8.gtex-interpret/1B.download-hematopoietic-data.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Download Publicly Available Hematopoietic Dataset\n",
    "\n",
    "**Gregory Way, 2018**\n",
    "\n",
    "Here, I download [GSE24759](https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE24759) which is associated with [Novershtern et al. 2011](https://doi.org/10.1016/j.cell.2011.01.004).\n",
    "\n",
    "This dataset includes 211 samples consisting of 38 distinct hematopoietic states in various stages of differentiation.\n",
    "\n",
    "We hypothesized that our constructed feature identified through our interpret compression approach would have higher activation patterns in Monocytes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import csv\n",
    "import pandas as pd\n",
    "from sklearn import preprocessing\n",
    "\n",
    "from scripts.utils import download_geo"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "base_url = 'ftp://ftp.ncbi.nlm.nih.gov/geo/series/GSE24nnn/GSE24759/suppl/'\n",
    "name = 'GSE24759_data.sort.txt.gz'\n",
    "directory = 'download'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "download_geo(base_url, name, directory)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "98c248be560f0f422dbb07f24083306a58d208806eeb4d9b4f7da4f41de80905  download/GSE24759_data.sort.txt.gz\r\n"
     ]
    }
   ],
   "source": [
    "path = 'download/GSE24759_data.sort.txt.gz'\n",
    "! sha256sum $path"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Process the Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(8968, 213)\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>A_Name</th>\n",
       "      <th>A_Desc</th>\n",
       "      <th>HSC1_1</th>\n",
       "      <th>HSC1_13</th>\n",
       "      <th>HSC1_14</th>\n",
       "      <th>HSC1_2</th>\n",
       "      <th>HSC1_3</th>\n",
       "      <th>HSC1_4</th>\n",
       "      <th>HSC1_6</th>\n",
       "      <th>HSC1_7</th>\n",
       "      <th>...</th>\n",
       "      <th>TCELLA7_6</th>\n",
       "      <th>TCELLA7_7</th>\n",
       "      <th>TCELLA7_8</th>\n",
       "      <th>TCELLA8_10</th>\n",
       "      <th>TCELLA8_11</th>\n",
       "      <th>TCELLA8_3</th>\n",
       "      <th>TCELLA8_5</th>\n",
       "      <th>TCELLA8_6</th>\n",
       "      <th>TCELLA8_7</th>\n",
       "      <th>TCELLA8_8</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>8563</td>\n",
       "      <td>THOC5</td>\n",
       "      <td>-0.230509</td>\n",
       "      <td>-0.233209</td>\n",
       "      <td>0.093491</td>\n",
       "      <td>-0.366109</td>\n",
       "      <td>-0.343909</td>\n",
       "      <td>0.081691</td>\n",
       "      <td>-0.156809</td>\n",
       "      <td>-0.364609</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.309709</td>\n",
       "      <td>0.421291</td>\n",
       "      <td>0.114291</td>\n",
       "      <td>-0.085009</td>\n",
       "      <td>0.128091</td>\n",
       "      <td>-0.463309</td>\n",
       "      <td>0.287691</td>\n",
       "      <td>0.388191</td>\n",
       "      <td>-0.509209</td>\n",
       "      <td>0.097591</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>8434</td>\n",
       "      <td>RECK</td>\n",
       "      <td>0.768078</td>\n",
       "      <td>0.172978</td>\n",
       "      <td>0.331778</td>\n",
       "      <td>0.628078</td>\n",
       "      <td>0.742478</td>\n",
       "      <td>-0.181122</td>\n",
       "      <td>0.597078</td>\n",
       "      <td>0.497378</td>\n",
