01_halicin_train.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Build a ChemProp model to predict antibiotic activity.  This notebook, and subsequent notebooks, are my attempt to reproduce the work in a recent paper from John Stokes and coworkers at the Broad Institute and MIT, [\"A Deep Learning Approach to Antibiotic Discovery\"](https://www.sciencedirect.com/science/article/pii/S0092867420301021)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Setup the appropriate imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "RDKit WARNING: [17:14:08] Enabling RDKit 2019.09.3 jupyter extensions\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "from rdkit import Chem\n",
    "from tqdm.notebook import tqdm"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Read the spreadsheet witht the data from Stokes et al"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_excel(\"mmc1.xlsx\",sheet_name=1,skiprows=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Look at the first few lines from the dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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>Mean_Inhibition</th>\n",
       "      <th>SMILES</th>\n",
       "      <th>Name</th>\n",
       "      <th>Activity</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.041572</td>\n",
       "      <td>Cc1cc(O)c(C(=O)NC(C(=O)NC2C(=O)N3C(C(=O)O)=C(C...</td>\n",
       "      <td>CEFPIRAMIDE</td>\n",
       "      <td>Active</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.041876</td>\n",
       "      <td>CON=C1CN(c2nc3c(cc2F)c(=O)c(C(=O)O)cn3C2CC2)CC...</td>\n",
       "      <td>GEMIFLOXACIN MESYLATE</td>\n",
       "      <td>Active</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.041916</td>\n",
       "      <td>CCC(C)CCCCC(=O)NC(CCN)C(=O)NC(C(=O)NC(CCN)C(=O...</td>\n",
       "      <td>POLYMYXIN B SULFATE</td>\n",
       "      <td>Active</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.041964</td>\n",
       "      <td>Cl.N=C(N)n1cccn1</td>\n",
       "      <td>PRAXADINE HYDROCHLORIDE</td>\n",
       "      <td>Active</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.042295</td>\n",
       "      <td>Cl.Cl.N=C(NCCCCCCNC(=N)NC(=N)Nc1ccc(Cl)cc1)NC(...</td>\n",
       "      <td>CHLORHEXIDINE DIHYDROCHLORIDE</td>\n",
       "      <td>Active</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Mean_Inhibition                                             SMILES  \\\n",
       "0         0.041572  Cc1cc(O)c(C(=O)NC(C(=O)NC2C(=O)N3C(C(=O)O)=C(C...   \n",
       "1         0.041876  CON=C1CN(c2nc3c(cc2F)c(=O)c(C(=O)O)cn3C2CC2)CC...   \n",
       "2         0.041916  CCC(C)CCCCC(=O)NC(CCN)C(=O)NC(C(=O)NC(CCN)C(=O...   \n",
       "3         0.041964                                   Cl.N=C(N)n1cccn1   \n",
       "4         0.042295  Cl.Cl.N=C(NCCCCCCNC(=N)NC(=N)Nc1ccc(Cl)cc1)NC(...   \n",
       "\n",
       "                            Name Activity  \n",
       "0                    CEFPIRAMIDE   Active  \n",
       "1          GEMIFLOXACIN MESYLATE   Active  \n",
       "2            POLYMYXIN B SULFATE   Active  \n",
       "3        PRAXADINE HYDROCHLORIDE   Active  \n",
       "4  CHLORHEXIDINE DIHYDROCHLORIDE   Active  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Make a barplot of the activity distribution, notice that the dataset is imbalanced."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "ax = df.Activity.value_counts().plot(kind=\"bar\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Create a new field which represents the activity as an integer (Active = 1, Inactive = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['active'] = [int(x == \"Active\") for x in df.Activity]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Mean_Inhibition</th>\n",
       "      <th>SMILES</th>\n",
       "      <th>Name</th>\n",
       "      <th>Activity</th>\n",
       "      <th>active</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.041572</td>\n",
       "      <td>Cc1cc(O)c(C(=O)NC(C(=O)NC2C(=O)N3C(C(=O)O)=C(C...</td>\n",
       "      <td>CEFPIRAMIDE</td>\n",
       "      <td>Active</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.041876</td>\n",
       "      <td>CON=C1CN(c2nc3c(cc2F)c(=O)c(C(=O)O)cn3C2CC2)CC...</td>\n",
       "      <td>GEMIFLOXACIN MESYLATE</td>\n",
       "      <td>Active</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.041916</td>\n",
       "      <td>CCC(C)CCCCC(=O)NC(CCN)C(=O)NC(C(=O)NC(CCN)C(=O...</td>\n",
       "      <td>POLYMYXIN B SULFATE</td>\n",
       "      <td>Active</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.041964</td>\n",
       "      <td>Cl.N=C(N)n1cccn1</td>\n",
       "      <td>PRAXADINE HYDROCHLORIDE</td>\n",
       "      <td>Active</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.042295</td>\n",
       "      <td>Cl.Cl.N=C(NCCCCCCNC(=N)NC(=N)Nc1ccc(Cl)cc1)NC(...</td>\n",
       "      <td>CHLORHEXIDINE DIHYDROCHLORIDE</td>\n",
       "      <td>Active</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Mean_Inhibition                                             SMILES  \\\n",
       "0         0.041572  Cc1cc(O)c(C(=O)NC(C(=O)NC2C(=O)N3C(C(=O)O)=C(C...   \n",
       "1         0.041876  CON=C1CN(c2nc3c(cc2F)c(=O)c(C(=O)O)cn3C2CC2)CC...   \n",
       "2         0.041916  CCC(C)CCCCC(=O)NC(CCN)C(=O)NC(C(=O)NC(CCN)C(=O...   \n",
       "3         0.041964                                   Cl.N=C(N)n1cccn1   \n",
       "4         0.042295  Cl.Cl.N=C(NCCCCCCNC(=N)NC(=N)Nc1ccc(Cl)cc1)NC(...   \n",
       "\n",
       "                            Name Activity  active  \n",
       "0                    CEFPIRAMIDE   Active       1  \n",
       "1          GEMIFLOXACIN MESYLATE   Active       1  \n",
       "2            POLYMYXIN B SULFATE   Active       1  \n",
       "3        PRAXADINE HYDROCHLORIDE   Active       1  \n",
       "4  CHLORHEXIDINE DIHYDROCHLORIDE   Active       1  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's double check to ensure that we have the correct number of active and inactive compounds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(Inactive    2215\n",
       " Active       120\n",
       " Name: Activity, dtype: int64,\n",
       " 0    2215\n",
       " 1     120\n",
       " Name: active, dtype: int64)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.Activity.value_counts(),df.active.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Write SMILES and active fields out in a csv file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "df[[\"SMILES\",\"active\"]].to_csv(\"mmcl.csv\",index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Uncomment the cell below to run the training.  I think I've reproduced the hyperparameters from the Cell paper, but I may have missed something or gotten something wrong"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "#%python ~/software/chemprop/train.py --data_path mmcl.csv --dataset_type classification --save_dir chemprop_train --gpu 0 --features_generator rdkit_2d_normalized --no_features_scaling --depth 5 --hidden_size 1600 --ffn_num_layers 1 --dropout 0.35 --ensemble_size 20"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Environment (conda_chemprop)",
   "language": "python",
   "name": "conda_chemprop"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}