From_cell_files.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-08-02T13:25:16.958768Z",
     "start_time": "2018-08-02T13:25:14.882714Z"
    }
   },
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%matplotlib inline\n",
    "%config IPCompleter.use_jedi = False\n",
    "\n",
    "import os\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "import pathlib\n",
    "\n",
    "from tqdm import tqdm_notebook\n",
    "import matplotlib.pyplot as plt\n",
    "%config InlineBackend.figure_format = 'png'\n",
    "plt.rcParams['pdf.fonttype'] = 'truetype'\n",
    "plt.rcParams['svg.fonttype'] = 'none'\n",
    "plt.rcParams['figure.dpi'] = 120"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from portraits.plotting import pca_plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext rpy2.ipython"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Example on Ulloa-Montoya GSE35640, GPL570"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Get the data from GEO"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "GSE = 'GSE35640'\n",
    "PLATFORM = 'GPL570'\n",
    "\n",
    "current_dir = pathlib.Path().parent.absolute()\n",
    "dir_to_process = str(current_dir / 'TMPDIR')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "if not os.path.exists(dir_to_process):\n",
    "    os.mkdir(dir_to_process)\n",
    "os.chdir(dir_to_process)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%bash\n",
    "wget ftp://ftp.ncbi.nlm.nih.gov/geo/series/GSE35nnn/GSE35640/suppl/GSE35640_RAW.tar "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%bash\n",
    "tar -xf GSE35640_RAW.tar \n",
    "rm GSE35640_RAW.tar "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "GSM872328_MAGE008_sample_1.CEL.gz   GSM872361_MAGE008_sample_34.CEL.gz\r\n",
      "GSM872329_MAGE008_sample_2.CEL.gz   GSM872362_MAGE008_sample_35.CEL.gz\r\n",
      "GSM872330_MAGE008_sample_3.CEL.gz   GSM872363_MAGE008_sample_36.CEL.gz\r\n",
      "GSM872331_MAGE008_sample_4.CEL.gz   GSM872364_MAGE008_sample_37.CEL.gz\r\n",
      "GSM872332_MAGE008_sample_5.CEL.gz   GSM872365_MAGE008_sample_38.CEL.gz\r\n",
      "GSM872333_MAGE008_sample_6.CEL.gz   GSM872366_MAGE008_sample_39.CEL.gz\r\n",
      "GSM872334_MAGE008_sample_7.CEL.gz   GSM872367_MAGE008_sample_40.CEL.gz\r\n",
      "GSM872335_MAGE008_sample_8.CEL.gz   GSM872368_MAGE008_sample_41.CEL.gz\r\n",
      "GSM872336_MAGE008_sample_9.CEL.gz   GSM872369_MAGE008_sample_42.CEL.gz\r\n",
      "GSM872337_MAGE008_sample_10.CEL.gz  GSM872370_MAGE008_sample_43.CEL.gz\r\n",
      "GSM872338_MAGE008_sample_11.CEL.gz  GSM872371_MAGE008_sample_44.CEL.gz\r\n",
      "GSM872339_MAGE008_sample_12.CEL.gz  GSM872372_MAGE008_sample_45.CEL.gz\r\n",
      "GSM872340_MAGE008_sample_13.CEL.gz  GSM872373_MAGE008_sample_46.CEL.gz\r\n",
      "GSM872341_MAGE008_sample_14.CEL.gz  GSM872374_MAGE008_sample_47.CEL.gz\r\n",
      "GSM872342_MAGE008_sample_15.CEL.gz  GSM872375_MAGE008_sample_48.CEL.gz\r\n",
      "GSM872343_MAGE008_sample_16.CEL.gz  GSM872376_MAGE008_sample_49.CEL.gz\r\n",
      "GSM872344_MAGE008_sample_17.CEL.gz  GSM872377_MAGE008_sample_50.CEL.gz\r\n",
      "GSM872345_MAGE008_sample_18.CEL.gz  GSM872378_MAGE008_sample_51.CEL.gz\r\n",
      "GSM872346_MAGE008_sample_19.CEL.gz  GSM872379_MAGE008_sample_52.CEL.gz\r\n",
      "GSM872347_MAGE008_sample_20.CEL.gz  GSM872380_MAGE008_sample_53.CEL.gz\r\n",
      "GSM872348_MAGE008_sample_21.CEL.gz  GSM872381_MAGE008_sample_54.CEL.gz\r\n",
      "GSM872349_MAGE008_sample_22.CEL.gz  GSM872382_MAGE008_sample_55.CEL.gz\r\n",
      "GSM872350_MAGE008_sample_23.CEL.gz  GSM872383_MAGE008_sample_56.CEL.gz\r\n",
      "GSM872351_MAGE008_sample_24.CEL.gz  GSM872384_MAGE008_sample_57.CEL.gz\r\n",
      "GSM872352_MAGE008_sample_25.CEL.gz  GSM872385_MAGE008_sample_58.CEL.gz\r\n",
      "GSM872353_MAGE008_sample_26.CEL.gz  GSM872386_MAGE008_sample_59.CEL.gz\r\n",
      "GSM872354_MAGE008_sample_27.CEL.gz  GSM872387_MAGE008_sample_60.CEL.gz\r\n",
      "GSM872355_MAGE008_sample_28.CEL.gz  GSM872388_MAGE008_sample_61.CEL.gz\r\n",
      "GSM872356_MAGE008_sample_29.CEL.gz  GSM872389_MAGE008_sample_62.CEL.gz\r\n",
      "GSM872357_MAGE008_sample_30.CEL.gz  GSM872390_MAGE008_sample_63.CEL.gz\r\n",
      "GSM872358_MAGE008_sample_31.CEL.gz  GSM872391_MAGE008_sample_64.CEL.gz\r\n",
      "GSM872359_MAGE008_sample_32.CEL.gz  GSM872392_MAGE008_sample_65.CEL.gz\r\n",
      "GSM872360_MAGE008_sample_33.CEL.gz\r\n"
     ]
    }
   ],
   "source": [
    "ls"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Extracting expression from CEL file"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In case of affy arrays without special probes we have to use justRMA<br>\n",
    "For GPL570/GPL96 we can use gcrma"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-08-02T13:25:21.735530Z",
     "start_time": "2018-08-02T13:25:16.965609Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "R[write to console]: Loading required package: BiocGenerics\n",
      "\n",
      "R[write to console]: Loading required package: parallel\n",
      "\n",
      "R[write to console]: \n",
      "Attaching package: ‘BiocGenerics’\n",
      "\n",
      "\n",
      "R[write to console]: The following objects are masked from ‘package:parallel’:\n",
      "\n",
      "    clusterApply, clusterApplyLB, clusterCall, clusterEvalQ,\n",
      "    clusterExport, clusterMap, parApply, parCapply, parLapply,\n",
      "    parLapplyLB, parRapply, parSapply, parSapplyLB\n",
      "\n",
      "\n",
      "R[write to console]: The following objects are masked from ‘package:stats’:\n",
