Biomarker Identification - Mirdametinib.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Biomarker Identification - Mirdametinib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "\n",
    "cwd = os.getcwd()\n",
    "DATA = cwd + '/data'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Association between MAP2K1/2 mutation and MEK inhibitor"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This analysis follows the steps in the [Barretina et al.](https://www.nature.com/articles/nature11003) paper to investigate Mirdametinib, a MEK inhibitor. \n",
    "\n",
    "The Cancer Dependency Map (DepMap) project is an umbrella project that aims to providing high quality genomic profilings of cancer cell lines (CCLE), their sensitivities to genetic perturbations (Achilies, DRIVE, DEMETER2, etc.), and their sensitivities to small molecule perturbations (PRISM). Mining these data allows us to better understand cancer biology, and potentially discover new genes for targeted therapies.\n",
    "\n",
    "\n",
    "The first dataset that I will be using is the drug response data from PRISM, which stands for Profiling Relative Inhibition Simultaneously in Mixtures and is a project within DepMap. It uses high-throughput multiplexed screening on hundreds of human cancer cell lines and allows researchers to validate the mechanisms of action (MoA) of drugs, or identify potential novel targets of existing drugs for drug repurposing.\n",
    "\n",
    "At the time when this dataset was published, Mirdametinib was called by its lab name `PD-0325901`. The drug response data can be loaded and only the rows having `name=='PD-0325901'` will be included. Only those rows with `passed_str_profiling==True` have been kept."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/crystalshin/opt/anaconda3/lib/python3.8/site-packages/IPython/core/interactiveshell.py:3146: DtypeWarning: Columns (14,15) have mixed types.Specify dtype option on import or set low_memory=False.\n",
      "  has_raised = await self.run_ast_nodes(code_ast.body, cell_name,\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(462, 20)\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>broad_id</th>\n",
       "      <th>depmap_id</th>\n",
       "      <th>ccle_name</th>\n",
       "      <th>screen_id</th>\n",
       "      <th>upper_limit</th>\n",
       "      <th>lower_limit</th>\n",
       "      <th>slope</th>\n",
       "      <th>r2</th>\n",
       "      <th>auc</th>\n",
       "      <th>ec50</th>\n",
       "      <th>ic50</th>\n",
       "      <th>name</th>\n",
       "      <th>moa</th>\n",
       "      <th>target</th>\n",
       "      <th>disease.area</th>\n",
       "      <th>indication</th>\n",
       "      <th>smiles</th>\n",
       "      <th>phase</th>\n",
       "      <th>passed_str_profiling</th>\n",
       "      <th>row_name</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>320706</th>\n",
       "      <td>BRD-K49865102-001-08-4</td>\n",
       "      <td>ACH-000879</td>\n",
       "      <td>MFE296_ENDOMETRIUM</td>\n",
       "      <td>HTS002</td>\n",
       "      <td>1</td>\n",
       "      <td>1.753852</td>\n",
       "      <td>-0.140424</td>\n",
       "      <td>0.049046</td>\n",
       "      <td>1.358403</td>\n",
       "      <td>0.037749</td>\n",
       "      <td>NaN</td>\n",
       "      <td>PD-0325901</td>\n",
       "      <td>MEK inhibitor</td>\n",
       "      <td>MAP2K1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>OC[C@@H](O)CONC(=O)c1ccc(F)c(F)c1Nc1ccc(I)cc1F</td>\n",
       "      <td>Phase 2</td>\n",
       "      <td>True</td>\n",
       "      <td>ACH-000879</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>320707</th>\n",
       "      <td>BRD-K49865102-001-08-4</td>\n",
       "      <td>ACH-000320</td>\n",
       "      <td>PSN1_PANCREAS</td>\n",
       "      <td>HTS002</td>\n",
       "      <td>1</td>\n",
       "      <td>0.017189</td>\n",
       "      <td>0.867499</td>\n",
       "      <td>0.768102</td>\n",
       "      <td>0.309807</td>\n",
       "      <td>0.009980</td>\n",
       "      <td>0.010391</td>\n",
       "      <td>PD-0325901</td>\n",
       "      <td>MEK inhibitor</td>\n",
       "      <td>MAP2K1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>OC[C@@H](O)CONC(=O)c1ccc(F)c(F)c1Nc1ccc(I)cc1F</td>\n",
       "      <td>Phase 2</td>\n",
       "      <td>True</td>\n",
       "      <td>ACH-000320</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>320708</th>\n",
       "      <td>BRD-K49865102-001-08-4</td>\n",
       "      <td>ACH-001145</td>\n",
       "      <td>OC316_OVARY</td>\n",
       "      <td>HTS002</td>\n",
       "      <td>1</td>\n",
       "      <td>0.240523</td>\n",
       "      <td>1.622713</td>\n",
       "      <td>0.687787</td>\n",
       "      <td>0.578996</td>\n",
       "      <td>0.046090</td>\n",
       "      <td>0.069049</td>\n",
       "      <td>PD-0325901</td>\n",
       "      <td>MEK inhibitor</td>\n",
       "      <td>MAP2K1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>OC[C@@H](O)CONC(=O)c1ccc(F)c(F)c1Nc1ccc(I)cc1F</td>\n",
       "      <td>Phase 2</td>\n",
       "      <td>True</td>\n",
       "      <td>ACH-001145</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>320709</th>\n",
       "      <td>BRD-K49865102-001-08-4</td>\n",
       "      <td>ACH-000873</td>\n",
       "      <td>KYSE270_OESOPHAGUS</td>\n",
       "      <td>HTS002</td>\n",
       "      <td>1</td>\n",
       "      <td>0.387321</td>\n",
