colab_MixTCRpred.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "#################################\n",
    "#Run this cell if using Google COLAB\n",
    "#################################\n",
    "\n",
    "#clone the repository and the data to run the notebook on Google Colab\n",
    "!git clone https://github.com/GfellerLab/MixTCRpred\n",
    "!pip install wget >> /dev/null \n",
    "!pip install pytorch_lightning >> /dev/null \n",
    "%cd MixTCRpred"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# MixTCRpred\n",
    "\n",
    "MixTCRpred is a sequence-based predictor for TCR-pMHC interactions. Currently, TCR binding predictions are available for 146 pMHCs. Robust predictions have been achieved in internal cross-validation for 43 pMHCs with more than 50 training TCRs.\n",
    "\n",
    "The file /pretrained_models/info_models.csv lists the available MixTCRpred pre-trained models and additional information."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "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>MixTCRpred_model_name</th>\n",
       "      <th>Peptide</th>\n",
       "      <th>Origin</th>\n",
       "      <th>MHC_class</th>\n",
       "      <th>MHC</th>\n",
       "      <th>Host_species</th>\n",
       "      <th>Number_training_abTCR</th>\n",
       "      <th>AUC_5fold</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>H2-IAb_DIYKGVYQFKSV</td>\n",
       "      <td>DIYKGVYQFKSV</td>\n",
       "      <td>Lymphocytic choriomeningitis mammarenavirus (L...</td>\n",
       "      <td>MHCII</td>\n",
       "      <td>H2-IAb</td>\n",
       "      <td>MusMusculus</td>\n",
       "      <td>3650</td>\n",
       "      <td>0.953809</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>A0201_GILGFVFTL</td>\n",
       "      <td>GILGFVFTL</td>\n",
       "      <td>Influenza A virus</td>\n",
       "      <td>MHCI</td>\n",
       "      <td>HLA-A*02:01</td>\n",
       "      <td>HomoSapiens</td>\n",
       "      <td>2079</td>\n",
       "      <td>0.943283</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>H2-Kb_SSYRRPVGI</td>\n",
       "      <td>SSYRRPVGI</td>\n",
       "      <td>Influenza A virus</td>\n",
       "      <td>MHCI</td>\n",
       "      <td>H2-Kb</td>\n",
       "      <td>MusMusculus</td>\n",
       "      <td>1158</td>\n",
       "      <td>0.938677</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>H2-Db_SSLENFRAYV</td>\n",
       "      <td>SSLENFRAYV</td>\n",
       "      <td>Influenza A virus</td>\n",
       "      <td>MHCI</td>\n",
       "      <td>H2-Db</td>\n",
       "      <td>MusMusculus</td>\n",
       "      <td>798</td>\n",
       "      <td>0.906340</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>A0201_LLWNGPMAV</td>\n",
       "      <td>LLWNGPMAV</td>\n",
       "      <td>Yellow fever virus (YFV)</td>\n",
       "      <td>MHCI</td>\n",
       "      <td>HLA-A*02:01</td>\n",
       "      <td>HomoSapiens</td>\n",
       "      <td>644</td>\n",
       "      <td>0.928752</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>141</th>\n",
       "      <td>DQ2_PQPELPYPQPE</td>\n",
       "      <td>PQPELPYPQPE</td>\n",
       "      <td>Triticum aestivum</td>\n",
       "      <td>MHCII</td>\n",
       "      <td>HLA-DQ2</td>\n",
       "      <td>HomoSapiens</td>\n",
       "      <td>10</td>\n",
       "      <td>0.992424</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>142</th>\n",
