pbmc_10.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# PBMC3k: 30% Cells: 1635 Features: 4121"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Preprocess"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['.ipynb_checkpoints',\n",
       " '.Rhistory',\n",
       " '1',\n",
       " '10',\n",
       " '30',\n",
       " '50',\n",
       " '60',\n",
       " '70',\n",
       " '90',\n",
       " '90new',\n",
       " 'barcodes.tsv',\n",
       " 'Comparison-sc10x-3c.ipynb',\n",
       " 'cooc',\n",
       " 'cooc.csv',\n",
       " 'cooc1',\n",
       " 'cooc_standard',\n",
       " 'filtered_feature_bc_matrix',\n",
       " 'genes.tsv',\n",
       " 'matrix.mtx',\n",
       " 'pbmc_orign.ipynb',\n",
       " 'R.txt',\n",
       " 'RaceID3.csv',\n",
       " 'RaceID3new.csv',\n",
       " 'sc10x-3c-sce.rds',\n",
       " 'sc10x-3c.rds',\n",
       " 'sc3_results.xls',\n",
       " 'sc3_results1.xls',\n",
       " 'server',\n",
       " 'seurat-R-stan.csv',\n",
       " 'seurat-R.csv',\n",
       " 'seurat-R05.csv',\n",
       " 'seurat-R1.csv',\n",
       " 'seurat-R8.csv',\n",
       " 'seurat-Rrs1.csv',\n",
       " 'Single-Cell-RNA-seq-Benchmark',\n",
       " 'web_summary.html']"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import os\n",
    "path = 'C:/Users/jzhou417/Desktop/pbmc3k/'\n",
    "if os.path.exists(path):\n",
    "    files = os.listdir(path)\n",
    "files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": true
   },
   "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>barcode</th>\n",
       "      <th>seurat_clusters</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>AAACATACAACCAC</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>AAACATTGAGCTAC</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>AAACATTGATCAGC</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>AAACCGTGCTTCCG</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>AAACCGTGTATGCG</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>AAACGCACTGGTAC</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>AAACGCTGACCAGT</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>AAACGCTGGTTCTT</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>AAACGCTGTAGCCA</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>AAACGCTGTTTCTG</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          barcode  seurat_clusters\n",
       "0  AAACATACAACCAC                3\n",
       "1  AAACATTGAGCTAC                4\n",
       "2  AAACATTGATCAGC                1\n",
       "3  AAACCGTGCTTCCG                5\n",
       "4  AAACCGTGTATGCG                6\n",
       "5  AAACGCACTGGTAC                1\n",
       "6  AAACGCTGACCAGT                3\n",
       "7  AAACGCTGGTTCTT                3\n",
       "8  AAACGCTGTAGCCA                3\n",
       "9  AAACGCTGTTTCTG                5"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "table = pd.read_csv(os.path.join(path, 'seurat-R-stan.csv'))\n",
    "table.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "R_10x = pd.read_csv('seurat-Rrs2.csv')\n",
    "R_10x[\"barcode\"] = R_10x['Unnamed: 0']\n",
    "R_10x.pop('Unnamed: 0')\n",
    "newdata = table.merge(R_10x, on='barcode')\n",
    "newdata['target'] = newdata['seurat_clusters_x'].factorize()[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "sc3 = pd.read_excel('sc3_results.xls')\n",
    "sc3_r = table.merge(sc3, on=\"barcode\")\n",
    "sc3_r['target'] = sc3_r['seurat_clusters'].factorize()[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>barcode</th>\n",
       "      <th>seurat_clusters</th>\n",
       "      <th>sc3_10_clusters</th>\n",
       "      <th>sc3_10_log2_outlier_score</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>AAACATACAACCAC</td>\n",
       "      <td>3</td>\n",
       "      <td>7</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>AAACATTGAGCTAC</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>AAACATTGATCAGC</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>AAACCGTGCTTCCG</td>\n",
       "      <td>5</td>\n",
       "      <td>8</td>\n",
       "      <td>0.0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>AAACCGTGTATGCG</td>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>AAACGCACTGGTAC</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>AAACGCTGACCAGT</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>AAACGCTGGTTCTT</td>\n",
       "      <td>3</td>\n",
       "      <td>7</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>AAACGCTGTAGCCA</td>\n",
       "      <td>3</td>\n",
       "      <td>7</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>AAACTTGAAAAACG</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          barcode  seurat_clusters  sc3_10_clusters  \\\n",
       "0  AAACATACAACCAC                3                7   \n",
       "1  AAACATTGAGCTAC                4                2   \n",
       "2  AAACATTGATCAGC                1                7   \n",