       "      <td>3.827219</td>\n",
       "      <td>0.586254</td>\n",
       "      <td>0.734215</td>\n",
       "      <td>0.148506</td>\n",
       "      <td>0.219195</td>\n",
       "      <td>PD-0325901</td>\n",
       "      <td>MEK inhibitor</td>\n",
       "      <td>MAP2K1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>OC[C@@H](O)CONC(=O)c1ccc(F)c(F)c1Nc1ccc(I)cc1F</td>\n",
       "      <td>Phase 2</td>\n",
       "      <td>True</td>\n",
       "      <td>ACH-000873</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>320710</th>\n",
       "      <td>BRD-K49865102-001-08-4</td>\n",
       "      <td>ACH-000855</td>\n",
       "      <td>KYSE150_OESOPHAGUS</td>\n",
       "      <td>HTS002</td>\n",
       "      <td>1</td>\n",
       "      <td>0.350424</td>\n",
       "      <td>0.754771</td>\n",
       "      <td>0.680582</td>\n",
       "      <td>0.722535</td>\n",
       "      <td>0.164757</td>\n",
       "      <td>0.815145</td>\n",
       "      <td>PD-0325901</td>\n",
       "      <td>MEK inhibitor</td>\n",
       "      <td>MAP2K1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>OC[C@@H](O)CONC(=O)c1ccc(F)c(F)c1Nc1ccc(I)cc1F</td>\n",
       "      <td>Phase 2</td>\n",
       "      <td>True</td>\n",
       "      <td>ACH-000855</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      broad_id   depmap_id           ccle_name screen_id  \\\n",
       "320706  BRD-K49865102-001-08-4  ACH-000879  MFE296_ENDOMETRIUM    HTS002   \n",
       "320707  BRD-K49865102-001-08-4  ACH-000320       PSN1_PANCREAS    HTS002   \n",
       "320708  BRD-K49865102-001-08-4  ACH-001145         OC316_OVARY    HTS002   \n",
       "320709  BRD-K49865102-001-08-4  ACH-000873  KYSE270_OESOPHAGUS    HTS002   \n",
       "320710  BRD-K49865102-001-08-4  ACH-000855  KYSE150_OESOPHAGUS    HTS002   \n",
       "\n",
       "        upper_limit  lower_limit     slope        r2       auc      ec50  \\\n",
       "320706            1     1.753852 -0.140424  0.049046  1.358403  0.037749   \n",
       "320707            1     0.017189  0.867499  0.768102  0.309807  0.009980   \n",
       "320708            1     0.240523  1.622713  0.687787  0.578996  0.046090   \n",
       "320709            1     0.387321  3.827219  0.586254  0.734215  0.148506   \n",
       "320710            1     0.350424  0.754771  0.680582  0.722535  0.164757   \n",
       "\n",
       "            ic50        name            moa  target disease.area indication  \\\n",
       "320706       NaN  PD-0325901  MEK inhibitor  MAP2K1          NaN        NaN   \n",
       "320707  0.010391  PD-0325901  MEK inhibitor  MAP2K1          NaN        NaN   \n",
       "320708  0.069049  PD-0325901  MEK inhibitor  MAP2K1          NaN        NaN   \n",
       "320709  0.219195  PD-0325901  MEK inhibitor  MAP2K1          NaN        NaN   \n",
       "320710  0.815145  PD-0325901  MEK inhibitor  MAP2K1          NaN        NaN   \n",
       "\n",
       "                                                smiles    phase  \\\n",
       "320706  OC[C@@H](O)CONC(=O)c1ccc(F)c(F)c1Nc1ccc(I)cc1F  Phase 2   \n",
       "320707  OC[C@@H](O)CONC(=O)c1ccc(F)c(F)c1Nc1ccc(I)cc1F  Phase 2   \n",
       "320708  OC[C@@H](O)CONC(=O)c1ccc(F)c(F)c1Nc1ccc(I)cc1F  Phase 2   \n",
       "320709  OC[C@@H](O)CONC(=O)c1ccc(F)c(F)c1Nc1ccc(I)cc1F  Phase 2   \n",
       "320710  OC[C@@H](O)CONC(=O)c1ccc(F)c(F)c1Nc1ccc(I)cc1F  Phase 2   \n",
       "\n",
       "        passed_str_profiling    row_name  \n",
       "320706                  True  ACH-000879  \n",
       "320707                  True  ACH-000320  \n",
       "320708                  True  ACH-001145  \n",
       "320709                  True  ACH-000873  \n",
       "320710                  True  ACH-000855  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "prism = pd.read_csv(DATA + \"/secondary-screen-dose-response-curve-parameters.csv\")\n",
    "prism = prism[prism.passed_str_profiling==True]\n",
    "prism = prism[prism.name==\"PD-0325901\"]\n",
    "print(prism.shape)\n",
    "prism.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this dataset, each row is a cell line-drug pair, where `depmap_id` column indicates the cell line ID as in the cell line info table above. The compound (drug) name can be found in the `name` column. If the molecule is designed to target a certain gene, it will be annotated in the `target` and `moa` columns. This allows us to investigate how inhibition of a certain pathway affect the viability of cancer cell lines with a certain mutation profiles.\n",
    "\n",
    "As we can see in the above table, we have selected out 462 cell lines that have sensitivity data for `PD-0325901`."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Association between MEK1/2 mutation and PD-0325901"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that we have identified which cell lines have sensitivity data for `PD-0325901`, we can now study the correlation between `PD-0325901` and its target genes `MEK1 (MAP2K1)` and `MEK2 (MAP2K2)`. The goal is to find in which specific genomic feature such as mutations, copy number variation, or over/under expression the drug will have the highest potency. \n",
    "\n",
    "We can load in the CCLE mutation MAF file for this analysis. Each cancer cell line is derived from a tumor, and knowing the information from the source of the cell line oftentimes could help the drug developers identify the indications ofinterest by answering questions like: in which cancer types are BRCA1/BRCA2 mutations most prevalent? \n",