       "      <td>H2-Kb_KGYVYQGL</td>\n",
       "      <td>KGYVYQGL</td>\n",
       "      <td>Vesicular stomatitis virus (VSV)</td>\n",
       "      <td>MHCII</td>\n",
       "      <td>H2-Kb</td>\n",
       "      <td>MusMusculus</td>\n",
       "      <td>10</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>143</th>\n",
       "      <td>B4402_QELIRQGTDY</td>\n",
       "      <td>QELIRQGTDY</td>\n",
       "      <td>Severe acute respiratory syndrome coronavirus ...</td>\n",
       "      <td>MHCI</td>\n",
       "      <td>HLA-B*44:02</td>\n",
       "      <td>HomoSapiens</td>\n",
       "      <td>10</td>\n",
       "      <td>0.933333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>144</th>\n",
       "      <td>A0201_RLITGRLQSL</td>\n",
       "      <td>RLITGRLQSL</td>\n",
       "      <td>Severe acute respiratory syndrome coronavirus ...</td>\n",
       "      <td>MHCI</td>\n",
       "      <td>HLA-A*02:01</td>\n",
       "      <td>HomoSapiens</td>\n",
       "      <td>10</td>\n",
       "      <td>0.720833</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>145</th>\n",
       "      <td>A0201_YVLTWIVGA</td>\n",
       "      <td>YVLTWIVGA</td>\n",
       "      <td>Homo sapiens (Neoepitope)</td>\n",
       "      <td>MHCI</td>\n",
       "      <td>HLA-A*02:01</td>\n",
       "      <td>HomoSapiens</td>\n",
       "      <td>10</td>\n",
       "      <td>0.774592</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>146 rows × 8 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      MixTCRpred_model_name       Peptide  \\\n",
       "0    H2-IAb_DIYKGVYQFKSV  DIYKGVYQFKSV   \n",
       "1        A0201_GILGFVFTL     GILGFVFTL   \n",
       "2        H2-Kb_SSYRRPVGI     SSYRRPVGI   \n",
       "3       H2-Db_SSLENFRAYV    SSLENFRAYV   \n",
       "4        A0201_LLWNGPMAV     LLWNGPMAV   \n",
       "..                   ...           ...   \n",
       "141      DQ2_PQPELPYPQPE   PQPELPYPQPE   \n",
       "142       H2-Kb_KGYVYQGL      KGYVYQGL   \n",
       "143     B4402_QELIRQGTDY    QELIRQGTDY   \n",
       "144     A0201_RLITGRLQSL    RLITGRLQSL   \n",
       "145      A0201_YVLTWIVGA     YVLTWIVGA   \n",
       "\n",
       "                                                Origin MHC_class          MHC  \\\n",
       "0    Lymphocytic choriomeningitis mammarenavirus (L...     MHCII       H2-IAb   \n",
       "1                                   Influenza A virus       MHCI  HLA-A*02:01   \n",
       "2                                   Influenza A virus       MHCI        H2-Kb   \n",
       "3                                   Influenza A virus       MHCI        H2-Db   \n",
       "4                             Yellow fever virus (YFV)      MHCI  HLA-A*02:01   \n",
       "..                                                 ...       ...          ...   \n",
       "141                                  Triticum aestivum     MHCII      HLA-DQ2   \n",
       "142                   Vesicular stomatitis virus (VSV)     MHCII        H2-Kb   \n",
       "143  Severe acute respiratory syndrome coronavirus ...      MHCI  HLA-B*44:02   \n",
       "144  Severe acute respiratory syndrome coronavirus ...      MHCI  HLA-A*02:01   \n",
       "145                          Homo sapiens (Neoepitope)      MHCI  HLA-A*02:01   \n",
       "\n",
       "    Host_species  Number_training_abTCR  AUC_5fold  \n",
       "0    MusMusculus                   3650   0.953809  \n",