      "\n",
      "    IQR, mad, sd, var, xtabs\n",
      "\n",
      "\n",
      "R[write to console]: The following objects are masked from ‘package:base’:\n",
      "\n",
      "    anyDuplicated, append, as.data.frame, basename, cbind, colnames,\n",
      "    dirname, do.call, duplicated, eval, evalq, Filter, Find, get, grep,\n",
      "    grepl, intersect, is.unsorted, lapply, Map, mapply, match, mget,\n",
      "    order, paste, pmax, pmax.int, pmin, pmin.int, Position, rank,\n",
      "    rbind, Reduce, rownames, sapply, setdiff, sort, table, tapply,\n",
      "    union, unique, unsplit, which, which.max, which.min\n",
      "\n",
      "\n",
      "R[write to console]: Loading required package: Biobase\n",
      "\n",
      "R[write to console]: Welcome to Bioconductor\n",
      "\n",
      "    Vignettes contain introductory material; view with\n",
      "    'browseVignettes()'. To cite Bioconductor, see\n",
      "    'citation(\"Biobase\")', and for packages 'citation(\"pkgname\")'.\n",
      "\n",
      "\n",
      "R[write to console]: Loading required package: AnnotationDbi\n",
      "\n",
      "R[write to console]: Loading required package: stats4\n",
      "\n",
      "R[write to console]: Loading required package: IRanges\n",
      "\n",
      "R[write to console]: Loading required package: S4Vectors\n",
      "\n",
      "R[write to console]: \n",
      "Attaching package: ‘S4Vectors’\n",
      "\n",
      "\n",
      "R[write to console]: The following object is masked from ‘package:base’:\n",
      "\n",
      "    expand.grid\n",
      "\n",
      "\n",
      "R[write to console]: Loading required package: XML\n",
      "\n",
      "R[write to console]: \n",
      "Attaching package: ‘XML’\n",
      "\n",
      "\n",
      "R[write to console]: The following object is masked from ‘package:tools’:\n",
      "\n",
      "    toHTML\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "%%R\n",
    "\n",
    "# Loading required R packages\n",
    "library(affy)\n",
    "library(annotate)\n",
    "library(gcrma)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-08-02T14:25:38.950721Z",
     "start_time": "2018-08-02T14:19:31.888815Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "R[write to console]: \n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Adjusting for optical effect.................................................................Done.\n",
      "Computing affinities.Done.\n",
      "Adjusting for non-specific binding.................................................................Done.\n",
      "Normalizing\n",
      "Calculating Expression\n"
     ]
    }
   ],
   "source": [
    "%%R -i dir_to_process -o normalized_expression\n",
    "\n",
    "# Bulk read cell files\n",
    "raw_expression <- ReadAffy(celfile.path = dir_to_process)\n",
    "\n",
    "# Normalize expression using RMA\n",
    "rma_normalized <- gcrma(raw_expression)\n",
    "\n",
    "# Retrieving expressions from dataset\n",
    "normalized_expression <- as.data.frame(exprs(rma_normalized))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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>GSM872328_MAGE008_sample_1.CEL.gz</th>\n",
       "      <th>GSM872329_MAGE008_sample_2.CEL.gz</th>\n",
       "      <th>GSM872330_MAGE008_sample_3.CEL.gz</th>\n",
       "      <th>GSM872331_MAGE008_sample_4.CEL.gz</th>\n",
       "      <th>GSM872332_MAGE008_sample_5.CEL.gz</th>\n",
       "      <th>GSM872333_MAGE008_sample_6.CEL.gz</th>\n",
       "      <th>GSM872334_MAGE008_sample_7.CEL.gz</th>\n",
       "      <th>GSM872335_MAGE008_sample_8.CEL.gz</th>\n",
       "      <th>GSM872336_MAGE008_sample_9.CEL.gz</th>\n",
       "      <th>GSM872337_MAGE008_sample_10.CEL.gz</th>\n",
       "      <th>...</th>\n",
       "      <th>GSM872383_MAGE008_sample_56.CEL.gz</th>\n",
       "      <th>GSM872384_MAGE008_sample_57.CEL.gz</th>\n",
       "      <th>GSM872385_MAGE008_sample_58.CEL.gz</th>\n",
       "      <th>GSM872386_MAGE008_sample_59.CEL.gz</th>\n",
       "      <th>GSM872387_MAGE008_sample_60.CEL.gz</th>\n",
       "      <th>GSM872388_MAGE008_sample_61.CEL.gz</th>\n",
       "      <th>GSM872389_MAGE008_sample_62.CEL.gz</th>\n",
       "      <th>GSM872390_MAGE008_sample_63.CEL.gz</th>\n",
       "      <th>GSM872391_MAGE008_sample_64.CEL.gz</th>\n",
       "      <th>GSM872392_MAGE008_sample_65.CEL.gz</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1007_s_at</th>\n",
       "      <td>9.811332</td>\n",
       "      <td>6.470581</td>\n",
       "      <td>9.109760</td>\n",
       "      <td>7.040136</td>\n",
       "      <td>8.716385</td>\n",
       "      <td>4.848676</td>\n",
       "      <td>7.186724</td>\n",
       "      <td>8.631181</td>\n",
       "      <td>5.618537</td>\n",
       "      <td>7.429797</td>\n",
       "      <td>...</td>\n",
       "      <td>6.282597</td>\n",
       "      <td>7.012282</td>\n",
       "      <td>7.528732</td>\n",
       "      <td>7.037582</td>\n",
       "      <td>8.804607</td>\n",
       "      <td>8.193118</td>\n",
       "      <td>8.452822</td>\n",
       "      <td>7.525328</td>\n",
       "      <td>4.722824</td>\n",
       "      <td>9.217646</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1053_at</th>\n",
       "      <td>5.663456</td>\n",
       "      <td>6.539775</td>\n",
       "      <td>5.568130</td>\n",
       "      <td>5.744605</td>\n",
       "      <td>4.690514</td>\n",
       "      <td>5.922860</td>\n",
       "      <td>5.699696</td>\n",