       "3  AAACCGTGCTTCCG                5                8   \n",
       "4  AAACCGTGTATGCG                6                1   \n",
       "5  AAACGCACTGGTAC                1                6   \n",
       "6  AAACGCTGACCAGT                3                1   \n",
       "7  AAACGCTGGTTCTT                3                7   \n",
       "8  AAACGCTGTAGCCA                3                7   \n",
       "9  AAACTTGAAAAACG                4                2   \n",
       "\n",
       "   sc3_10_log2_outlier_score  target  \n",
       "0                        0.0       0  \n",
       "1                        0.0       1  \n",
       "2                        0.0       2  \n",
       "3                        0.0       3  \n",
       "4                        0.0       4  \n",
       "5                        0.0       2  \n",
       "6                        0.0       0  \n",
       "7                        0.0       0  \n",
       "8                        0.0       0  \n",
       "9                        0.0       1  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sc3_r.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "cooc = pd.read_csv('cooc.csv').T.reset_index()[1:]\n",
    "cooc[\"barcode\"] = cooc[\"index\"]\n",
    "cooc[\"cooc_clusters\"] = cooc.loc[:,0]\n",
    "cooc.pop(\"index\")\n",
    "cooc.pop(0)\n",
    "cooc_data = table.merge(cooc, on=\"barcode\")\n",
    "cooc_data['target'] = cooc_data['seurat_clusters'].factorize()[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>barcode</th>\n",
       "      <th>seurat_clusters</th>\n",
       "      <th>cooc_clusters</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>AAACATACAACCAC</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>AAACATTGAGCTAC</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>AAACATTGATCAGC</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>AAACCGTGCTTCCG</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>AAACCGTGTATGCG</td>\n",
       "      <td>6</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>AAACGCACTGGTAC</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>AAACGCTGACCAGT</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>AAACGCTGGTTCTT</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>AAACGCTGTAGCCA</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>AAACGCTGTTTCTG</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          barcode  seurat_clusters cooc_clusters  target\n",
       "0  AAACATACAACCAC                3             2       0\n",
       "1  AAACATTGAGCTAC                4             1       1\n",
       "2  AAACATTGATCAGC                1             2       2\n",
       "3  AAACCGTGCTTCCG                5             1       3\n",
       "4  AAACCGTGTATGCG                6             2       4\n",
       "5  AAACGCACTGGTAC                1             2       2\n",
       "6  AAACGCTGACCAGT                3             2       0\n",
       "7  AAACGCTGGTTCTT                3             2       0\n",
       "8  AAACGCTGTAGCCA                3             2       0\n",
       "9  AAACGCTGTTTCTG                5             2       3"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cooc_data.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "race = pd.read_csv('RaceID3.csv')\n",
    "race[\"barcode\"] = race[\"Unnamed: 0\"]\n",
    "race.pop(\"Unnamed: 0\")\n",
    "race_data = table.merge(race, on=\"barcode\")\n",
    "race_data['target'] = race_data['seurat_clusters'].factorize()[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "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>barcode</th>\n",
       "      <th>seurat_clusters</th>\n",
       "      <th>V1</th>\n",
       "      <th>target</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>AAACATACAACCAC</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>AAACATTGAGCTAC</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>AAACATTGATCAGC</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>AAACCGTGCTTCCG</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>AAACCGTGTATGCG</td>\n",
       "      <td>6</td>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>AAACGCACTGGTAC</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>AAACGCTGACCAGT</td>\n",
       "      <td>3</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>AAACGCTGGTTCTT</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>AAACGCTGTAGCCA</td>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>AAACTTGAAAAACG</td>\n",
       "      <td>4</td>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          barcode  seurat_clusters  V1  target\n",
       "0  AAACATACAACCAC                3   1       0\n",
       "1  AAACATTGAGCTAC                4   1       1\n",
       "2  AAACATTGATCAGC                1   2       2\n",
       "3  AAACCGTGCTTCCG                5   3       3\n",
       "4  AAACCGTGTATGCG                6   4       4\n",
       "5  AAACGCACTGGTAC                1   5       2\n",