       "1    HomoSapiens                   2079   0.943283  \n",
       "2    MusMusculus                   1158   0.938677  \n",
       "3    MusMusculus                    798   0.906340  \n",
       "4    HomoSapiens                    644   0.928752  \n",
       "..           ...                    ...        ...  \n",
       "141  HomoSapiens                     10   0.992424  \n",
       "142  MusMusculus                     10   1.000000  \n",
       "143  HomoSapiens                     10   0.933333  \n",
       "144  HomoSapiens                     10   0.720833  \n",
       "145  HomoSapiens                     10   0.774592  \n",
       "\n",
       "[146 rows x 8 columns]"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import wget\n",
    "import os\n",
    "df_info = pd.read_csv(\"./pretrained_models/info_models.csv\")\n",
    "df_info"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The input of MixTCRpred are paired TCRs (V and J genes and CDR3 alpha and beta sequence) like those in the file ./test/test.out. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>cdr3_TRA</th>\n",
       "      <th>cdr3_TRB</th>\n",
       "      <th>TRAV</th>\n",
       "      <th>TRAJ</th>\n",
       "      <th>TRBV</th>\n",
       "      <th>TRBJ</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CARGSNYNVLYF</td>\n",
       "      <td>CASRGQSQNTLYF</td>\n",
       "      <td>TRAV14D</td>\n",
       "      <td>TRAJ21</td>\n",
       "      <td>TRBV13-3</td>\n",
       "      <td>TRBJ2-4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CAMSAIMNRDDKIIF</td>\n",
       "      <td>CASRPNPGQGSYEQYF</td>\n",
       "      <td>TRAV17</td>\n",
       "      <td>TRAJ9</td>\n",
       "      <td>TRBV30</td>\n",
       "      <td>TRBJ2-7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CAVQRGGQKLLF</td>\n",
       "      <td>CASSPPQRLQETQYF</td>\n",
       "      <td>TRAV19</td>\n",
       "      <td>TRAJ10</td>\n",
       "      <td>TRBV19</td>\n",
       "      <td>TRBJ1-2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CAASIVWGSNFGNEKLTF</td>\n",
       "      <td>CASRTGDGQPQHF</td>\n",
       "      <td>TRAV14/DV4</td>\n",
       "      <td>TRAJ26</td>\n",
       "      <td>TRBV10-3</td>\n",
       "      <td>TRBJ2-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>CAAGSYNFNKFYF</td>\n",
       "      <td>CASSLSGGRTEAFF</td>\n",
       "      <td>TRAV12-3</td>\n",
       "      <td>TRAJ10</td>\n",
       "      <td>TRBV9</td>\n",
       "      <td>TRBJ2-7</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>998</th>\n",
       "      <td>CAGRDYGGATNKLIF</td>\n",
       "      <td>CSVRLVSKNIQYF</td>\n",
       "      <td>TRAV17</td>\n",
       "      <td>TRAJ9</td>\n",
       "      <td>TRBV12-3</td>\n",
       "      <td>TRBJ2-2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>999</th>\n",
       "      <td>CAVRDKGTGGFKTIF</td>\n",
       "      <td>CASSLTGTGAQEQYF</td>\n",
       "      <td>TRAV14/DV4</td>\n",
       "      <td>TRAJ17</td>\n",
       "      <td>TRBV6-6</td>\n",
       "      <td>TRBJ2-3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1000</th>\n",
       "      <td>CAGTDRGSTLGRLYF</td>\n",
       "      <td>CASSQVKVSSYNEQFF</td>\n",
       "      <td>TRAV14/DV4</td>\n",
       "      <td>TRAJ3</td>\n",
       "      <td>TRBV7-9</td>\n",
       "      <td>TRBJ2-7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1001</th>\n",