    "\n",
    "When we load in the mutation dataset, we will make sure that we load in the rows that meet certain conditions. First, the `isDeleterious` column indicates if the mutation is deleterious. The `Variant_annotation` column indicates whether the mutation is damaging. The `isCOSMIChotspot` column shows whether the mutation has been observed as a mutation hotspot in the COSMIC (Catalogue of Somatic Mutations in Cancer) database. The union of the three columns would be used to filter out those mutations that are potentially associated with damaging the target genes. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/crystalshin/opt/anaconda3/lib/python3.8/site-packages/IPython/core/interactiveshell.py:3146: DtypeWarning: Columns (3,19,22,27,28,29,30,31) have mixed types.Specify dtype option on import or set low_memory=False.\n",
      "  has_raised = await self.run_ast_nodes(code_ast.body, cell_name,\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>Hugo_Symbol</th>\n",
       "      <th>Entrez_Gene_Id</th>\n",
       "      <th>NCBI_Build</th>\n",
       "      <th>Chromosome</th>\n",
       "      <th>Start_position</th>\n",
       "      <th>End_position</th>\n",
       "      <th>Strand</th>\n",
       "      <th>Variant_Classification</th>\n",
       "      <th>Variant_Type</th>\n",
       "      <th>Reference_Allele</th>\n",
       "      <th>...</th>\n",
       "      <th>isCOSMIChotspot</th>\n",
       "      <th>COSMIChsCnt</th>\n",
       "      <th>ExAC_AF</th>\n",
       "      <th>Variant_annotation</th>\n",
       "      <th>CGA_WES_AC</th>\n",
       "      <th>HC_AC</th>\n",
       "      <th>RD_AC</th>\n",
       "      <th>RNAseq_AC</th>\n",
       "      <th>SangerWES_AC</th>\n",
       "      <th>WGS_AC</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>VPS13D</td>\n",
       "      <td>55187</td>\n",
       "      <td>37</td>\n",
       "      <td>1</td>\n",
       "      <td>12359347</td>\n",
       "      <td>12359347</td>\n",
       "      <td>+</td>\n",
       "      <td>Nonsense_Mutation</td>\n",
       "      <td>SNP</td>\n",
       "      <td>C</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>damaging</td>\n",
       "      <td>34:213</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>34:221</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>AADACL4</td>\n",
       "      <td>343066</td>\n",
       "      <td>37</td>\n",
       "      <td>1</td>\n",
       "      <td>12726308</td>\n",
       "      <td>12726322</td>\n",
       "      <td>+</td>\n",
       "      <td>In_Frame_Del</td>\n",
       "      <td>DEL</td>\n",
       "      <td>CTGGCGTGACGCCAT</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>3.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>other non-conserving</td>\n",
       "      <td>57:141</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>9:0</td>\n",
       "      <td>28:32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>IFNLR1</td>\n",
       "      <td>163702</td>\n",
       "      <td>37</td>\n",
       "      <td>1</td>\n",
       "      <td>24484172</td>\n",
       "      <td>24484172</td>\n",
       "      <td>+</td>\n",
       "      <td>Silent</td>\n",
       "      <td>SNP</td>\n",
       "      <td>G</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>silent</td>\n",
       "      <td>118:0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>10:0</td>\n",
       "      <td>118:0</td>\n",
       "      <td>18:0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>TMEM57</td>\n",
       "      <td>55219</td>\n",
       "      <td>37</td>\n",
       "      <td>1</td>\n",
       "      <td>25785018</td>\n",
       "      <td>25785019</td>\n",
       "      <td>+</td>\n",
       "      <td>Frame_Shift_Ins</td>\n",
       "      <td>INS</td>\n",
       "      <td>-</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>damaging</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>6:28</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ZSCAN20</td>\n",
       "      <td>7579</td>\n",
       "      <td>37</td>\n",
       "      <td>1</td>\n",
       "      <td>33954141</td>\n",
       "      <td>33954141</td>\n",
       "      <td>+</td>\n",
       "      <td>Missense_Mutation</td>\n",
       "      <td>SNP</td>\n",
       "      <td>T</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>other non-conserving</td>\n",
       "      <td>28:62</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>27:61</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 32 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  Hugo_Symbol  Entrez_Gene_Id  NCBI_Build Chromosome  Start_position  \\\n",
       "0      VPS13D           55187          37          1        12359347   \n",
       "1     AADACL4          343066          37          1        12726308   \n",
       "2      IFNLR1          163702          37          1        24484172   \n",
       "3      TMEM57           55219          37          1        25785018   \n",
       "4     ZSCAN20            7579          37          1        33954141   \n",
       "\n",
       "   End_position Strand Variant_Classification Variant_Type Reference_Allele  \\\n",
       "0      12359347      +      Nonsense_Mutation          SNP                C   \n",
       "1      12726322      +           In_Frame_Del          DEL  CTGGCGTGACGCCAT   \n",
       "2      24484172      +                 Silent          SNP                G   \n",