       "6  AAACGCTGACCAGT                3   6       0\n",
       "7  AAACGCTGGTTCTT                3   2       0\n",
       "8  AAACGCTGTAGCCA                3   5       0\n",
       "9  AAACTTGAAAAACG                4   5       1"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "race_data.head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Cluster Count"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "def cluster_counts(df, cluster_name, name):\n",
    "    print(\"Cluster counts for {}\".format(name))\n",
    "    groups = df.groupby(by = cluster_name)\n",
    "    return groups[\"target\"].describe()[['count']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cluster counts for 3. sc3 10x\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>count</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>sc3_10_clusters</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>282.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>233.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>27.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>12.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>249.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>360.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>682.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>256.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>171.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>10</td>\n",
       "      <td>111.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 count\n",
       "sc3_10_clusters       \n",
       "1                282.0\n",
       "2                233.0\n",
       "3                 27.0\n",
       "4                 12.0\n",
       "5                249.0\n",
       "6                360.0\n",
       "7                682.0\n",
       "8                256.0\n",
       "9                171.0\n",
       "10               111.0"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cluster_counts(sc3_r, 'sc3_10_clusters', '3. sc3 10x')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cluster counts for 4. cooc 10x\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>count</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>seurat_clusters</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>589.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>508.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>463.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>348.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>344.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>168.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>144.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>36.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>14.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 count\n",
       "seurat_clusters       \n",
       "0                589.0\n",
       "1                508.0\n",
       "2                463.0\n",
       "3                348.0\n",
       "4                344.0\n",
       "5                168.0\n",
       "6                144.0\n",
       "7                 36.0\n",
       "8                 14.0"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cluster_counts(cooc_data, 'seurat_clusters', '4. cooc 10x')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cluster counts for 4. cooc 10x\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>count</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cooc_clusters</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>576.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>2038.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                count\n",
       "cooc_clusters        \n",
       "1               576.0\n",
       "2              2038.0"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cluster_counts(cooc_data, 'cooc_clusters', '4. cooc 10x')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Compare with Seurat"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def reorder(df):\n",
    "    \n",
    "    dic = {i: np.argmax(df.loc[:,i].values) for i in range(1, len(df.columns)+1)}\n",
    "    dic = sorted(dic.items(), key = lambda x: x[1])\n",
    "    df_new = pd.concat([df.loc[:,col] for col, target in dic],axis=1)\n",
    "    \n",
    "    # Normalize\n",
    "    #df_new = df_new.apply(lambda x: x/x.sum(), axis=0)\n",
    "    \n",
    "    return df_new\n",
    "\n",
    "def reorder_seurat(df):\n",
    "    \n",
    "    dic = {i: np.argmax(df.loc[:,i].values) for i in range(1, len(df.columns))}\n",
    "    dic = sorted(dic.items(), key = lambda x: x[1])\n",
    "    df_new = pd.concat([df.loc[:,col] for col, target in dic],axis=1)\n",
    "    \n",
    "    # Normalize\n",
    "    #df_new = df_new.apply(lambda x: x/x.sum(), axis=0)\n",
    "    \n",
    "    return df_new"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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>seurat_clusters_y</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>sum</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>target</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>21</td>\n",