       "      <td>5.120274</td>\n",
       "      <td>5.393646</td>\n",
       "      <td>6.704532</td>\n",
       "      <td>...</td>\n",
       "      <td>5.991613</td>\n",
       "      <td>6.753494</td>\n",
       "      <td>5.942335</td>\n",
       "      <td>6.333092</td>\n",
       "      <td>6.062780</td>\n",
       "      <td>6.359215</td>\n",
       "      <td>6.071876</td>\n",
       "      <td>7.140830</td>\n",
       "      <td>7.190948</td>\n",
       "      <td>5.611007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>117_at</th>\n",
       "      <td>8.392470</td>\n",
       "      <td>6.080836</td>\n",
       "      <td>6.585071</td>\n",
       "      <td>6.091045</td>\n",
       "      <td>3.409561</td>\n",
       "      <td>3.789047</td>\n",
       "      <td>5.666034</td>\n",
       "      <td>4.964667</td>\n",
       "      <td>4.473665</td>\n",
       "      <td>4.441460</td>\n",
       "      <td>...</td>\n",
       "      <td>3.295174</td>\n",
       "      <td>5.637473</td>\n",
       "      <td>3.203354</td>\n",
       "      <td>3.469941</td>\n",
       "      <td>3.679389</td>\n",
       "      <td>8.969278</td>\n",
       "      <td>4.341028</td>\n",
       "      <td>5.736370</td>\n",
       "      <td>4.805148</td>\n",
       "      <td>10.329342</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121_at</th>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.344876</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.387679</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.454047</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>3.128001</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>...</td>\n",
       "      <td>2.320199</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.411106</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.380717</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.882447</td>\n",
       "      <td>2.433872</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1255_g_at</th>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>...</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.349209</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 65 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           GSM872328_MAGE008_sample_1.CEL.gz  \\\n",
       "1007_s_at                           9.811332   \n",
       "1053_at                             5.663456   \n",
       "117_at                              8.392470   \n",
       "121_at                              2.390067   \n",
       "1255_g_at                           2.223265   \n",
       "\n",
       "           GSM872329_MAGE008_sample_2.CEL.gz  \\\n",
       "1007_s_at                           6.470581   \n",
       "1053_at                             6.539775   \n",
       "117_at                              6.080836   \n",
       "121_at                              2.344876   \n",
       "1255_g_at                           2.223265   \n",
       "\n",
       "           GSM872330_MAGE008_sample_3.CEL.gz  \\\n",
       "1007_s_at                           9.109760   \n",
       "1053_at                             5.568130   \n",
       "117_at                              6.585071   \n",
       "121_at                              2.390067   \n",
       "1255_g_at                           2.223265   \n",
       "\n",
       "           GSM872331_MAGE008_sample_4.CEL.gz  \\\n",
       "1007_s_at                           7.040136   \n",
       "1053_at                             5.744605   \n",
       "117_at                              6.091045   \n",
       "121_at                              2.387679   \n",
       "1255_g_at                           2.223265   \n",
       "\n",
       "           GSM872332_MAGE008_sample_5.CEL.gz  \\\n",
       "1007_s_at                           8.716385   \n",
       "1053_at                             4.690514   \n",
       "117_at                              3.409561   \n",
       "121_at                              2.390067   \n",
       "1255_g_at                           2.223265   \n",
       "\n",
       "           GSM872333_MAGE008_sample_6.CEL.gz  \\\n",
       "1007_s_at                           4.848676   \n",
       "1053_at                             5.922860   \n",
       "117_at                              3.789047   \n",
       "121_at                              2.390067   \n",
       "1255_g_at                           2.223265   \n",
       "\n",
       "           GSM872334_MAGE008_sample_7.CEL.gz  \\\n",
       "1007_s_at                           7.186724   \n",
       "1053_at                             5.699696   \n",
       "117_at                              5.666034   \n",
       "121_at                              2.454047   \n",
       "1255_g_at                           2.223265   \n",
       "\n",
       "           GSM872335_MAGE008_sample_8.CEL.gz  \\\n",
       "1007_s_at                           8.631181   \n",
       "1053_at                             5.120274   \n",
       "117_at                              4.964667   \n",
       "121_at                              2.390067   \n",
       "1255_g_at                           2.223265   \n",
       "\n",
       "           GSM872336_MAGE008_sample_9.CEL.gz  \\\n",
       "1007_s_at                           5.618537   \n",
       "1053_at                             5.393646   \n",
       "117_at                              4.473665   \n",
       "121_at                              3.128001   \n",
       "1255_g_at                           2.223265   \n",