       "3      25785019      +        Frame_Shift_Ins          INS                -   \n",
       "4      33954141      +      Missense_Mutation          SNP                T   \n",
       "\n",
       "   ... isCOSMIChotspot COSMIChsCnt ExAC_AF    Variant_annotation CGA_WES_AC  \\\n",
       "0  ...           False         0.0     NaN              damaging     34:213   \n",
       "1  ...           False         3.0     NaN  other non-conserving     57:141   \n",
       "2  ...           False         0.0     NaN                silent      118:0   \n",
       "3  ...           False         0.0     NaN              damaging        NaN   \n",
       "4  ...           False         0.0     NaN  other non-conserving      28:62   \n",
       "\n",
       "  HC_AC RD_AC RNAseq_AC SangerWES_AC WGS_AC  \n",
       "0   NaN   NaN       NaN       34:221    NaN  \n",
       "1   NaN   NaN       NaN          9:0  28:32  \n",
       "2   NaN   NaN      10:0        118:0   18:0  \n",
       "3   NaN   NaN      6:28          NaN    NaN  \n",
       "4   NaN   NaN       NaN        27:61    NaN  \n",
       "\n",
       "[5 rows x 32 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "maf = pd.read_csv(DATA + \"/CCLE_mutations.csv\")\n",
    "maf.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Hugo_Symbol</th>\n",
       "      <th>Entrez_Gene_Id</th>\n",
       "      <th>NCBI_Build</th>\n",
       "      <th>Chromosome</th>\n",
       "      <th>Start_position</th>\n",
       "      <th>End_position</th>\n",
       "      <th>Strand</th>\n",
       "      <th>Variant_Classification</th>\n",
       "      <th>Variant_Type</th>\n",
       "      <th>Reference_Allele</th>\n",
       "      <th>...</th>\n",
       "      <th>isCOSMIChotspot</th>\n",
       "      <th>COSMIChsCnt</th>\n",
       "      <th>ExAC_AF</th>\n",
       "      <th>Variant_annotation</th>\n",
       "      <th>CGA_WES_AC</th>\n",
       "      <th>HC_AC</th>\n",
       "      <th>RD_AC</th>\n",
       "      <th>RNAseq_AC</th>\n",
       "      <th>SangerWES_AC</th>\n",
       "      <th>WGS_AC</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>179706</th>\n",
       "      <td>MAP2K2</td>\n",
       "      <td>5605</td>\n",
       "      <td>37</td>\n",
       "      <td>19</td>\n",
       "      <td>4110508</td>\n",
       "      <td>4110508</td>\n",
       "      <td>+</td>\n",
       "      <td>Splice_Site</td>\n",
       "      <td>SNP</td>\n",
       "      <td>A</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>damaging</td>\n",
       "      <td>43:115</td>\n",
       "      <td>98:184</td>\n",
       "      <td>NaN</td>\n",
       "      <td>220:331</td>\n",
       "      <td>14:26</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>274669</th>\n",
       "      <td>MAP2K1</td>\n",
       "      <td>5604</td>\n",
       "      <td>37</td>\n",
       "      <td>15</td>\n",
       "      <td>66777530</td>\n",
       "      <td>66777530</td>\n",
       "      <td>+</td>\n",
       "      <td>Splice_Site</td>\n",
       "      <td>SNP</td>\n",
       "      <td>G</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>damaging</td>\n",
       "      <td>26:11</td>\n",
       "      <td>42:35</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5:0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>750333</th>\n",
       "      <td>MAP2K1</td>\n",
       "      <td>5604</td>\n",
       "      <td>37</td>\n",
       "      <td>15</td>\n",
       "      <td>66779629</td>\n",
       "      <td>66779629</td>\n",
       "      <td>+</td>\n",
       "      <td>Splice_Site</td>\n",
       "      <td>SNP</td>\n",
       "      <td>A</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>damaging</td>\n",
       "      <td>25:45</td>\n",
       "      <td>115:136</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>758884</th>\n",
       "      <td>MAP2K1</td>\n",
       "      <td>5604</td>\n",
       "      <td>37</td>\n",
       "      <td>15</td>\n",
       "      <td>66737045</td>\n",
       "      <td>66737045</td>\n",
       "      <td>+</td>\n",
       "      <td>Splice_Site</td>\n",
       "      <td>SNP</td>\n",
       "      <td>G</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>damaging</td>\n",
       "      <td>53:30</td>\n",
       "      <td>29:16</td>\n",
       "      <td>NaN</td>\n",
       "      <td>185:149</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>814497</th>\n",
       "      <td>MAP2K1</td>\n",
       "      <td>5604</td>\n",
       "      <td>37</td>\n",
       "      <td>15</td>\n",
       "      <td>66737045</td>\n",
       "      <td>66737045</td>\n",
       "      <td>+</td>\n",
       "      <td>Splice_Site</td>\n",
       "      <td>SNP</td>\n",
       "      <td>G</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>damaging</td>\n",
       "      <td>63:86</td>\n",
       "      <td>63:59</td>\n",
       "      <td>NaN</td>\n",
       "      <td>331:417</td>\n",
       "      <td>67:84</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>966081</th>\n",
       "      <td>MAP2K1</td>\n",
       "      <td>5604</td>\n",
       "      <td>37</td>\n",
       "      <td>15</td>\n",
       "      <td>66782055</td>\n",
       "      <td>66782055</td>\n",
       "      <td>+</td>\n",
       "      <td>Splice_Site</td>\n",
       "      <td>SNP</td>\n",
       "      <td>G</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>damaging</td>\n",