       "      <td>195</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>73</td>\n",
       "      <td>0</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>308</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>298</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>307</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>250</td>\n",
       "      <td>38</td>\n",
       "      <td>98</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>116</td>\n",
       "      <td>0</td>\n",
       "      <td>507</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>40</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "      <td>109</td>\n",
       "      <td>154</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>134</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>141</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>207</td>\n",
       "      <td>300</td>\n",
       "      <td>27</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>542</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>258</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>108</td>\n",
       "      <td>0</td>\n",
       "      <td>19</td>\n",
       "      <td>386</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>32</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>35</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "seurat_clusters_y    0    1    2    3    4    5    6    7    8  sum\n",
       "target                                                             \n",
       "0                    9   21  195    0    0   73    0   10    0  308\n",
       "1                    0    1    0    0  298    1    1    6    0  307\n",
       "2                  250   38   98    0    0    4    1  116    0  507\n",
       "3                    0    0    0   40    0    0    5    0  109  154\n",
       "4                    0    0    5    0    0  134    0    2    0  141\n",
       "5                  207  300   27    0    0    0    0    8    0  542\n",
       "6                    0    0    0  258    0    1  108    0   19  386\n",
       "7                    0    0    0    2    0    0   32    1    0   35\n",
       "8                    0    0    0    1    1    1    0    0    0    3"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_r = pd.crosstab(newdata['target'], newdata['seurat_clusters_y'])\n",
    "df_r_s = df_r.copy()\n",
    "df_r_s[\"sum\"] = df_r.sum(axis=1)\n",
    "df_r_s"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0xb6f16c8>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_r = reorder_seurat(df_r)\n",
    "sns.heatmap(df_r, cmap=\"YlGnBu\", annot=True, fmt=\"d\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "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>sc3_10_clusters</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>sum</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>target</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>158</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>10</td>\n",
       "      <td>14</td>\n",
       "      <td>49</td>\n",
       "      <td>74</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>308</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>223</td>\n",
       "      <td>17</td>\n",
       "      <td>0</td>\n",
       "      <td>19</td>\n",
       "      <td>1</td>\n",
       "      <td>44</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>307</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>83</td>\n",
       "      <td>153</td>\n",
       "      <td>267</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>507</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>8</td>\n",
       "      <td>10</td>\n",
       "      <td>130</td>\n",
       "      <td>2</td>\n",
       "      <td>154</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>122</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>15</td>\n",
       "      <td>141</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>131</td>\n",
       "      <td>152</td>\n",
       "      <td>258</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>542</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>27</td>\n",
       "      <td>223</td>\n",
       "      <td>41</td>\n",
       "      <td>89</td>\n",
       "      <td>386</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>18</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>35</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "sc3_10_clusters    1    2   3   4    5    6    7    8    9  10  sum\n",
       "target                                                             \n",
       "0                158    0   0  10   14   49   74    0    0   3  308\n",
       "1                  0  223  17   0   19    1   44    3    0   0  307\n",
       "2                  0    3   1   0   83  153  267    0    0   0  507\n",