       "      <td>...</td>\n",
       "      <td>0.684478</td>\n",
       "      <td>0.335278</td>\n",
       "      <td>0.559278</td>\n",
       "      <td>0.647978</td>\n",
       "      <td>-0.390622</td>\n",
       "      <td>0.931678</td>\n",
       "      <td>0.648878</td>\n",
       "      <td>0.619878</td>\n",
       "      <td>0.614878</td>\n",
       "      <td>0.746678</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 213 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   A_Name A_Desc    HSC1_1   HSC1_13   HSC1_14    HSC1_2    HSC1_3    HSC1_4  \\\n",
       "0    8563  THOC5 -0.230509 -0.233209  0.093491 -0.366109 -0.343909  0.081691   \n",
       "1    8434   RECK  0.768078  0.172978  0.331778  0.628078  0.742478 -0.181122   \n",
       "\n",
       "     HSC1_6    HSC1_7    ...      TCELLA7_6  TCELLA7_7  TCELLA7_8  TCELLA8_10  \\\n",
       "0 -0.156809 -0.364609    ...      -0.309709   0.421291   0.114291   -0.085009   \n",
       "1  0.597078  0.497378    ...       0.684478   0.335278   0.559278    0.647978   \n",
       "\n",
       "   TCELLA8_11  TCELLA8_3  TCELLA8_5  TCELLA8_6  TCELLA8_7  TCELLA8_8  \n",
       "0    0.128091  -0.463309   0.287691   0.388191  -0.509209   0.097591  \n",
       "1   -0.390622   0.931678   0.648878   0.619878   0.614878   0.746678  \n",
       "\n",
       "[2 rows x 213 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load Additional File 3\n",
    "geo_df = pd.read_table(path)\n",
    "\n",
    "print(geo_df.shape)\n",
    "geo_df.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Update Gene Names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load curated gene names from versioned resource \n",
    "commit = '721204091a96e55de6dcad165d6d8265e67e2a48'\n",
    "url = 'https://raw.githubusercontent.com/cognoma/genes/{}/data/genes.tsv'.format(commit)\n",
    "gene_df = pd.read_table(url)\n",
    "\n",
    "# Only consider protein-coding genes\n",
    "gene_df = (\n",
    "    gene_df.query(\"gene_type == 'protein-coding'\")\n",
    ")\n",
    "\n",
    "symbol_to_entrez = dict(zip(gene_df.symbol,\n",
    "                            gene_df.entrez_gene_id))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Add alternative symbols to entrez mapping dictionary\n",
    "gene_df = gene_df.dropna(axis='rows', subset=['synonyms'])\n",
    "gene_df.synonyms = gene_df.synonyms.str.split('|')\n",
    "\n",
    "all_syn = (\n",
    "    gene_df.apply(lambda x: pd.Series(x.synonyms), axis=1)\n",
    "    .stack()\n",
    "    .reset_index(level=1, drop=True)\n",
    ")\n",
    "\n",
    "# Name the synonym series and join with rest of genes\n",
    "all_syn.name = 'all_synonyms'\n",
    "gene_with_syn_df = gene_df.join(all_syn)\n",
    "\n",
    "# Remove rows that have redundant symbols in all_synonyms\n",
    "gene_with_syn_df = (\n",
    "    gene_with_syn_df\n",
    "    \n",
    "    # Drop synonyms that are duplicated - can't be sure of mapping\n",
    "    .drop_duplicates(['all_synonyms'], keep=False)\n",
    "\n",
    "    # Drop rows in which the symbol appears in the list of synonyms\n",
    "    .query('symbol not in all_synonyms')\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create a synonym to entrez mapping and add to dictionary\n",
    "synonym_to_entrez = dict(zip(gene_with_syn_df.all_synonyms,\n",
    "                             gene_with_syn_df.entrez_gene_id))\n",
    "\n",
    "symbol_to_entrez.update(synonym_to_entrez)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load gene updater\n",