       "\n",
       "           GSM872337_MAGE008_sample_10.CEL.gz  ...  \\\n",
       "1007_s_at                            7.429797  ...   \n",
       "1053_at                              6.704532  ...   \n",
       "117_at                               4.441460  ...   \n",
       "121_at                               2.390067  ...   \n",
       "1255_g_at                            2.223265  ...   \n",
       "\n",
       "           GSM872383_MAGE008_sample_56.CEL.gz  \\\n",
       "1007_s_at                            6.282597   \n",
       "1053_at                              5.991613   \n",
       "117_at                               3.295174   \n",
       "121_at                               2.320199   \n",
       "1255_g_at                            2.223265   \n",
       "\n",
       "           GSM872384_MAGE008_sample_57.CEL.gz  \\\n",
       "1007_s_at                            7.012282   \n",
       "1053_at                              6.753494   \n",
       "117_at                               5.637473   \n",
       "121_at                               2.390067   \n",
       "1255_g_at                            2.223265   \n",
       "\n",
       "           GSM872385_MAGE008_sample_58.CEL.gz  \\\n",
       "1007_s_at                            7.528732   \n",
       "1053_at                              5.942335   \n",
       "117_at                               3.203354   \n",
       "121_at                               2.411106   \n",
       "1255_g_at                            2.223265   \n",
       "\n",
       "           GSM872386_MAGE008_sample_59.CEL.gz  \\\n",
       "1007_s_at                            7.037582   \n",
       "1053_at                              6.333092   \n",
       "117_at                               3.469941   \n",
       "121_at                               2.390067   \n",
       "1255_g_at                            2.223265   \n",
       "\n",
       "           GSM872387_MAGE008_sample_60.CEL.gz  \\\n",
       "1007_s_at                            8.804607   \n",
       "1053_at                              6.062780   \n",
       "117_at                               3.679389   \n",
       "121_at                               2.390067   \n",
       "1255_g_at                            2.223265   \n",
       "\n",
       "           GSM872388_MAGE008_sample_61.CEL.gz  \\\n",
       "1007_s_at                            8.193118   \n",
       "1053_at                              6.359215   \n",
       "117_at                               8.969278   \n",
       "121_at                               2.390067   \n",
       "1255_g_at                            2.223265   \n",
       "\n",
       "           GSM872389_MAGE008_sample_62.CEL.gz  \\\n",
       "1007_s_at                            8.452822   \n",
       "1053_at                              6.071876   \n",
       "117_at                               4.341028   \n",
       "121_at                               2.380717   \n",
       "1255_g_at                            2.349209   \n",
       "\n",
       "           GSM872390_MAGE008_sample_63.CEL.gz  \\\n",
       "1007_s_at                            7.525328   \n",
       "1053_at                              7.140830   \n",
       "117_at                               5.736370   \n",
       "121_at                               2.390067   \n",
       "1255_g_at                            2.223265   \n",
       "\n",
       "           GSM872391_MAGE008_sample_64.CEL.gz  \\\n",
       "1007_s_at                            4.722824   \n",
       "1053_at                              7.190948   \n",
       "117_at                               4.805148   \n",
       "121_at                               2.882447   \n",
       "1255_g_at                            2.223265   \n",
       "\n",
       "           GSM872392_MAGE008_sample_65.CEL.gz  \n",
       "1007_s_at                            9.217646  \n",
       "1053_at                              5.611007  \n",
       "117_at                              10.329342  \n",
       "121_at                               2.433872  \n",
       "1255_g_at                            2.223265  \n",
       "\n",
       "[5 rows x 65 columns]"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "normalized_expression.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Trim names "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-08-02T14:25:38.956494Z",
     "start_time": "2018-08-02T14:25:38.952824Z"
    }
   },
   "outputs": [],
   "source": [
    "normalized_expression.columns = normalized_expression.columns.to_series().apply(lambda x: x.split('_')[0]).values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='PCA 1 component 11% variance explained', ylabel='PCA 2 component 9% variance explained'>"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 600x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pca_plot(normalized_expression.T, legend='out')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-08-02T14:25:38.970277Z",
     "start_time": "2018-08-02T14:25:38.958537Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "normalized_expression.columns.value_counts().max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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>GSM872328</th>\n",
       "      <th>GSM872329</th>\n",
       "      <th>GSM872330</th>\n",
       "      <th>GSM872331</th>\n",
       "      <th>GSM872332</th>\n",
       "      <th>GSM872333</th>\n",
       "      <th>GSM872334</th>\n",