       "3                  1    0   1   0    0    2    8   10  130   2  154\n",
       "4                122    0   0   2    0    2    0    0    0  15  141\n",
       "5                  0    0   0   0  131  152  258    1    0   0  542\n",
       "6                  1    2   0   0    2    1   27  223   41  89  386\n",
       "7                  0    5   8   0    0    0    2   18    0   2   35\n",
       "8                  0    0   0   0    0    0    2    1    0   0    3"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_sc3 = pd.crosstab(sc3_r['target'], sc3_r['sc3_10_clusters'])\n",
    "df_sc3_s = df_sc3.copy()\n",
    "df_sc3_s[\"sum\"] = df_sc3_s.sum(axis=1)\n",
    "df_sc3_s"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0xbd4c908>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_sc3 = reorder(df_sc3)\n",
    "sns.heatmap(df_sc3, cmap=\"YlGnBu\", annot=True, fmt=\"d\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "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>cooc_clusters</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>sum</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>target</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>347</td>\n",
       "      <td>348</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>8</td>\n",
       "      <td>336</td>\n",
       "      <td>344</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>504</td>\n",
       "      <td>508</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>158</td>\n",
       "      <td>10</td>\n",
       "      <td>168</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>144</td>\n",
       "      <td>144</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "      <td>589</td>\n",
       "      <td>589</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>371</td>\n",
       "      <td>92</td>\n",
       "      <td>463</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>1</td>\n",
       "      <td>13</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>33</td>\n",
       "      <td>3</td>\n",
       "      <td>36</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "cooc_clusters    1    2  sum\n",
       "target                      \n",
       "0                1  347  348\n",
       "1                8  336  344\n",
       "2                4  504  508\n",
       "3              158   10  168\n",
       "4                0  144  144\n",
       "5                0  589  589\n",
       "6              371   92  463\n",
       "7                1   13   14\n",
       "8               33    3   36"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_cooc = pd.crosstab(cooc_data['target'], cooc_data[cooc_data['cooc_clusters']<20]['cooc_clusters'])\n",
    "df_cooc_s = df_cooc.copy()\n",
    "df_cooc_s[\"sum\"] = df_cooc.sum(axis=1)\n",
    "df_cooc_s"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0xc4b4188>"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_cooc = reorder(df_cooc)\n",
    "sns.heatmap(df_cooc, cmap=\"YlGnBu\", annot=True, fmt=\"d\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "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>V1</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>11</th>\n",
       "      <th>12</th>\n",
       "      <th>sum</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>target</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>24</td>\n",
       "      <td>35</td>\n",
       "      <td>8</td>\n",
       "      <td>4</td>\n",
       "      <td>62</td>\n",
       "      <td>46</td>\n",
       "      <td>18</td>\n",
       "      <td>94</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>299</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>19</td>\n",
       "      <td>26</td>\n",
       "      <td>10</td>\n",
       "      <td>0</td>\n",
       "      <td>88</td>\n",
       "      <td>63</td>\n",
       "      <td>46</td>\n",
       "      <td>0</td>\n",
       "      <td>37</td>\n",
       "      <td>11</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>49</td>\n",
       "      <td>46</td>\n",
       "      <td>9</td>\n",
       "      <td>0</td>\n",
       "      <td>280</td>\n",
       "      <td>81</td>\n",
       "      <td>22</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>15</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>503</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "      <td>88</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>9</td>\n",
       "      <td>36</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>150</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>88</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>47</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>141</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>76</td>\n",
       "      <td>46</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>305</td>\n",
       "      <td>70</td>\n",