    "commit = '721204091a96e55de6dcad165d6d8265e67e2a48'\n",
    "url = 'https://raw.githubusercontent.com/cognoma/genes/{}/data/updater.tsv'.format(commit)\n",
    "updater_df = pd.read_table(url)\n",
    "old_to_new_entrez = dict(zip(updater_df.old_entrez_gene_id,\n",
    "                             updater_df.new_entrez_gene_id))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Update the symbol column to entrez_gene_id\n",
    "geo_map = geo_df.A_Desc.replace(symbol_to_entrez)\n",
    "geo_map = geo_map.replace(old_to_new_entrez)\n",
    "geo_df.index = geo_map\n",
    "geo_df.index.name = 'entrez_gene_id'\n",
    "geo_df = geo_df.drop(['A_Name', 'A_Desc'], axis='columns')\n",
    "geo_df = geo_df.loc[geo_df.index.isin(symbol_to_entrez.values()), :]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Scale Data and Output to File"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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>entrez_gene_id</th>\n",
       "      <th>2</th>\n",
       "      <th>9</th>\n",
       "      <th>14</th>\n",
       "      <th>16</th>\n",
       "      <th>18</th>\n",
       "      <th>19</th>\n",
       "      <th>20</th>\n",
       "      <th>22</th>\n",
       "      <th>23</th>\n",
       "      <th>25</th>\n",
       "      <th>...</th>\n",
       "      <th>100137049</th>\n",
       "      <th>100462977</th>\n",
       "      <th>100505741</th>\n",
       "      <th>100507436</th>\n",
       "      <th>100527963</th>\n",
       "      <th>100652748</th>\n",
       "      <th>102723547</th>\n",
       "      <th>105369230</th>\n",
       "      <th>107984923</th>\n",
       "      <th>107987479</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>BASO1_3</th>\n",
       "      <td>0.389491</td>\n",
       "      <td>0.342512</td>\n",
       "      <td>0.173372</td>\n",
       "      <td>0.196964</td>\n",
       "      <td>0.431818</td>\n",
       "      <td>0.405355</td>\n",
       "      <td>0.200679</td>\n",
       "      <td>0.580138</td>\n",
       "      <td>0.457497</td>\n",
       "      <td>0.475122</td>\n",
       "      <td>...</td>\n",
       "      <td>0.150598</td>\n",
       "      <td>0.297524</td>\n",
       "      <td>0.833020</td>\n",
       "      <td>0.410220</td>\n",
       "      <td>0.379413</td>\n",
       "      <td>0.271206</td>\n",
       "      <td>0.467457</td>\n",
       "      <td>0.842266</td>\n",
       "      <td>0.441286</td>\n",
       "      <td>0.384583</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BASO1_4</th>\n",
       "      <td>0.308525</td>\n",
       "      <td>0.233696</td>\n",
       "      <td>0.281117</td>\n",
       "      <td>0.385034</td>\n",
       "      <td>0.540101</td>\n",
       "      <td>0.317294</td>\n",
       "      <td>0.262179</td>\n",
       "      <td>0.688354</td>\n",
       "      <td>0.453625</td>\n",
       "      <td>0.586759</td>\n",
       "      <td>...</td>\n",
       "      <td>0.268785</td>\n",
       "      <td>0.490944</td>\n",
       "      <td>0.770530</td>\n",
       "      <td>0.495008</td>\n",
       "      <td>0.660851</td>\n",
       "      <td>0.329928</td>\n",
       "      <td>0.365230</td>\n",
       "      <td>0.923228</td>\n",
       "      <td>0.487514</td>\n",
       "      <td>0.371348</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BASO1_5</th>\n",
       "      <td>0.318499</td>\n",
       "      <td>0.304720</td>\n",
       "      <td>0.286889</td>\n",
       "      <td>0.198441</td>\n",
       "      <td>0.710227</td>\n",
