notebooks/14_full_logP_dataset_analysis.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Full LogP only dataset normalizing summary\n",
    "- The dataset size was 14111\n",
    "- There were 13777 unique smiles\n",
    "- 18 molecules had variation > 1, so they were dropped.\n",
    "- The rest were averaged.\n",
    "- All in all there are 13759 unique records left."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import os\n",
    "import seaborn as sns\n",
    "from pathlib import Path\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "DATA_PATH = \"../data/3_final_data/\""
   ]
  },
  {
   "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>smiles</th>\n",
       "      <th>logP</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>COc1ccccc1[C@@H]1CC(=O)Nc2cc(C)c(C)cc21</td>\n",
       "      <td>4.17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>COc1ccc2c(c1)O[C@](O)(C(F)(F)F)CC2=O</td>\n",
       "      <td>2.79</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CC(C)(C)C1(C)OC(C(N)=O)C(C(N)=O)O1</td>\n",
       "      <td>1.60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CCOc1cc(-c2nnc(-c3ccco3)o2)cc(OCC)c1OCC</td>\n",
       "      <td>3.96</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>C=C(c1ccc(N(C)C)cc1)c1ccc(N(C)C)cc1</td>\n",
       "      <td>5.30</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                    smiles  logP\n",
       "0  COc1ccccc1[C@@H]1CC(=O)Nc2cc(C)c(C)cc21  4.17\n",
       "1     COc1ccc2c(c1)O[C@](O)(C(F)(F)F)CC2=O  2.79\n",
       "2       CC(C)(C)C1(C)OC(C(N)=O)C(C(N)=O)O1  1.60\n",
       "3  CCOc1cc(-c2nnc(-c3ccco3)o2)cc(OCC)c1OCC  3.96\n",
       "4      C=C(c1ccc(N(C)C)cc1)c1ccc(N(C)C)cc1  5.30"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logp_data = pd.read_csv(DATA_PATH + 'logP.csv')\n",
    "logp_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(14111, 2)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logp_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "13777"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logp_data.smiles.nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>logP</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>smiles</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>CCC1(c2ccccc2)C(=O)NC(=O)N1S(=O)(=O)c1ccccc1</th>\n",
       "      <td>11.907200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>COC(=O)c1ccccc1S(=O)(=O)NC(=O)N=c1nc(C)[nH]c(OC)n1</th>\n",
       "      <td>7.761800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>O=C(Nc1ccc([N+](=O)[O-])cc1Cl)c1cc(Cl)ccc1O</th>\n",
       "      <td>5.088050</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>O=C(O)c1ccc[nH]c1=Nc1cccc(C(F)(F)F)c1</th>\n",
       "      <td>4.032800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CN1C(C(=O)N=c2cccc[nH]2)C(=O)c2ccccc2S1(=O)=O</th>\n",
       "      <td>3.934012</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CC(Oc1ccc(Cl)cc1Cl)C(=O)O</th>\n",
       "      <td>3.396400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CC1(C)SC2C(NC(=O)C(N)c3ccc(O)cc3)C(=O)N2C1C(=O)O</th>\n",
       "      <td>3.328200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CC(C)C(N)C(=O)NC(C(=O)O)C(C)C</th>\n",
       "      <td>3.125000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CC(C)CCn1c(=NCCO)[nH]c2c1c(=O)n(C)c(=O)n2C</th>\n",
       "      <td>2.952450</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CCC(C)C(NC(=O)C(N)CC(C)C)C(=O)O</th>\n",
       "      <td>2.691200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CC(C)CC(N)C(=O)NC(C(=O)O)C(C)C</th>\n",
       "      <td>2.668050</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CSCCC(N)C(=O)NC(C(=O)O)C(C)C</th>\n",
       "      <td>2.531250</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CCOC(=O)c1ccccc1S(=O)(=O)NC(=O)N=c1nc(Cl)cc(OC)[nH]1</th>\n",