       "      <th>GSM872335</th>\n",
       "      <th>GSM872336</th>\n",
       "      <th>GSM872337</th>\n",
       "      <th>...</th>\n",
       "      <th>GSM872383</th>\n",
       "      <th>GSM872384</th>\n",
       "      <th>GSM872385</th>\n",
       "      <th>GSM872386</th>\n",
       "      <th>GSM872387</th>\n",
       "      <th>GSM872388</th>\n",
       "      <th>GSM872389</th>\n",
       "      <th>GSM872390</th>\n",
       "      <th>GSM872391</th>\n",
       "      <th>GSM872392</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1007_s_at</th>\n",
       "      <td>9.811332</td>\n",
       "      <td>6.470581</td>\n",
       "      <td>9.109760</td>\n",
       "      <td>7.040136</td>\n",
       "      <td>8.716385</td>\n",
       "      <td>4.848676</td>\n",
       "      <td>7.186724</td>\n",
       "      <td>8.631181</td>\n",
       "      <td>5.618537</td>\n",
       "      <td>7.429797</td>\n",
       "      <td>...</td>\n",
       "      <td>6.282597</td>\n",
       "      <td>7.012282</td>\n",
       "      <td>7.528732</td>\n",
       "      <td>7.037582</td>\n",
       "      <td>8.804607</td>\n",
       "      <td>8.193118</td>\n",
       "      <td>8.452822</td>\n",
       "      <td>7.525328</td>\n",
       "      <td>4.722824</td>\n",
       "      <td>9.217646</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1053_at</th>\n",
       "      <td>5.663456</td>\n",
       "      <td>6.539775</td>\n",
       "      <td>5.568130</td>\n",
       "      <td>5.744605</td>\n",
       "      <td>4.690514</td>\n",
       "      <td>5.922860</td>\n",
       "      <td>5.699696</td>\n",
       "      <td>5.120274</td>\n",
       "      <td>5.393646</td>\n",
       "      <td>6.704532</td>\n",
       "      <td>...</td>\n",
       "      <td>5.991613</td>\n",
       "      <td>6.753494</td>\n",
       "      <td>5.942335</td>\n",
       "      <td>6.333092</td>\n",
       "      <td>6.062780</td>\n",
       "      <td>6.359215</td>\n",
       "      <td>6.071876</td>\n",
       "      <td>7.140830</td>\n",
       "      <td>7.190948</td>\n",
       "      <td>5.611007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>117_at</th>\n",
       "      <td>8.392470</td>\n",
       "      <td>6.080836</td>\n",
       "      <td>6.585071</td>\n",
       "      <td>6.091045</td>\n",
       "      <td>3.409561</td>\n",
       "      <td>3.789047</td>\n",
       "      <td>5.666034</td>\n",
       "      <td>4.964667</td>\n",
       "      <td>4.473665</td>\n",
       "      <td>4.441460</td>\n",
       "      <td>...</td>\n",
       "      <td>3.295174</td>\n",
       "      <td>5.637473</td>\n",
       "      <td>3.203354</td>\n",
       "      <td>3.469941</td>\n",
       "      <td>3.679389</td>\n",
       "      <td>8.969278</td>\n",
       "      <td>4.341028</td>\n",
       "      <td>5.736370</td>\n",
       "      <td>4.805148</td>\n",
       "      <td>10.329342</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121_at</th>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.344876</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.387679</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.454047</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>3.128001</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>...</td>\n",
       "      <td>2.320199</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.411106</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.380717</td>\n",
       "      <td>2.390067</td>\n",
       "      <td>2.882447</td>\n",
       "      <td>2.433872</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1255_g_at</th>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>...</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.349209</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 65 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "           GSM872328  GSM872329  GSM872330  GSM872331  GSM872332  GSM872333  \\\n",
       "1007_s_at   9.811332   6.470581   9.109760   7.040136   8.716385   4.848676   \n",
       "1053_at     5.663456   6.539775   5.568130   5.744605   4.690514   5.922860   \n",
       "117_at      8.392470   6.080836   6.585071   6.091045   3.409561   3.789047   \n",
       "121_at      2.390067   2.344876   2.390067   2.387679   2.390067   2.390067   \n",
       "1255_g_at   2.223265   2.223265   2.223265   2.223265   2.223265   2.223265   \n",
       "\n",
       "           GSM872334  GSM872335  GSM872336  GSM872337  ...  GSM872383  \\\n",
       "1007_s_at   7.186724   8.631181   5.618537   7.429797  ...   6.282597   \n",
       "1053_at     5.699696   5.120274   5.393646   6.704532  ...   5.991613   \n",
       "117_at      5.666034   4.964667   4.473665   4.441460  ...   3.295174   \n",
       "121_at      2.454047   2.390067   3.128001   2.390067  ...   2.320199   \n",
       "1255_g_at   2.223265   2.223265   2.223265   2.223265  ...   2.223265   \n",
       "\n",
       "           GSM872384  GSM872385  GSM872386  GSM872387  GSM872388  GSM872389  \\\n",
       "1007_s_at   7.012282   7.528732   7.037582   8.804607   8.193118   8.452822   \n",
       "1053_at     6.753494   5.942335   6.333092   6.062780   6.359215   6.071876   \n",
       "117_at      5.637473   3.203354   3.469941   3.679389   8.969278   4.341028   \n",
       "121_at      2.390067   2.411106   2.390067   2.390067   2.390067   2.380717   \n",
       "1255_g_at   2.223265   2.223265   2.223265   2.223265   2.223265   2.349209   \n",
       "\n",
       "           GSM872390  GSM872391  GSM872392  \n",