       "      <td>0.337265</td>\n",
       "      <td>0.185538</td>\n",
       "      <td>0.729035</td>\n",
       "      <td>0.280381</td>\n",
       "      <td>0.273253</td>\n",
       "      <td>...</td>\n",
       "      <td>0.127869</td>\n",
       "      <td>0.139705</td>\n",
       "      <td>0.951369</td>\n",
       "      <td>0.641621</td>\n",
       "      <td>0.542492</td>\n",
       "      <td>0.176403</td>\n",
       "      <td>0.720273</td>\n",
       "      <td>0.784543</td>\n",
       "      <td>0.223465</td>\n",
       "      <td>0.468075</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BASO1_6</th>\n",
       "      <td>0.478534</td>\n",
       "      <td>0.254761</td>\n",
       "      <td>0.094522</td>\n",
       "      <td>0.142994</td>\n",
       "      <td>0.649876</td>\n",
       "      <td>0.320514</td>\n",
       "      <td>0.205943</td>\n",
       "      <td>0.794756</td>\n",
       "      <td>0.427135</td>\n",
       "      <td>0.567184</td>\n",
       "      <td>...</td>\n",
       "      <td>0.177826</td>\n",
       "      <td>0.333096</td>\n",
       "      <td>0.883992</td>\n",
       "      <td>0.452337</td>\n",
       "      <td>0.352422</td>\n",
       "      <td>0.334565</td>\n",
       "      <td>0.417169</td>\n",
       "      <td>0.866243</td>\n",
       "      <td>0.286279</td>\n",
       "      <td>0.314256</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BASO1_7</th>\n",
       "      <td>0.315282</td>\n",
       "      <td>0.447026</td>\n",
       "      <td>0.121143</td>\n",
       "      <td>0.196732</td>\n",
       "      <td>0.505682</td>\n",
       "      <td>0.303111</td>\n",
       "      <td>0.169100</td>\n",
       "      <td>0.562777</td>\n",
       "      <td>0.388825</td>\n",
       "      <td>0.266749</td>\n",
       "      <td>...</td>\n",
       "      <td>0.246057</td>\n",
       "      <td>0.343073</td>\n",
       "      <td>0.855247</td>\n",
       "      <td>0.627348</td>\n",
       "      <td>0.346892</td>\n",
       "      <td>0.298343</td>\n",
       "      <td>0.221455</td>\n",
       "      <td>0.626021</td>\n",
       "      <td>0.431160</td>\n",
       "      <td>0.387751</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 8190 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "entrez_gene_id         2         9        14        16        18        19  \\\n",
       "BASO1_3         0.389491  0.342512  0.173372  0.196964  0.431818  0.405355   \n",
       "BASO1_4         0.308525  0.233696  0.281117  0.385034  0.540101  0.317294   \n",
       "BASO1_5         0.318499  0.304720  0.286889  0.198441  0.710227  0.337265   \n",
       "BASO1_6         0.478534  0.254761  0.094522  0.142994  0.649876  0.320514   \n",
       "BASO1_7         0.315282  0.447026  0.121143  0.196732  0.505682  0.303111   \n",
       "\n",
       "entrez_gene_id        20        22        23        25    ...      100137049  \\\n",
       "BASO1_3         0.200679  0.580138  0.457497  0.475122    ...       0.150598   \n",
       "BASO1_4         0.262179  0.688354  0.453625  0.586759    ...       0.268785   \n",
       "BASO1_5         0.185538  0.729035  0.280381  0.273253    ...       0.127869   \n",
       "BASO1_6         0.205943  0.794756  0.427135  0.567184    ...       0.177826   \n",
       "BASO1_7         0.169100  0.562777  0.388825  0.266749    ...       0.246057   \n",
       "\n",
       "entrez_gene_id  100462977  100505741  100507436  100527963  100652748  \\\n",
       "BASO1_3          0.297524   0.833020   0.410220   0.379413   0.271206   \n",