       "      <td>CAATYGTNAGKSTF</td>\n",
       "      <td>CASSQNGVGGEQYF</td>\n",
       "      <td>TRAV8-2</td>\n",
       "      <td>TRAJ39</td>\n",
       "      <td>TRBV5-5</td>\n",
       "      <td>TRBJ2-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1002</th>\n",
       "      <td>CVVSDRGSTLGRLYF</td>\n",
       "      <td>CASSDPRPGGNSPLHF</td>\n",
       "      <td>TRAV17</td>\n",
       "      <td>TRAJ58</td>\n",
       "      <td>TRBV5-1</td>\n",
       "      <td>TRBJ2-1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1003 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                cdr3_TRA          cdr3_TRB        TRAV    TRAJ      TRBV  \\\n",
       "0           CARGSNYNVLYF     CASRGQSQNTLYF     TRAV14D  TRAJ21  TRBV13-3   \n",
       "1        CAMSAIMNRDDKIIF  CASRPNPGQGSYEQYF      TRAV17   TRAJ9    TRBV30   \n",
       "2           CAVQRGGQKLLF   CASSPPQRLQETQYF      TRAV19  TRAJ10    TRBV19   \n",
       "3     CAASIVWGSNFGNEKLTF     CASRTGDGQPQHF  TRAV14/DV4  TRAJ26  TRBV10-3   \n",
       "4          CAAGSYNFNKFYF    CASSLSGGRTEAFF    TRAV12-3  TRAJ10     TRBV9   \n",
       "...                  ...               ...         ...     ...       ...   \n",
       "998      CAGRDYGGATNKLIF     CSVRLVSKNIQYF      TRAV17   TRAJ9  TRBV12-3   \n",
       "999      CAVRDKGTGGFKTIF   CASSLTGTGAQEQYF  TRAV14/DV4  TRAJ17   TRBV6-6   \n",
       "1000     CAGTDRGSTLGRLYF  CASSQVKVSSYNEQFF  TRAV14/DV4   TRAJ3   TRBV7-9   \n",
       "1001      CAATYGTNAGKSTF    CASSQNGVGGEQYF     TRAV8-2  TRAJ39   TRBV5-5   \n",
       "1002     CVVSDRGSTLGRLYF  CASSDPRPGGNSPLHF      TRAV17  TRAJ58   TRBV5-1   \n",
       "\n",
       "         TRBJ  \n",
       "0     TRBJ2-4  \n",
       "1     TRBJ2-7  \n",
       "2     TRBJ1-2  \n",
       "3     TRBJ2-1  \n",
       "4     TRBJ2-7  \n",
       "...       ...  \n",
       "998   TRBJ2-2  \n",
       "999   TRBJ2-3  \n",
       "1000  TRBJ2-7  \n",
       "1001  TRBJ2-1  \n",
       "1002  TRBJ2-1  \n",
       "\n",
       "[1003 rows x 6 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "input_tcrs = pd.read_csv(\"./test/test.csv\")\n",
    "input_tcrs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To run MixTCRpred, 3 arguments are required:  \n",
    "1. --model (the MixTCRpred model for the target pMHC). \n",
    "2. --input (the list of TCRs)  \n",
    "3. --output (output file) the file where the results will be stored.  \n",
    "\n",
    "E.g., Which TCRs in ./test/test.csv are more likely to target the HLA-A\\*02:01 GILGFVFTL epitope (MixTCRpred model = A0201_GILGFVFTL)? We store the results in ./test/out_A0201_GILGFVFTL.csv.\n",
    "The code to run is:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "#########################################################################\n",
      "###### MixTCRpred: a sequence-based predictor of TCR-pMHC interaction  ####\n",
      "#########################################################################\n",
      "MixTCRpred model A0201_GILGFVFTL \n",
      "Computing binding predictions for ./test/test.csv\n",
      "Testing DataLoader 0: 100%|████████████████| 1003/1003 [00:05<00:00, 169.60it/s]\n",
      "The results are stored in ./test/out_A0201_GILGFVFTL.csv.\n"
     ]
    }
   ],
   "source": [
    "!python MixTCRpred.py --model A0201_GILGFVFTL --input ./test/test.csv --output ./test/out_A0201_GILGFVFTL.csv\n",