       "      <td>127:490</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>6 rows × 32 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       Hugo_Symbol  Entrez_Gene_Id  NCBI_Build Chromosome  Start_position  \\\n",
       "179706      MAP2K2            5605          37         19         4110508   \n",
       "274669      MAP2K1            5604          37         15        66777530   \n",
       "750333      MAP2K1            5604          37         15        66779629   \n",
       "758884      MAP2K1            5604          37         15        66737045   \n",
       "814497      MAP2K1            5604          37         15        66737045   \n",
       "966081      MAP2K1            5604          37         15        66782055   \n",
       "\n",
       "        End_position Strand Variant_Classification Variant_Type  \\\n",
       "179706       4110508      +            Splice_Site          SNP   \n",
       "274669      66777530      +            Splice_Site          SNP   \n",
       "750333      66779629      +            Splice_Site          SNP   \n",
       "758884      66737045      +            Splice_Site          SNP   \n",
       "814497      66737045      +            Splice_Site          SNP   \n",
       "966081      66782055      +            Splice_Site          SNP   \n",
       "\n",
       "       Reference_Allele  ... isCOSMIChotspot COSMIChsCnt ExAC_AF  \\\n",
       "179706                A  ...           False         0.0     NaN   \n",
       "274669                G  ...           False         0.0     NaN   \n",
       "750333                A  ...           False         0.0     NaN   \n",
       "758884                G  ...           False         0.0     NaN   \n",
       "814497                G  ...           False         0.0     NaN   \n",
       "966081                G  ...           False         0.0     NaN   \n",
       "\n",
       "       Variant_annotation CGA_WES_AC    HC_AC RD_AC RNAseq_AC SangerWES_AC  \\\n",
       "179706           damaging     43:115   98:184   NaN   220:331        14:26   \n",
       "274669           damaging      26:11    42:35   NaN       5:0          NaN   \n",
       "750333           damaging      25:45  115:136   NaN       NaN          NaN   \n",
       "758884           damaging      53:30    29:16   NaN   185:149          NaN   \n",
       "814497           damaging      63:86    63:59   NaN   331:417        67:84   \n",
       "966081           damaging    127:490      NaN   NaN       NaN          NaN   \n",
       "\n",
       "       WGS_AC  \n",
       "179706    NaN  \n",
       "274669    NaN  \n",
       "750333    NaN  \n",
       "758884    NaN  \n",
       "814497    NaN  \n",
       "966081    NaN  \n",
       "\n",
       "[6 rows x 32 columns]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mek = maf[((maf.Hugo_Symbol == \"MAP2K1\") | (maf.Hugo_Symbol == \"MAP2K2\"))&\n",
    "           (maf.isDeleterious | (maf.Variant_annotation == \"damaging\") | maf.isCOSMIChotspot)]\n",
    "mek"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We were able to identify 6 cell lines that are either deliterious, damaging or appear in COSMIC hotspot database. \n",
    "\n",
    "Now we can use the PRISM data to extract the drug screening result. To do this, we would select for all rows that have `depmap_id` appear in the mutation data's `DepMap_ID`. Then we would extract the columns `auc` (auc, which is the inverse of activity area), `ec50`, and `ic50`. Also, we would add a column indicating whether there is a mutation in MEK genes for each cell line. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": false
   },
   "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>depmap_id</th>\n",
       "      <th>auc</th>\n",
       "      <th>ec50</th>\n",
       "      <th>ic50</th>\n",
       "      <th>MEK_mut</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>320706</th>\n",
       "      <td>ACH-000879</td>\n",
       "      <td>1.358403</td>\n",
       "      <td>0.037749</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>320707</th>\n",
       "      <td>ACH-000320</td>\n",
       "      <td>0.309807</td>\n",
       "      <td>0.009980</td>\n",
       "      <td>0.010391</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>320708</th>\n",
       "      <td>ACH-001145</td>\n",
       "      <td>0.578996</td>\n",
       "      <td>0.046090</td>\n",
       "      <td>0.069049</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>320709</th>\n",
       "      <td>ACH-000873</td>\n",
       "      <td>0.734215</td>\n",
       "      <td>0.148506</td>\n",
       "      <td>0.219195</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>320710</th>\n",
       "      <td>ACH-000855</td>\n",
       "      <td>0.722535</td>\n",
       "      <td>0.164757</td>\n",
       "      <td>0.815145</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         depmap_id       auc      ec50      ic50  MEK_mut\n",
       "320706  ACH-000879  1.358403  0.037749       NaN    False\n",
       "320707  ACH-000320  0.309807  0.009980  0.010391    False\n",
       "320708  ACH-001145  0.578996  0.046090  0.069049    False\n",
       "320709  ACH-000873  0.734215  0.148506  0.219195    False\n",
       "320710  ACH-000855  0.722535  0.164757  0.815145    False"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plotdata = prism[\n",
    "    (prism.passed_str_profiling) & \n",
    "    (prism.name == \"PD-0325901\") & \n",
    "    prism.depmap_id.isin(set(maf.DepMap_ID))\n",
    "][\n",