       "BASO1_4          0.490944   0.770530   0.495008   0.660851   0.329928   \n",
       "BASO1_5          0.139705   0.951369   0.641621   0.542492   0.176403   \n",
       "BASO1_6          0.333096   0.883992   0.452337   0.352422   0.334565   \n",
       "BASO1_7          0.343073   0.855247   0.627348   0.346892   0.298343   \n",
       "\n",
       "entrez_gene_id  102723547  105369230  107984923  107987479  \n",
       "BASO1_3          0.467457   0.842266   0.441286   0.384583  \n",
       "BASO1_4          0.365230   0.923228   0.487514   0.371348  \n",
       "BASO1_5          0.720273   0.784543   0.223465   0.468075  \n",
       "BASO1_6          0.417169   0.866243   0.286279   0.314256  \n",
       "BASO1_7          0.221455   0.626021   0.431160   0.387751  \n",
       "\n",
       "[5 rows x 8190 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Scale RNAseq data using zero-one normalization\n",
    "geo_scaled_zeroone_df = preprocessing.MinMaxScaler().fit_transform(geo_df.transpose())\n",
    "geo_scaled_zeroone_df = (\n",
    "    pd.DataFrame(geo_scaled_zeroone_df,\n",
    "                 columns=geo_df.index,\n",
    "                 index=geo_df.columns)\n",
    "    .sort_index(axis='columns')\n",
    "    .sort_index(axis='rows')\n",
    ")\n",
    "\n",
    "geo_scaled_zeroone_df.columns = geo_scaled_zeroone_df.columns.astype(str)\n",
    "geo_scaled_zeroone_df = geo_scaled_zeroone_df.loc[:, ~geo_scaled_zeroone_df.columns.duplicated(keep='first')]\n",
    "\n",
    "os.makedirs('data', exist_ok=True)\n",
    "\n",
    "geo_scaled_zeroone_df.columns = geo_scaled_zeroone_df.columns.astype(str)\n",
    "geo_scaled_zeroone_df = geo_scaled_zeroone_df.loc[:, ~geo_scaled_zeroone_df.columns.duplicated(keep='first')]\n",
    "\n",
    "file = os.path.join('data', 'GSE24759_processed_matrix.tsv.gz')\n",
    "geo_scaled_zeroone_df.to_csv(file, sep='\\t', compression='gzip')\n",
    "\n",
    "geo_scaled_zeroone_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Process Cell-Type Classification\n",
    "\n",
    "Data acquired from Supplementary Table 1 of [Novershtern et al. 2011](https://doi.org/10.1016/j.cell.2011.01.004)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "cell_class = {\n",
    "    # Hematopoietic Stem Cells\n",
    "    'HSC1': ['HSC', 'Non Monocyte'],\n",
    "    'HSC2': ['HSC', 'Non Monocyte'],\n",
    "    'HSC3': ['HSC', 'Non Monocyte'],\n",
    "    \n",
    "    # Myeloid Progenitors\n",
    "    'CMP': ['Myeloid', 'Non Monocyte'],\n",
    "    'MEP': ['Myeloid', 'Non Monocyte'],\n",
    "    'GMP': ['Myeloid', 'Non Monocyte'],\n",
    "    \n",
    "    # Erythroid Populations\n",
    "    'ERY1': ['Erythroid', 'Non Monocyte'],\n",
    "    'ERY2': ['Erythroid', 'Non Monocyte'],\n",
    "    'ERY3': ['Erythroid', 'Non Monocyte'],\n",
    "    'ERY4': ['Erythroid', 'Non Monocyte'],\n",
    "    'ERY5': ['Erythroid', 'Non Monocyte'],\n",
    "    \n",
    "    # Megakaryocytic Populations\n",
    "    'MEGA1': ['Megakaryocytic', 'Non Monocyte'],\n",
    "    'MEGA2': ['Megakaryocytic', 'Non Monocyte'],\n",
    "    \n",
    "    # Granulocytic Populations\n",
    "    'GRAN1': ['Granulocytic', 'Non Monocyte'],\n",
    "    'GRAN2': ['Granulocytic', 'Non Monocyte'],\n",
    "    'GRAN3': ['Granulocytic', 'Non Monocyte'],\n",
    "    \n",
    "    # Monocyte Population (Note MONO1 is a CFU-M)\n",
    "    'MONO1': ['Monocyte', 'Non Monocyte'],\n",
    "    'MONO2': ['Monocyte', 'Monocyte'],\n",