    "#Please note the leading \"!\" which allows you to run bash code in the notebook"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Done! The results are sorted from the most probable binder to the least probable binder, and we can now proceed with their analysis. We provide both the raw MixTCRpred score and the % rank. In the paper, we observed that most true binders have a %rank < 0.1."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>cdr3_TRA</th>\n",
       "      <th>cdr3_TRB</th>\n",
       "      <th>TRAV</th>\n",
       "      <th>TRAJ</th>\n",
       "      <th>TRBV</th>\n",
       "      <th>TRBJ</th>\n",
       "      <th>score</th>\n",
       "      <th>perc_rank</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CAGGGSQGNLIF</td>\n",
       "      <td>CASSIRSSYEQYF</td>\n",
       "      <td>TRAV27</td>\n",
       "      <td>TRAJ42</td>\n",
       "      <td>TRBV19</td>\n",
       "      <td>TRBJ2-6</td>\n",
       "      <td>3.85309</td>\n",
       "      <td>0.00040</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CAGQYGGSQGNLIF</td>\n",
       "      <td>CASSIRSTDTQYF</td>\n",
       "      <td>TRAV12</td>\n",
       "      <td>TRAJ42</td>\n",
       "      <td>TRBV19</td>\n",
       "      <td>TRBJ2-3</td>\n",
       "      <td>3.57812</td>\n",
       "      <td>0.00164</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CAFINGSSNTGKLIF</td>\n",
       "      <td>CATSSFLAVSYEQYF</td>\n",
       "      <td>TRAV1-2</td>\n",
       "      <td>TRAJ29</td>\n",
       "      <td>TRBV3-1</td>\n",
       "      <td>TRBJ2-2</td>\n",
       "      <td>0.50054</td>\n",
       "      <td>26.01601</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CAASIDGRNNDMRF</td>\n",
       "      <td>CASSPFTGPPYEQYF</td>\n",
       "      <td>TRAV14/DV4</td>\n",
       "      <td>TRAJ45</td>\n",
       "      <td>TRBV19</td>\n",
       "      <td>TRBJ2-7</td>\n",
       "      <td>0.36692</td>\n",
       "      <td>31.19044</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>CAYRSPWGMGGSQGNLIF</td>\n",
       "      <td>CATSFMVQETQYF</td>\n",
       "      <td>TRAV13-1</td>\n",
       "      <td>TRAJ6</td>\n",
       "      <td>TRBV19</td>\n",
       "      <td>TRBJ1-4</td>\n",
       "      <td>0.04156</td>\n",
       "      <td>45.24820</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>998</th>\n",
       "      <td>CAEISFFSGGYNKLIF</td>\n",
       "      <td>CSALAGGLNTQYF</td>\n",
       "      <td>TRAV21</td>\n",
       "      <td>TRAJ49</td>\n",
       "      <td>TRBV9</td>\n",
       "      <td>TRBJ2-1</td>\n",
       "      <td>-5.66217</td>\n",
       "      <td>100.00000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>999</th>\n",
       "      <td>CAVTDGAGSYQLTF</td>\n",
       "      <td>CASSPSGITGELFF</td>\n",
       "      <td>TRAV2</td>\n",
       "      <td>TRAJ42</td>\n",
       "      <td>TRBV5-8</td>\n",
       "      <td>TRBJ1-1</td>\n",
       "      <td>-5.68870</td>\n",
       "      <td>100.00000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1000</th>\n",
       "      <td>CAVRWDTGNQFYF</td>\n",
       "      <td>CASSQTGRYQETQYF</td>\n",
       "      <td>TRAV8-4</td>\n",
       "      <td>TRAJ26</td>\n",
       "      <td>TRBV9</td>\n",
       "      <td>TRBJ1-5</td>\n",
       "      <td>-5.77388</td>\n",
       "      <td>100.00000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1001</th>\n",