       "      <td>2.289800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CC(C)CC(NC(=O)C(N)Cc1ccc(O)cc1)C(=O)O</th>\n",
       "      <td>2.142450</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>COc1nc(=NC(=O)NS(=O)(=O)c2ccccc2OCCCl)nc(C)[nH]1</th>\n",
       "      <td>2.121700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Cc1cnc2c(C(=O)O)c(Cl)ccc2c1</th>\n",
       "      <td>1.786050</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CC(CCc1ccccc1)NCC(O)c1ccc(O)c(C(N)=O)c1</th>\n",
       "      <td>1.548800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CC#CC#CC#CC=CC(=O)OC</th>\n",
       "      <td>1.022450</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Clc1cc(Cl)c(-c2cc(Cl)c(Cl)cc2Cl)cc1Cl</th>\n",
       "      <td>0.994050</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>COC(=O)c1sccc1S(=O)(=O)NC(=O)N=c1nc(C)[nH]c(OC)n1</th>\n",
       "      <td>0.924800</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                         logP\n",
       "smiles                                                       \n",
       "CCC1(c2ccccc2)C(=O)NC(=O)N1S(=O)(=O)c1ccccc1        11.907200\n",
       "COC(=O)c1ccccc1S(=O)(=O)NC(=O)N=c1nc(C)[nH]c(OC)n1   7.761800\n",
       "O=C(Nc1ccc([N+](=O)[O-])cc1Cl)c1cc(Cl)ccc1O          5.088050\n",
       "O=C(O)c1ccc[nH]c1=Nc1cccc(C(F)(F)F)c1                4.032800\n",
       "CN1C(C(=O)N=c2cccc[nH]2)C(=O)c2ccccc2S1(=O)=O        3.934012\n",
       "CC(Oc1ccc(Cl)cc1Cl)C(=O)O                            3.396400\n",
       "CC1(C)SC2C(NC(=O)C(N)c3ccc(O)cc3)C(=O)N2C1C(=O)O     3.328200\n",
       "CC(C)C(N)C(=O)NC(C(=O)O)C(C)C                        3.125000\n",
       "CC(C)CCn1c(=NCCO)[nH]c2c1c(=O)n(C)c(=O)n2C           2.952450\n",
       "CCC(C)C(NC(=O)C(N)CC(C)C)C(=O)O                      2.691200\n",
       "CC(C)CC(N)C(=O)NC(C(=O)O)C(C)C                       2.668050\n",
       "CSCCC(N)C(=O)NC(C(=O)O)C(C)C                         2.531250\n",
       "CCOC(=O)c1ccccc1S(=O)(=O)NC(=O)N=c1nc(Cl)cc(OC)...   2.289800\n",
       "CC(C)CC(NC(=O)C(N)Cc1ccc(O)cc1)C(=O)O                2.142450\n",
       "COc1nc(=NC(=O)NS(=O)(=O)c2ccccc2OCCCl)nc(C)[nH]1     2.121700\n",
       "Cc1cnc2c(C(=O)O)c(Cl)ccc2c1                          1.786050\n",
       "CC(CCc1ccccc1)NCC(O)c1ccc(O)c(C(N)=O)c1              1.548800\n",
       "CC#CC#CC#CC=CC(=O)OC                                 1.022450\n",
       "Clc1cc(Cl)c(-c2cc(Cl)c(Cl)cc2Cl)cc1Cl                0.994050\n",
       "COC(=O)c1sccc1S(=O)(=O)NC(=O)N=c1nc(C)[nH]c(OC)n1    0.924800"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grouped_logp_data = logp_data.groupby(['smiles']).agg('var').sort_values(by=['logP'], ascending=False)\n",
    "grouped_logp_data.head(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>logP</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>14111.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>2.127711</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>1.876122</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-4.650000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.890000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>2.030000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>3.260000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>9.960000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               logP\n",
       "count  14111.000000\n",
       "mean       2.127711\n",
       "std        1.876122\n",
       "min       -4.650000\n",
       "25%        0.890000\n",
       "50%        2.030000\n",
       "75%        3.260000\n",
       "max        9.960000"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logp_data.drop_duplicates(keep=False).describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "bad_smiles = ['CCC1(c2ccccc2)C(=O)NC(=O)N1S(=O)(=O)c1ccccc1',\n",