       "1007_s_at   7.525328   4.722824   9.217646  \n",
       "1053_at     7.140830   7.190948   5.611007  \n",
       "117_at      5.736370   4.805148  10.329342  \n",
       "121_at      2.390067   2.882447   2.433872  \n",
       "1255_g_at   2.223265   2.223265   2.223265  \n",
       "\n",
       "[5 rows x 65 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "normalized_expression.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%bash \n",
    "rm *.CEL.gz"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Converting probes to HUGO gene symbols"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Get chip SOFT file and make a 3-col file out of that.<br>\n",
    "1st column: probe id<br>\n",
    "2rd column: gene symbol column (as is with '///')<br>\n",
    "3nd column: entrez id (not needed for the study)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1007_s_at</td>\n",
       "      <td>DDR1///MIR4640</td>\n",
       "      <td>780///100616237</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1053_at</td>\n",
       "      <td>RFC2</td>\n",
       "      <td>5982</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>117_at</td>\n",
       "      <td>HSPA6</td>\n",
       "      <td>3310</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>121_at</td>\n",
       "      <td>PAX8</td>\n",
       "      <td>7849</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1255_g_at</td>\n",
       "      <td>GUCA1A</td>\n",
       "      <td>2978</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           0               1                2\n",
       "0  1007_s_at  DDR1///MIR4640  780///100616237\n",
       "1    1053_at            RFC2             5982\n",
       "2     117_at           HSPA6             3310\n",
       "3     121_at            PAX8             7849\n",
       "4  1255_g_at          GUCA1A             2978"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.read_csv('GPL570.3col', index_col=None, sep='\\t', header=None).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-08-02T14:25:49.021634Z",
     "start_time": "2018-08-02T14:25:49.016767Z"
    }
   },
   "outputs": [],
   "source": [
    "# from modules.mapping import get_gs_for_probes_from_3col, get_expressions_for_gs, get_database_dir, get_reverse_dictionary, get_expressions_list, defaultdict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_gs_for_probes_from_3col(platform_file, probe_list):\n",
    "    \"\"\"\n",
    "    Getting probe-gene symbol dictionary\n",
    "\n",
    "    :param platform_name: str, platform name\n",
    "    :param probe_list: list, list with probe names\n",
    "\n",
    "    :return: dict, dictionary with probe-gene symbol key-values\n",
    "    \"\"\"\n",
    "    try:\n",
    "        platform_data = pd.read_csv(platform_file, sep='\\t', header=None, index_col=0, na_values=[\"NONE\"])\n",
    "    except Exception as e:\n",
    "        logging.warning(f\"Failed to read mapping 3col-file: {str(e)}\")\n",
    "        return None\n",
    "\n",
    "    dict_raw_name_id = dict()\n",
    "    not_found_probes_amount = len(set(probe_list).difference(platform_data.index))\n",
    "    if not_found_probes_amount:\n",
    "        logging.warn(f'{not_found_probes_amount} probes not found or format is not correct.')\n",
    "        return dict()\n",
    "\n",
    "    result = platform_data[1].loc[probe_list].dropna().astype(str).apply(\n",
    "        lambda x: x.strip().replace(\" \", \"\").split(\"///\")).to_dict()\n",
    "\n",
    "    return result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-08-02T14:25:50.137216Z",
     "start_time": "2018-08-02T14:25:49.716242Z"
    },
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "probes_gs_dict = get_gs_for_probes_from_3col('GPL570.3col', normalized_expression.index.tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1007_s_at                 [DDR1, MIR4640]\n",
       "1053_at                            [RFC2]\n",
       "117_at                            [HSPA6]\n",
       "121_at                             [PAX8]\n",
       "1255_g_at                        [GUCA1A]\n",
       "1294_at                   [MIR5193, UBA7]\n",
       "1316_at                            [THRA]\n",
       "1320_at                          [PTPN21]\n",
       "1405_i_at                          [CCL5]\n",
       "1431_at                          [CYP2E1]\n",
       "1438_at                           [EPHB3]\n",
       "1487_at                           [ESRRA]\n",
       "1494_f_at                        [CYP2A6]\n",
       "1552256_a_at                     [SCARB1]\n",
       "1552257_a_at                     [TTLL12]\n",
       "1552258_at      [LINC00152, LOC101930489]\n",
       "1552261_at                        [WFDC2]\n",
       "1552263_at                        [MAPK1]\n",
       "1552264_a_at                      [MAPK1]\n",
       "1552266_at                       [ADAM32]\n",
       "dtype: object"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.Series(probes_gs_dict).head(n=20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_expressions_list(probes_list, probes_value_table, method='max'):\n",
    "    \"\"\"\n",
    "    Returns list of expressions for matching gene_symbol\n",
    "\n",
    "    :param probes_list: list, list of probes (for matching gene_symbol)\n",
    "    :param probes_value_table: pd.DataFrame, matching table for probe_id and expression values (for each sample)\n",
    "    :param method: str, getting expressions method (max / med )\n",
    "\n",