       "      <td>CAVSPTGRRALTF</td>\n",
       "      <td>CATSPGQNTGELFF</td>\n",
       "      <td>TRAV2</td>\n",
       "      <td>TRAJ52</td>\n",
       "      <td>TRBV10-3</td>\n",
       "      <td>TRBJ1-1</td>\n",
       "      <td>-6.11638</td>\n",
       "      <td>100.00000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1002</th>\n",
       "      <td>CALTKPGAGSYQLTF</td>\n",
       "      <td>CASRPTRRVEKLFF</td>\n",
       "      <td>TRAV9-2</td>\n",
       "      <td>TRAJ9</td>\n",
       "      <td>TRBV15</td>\n",
       "      <td>TRBJ1-1</td>\n",
       "      <td>-6.18402</td>\n",
       "      <td>100.00000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1003 rows × 8 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                cdr3_TRA         cdr3_TRB        TRAV    TRAJ      TRBV  \\\n",
       "0           CAGGGSQGNLIF    CASSIRSSYEQYF      TRAV27  TRAJ42    TRBV19   \n",
       "1         CAGQYGGSQGNLIF    CASSIRSTDTQYF      TRAV12  TRAJ42    TRBV19   \n",
       "2        CAFINGSSNTGKLIF  CATSSFLAVSYEQYF     TRAV1-2  TRAJ29   TRBV3-1   \n",
       "3         CAASIDGRNNDMRF  CASSPFTGPPYEQYF  TRAV14/DV4  TRAJ45    TRBV19   \n",
       "4     CAYRSPWGMGGSQGNLIF    CATSFMVQETQYF    TRAV13-1   TRAJ6    TRBV19   \n",
       "...                  ...              ...         ...     ...       ...   \n",
       "998     CAEISFFSGGYNKLIF    CSALAGGLNTQYF      TRAV21  TRAJ49     TRBV9   \n",
       "999       CAVTDGAGSYQLTF   CASSPSGITGELFF       TRAV2  TRAJ42   TRBV5-8   \n",
       "1000       CAVRWDTGNQFYF  CASSQTGRYQETQYF     TRAV8-4  TRAJ26     TRBV9   \n",
       "1001       CAVSPTGRRALTF   CATSPGQNTGELFF       TRAV2  TRAJ52  TRBV10-3   \n",
       "1002     CALTKPGAGSYQLTF   CASRPTRRVEKLFF     TRAV9-2   TRAJ9    TRBV15   \n",
       "\n",
       "         TRBJ    score  perc_rank  \n",
       "0     TRBJ2-6  3.85309    0.00040  \n",
       "1     TRBJ2-3  3.57812    0.00164  \n",
       "2     TRBJ2-2  0.50054   26.01601  \n",
       "3     TRBJ2-7  0.36692   31.19044  \n",
       "4     TRBJ1-4  0.04156   45.24820  \n",
       "...       ...      ...        ...  \n",
       "998   TRBJ2-1 -5.66217  100.00000  \n",
       "999   TRBJ1-1 -5.68870  100.00000  \n",
       "1000  TRBJ1-5 -5.77388  100.00000  \n",
       "1001  TRBJ1-1 -6.11638  100.00000  \n",
       "1002  TRBJ1-1 -6.18402  100.00000  \n",
       "\n",
       "[1003 rows x 8 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_GIL = pd.read_csv(\"./test/out_A0201_GILGFVFTL.csv\", comment = '#')\n",
    "results_GIL"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Currently, 146 pretrained models are available, but not stored on GitHub. To make predictions for another pMHC we first need to download the corresponding MixTCRpred model.  \n",
    "E.g., to identify TCRs specific for the HLA-A\\*02:01,YLQPRTFLL epitope from the SARS-CoV-2 (MixTCRpred model_name A0201_YLQPRTFLL), we can first download it by running:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MixTCRpred model already downloaded\n"
     ]
    }
   ],
   "source": [
    "!python MixTCRpred.py --download A0201_YLQPRTFLL"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " and then we can make predictions for the new pMHC."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "#########################################################################\n",
      "###### MixTCRpred: a sequence-based predictor of TCR-pMHC interaction  ####\n",