    "    [\"depmap_id\", \"auc\", \"ec50\", \"ic50\",]\n",
    "].copy()\n",
    "\n",
    "plotdata[\"MEK_mut\"] = False\n",
    "plotdata.loc[plotdata.depmap_id.isin(set(mek.DepMap_ID)), \"MEK_mut\"] = True\n",
    "plotdata.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Using the table created, we can create a boxplot showing the area under the curve between the cell lines that have at least one damaging mutation in `MAP2K1` or `MAP2K2` gene. We can also perform a 2-sided Mann-Whitney U test to see if the difference is statistically significant."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/crystalshin/opt/anaconda3/lib/python3.8/site-packages/seaborn/categorical.py:1296: UserWarning: 18.7% of the points cannot be placed; you may want to decrease the size of the markers or use stripplot.\n",
      "  warnings.warn(msg, UserWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Text(1, 1.6, 'M-W U P-value = 0.7351')"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from scipy.stats import mannwhitneyu\n",
    "import seaborn as sns\n",
    "import numpy as np\n",
    "\n",
    "mwu, pval = mannwhitneyu(plotdata.loc[plotdata.MEK_mut, \"auc\"],\n",
    "                         plotdata.loc[~plotdata.MEK_mut, \"auc\"], alternative=\"two-sided\")\n",
    "ax = sns.boxplot(x=\"MEK_mut\", y=\"auc\", data=plotdata, fliersize=0)\n",
    "ax = sns.swarmplot(x=\"MEK_mut\", y=\"auc\", data=plotdata, color=\"k\",\n",
    "                  alpha=0.5, ax=ax)\n",
    "\n",
    "ax.text(1, 1.6, \"M-W U P-value = {}\".format(np.round(pval, 4)), ha=\"center\", va=\"center\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plotdata.MEK_mut.sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "From the boxplot, we do not see much difference between the two groups and the p-value was 0.74, supporting that it is not statistically significant. Also, there are only 3 cell lines that have mutations in MEK genes."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Identifying Other Predictive Features for Mirdametinib"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Since mutation in MEK genes were not highly associated with Mirdametinib, we could try to identify other predictive features. Here, we will only investigate mutation features' association with the response. We will create a `feat` matrix with pivoted mutation data, subset the feature matrix to only the cell lines that has auc values from `PD-0325901`, then fit it to the `SelectFromModel` feature selector with `RandomForestRegressor` estimator."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "scrolled": false
   },
   "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>Hugo_Symbol</th>\n",
       "      <th>A1BG_mut</th>\n",
       "      <th>A1CF_mut</th>\n",
       "      <th>A2M_mut</th>\n",
       "      <th>A2ML1_mut</th>\n",
       "      <th>A3GALT2_mut</th>\n",
       "      <th>A4GALT_mut</th>\n",
       "      <th>A4GNT_mut</th>\n",
       "      <th>AAAS_mut</th>\n",
       "      <th>AACS_mut</th>\n",
       "      <th>AADAC_mut</th>\n",
       "      <th>...</th>\n",
       "      <th>ZWINT_mut</th>\n",
       "      <th>ZXDA_mut</th>\n",
       "      <th>ZXDB_mut</th>\n",
       "      <th>ZXDC_mut</th>\n",
       "      <th>ZYG11A_mut</th>\n",
       "      <th>ZYG11B_mut</th>\n",
       "      <th>ZYX_mut</th>\n",
       "      <th>ZZEF1_mut</th>\n",
       "      <th>ZZZ3_mut</th>\n",
       "      <th>hsa-mir-1199_mut</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DepMap_ID</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",
       "      <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",
       "      <th>ACH-000001</th>\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>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-000002</th>\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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-000003</th>\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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-000004</th>\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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-000005</th>\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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-002508</th>\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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-002509</th>\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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-002510</th>\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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-002511</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</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>...</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-002512</th>\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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</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>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1745 rows × 18135 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "Hugo_Symbol  A1BG_mut  A1CF_mut  A2M_mut  A2ML1_mut  A3GALT2_mut  A4GALT_mut  \\\n",
       "DepMap_ID                                                                      \n",
       "ACH-000001          0         0        0          0            0           0   \n",
       "ACH-000002          0         0        0          0            0           0   \n",
       "ACH-000003          0         0        0          0            0           0   \n",