    "    :return: list, list of expressions for matching gene_symbol\n",
    "    \"\"\"\n",
    "\n",
    "    def average_expression(expressions_list):\n",
    "        return sum(expressions_list) / len(expressions_list)\n",
    "\n",
    "    probes_avg_expr_dict = {}\n",
    "    # count average for all gsms\n",
    "    for probe in probes_list:\n",
    "        probes_avg_expr_dict[probe] = average_expression(probes_value_table.loc[probe, :])\n",
    "    probe_res = ''\n",
    "    if method == 'max':\n",
    "        # choose probe-id with max average value\n",
    "        probe_res = max(probes_avg_expr_dict, key=probes_avg_expr_dict.get)\n",
    "\n",
    "    elif method == 'med':\n",
    "        # choose probe with median value\n",
    "        # if 2 probes choose probe with max average value\n",
    "\n",
    "        # sort dict by values and return list of keys\n",
    "        sorted_probes_list = sorted(probes_avg_expr_dict, key=probes_avg_expr_dict.get)\n",
    "        if len(probes_list) % 2 == 0:\n",
    "            probe_res = max(sorted_probes_list[len(probes_list) / 2 - 1],\n",
    "                            sorted_probes_list[len(probes_list) / 2],\n",
    "                            key=probes_avg_expr_dict.get)\n",
    "        else:\n",
    "            probe_res = sorted_probes_list[len(probes_list) / 2]\n",
    "    return probes_value_table.loc[probe_res, :]\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_expressions_for_gs(probes_gs_dict, probes_value_table, gs_sel_alg='max'):\n",
    "    \"\"\"\n",
    "    Getting genes/samples expression table\n",
    "\n",
    "    :param probes_gs_dict: dict, dictionary with probe-gene symbol key-values\n",
    "    :param probes_value_table: pd.DataFrame, probes/samples expression transformed dataframe\n",
    "    :param gs_sel_alg: str, getting expressions method (max / med )\n",
    "\n",
    "    :return: gs_expr_table: pd.DataFrame, genes/samples expression table\n",
    "    \"\"\"\n",
    "    import logging\n",
    "    \n",
    "    def get_reverse_dictionary(probes_gs_dict):\n",
    "        from collections import defaultdict\n",
    "        gs_probes_dict = defaultdict(list)\n",
    "        for probe, gs_list in probes_gs_dict.items():\n",
    "            for gs in gs_list:\n",
    "                gs_probes_dict[gs].append(probe)\n",
    "        return gs_probes_dict\n",
    "    \n",
    "    gs_expr_table = pd.DataFrame()\n",
    "\n",
    "    logging.info(\"Making list of probes for each of gene-symbols ...\")\n",
    "    gs_probes_dict = get_reverse_dictionary(probes_gs_dict)\n",
    "\n",
    "    logging.info(\"Making expression list for each of gene-symbols ...\")\n",
    "    for gs, probe_list in gs_probes_dict.items():\n",
    "        # print(probe_list[:30])\n",
    "        if (len(probe_list) == 1):\n",
    "            gs_expr_table[gs] = probes_value_table.loc[probe_list[0], :]\n",
    "        else:\n",
    "            gs_expr_table[gs] = get_expressions_list(\n",
    "                probes_list=probe_list,\n",
    "                probes_value_table=probes_value_table,\n",
    "                method=gs_sel_alg\n",
    "            )\n",
    "\n",
    "    logging.info(f'Final expression samples/gene symbols table shape: {gs_expr_table.shape}')\n",
    "\n",
    "    return gs_expr_table"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-08-02T14:26:46.823094Z",
     "start_time": "2018-08-02T14:25:52.167425Z"
    }
   },
   "outputs": [],
   "source": [
    "annotated_expression = get_expressions_for_gs(pd.Series(probes_gs_dict),\n",
    "                                              normalized_expression, 'max').T.sort_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-08-02T14:26:47.343521Z",
     "start_time": "2018-08-02T14:26:46.953242Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(annotated_expression.T.mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "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>GSM872328</th>\n",
       "      <th>GSM872329</th>\n",
       "      <th>GSM872330</th>\n",
       "      <th>GSM872331</th>\n",
       "      <th>GSM872332</th>\n",
       "      <th>GSM872333</th>\n",
       "      <th>GSM872334</th>\n",
       "      <th>GSM872335</th>\n",
       "      <th>GSM872336</th>\n",
       "      <th>GSM872337</th>\n",
       "      <th>...</th>\n",
       "      <th>GSM872383</th>\n",
       "      <th>GSM872384</th>\n",
       "      <th>GSM872385</th>\n",
       "      <th>GSM872386</th>\n",
       "      <th>GSM872387</th>\n",
       "      <th>GSM872388</th>\n",
       "      <th>GSM872389</th>\n",
       "      <th>GSM872390</th>\n",
       "      <th>GSM872391</th>\n",
       "      <th>GSM872392</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>A1BG</th>\n",
       "      <td>3.076475</td>\n",
       "      <td>2.987292</td>\n",
       "      <td>2.987292</td>\n",
       "      <td>3.165908</td>\n",
       "      <td>2.727649</td>\n",
       "      <td>3.655183</td>\n",
       "      <td>2.987292</td>\n",
       "      <td>2.987292</td>\n",
       "      <td>3.574711</td>\n",
       "      <td>6.003634</td>\n",
       "      <td>...</td>\n",
       "      <td>4.334097</td>\n",
       "      <td>3.014834</td>\n",
       "      <td>2.987292</td>\n",
       "      <td>3.565207</td>\n",
       "      <td>3.566093</td>\n",
       "      <td>2.987292</td>\n",
       "      <td>2.987292</td>\n",
       "      <td>2.898243</td>\n",
       "      <td>3.057402</td>\n",
