999993                            1.0 -0.999993   \n",
       "\n",
       "    Standard-Deviation-DDG-PO4  Ratio-kcat/Km  \\\n",
       "0                          1.0       0.000013   \n",
       "1                          1.0       0.000026   \n",
       "2                          1.0       0.007728   \n",
       "3                          1.0       0.000007   \n",
       "4                          1.0       0.000007   \n",
       "\n",
       "   Standard-Deviation-Ratio-kcat/Km Comment    mut  protbert_score  ...  \\\n",
       "0                               1.0       *   E33Q        0.998343  ...   \n",
       "1                               1.0       *   E81K        0.994183  ...   \n",
       "2                               1.0       *  L121I        0.999190  ...   \n",
       "3                               1.0       *  R135K        0.997200  ...   \n",
       "4                               1.0       *  D162G        0.998402  ...   \n",
       "\n",
       "   el2_score  protbert_ratio  proteinsolver_ratio  el2_ratio  \\\n",
       "0   1.582201        0.000155             0.519324   0.316016   \n",
       "1   1.630727        0.000131             0.090255   0.306612   \n",
       "2   1.390088        0.000108             0.157240   0.359689   \n",
       "3   1.664755        0.000537             0.310411   0.300345   \n",
       "4   2.243989        0.000142             0.053119   0.222818   \n",
       "\n",
       "   rosetta_dg_score  rosetta_dg_ratio  pred_7f9826be_optimized_score  \\\n",
       "0         -2.457000          0.996143                       0.999970   \n",
       "1          2.456000          1.003807                       0.999963   \n",
       "2          2.965667          1.004512                       0.994864   \n",
       "3          2.395333          1.003528                       0.999991   \n",
       "4          3.375667          1.005196                       0.999963   \n",
       "\n",
       "   pred_7f9826be_optimized_ratio  pred_900500fe_optimized_score  \\\n",
       "0                       0.000030                       0.999987   \n",
       "1                       0.000037                       0.999974   \n",
       "2                       0.005136                       0.992272   \n",
       "3                       0.000009                       0.999993   \n",
       "4                       0.000037                       0.999993   \n",
       "\n",
       "   pred_900500fe_optimized_ratio  \n",
       "0                       0.000013  \n",
       "1                       0.000026  \n",
       "2                       0.007728  \n",
       "3                       0.000007  \n",
       "4                       0.000007  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The file's format is valid! You are good to submit now!\n"
     ]
    }
   ],
   "source": [
    "result_dfs = {}\n",
    "for i, metric in enumerate(\n",
    "    [\n",
    "        \"el2\",\n",
    "        \"proteinsolver\",\n",
    "        \"protbert\",\n",
    "        \"rosetta_dg\",\n",
    "        \"pred_7f9826be_optimized\",\n",
    "        \"pred_900500fe_optimized\",\n",
    "    ]\n",
    "):\n",
    "    result_dfs[metric] = final_result_df.copy()\n",
    "    result_dfs[metric][\"DDG-NotPO4\"] = -result_dfs[metric][f\"{metric}_score\"].values\n",
    "    result_dfs[metric][\"Standard-Deviation-DDG-NotPO4\"] = 1.0\n",
    "    result_dfs[metric][\"DDG-PO4\"] = -result_dfs[metric][f\"{metric}_score\"].values\n",
    "    result_dfs[metric][\" Standard-Deviation-DDG-PO4\"] = 1.0\n",
    "    result_dfs[metric][\"Ratio-kcat/Km\"] = result_dfs[metric][f\"{metric}_ratio\"].values\n",
    "    result_dfs[metric][\"Standard-Deviation-Ratio-kcat/Km\"] = 1.0\n",
    "    result_dfs[metric][\"Comment\"] = \"*\"\n",
    "\n",
    "    print(metric)\n",
    "    display(result_dfs[metric].head())\n",
    "\n",
    "    output_file = output_dir.joinpath(f\"strokach_modelnumber_{i + 1}.tsv\")\n",
    "    result_dfs[metric][result_template_df.columns].to_csv(output_file, sep=\"\\t\", index=False)\n",
    "    !python {NOTEBOOK_DIR}/validation.py {output_file}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "75c1fbe0-299d-4f43-bb04-7e22f1adb1b9",
   "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}