       "ACH-000004          0         0        0          0            0           0   \n",
       "ACH-000005          0         0        0          0            0           0   \n",
       "...               ...       ...      ...        ...          ...         ...   \n",
       "ACH-002508          0         0        0          0            0           0   \n",
       "ACH-002509          0         0        0          0            0           0   \n",
       "ACH-002510          0         0        0          0            0           0   \n",
       "ACH-002511          0         0        0          1            0           0   \n",
       "ACH-002512          0         0        0          0            0           0   \n",
       "\n",
       "Hugo_Symbol  A4GNT_mut  AAAS_mut  AACS_mut  AADAC_mut  ...  ZWINT_mut  \\\n",
       "DepMap_ID                                              ...              \n",
       "ACH-000001           0         1         0          0  ...          0   \n",
       "ACH-000002           0         0         0          0  ...          1   \n",
       "ACH-000003           0         0         0          0  ...          0   \n",
       "ACH-000004           0         0         0          0  ...          0   \n",
       "ACH-000005           0         0         0          0  ...          0   \n",
       "...                ...       ...       ...        ...  ...        ...   \n",
       "ACH-002508           0         0         0          0  ...          0   \n",
       "ACH-002509           0         0         0          0  ...          0   \n",
       "ACH-002510           0         0         0          0  ...          0   \n",
       "ACH-002511           0         0         0          0  ...          0   \n",
       "ACH-002512           0         0         0          0  ...          0   \n",
       "\n",
       "Hugo_Symbol  ZXDA_mut  ZXDB_mut  ZXDC_mut  ZYG11A_mut  ZYG11B_mut  ZYX_mut  \\\n",
       "DepMap_ID                                                                    \n",
       "ACH-000001          0         0         0           0           0        0   \n",
       "ACH-000002          0         0         0           0           0        0   \n",
       "ACH-000003          0         0         0           0           0        0   \n",
       "ACH-000004          0         0         0           0           0        0   \n",
       "ACH-000005          0         0         0           0           0        0   \n",
       "...               ...       ...       ...         ...         ...      ...   \n",
       "ACH-002508          0         0         0           0           0        0   \n",
       "ACH-002509          0         0         0           0           0        0   \n",
       "ACH-002510          0         0         0           0           0        0   \n",
       "ACH-002511          0         0         0           0           0        0   \n",
       "ACH-002512          0         0         0           0           0        0   \n",
       "\n",
       "Hugo_Symbol  ZZEF1_mut  ZZZ3_mut  hsa-mir-1199_mut  \n",
       "DepMap_ID                                           \n",
       "ACH-000001           0         0                 0  \n",
       "ACH-000002           0         0                 0  \n",
       "ACH-000003           0         0                 0  \n",
       "ACH-000004           0         0                 0  \n",
       "ACH-000005           0         0                 0  \n",
       "...                ...       ...               ...  \n",
       "ACH-002508           0         0                 0  \n",
       "ACH-002509           0         0                 0  \n",
       "ACH-002510           0         0                 0  \n",
       "ACH-002511           0         0                 0  \n",
       "ACH-002512           0         0                 0  \n",
       "\n",
       "[1745 rows x 18135 columns]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "maf_nn = maf[(maf.isDeleterious | (maf.Variant_annotation == \"damaging\") | maf.isCOSMIChotspot.astype(bool))]\n",
    "mut_mat = maf_nn.pivot_table(\n",
    "    index=\"Hugo_Symbol\", columns=\"DepMap_ID\", values=\"Start_position\",\n",
    "    # the values can be anything, as we just want to convert it to number of mutations\n",
    "    aggfunc=len, fill_value=0\n",
    ")\n",
    "mut_mat.index = mut_mat.index.to_series() + \"_mut\"\n",
    "feat = mut_mat.T\n",
    "feat"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Like we did previously, we can extract just the cell lines from the `feat` matrix."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "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>auc</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>depmap_id</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ACH-000879</th>\n",
       "      <td>1.358403</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-000320</th>\n",
       "      <td>0.309807</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-001145</th>\n",
       "      <td>0.578996</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-000873</th>\n",
       "      <td>0.734215</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-000855</th>\n",
       "      <td>0.722535</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-000977</th>\n",
       "      <td>0.835512</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-000397</th>\n",