      "#########################################################################\n",
      "MixTCRpred model A0201_YLQPRTFLL \n",
      "Computing binding predictions for ./test/test.csv\n",
      "Testing DataLoader 0: 100%|████████████████| 1003/1003 [00:05<00:00, 197.52it/s]\n",
      "The results are stored in ./test/out_A0201_YLQPRTFLL.csv.\n"
     ]
    }
   ],
   "source": [
    "!python MixTCRpred.py --model A0201_YLQPRTFLL --input ./test/test.csv --output ./test/out_A0201_YLQPRTFLL.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>cdr3_TRA</th>\n",
       "      <th>cdr3_TRB</th>\n",
       "      <th>TRAV</th>\n",
       "      <th>TRAJ</th>\n",
       "      <th>TRBV</th>\n",
       "      <th>TRBJ</th>\n",
       "      <th>score</th>\n",
       "      <th>perc_rank</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CAANRDDKIIF</td>\n",
       "      <td>CSVEFKSRAGELFF</td>\n",
       "      <td>TRAV12-1</td>\n",
       "      <td>TRAJ53</td>\n",
       "      <td>TRBV11-3</td>\n",
       "      <td>TRBJ2-7</td>\n",
       "      <td>2.84108</td>\n",
       "      <td>0.39592</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CAENLGENQFYF</td>\n",
       "      <td>CATIDTNTGELFF</td>\n",
       "      <td>TRAV13-2</td>\n",
       "      <td>TRAJ47</td>\n",
       "      <td>TRBV4-1</td>\n",
       "      <td>TRBJ1-5</td>\n",
       "      <td>2.46276</td>\n",
       "      <td>0.88953</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CVVSLRDNYGQNFVF</td>\n",
       "      <td>CASSDTDTGELFF</td>\n",
       "      <td>TRAV19</td>\n",
       "      <td>TRAJ42</td>\n",
       "      <td>TRBV7-8</td>\n",
       "      <td>TRBJ2-3</td>\n",
       "      <td>1.71331</td>\n",
       "      <td>4.08557</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CAARGFQKLVF</td>\n",
       "      <td>CSVDRTNEKLFF</td>\n",
       "      <td>TRAV21</td>\n",
       "      <td>TRAJ29</td>\n",
       "      <td>TRBV27</td>\n",
       "      <td>TRBJ2-5</td>\n",
       "      <td>1.14473</td>\n",
       "      <td>11.29627</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>CAVYEDDKIIF</td>\n",
       "      <td>CASSLGTDGNEQFF</td>\n",
       "      <td>TRAV1-2</td>\n",
       "      <td>TRAJ23</td>\n",
       "      <td>TRBV6-5</td>\n",
       "      <td>TRBJ2-2</td>\n",
       "      <td>0.78418</td>\n",
       "      <td>19.61315</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>998</th>\n",
       "      <td>CIVRVGGSSNTGKLIF</td>\n",
       "      <td>CASSQGWGAEGNTIYF</td>\n",
       "      <td>TRAV19</td>\n",
       "      <td>TRAJ45</td>\n",
       "      <td>TRBV19</td>\n",
       "      <td>TRBJ2-7</td>\n",
       "      <td>-3.60105</td>\n",
       "      <td>99.99661</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>999</th>\n",
       "      <td>CAASTPSGGGADGLTF</td>\n",
       "      <td>CASSQGDQHTDTQYF</td>\n",
       "      <td>TRAV6</td>\n",
       "      <td>TRAJ26</td>\n",
       "      <td>TRBV24-1</td>\n",
       "      <td>TRBJ2-1</td>\n",
       "      <td>-3.69587</td>\n",
       "      <td>99.99787</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1000</th>\n",
       "      <td>CAVRIYNAGNNRKLIW</td>\n",
       "      <td>CASSPIDGYGYTF</td>\n",
       "      <td>TRAV17</td>\n",
       "      <td>TRAJ16</td>\n",
       "      <td>TRBV28</td>\n",
       "      <td>TRBJ1-1</td>\n",
       "      <td>-3.75208</td>\n",