    "    \n",
    "    # Basophil Population\n",
    "    'BASO1': ['Basophil', 'Non Monocyte'],\n",
    "    \n",
    "    # Eosinophil Population\n",
    "    'EOS2': ['Eosinophil', 'Non Monocyte'],\n",
    "    \n",
    "    # B Lymphoid Progenitors\n",
    "    'PRE_BCELL2': ['B Lymphoid Progenitor', 'Non Monocyte'],\n",
    "    'PRE_BCELL3': ['B Lymphoid Progenitor', 'Non Monocyte'],\n",
    "    \n",
    "    # Naive Lymphoid Progenitors\n",
    "    'BCELLA1': ['Naive Lymphoid', 'Non Monocyte'],\n",
    "    'TCELLA6': ['Naive Lymphoid', 'Non Monocyte'],\n",
    "    'TCELLA2': ['Naive Lymphoid', 'Non Monocyte'],\n",
    "    \n",
    "    # Differentiated B Cells\n",
    "    'BCELLA2': ['Differentiated B Cell', 'Non Monocyte'],\n",
    "    'BCELLA3': ['Differentiated B Cell', 'Non Monocyte'],\n",
    "    'BCELLA4': ['Differentiated B Cell', 'Non Monocyte'],\n",
    "    \n",
    "    # Differentiated T Cells\n",
    "    'TCELLA7': ['Differentiated T Cell', 'Non Monocyte'],\n",
    "    'TCELLA8': ['Differentiated T Cell', 'Non Monocyte'],\n",
    "    'TCELLA1': ['Differentiated T Cell', 'Non Monocyte'],\n",
    "    'TCELLA3': ['Differentiated T Cell', 'Non Monocyte'],\n",
    "    'TCELLA4': ['Differentiated T Cell', 'Non Monocyte'],\n",
    "    \n",
    "    # Natural Killer Population\n",
    "    'NKA1': ['NK Cell', 'Non Monocyte'],\n",
    "    'NKA2': ['NK Cell', 'Non Monocyte'],\n",
    "    'NKA3': ['NK Cell', 'Non Monocyte'],\n",
    "    'NKA4': ['NK Cell', 'Non Monocyte'],\n",
    "    \n",
    "    # Dendritic Cell\n",
    "    'DENDA1': ['Dendritic', 'Non Monocyte'],\n",
    "    'DENDA2': ['Dendritic', 'Non Monocyte'],\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "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>label</th>\n",
       "      <th>classification</th>\n",
       "      <th>monocyte</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>HSC1</td>\n",
       "      <td>HSC</td>\n",
       "      <td>Non Monocyte</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>HSC2</td>\n",
       "      <td>HSC</td>\n",
       "      <td>Non Monocyte</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HSC3</td>\n",
       "      <td>HSC</td>\n",
       "      <td>Non Monocyte</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CMP</td>\n",
       "      <td>Myeloid</td>\n",
       "      <td>Non Monocyte</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>MEP</td>\n",
       "      <td>Myeloid</td>\n",
       "      <td>Non Monocyte</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  label classification      monocyte\n",
       "0  HSC1            HSC  Non Monocyte\n",
       "1  HSC2            HSC  Non Monocyte\n",
       "2  HSC3            HSC  Non Monocyte\n",
       "3   CMP        Myeloid  Non Monocyte\n",
       "4   MEP        Myeloid  Non Monocyte"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Write data to file\n",
    "cell_class_df = (\n",
    "    pd.DataFrame.from_dict(cell_class)\n",
    "    .transpose()\n",
    "    .reset_index()\n",
    "    .rename(columns={'index': 'label', 0: 'classification', 1: 'monocyte'})\n",
    ")\n",
    "\n",
    "cell_class_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "file = os.path.join('results', 'cell-type-classification.tsv')\n",
    "cell_class_df.to_csv(file, sep='\\t', index=False)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:biobombe]",
   "language": "python",
   "name": "conda-env-biobombe-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}