       "      <td>1.103102</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-000565</th>\n",
       "      <td>0.409922</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-000209</th>\n",
       "      <td>1.101516</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ACH-000927</th>\n",
       "      <td>0.982752</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>462 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                 auc\n",
       "depmap_id           \n",
       "ACH-000879  1.358403\n",
       "ACH-000320  0.309807\n",
       "ACH-001145  0.578996\n",
       "ACH-000873  0.734215\n",
       "ACH-000855  0.722535\n",
       "...              ...\n",
       "ACH-000977  0.835512\n",
       "ACH-000397  1.103102\n",
       "ACH-000565  0.409922\n",
       "ACH-000209  1.101516\n",
       "ACH-000927  0.982752\n",
       "\n",
       "[462 rows x 1 columns]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "target = prism.loc[\n",
    "    (prism.passed_str_profiling) & \n",
    "    (prism.name == \"PD-0325901\"),\n",
    "    [\"auc\", \"depmap_id\"]\n",
    "].set_index(\"depmap_id\", drop=True)\n",
    "target"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(462, 18135)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# find common samples\n",
    "common_celllines = list(set(target.index) & set(feat.index))\n",
    "target = target.loc[common_celllines]\n",
    "feat = feat.loc[common_celllines]\n",
    "print(feat.shape)\n",
    "np.all(feat.index == target.index)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Using the `RandomForestRegressor`, we can fit a model to select the feature with the highest importance. The target variable for the model to learn is the `auc`. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/crystalshin/opt/anaconda3/lib/python3.8/site-packages/sklearn/feature_selection/_from_model.py:224: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
      "  self.estimator_.fit(X, y, **fit_params)\n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn.feature_selection import SelectFromModel\n",
    "\n",
    "feat_sel = SelectFromModel(\n",
    "    RandomForestRegressor(random_state=4060)\n",
    ").fit(feat, target)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Within the feature selector, there is an `estimator_` attribute that contains the fitted random forest model. The `feature_importance_` attribute contains information about the importance of each feature in the model. We can create a new table listing features by its importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "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>feature</th>\n",
       "      <th>importance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1617</th>\n",
       "      <td>BRAF_mut</td>\n",
       "      <td>0.057794</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8153</th>\n",
       "      <td>KRAS_mut</td>\n",
       "      <td>0.043856</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4224</th>\n",
       "      <td>DET1_mut</td>\n",
       "      <td>0.021713</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4515</th>\n",
       "      <td>DPY19L1_mut</td>\n",
       "      <td>0.021441</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10365</th>\n",
       "      <td>NRAS_mut</td>\n",
       "      <td>0.019281</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7742</th>\n",
       "      <td>KBTBD3_mut</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7743</th>\n",
       "      <td>KBTBD4_mut</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7744</th>\n",
       "      <td>KBTBD6_mut</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7745</th>\n",
       "      <td>KBTBD7_mut</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18134</th>\n",
       "      <td>hsa-mir-1199_mut</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>18135 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                feature  importance\n",
       "1617           BRAF_mut    0.057794\n",
       "8153           KRAS_mut    0.043856\n",
       "4224           DET1_mut    0.021713\n",
       "4515        DPY19L1_mut    0.021441\n",
       "10365          NRAS_mut    0.019281\n",
       "...                 ...         ...\n",
       "7742         KBTBD3_mut    0.000000\n",
       "7743         KBTBD4_mut    0.000000\n",
       "7744         KBTBD6_mut    0.000000\n",
       "7745         KBTBD7_mut    0.000000\n",
       "18134  hsa-mir-1199_mut    0.000000\n",
       "\n",
       "[18135 rows x 2 columns]"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_importance = pd.DataFrame({\n",
    "    \"feature\": feat.columns,\n",
    "    \"importance\": feat_sel.estimator_.feature_importances_\n",
    "}).sort_values(\"importance\", ascending=False)\n",
    "feature_importance"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this case, the top predictive mutation feature was the BRAF mutation. BRAF gene works together with MEK gene to provide instructions for producing MEK1, which is a protein kinase that controls proper cell growth and division. Thus it makes sense that this gene might be of interest to us."
   ]
  },
  {
   "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": "3.8.5"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": false,
   "sideBar": true,
   "skip_h1_title": true,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "288px"
   },
   "toc_section_display": true,
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}