       "      <td>22</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>540</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>17</td>\n",
       "      <td>3</td>\n",
       "      <td>68</td>\n",
       "      <td>0</td>\n",
       "      <td>49</td>\n",
       "      <td>47</td>\n",
       "      <td>189</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>377</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>10</td>\n",
       "      <td>7</td>\n",
       "      <td>16</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>35</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "V1       1   2   3   4    5   6    7   8   9  10  11  12  sum\n",
       "target                                                       \n",
       "0       24  35   8   4   62  46   18  94   0   7   1   0  299\n",
       "1       19  26  10   0   88  63   46   0  37  11   0   0  300\n",
       "2       49  46   9   0  280  81   22   0   0  15   1   0  503\n",
       "3        5   0  88   0    9   9   36   0   0   2   1   0  150\n",
       "4        1   0   0  88    1   2    1  47   0   0   0   1  141\n",
       "5       76  46   1   0  305  70   22   0   0  20   0   0  540\n",
       "6       17   3  68   0   49  47  189   0   0   0   2   2  377\n",
       "7        1   0   1   0   10   7   16   0   0   0   0   0   35\n",
       "8        0   0   0   0    1   1    0   0   0   0   0   0    2"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_race = pd.crosstab(race_data['target'], race_data[race_data['V1']<13]['V1'])\n",
    "df_race_s = df_race.copy()\n",
    "df_race_s[\"sum\"] = df_race.sum(axis=1)\n",
    "df_race_s"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0xc5809c8>"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_race = reorder(df_race)\n",
    "df_race = df_race.drop(columns = [9,10])\n",
    "sns.heatmap(df_race, cmap=\"YlGnBu\", annot=True, fmt=\"d\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Calculating the metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "def calc_h_accuracy(df):\n",
    "    df_norm = df.apply(lambda x: x/x.sum(), axis=0)\n",
    "    df_entropy = df_norm.apply(lambda x: x*np.log2(x+0.0001))\n",
    "    return -np.sum(df_entropy.sum(axis=0)) / len(df.columns)\n",
    "\n",
    "def calc_h_purity(df):\n",
    "    df_norm = df.apply(lambda x: x/x.sum(), axis=1)\n",
    "    df_entropy = df_norm.apply(lambda x: x*np.log2(x+0.0001))\n",
    "    return -np.sum(df_entropy.sum(axis=0)) / len(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "======pbmc3k results======\n",
      "The entropy of cluster accuracy for seurat R: 0.84713\n",
      "The entropy of cluster accuracy for sc3: 1.09894\n",
      "The entropy of cluster accuracy for cooc: 1.88581\n",
      "The entropy of cluster accuracy for RaceID3: 1.69279\n"
     ]
    }
   ],
   "source": [
    "print(\"======pbmc3k results======\")\n",
    "print(\"The entropy of cluster accuracy for seurat R: {:.5f}\".format(calc_h_accuracy(df_r)))\n",
    "print(\"The entropy of cluster accuracy for sc3: {:.5f}\".format(calc_h_accuracy(df_sc3)))\n",
    "print(\"The entropy of cluster accuracy for cooc: {:.5f}\".format(calc_h_accuracy(df_cooc)))\n",
    "print(\"The entropy of cluster accuracy for RaceID3: {:.5f}\".format(calc_h_accuracy(df_race)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "======pbmc3k results======\n",
      "The entropy of cluster purity for seurat R: 0.91956\n",
      "The entropy of cluster purity for sc3: 1.36289\n",
      "The entropy of cluster purity for cooc: 0.23130\n",
      "The entropy of cluster purity for RaceID3: 1.80185\n"
     ]
    }
   ],
   "source": [
    "print(\"======pbmc3k results======\")\n",
    "print(\"The entropy of cluster purity for seurat R: {:.5f}\".format(calc_h_purity(df_r)))\n",
    "print(\"The entropy of cluster purity for sc3: {:.5f}\".format(calc_h_purity(df_sc3)))\n",
    "print(\"The entropy of cluster purity for cooc: {:.5f}\".format(calc_h_purity(df_cooc)))\n",
    "print(\"The entropy of cluster purity for RaceID3: {:.5f}\".format(calc_h_purity(df_race)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "======pbmc3k cell lines results======\n",
      "The ARI for seurat R (resolution=1): 0.47975\n",
      "The ARI for sc3: 0.30933\n",
      "The ARI for cooc: 0.14438\n",
      "The ARI for RaceID3: 0.16011\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics.cluster import adjusted_rand_score\n",
    "print(\"======pbmc3k cell lines results======\")\n",
    "print(\"The ARI for seurat R (resolution=1): {:.5f}\".format(adjusted_rand_score(newdata['target'], newdata['seurat_clusters_y'])))\n",
    "print(\"The ARI for sc3: {:.5f}\".format(adjusted_rand_score(sc3_r['target'], sc3_r['sc3_10_clusters'])))\n",
    "print(\"The ARI for cooc: {:.5f}\".format(adjusted_rand_score(cooc_data['target'], cooc_data['cooc_clusters'])))\n",
    "print(\"The ARI for RaceID3: {:.5f}\".format(adjusted_rand_score(race_data['target'], race_data['V1'])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}