{
 "cells": [
  {
   "cell_type": "markdown",
   "source": [
    "# Prepare REPTILE"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "cf931148",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-05T03:45:27.379905Z",
     "start_time": "2022-01-05T03:45:25.690665Z"
    }
   },
   "outputs": [],
   "source": [
    "import pathlib\n",
    "import pandas as pd\n",
    "from ALLCools.mcds import RegionDS"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "12d7c495",
   "metadata": {},
   "source": [
    "## Prepare Files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "34cf5f76",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-05T03:45:38.509598Z",
     "start_time": "2022-01-05T03:45:27.381912Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using dmr as region_dim\n"
     ]
    }
   ],
   "source": [
    "# save DMR bed\n",
    "dmr_ds = RegionDS.open('../RegionDS/HIP')\n",
    "dmr_ds.get_bed().to_csv('DMR.bed', sep='\\t', index=None, header=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "7b821045",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-05T03:45:38.536113Z",
     "start_time": "2022-01-05T03:45:38.512926Z"
    }
   },
   "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>mC</th>\n",
       "      <th>ATAC</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ASC</th>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CA1</th>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CA23</th>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CGE</th>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DG</th>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MGC</th>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MGE</th>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NonN</th>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ODC</th>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>OPC</th>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/HIPBulk...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mESC</th>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/REPTILE...</td>\n",
       "      <td>/home/hanliu/project/allcools_doc/data/REPTILE...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                     mC  \\\n",
       "ASC   /home/hanliu/project/allcools_doc/data/HIPBulk...   \n",
       "CA1   /home/hanliu/project/allcools_doc/data/HIPBulk...   \n",
       "CA23  /home/hanliu/project/allcools_doc/data/HIPBulk...   \n",
       "CGE   /home/hanliu/project/allcools_doc/data/HIPBulk...   \n",
       "DG    /home/hanliu/project/allcools_doc/data/HIPBulk...   \n",
       "MGC   /home/hanliu/project/allcools_doc/data/HIPBulk...   \n",
       "MGE   /home/hanliu/project/allcools_doc/data/HIPBulk...   \n",
       "NonN  /home/hanliu/project/allcools_doc/data/HIPBulk...   \n",
       "ODC   /home/hanliu/project/allcools_doc/data/HIPBulk...   \n",
       "OPC   /home/hanliu/project/allcools_doc/data/HIPBulk...   \n",
       "mESC  /home/hanliu/project/allcools_doc/data/REPTILE...   \n",
       "\n",
       "                                                   ATAC  \n",
       "ASC   /home/hanliu/project/allcools_doc/data/HIPBulk...  \n",
       "CA1   /home/hanliu/project/allcools_doc/data/HIPBulk...  \n",
       "CA23  /home/hanliu/project/allcools_doc/data/HIPBulk...  \n",
       "CGE   /home/hanliu/project/allcools_doc/data/HIPBulk...  \n",
       "DG    /home/hanliu/project/allcools_doc/data/HIPBulk...  \n",
       "MGC   /home/hanliu/project/allcools_doc/data/HIPBulk...  \n",
       "MGE   /home/hanliu/project/allcools_doc/data/HIPBulk...  \n",
       "NonN  /home/hanliu/project/allcools_doc/data/HIPBulk...  \n",
       "ODC   /home/hanliu/project/allcools_doc/data/HIPBulk...  \n",
       "OPC   /home/hanliu/project/allcools_doc/data/HIPBulk...  \n",
       "mESC  /home/hanliu/project/allcools_doc/data/REPTILE...  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# save bigwig table\n",
    "mc_table = {\n",
    "    p.name.split('.')[0].split('_')[-1].split('-')[0]: str(p)\n",
    "    for p in pathlib.Path(\n",
    "        '/home/hanliu/project/allcools_doc/data/HIPBulk/mc_bulk/').glob(\n",
    "            'snmC_*/snmC_*.CGN-both.frac.bw')\n",
    "}\n",
    "mc_table['mESC'] = '/home/hanliu/project/allcools_doc/data/REPTILE/mESC_Meth.bw'\n",
    "\n",
    "atac_table = {\n",
    "    p.name.split('.')[0].split('_')[-1]: str(p)\n",
    "    for p in pathlib.Path(\n",
    "        '/home/hanliu/project/allcools_doc/data/HIPBulk/atac_bulk/').glob(\n",
    "            'HIP_snATAC_*.bw')\n",
    "}\n",
    "atac_table['mESC'] = '/home/hanliu/project/allcools_doc/data/REPTILE/mESC_ATAC.bw'\n",
    "bigwig_table = pd.DataFrame({'mC': mc_table, 'ATAC': atac_table})\n",
    "bigwig_table.to_csv('bigwigs.csv')\n",
    "bigwig_table"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2f1a0736",
   "metadata": {},
   "source": [
    "## Final Parameters User Needs to Provide"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "1140290e",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-01-05T03:45:38.540548Z",
     "start_time": "2022-01-05T03:45:38.537144Z"
    }
   },
   "outputs": [],
   "source": [
    "parameters = dict(\n",
    "    output_path=\n",
    "    '/home/hanliu/project/allcools_doc/cluster_level/REPTILE/REPTILE',\n",
    "    # four column bed file, last column is id\n",
    "    train_regions=\n",
    "    '/home/hanliu/project/allcools_doc/data/REPTILE/mESC_region_for_train.bed',\n",
    "    dmr_regions=\n",
    "    '/home/hanliu/project/allcools_doc/cluster_level/REPTILE/DMR.bed',\n",
    "\n",
    "    # id matched training region labels\n",
    "    train_region_labels=\n",
    "    '/home/hanliu/project/allcools_doc/data/REPTILE/mESC_region_for_train_label.tsv',\n",
    "    train_sample='mESC',\n",
    "\n",
    "    # sample (including the traning sample) by column (modalities) bigwig path table\n",
    "    bigwig_table=\n",
    "    '/home/hanliu/project/allcools_doc/cluster_level/REPTILE/bigwigs.csv',\n",
    "    chrom_size_path='/home/hanliu/ref/mouse/genome/mm10.main.chrom.sizes',\n",
    "    window_size=2000,\n",
    "    step_size=200,\n",
    "    dmr_slop=150)\n",
    "\n",
    "import json\n",
    "with open('parameters.json', 'w') as f:\n",
    "    json.dump(parameters, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "09a18c29",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b35e56fb",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d76ee944",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dfe343b1",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ad42d900",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "86a99044",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c2741ada",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4ad6f5d9",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3adfe404",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "53ce8e18",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c9fccf45",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ef059fa5",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "hide_input": false,
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.6"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": true,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}