       "      <td>2.987292</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A1BG-AS1</th>\n",
       "      <td>5.749817</td>\n",
       "      <td>4.904232</td>\n",
       "      <td>4.387495</td>\n",
       "      <td>4.360151</td>\n",
       "      <td>4.243835</td>\n",
       "      <td>4.355736</td>\n",
       "      <td>4.493685</td>\n",
       "      <td>4.477549</td>\n",
       "      <td>4.594941</td>\n",
       "      <td>5.299739</td>\n",
       "      <td>...</td>\n",
       "      <td>4.309564</td>\n",
       "      <td>4.596462</td>\n",
       "      <td>4.635358</td>\n",
       "      <td>4.596462</td>\n",
       "      <td>4.605628</td>\n",
       "      <td>4.596462</td>\n",
       "      <td>4.596462</td>\n",
       "      <td>4.638988</td>\n",
       "      <td>4.412014</td>\n",
       "      <td>4.741860</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A1CF</th>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>...</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "      <td>2.223265</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A2M</th>\n",
       "      <td>11.727785</td>\n",
       "      <td>11.030724</td>\n",
       "      <td>11.897919</td>\n",
       "      <td>10.843243</td>\n",
       "      <td>12.948972</td>\n",
       "      <td>12.603923</td>\n",
       "      <td>13.163574</td>\n",
       "      <td>11.997735</td>\n",
       "      <td>12.926909</td>\n",
       "      <td>12.836881</td>\n",
       "      <td>...</td>\n",
       "      <td>12.034571</td>\n",
       "      <td>12.973060</td>\n",
       "      <td>12.725227</td>\n",
       "      <td>10.293444</td>\n",
       "      <td>13.931666</td>\n",
       "      <td>13.758073</td>\n",
       "      <td>12.477485</td>\n",
       "      <td>10.576160</td>\n",
       "      <td>12.909493</td>\n",
       "      <td>12.791211</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A2M-AS1</th>\n",
       "      <td>6.441273</td>\n",
       "      <td>6.099895</td>\n",
       "      <td>6.603606</td>\n",
       "      <td>5.386214</td>\n",
       "      <td>7.310987</td>\n",
       "      <td>7.769239</td>\n",
       "      <td>7.290848</td>\n",
       "      <td>6.464501</td>\n",
       "      <td>7.775331</td>\n",
       "      <td>7.137957</td>\n",
       "      <td>...</td>\n",
       "      <td>8.235774</td>\n",
       "      <td>7.260537</td>\n",
       "      <td>7.984555</td>\n",
       "      <td>4.548931</td>\n",
       "      <td>8.101128</td>\n",
       "      <td>8.616162</td>\n",
       "      <td>6.640015</td>\n",
       "      <td>5.115434</td>\n",
       "      <td>7.022282</td>\n",
       "      <td>8.439014</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 65 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          GSM872328  GSM872329  GSM872330  GSM872331  GSM872332  GSM872333  \\\n",
       "A1BG       3.076475   2.987292   2.987292   3.165908   2.727649   3.655183   \n",
       "A1BG-AS1   5.749817   4.904232   4.387495   4.360151   4.243835   4.355736   \n",
       "A1CF       2.223265   2.223265   2.223265   2.223265   2.223265   2.223265   \n",
       "A2M       11.727785  11.030724  11.897919  10.843243  12.948972  12.603923   \n",
       "A2M-AS1    6.441273   6.099895   6.603606   5.386214   7.310987   7.769239   \n",
       "\n",
       "          GSM872334  GSM872335  GSM872336  GSM872337  ...  GSM872383  \\\n",
       "A1BG       2.987292   2.987292   3.574711   6.003634  ...   4.334097   \n",
       "A1BG-AS1   4.493685   4.477549   4.594941   5.299739  ...   4.309564   \n",
       "A1CF       2.223265   2.223265   2.223265   2.223265  ...   2.223265   \n",
       "A2M       13.163574  11.997735  12.926909  12.836881  ...  12.034571   \n",
       "A2M-AS1    7.290848   6.464501   7.775331   7.137957  ...   8.235774   \n",
       "\n",
       "          GSM872384  GSM872385  GSM872386  GSM872387  GSM872388  GSM872389  \\\n",
       "A1BG       3.014834   2.987292   3.565207   3.566093   2.987292   2.987292   \n",
       "A1BG-AS1   4.596462   4.635358   4.596462   4.605628   4.596462   4.596462   \n",
       "A1CF       2.223265   2.223265   2.223265   2.223265   2.223265   2.223265   \n",
       "A2M       12.973060  12.725227  10.293444  13.931666  13.758073  12.477485   \n",
       "A2M-AS1    7.260537   7.984555   4.548931   8.101128   8.616162   6.640015   \n",
       "\n",
       "          GSM872390  GSM872391  GSM872392  \n",
       "A1BG       2.898243   3.057402   2.987292  \n",
       "A1BG-AS1   4.638988   4.412014   4.741860  \n",
       "A1CF       2.223265   2.223265   2.223265  \n",
       "A2M       10.576160  12.909493  12.791211  \n",
       "A2M-AS1    5.115434   7.022282   8.439014  \n",
       "\n",
       "[5 rows x 65 columns]"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "annotated_expression.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='PCA 1 component 14% variance explained', ylabel='PCA 2 component 10% variance explained'>"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 600x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pca_plot(annotated_expression.T)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "annotated_expression.to_csv('./expressions.tsv.gz', sep='\\t', compression='gzip')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "pbs",
   "language": "python",
   "name": "pbs"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.5"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "384px"
   },
   "toc_section_display": true,
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}