       "      <td>99.99839</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1001</th>\n",
       "      <td>CAVGGYGGSQGNLIF</td>\n",
       "      <td>CASSLWRGLSAGNTIYF</td>\n",
       "      <td>TRAV26-2</td>\n",
       "      <td>TRAJ47</td>\n",
       "      <td>TRBV27</td>\n",
       "      <td>TRBJ2-3</td>\n",
       "      <td>-3.77857</td>\n",
       "      <td>99.99859</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1002</th>\n",
       "      <td>CILSLRYRVNRKLTF</td>\n",
       "      <td>CASSYFRRRERSVAAQYF</td>\n",
       "      <td>TRAV26-1</td>\n",
       "      <td>TRAJ45</td>\n",
       "      <td>TRBV5-5</td>\n",
       "      <td>TRBJ2-7</td>\n",
       "      <td>-3.86357</td>\n",
       "      <td>99.99909</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1003 rows × 8 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              cdr3_TRA            cdr3_TRB      TRAV    TRAJ      TRBV  \\\n",
       "0          CAANRDDKIIF      CSVEFKSRAGELFF  TRAV12-1  TRAJ53  TRBV11-3   \n",
       "1         CAENLGENQFYF       CATIDTNTGELFF  TRAV13-2  TRAJ47   TRBV4-1   \n",
       "2      CVVSLRDNYGQNFVF       CASSDTDTGELFF    TRAV19  TRAJ42   TRBV7-8   \n",
       "3          CAARGFQKLVF        CSVDRTNEKLFF    TRAV21  TRAJ29    TRBV27   \n",
       "4          CAVYEDDKIIF      CASSLGTDGNEQFF   TRAV1-2  TRAJ23   TRBV6-5   \n",
       "...                ...                 ...       ...     ...       ...   \n",
       "998   CIVRVGGSSNTGKLIF    CASSQGWGAEGNTIYF    TRAV19  TRAJ45    TRBV19   \n",
       "999   CAASTPSGGGADGLTF     CASSQGDQHTDTQYF     TRAV6  TRAJ26  TRBV24-1   \n",
       "1000  CAVRIYNAGNNRKLIW       CASSPIDGYGYTF    TRAV17  TRAJ16    TRBV28   \n",
       "1001   CAVGGYGGSQGNLIF   CASSLWRGLSAGNTIYF  TRAV26-2  TRAJ47    TRBV27   \n",
       "1002   CILSLRYRVNRKLTF  CASSYFRRRERSVAAQYF  TRAV26-1  TRAJ45   TRBV5-5   \n",
       "\n",
       "         TRBJ    score  perc_rank  \n",
       "0     TRBJ2-7  2.84108    0.39592  \n",
       "1     TRBJ1-5  2.46276    0.88953  \n",
       "2     TRBJ2-3  1.71331    4.08557  \n",
       "3     TRBJ2-5  1.14473   11.29627  \n",
       "4     TRBJ2-2  0.78418   19.61315  \n",
       "...       ...      ...        ...  \n",
       "998   TRBJ2-7 -3.60105   99.99661  \n",
       "999   TRBJ2-1 -3.69587   99.99787  \n",
       "1000  TRBJ1-1 -3.75208   99.99839  \n",
       "1001  TRBJ2-3 -3.77857   99.99859  \n",
       "1002  TRBJ2-7 -3.86357   99.99909  \n",
       "\n",
       "[1003 rows x 8 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_LLW = pd.read_csv(\"./test/out_A0201_YLQPRTFLL.csv\", comment = '#')\n",
    "results_LLW"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If you want to test your own set of TCRs you can upload it to Google Colab by clicking on the \"Upload to session storage\" button in the top left corner, which will upload your file in the '/content/' folder. \n",
    "\n",
    " Next, you can run MixTCRpred specifying the path to your TCRs file (e.g. --input /content/your_file.csv). Refer to the ./test/test.csv file for the correct input format.\n",
    "\n",
    "Please note that Google Colab sessions have a time limit of 12 hours, after which they will automatically shut down. Download your results to avoid losing your work."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![](./upload_TCR_file.png)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}