    "'COC(=O)c1ccccc1S(=O)(=O)NC(=O)N=c1nc(C)[nH]c(OC)n1',\n",
    "'O=C(Nc1ccc([N+](=O)[O-])cc1Cl)c1cc(Cl)ccc1O',\n",
    "'O=C(O)c1ccc[nH]c1=Nc1cccc(C(F)(F)F)c1',\n",
    "'CN1C(C(=O)N=c2cccc[nH]2)C(=O)c2ccccc2S1(=O)=O',\n",
    "'CC(Oc1ccc(Cl)cc1Cl)C(=O)O',\n",
    "'CC1(C)SC2C(NC(=O)C(N)c3ccc(O)cc3)C(=O)N2C1C(=O)O',\n",
    "'CC(C)C(N)C(=O)NC(C(=O)O)C(C)C',\n",
    "'CC(C)CCn1c(=NCCO)[nH]c2c1c(=O)n(C)c(=O)n2C',\n",
    "'CCC(C)C(NC(=O)C(N)CC(C)C)C(=O)O',\n",
    "'CC(C)CC(N)C(=O)NC(C(=O)O)C(C)C',\n",
    "'CSCCC(N)C(=O)NC(C(=O)O)C(C)C',\n",
    "'CCOC(=O)c1ccccc1S(=O)(=O)NC(=O)N=c1nc(Cl)cc(OC)[nH]1',\n",
    "'CC(C)CC(NC(=O)C(N)Cc1ccc(O)cc1)C(=O)O',\n",
    "'COc1nc(=NC(=O)NS(=O)(=O)c2ccccc2OCCCl)nc(C)[nH]1',\n",
    "'Cc1cnc2c(C(=O)O)c(Cl)ccc2c1',\n",
    "'CC(CCc1ccccc1)NCC(O)c1ccc(O)c(C(N)=O)c1',\n",
    "'CC#CC#CC#CC=CC(=O)OC']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(14073, 2)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logp_verified_data = logp_data[~logp_data.smiles.isin(bad_smiles)]\n",
    "logp_verified_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>smiles</th>\n",
       "      <th>logP</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>BrC(Br)(Br)Br</td>\n",
       "      <td>3.42</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>BrC1C2CC3C(C2)C13</td>\n",
       "      <td>3.11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>BrC1CC2CCC1C2</td>\n",
       "      <td>3.54</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>BrC1CCCCC1</td>\n",
       "      <td>3.20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>BrC=C(Br)Br</td>\n",
       "      <td>3.20</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              smiles  logP\n",
       "0      BrC(Br)(Br)Br  3.42\n",
       "1  BrC1C2CC3C(C2)C13  3.11\n",
       "2      BrC1CC2CCC1C2  3.54\n",
       "3         BrC1CCCCC1  3.20\n",
       "4        BrC=C(Br)Br  3.20"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logp_verified_data = logp_verified_data.groupby(['smiles']).mean().reset_index()\n",
    "logp_verified_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(13759, 2)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logp_verified_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "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>logP</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>13759.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>2.126042</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>1.866106</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>-4.650000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.890000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>2.030000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>3.250000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>9.960000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               logP\n",
       "count  13759.000000\n",
       "mean       2.126042\n",
       "std        1.866106\n",
       "min       -4.650000\n",
       "25%        0.890000\n",
       "50%        2.030000\n",
       "75%        3.250000\n",
       "max        9.960000"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logp_verified_data.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x206639b3fd0>"
      ]
     },
     "execution_count": 12,
     "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": [
    "sns.distplot(logp_verified_data.logP)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "logp_verified_data.to_csv(\"../data/3_final_data/logp_mean.csv\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "chemvae",
   "language": "python",
   "name": "chemvae"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}