Plots_Generated_Vs_Experimental.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dd1b4906-18b5-4788-acab-be88885e028a",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from tqdm import tqdm, trange\n",
    "import pickle\n",
    "from rdkit import Chem, DataStructs\n",
    "import seaborn as sns\n",
    "import pandas as pd\n",
    "import random\n",
    "import re\n",
    "\n",
    "\n",
    "# For plotting\n",
    "import plotly.io as plt_io\n",
    "import plotly.graph_objects as go\n",
    "\n",
    "# Dimensionality reduction and clustering\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.manifold import TSNE\n",
    "import umap\n",
    "from sklearn.discriminant_analysis import LinearDiscriminantAnalysis as LDA\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.cluster import KMeans\n",
    "from sklearn.metrics import silhouette_samples, silhouette_score\n",
    "\n",
    "import umap\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.cm as cm\n",
    "%matplotlib inline\n",
    "\n",
    "from rdkit import Chem\n",
    "from rdkit.Chem import Draw, rdmolops\n",
    "from rdkit.Chem.Draw import IPythonConsole\n",
    "from rdkit.Chem import AllChem\n",
    "from rdkit import RDLogger\n",
    "RDLogger.DisableLog('rdApp.*')  # switch off RDKit warning messages\n",
    "\n",
    "from rdkit import Chem\n",
    "from rdkit.Chem import Descriptors\n",
    "from rdkit.Chem import Lipinski\n",
    "from rdkit.Chem.AllChem import CalcNumAtomStereoCenters\n",
    "from rdkit.Chem.GraphDescriptors import BertzCT\n",
    "from rdkit.Chem.Scaffolds.MurckoScaffold import MurckoScaffoldSmiles\n",
    "from rdkit.Chem import AllChem, rdmolops\n",
    "from rdkit.DataStructs import FingerprintSimilarity\n",
    "from rdkit.Chem import MACCSkeys\n",
    "\n",
    "# suppress Chem.MolFromSmiles error output\n",
    "from rdkit import rdBase\n",
    "import sys\n",
    "import os\n",
    "rdBase.DisableLog('rdApp.error')\n",
    "# import from rdkit.Contrib module\n",
    "from rdkit.Chem import RDConfig\n",
    "sys.path.append(os.path.join(RDConfig.RDContribDir, 'SA_Score'))\n",
    "import sascorer\n",
    "sys.path.append(os.path.join(RDConfig.RDContribDir, 'NP_Score'))\n",
    "import npscorer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "66e9a77e-5d2d-460a-ada7-e7ca1f626ae3",
   "metadata": {},
   "outputs": [],
   "source": [
    "exp_smiles = pd.read_csv('./data/Experimental_Ligands.csv')\n",
    "print('Dataset:', exp_smiles.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "71b29c61-dd38-46d7-b223-9e7f3b7d1885",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_generated = pd.read_csv('./results/generative_model/Generated_Practical_Set.csv')\n",
    "df_generated = df_generated[['smiles']]\n",
    "df_generated"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "39e82769-835a-47de-9cd2-5d05de3cfb66",
   "metadata": {},
   "outputs": [],
   "source": [
    "exp_smiles_list = list(exp_smiles['smiles'])\n",
    "print('Number of real molecules :', len(exp_smiles_list))\n",
    "df_generated_list = list(df_generated['smiles'])\n",
    "print('Number of generated molecules :', len(df_generated_list))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f5e9321e-86c9-49aa-85a2-0e2857073a4d",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "# Tanimoto Similarity\n",
    "mol_real = [Chem.MolFromSmiles(x) for x in exp_smiles_list]\n",
    "#fp_real = [AllChem.GetMorganFingerprintAsBitVect(y, 2) for y in mol_real]\n",
    "fp_real = [Chem.RDKFingerprint(y) for y in mol_real]\n",
    "fp_real = [MACCSkeys.GenMACCSKeys(y) for y in mol_real]\n",
    "\n",
    "mol_gen = [Chem.MolFromSmiles(x) for x in df_generated_list if x is not None]\n",
    "#fp_gen = [AllChem.GetMorganFingerprintAsBitVect(y,2) for y in mol_gen]\n",
    "fp_gen = [Chem.RDKFingerprint(y) for y in mol_gen]\n",
    "fp_gen = [MACCSkeys.GenMACCSKeys(y) for y in mol_gen]\n",
    "\n",
    "similarities_ij= []\n",
    "\n",
    "for i in fp_real:\n",
    "    for j in fp_gen:\n",
    "        similarities_ij.append(DataStructs.FingerprintSimilarity(i, j))\n",
    "\n",
    "similarities_ii= []        \n",
    "for i in fp_real:\n",
    "    for j in fp_real:\n",
    "        similarities_ii.append(DataStructs.FingerprintSimilarity(i, j))\n",
    "        \n",
    "similarities_jj= []        \n",
    "for i in fp_gen:\n",
    "    for j in fp_gen:\n",
    "        similarities_jj.append(DataStructs.FingerprintSimilarity(i, j))\n",
    "\n",
    "similarities_ij = np.asarray(similarities_ij, dtype=np.float32)\n",
    "similarities_ij = np.sort(similarities_ij)\n",
    "\n",
    "similarities_ii = np.asarray(similarities_ii, dtype=np.float32)\n",
    "similarities_ii = np.sort(similarities_ii)\n",
    "\n",
    "similarities_jj = np.asarray(similarities_jj, dtype=np.float32)\n",
    "similarities_jj = np.sort(similarities_jj)\n",
    "\n",
    "# Plotting with shading\n",
    "plt.figure(figsize=(6, 4))\n",
    "\n",
    "\n",
    "p = sns.kdeplot(similarities_jj,  fill=True, #color='green',\n",
    "                label='Generated = ' + str(np.round(np.mean(similarities_jj), 2)), linewidth=3,\n",
    "                shade=True, alpha=.3)\n",
    "p = sns.kdeplot(similarities_ii,  fill=True, #color='red',\n",
    "                label='Experimental = ' + str(np.round(np.mean(similarities_ii), 2)), linewidth=3,\n",
    "                shade=True, alpha=.3)\n",
    "p = sns.kdeplot(similarities_ij,  fill=True, #color='blue',\n",
    "                label='Crossed = ' + str(np.round(np.mean(similarities_ij), 2)), linewidth=3,\n",
    "                shade=True, alpha=.3)\n",
    "\n",
    "p.legend(bbox_to_anchor=(1.00, 1.00), loc='best', fontsize=12)\n",
    "#p.set(xlabel='Tanimoto coefficient', title='Distribution of Tanimoto coefficient with mean value')\n",
    "#plt.tight_layout()\n",
    "p.set_xlabel('Tanimoto coefficient', fontsize=14)  # Adjust the font size as needed\n",
    "p.set_ylabel('Density', fontsize=14)  # Adjust the font size as needed\n",
    "plt.tight_layout()\n",
    "plt.savefig('./results/TANIMOTO_experimental_vs_generated.png',dpi = 600)\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d38b452f-9140-47cb-b0de-7668c84f2902",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "# Tanimoto Similarity\n",
    "mol_real = [Chem.MolFromSmiles(x) for x in exp_smiles_list]\n",
    "#fp_real = [AllChem.GetMorganFingerprintAsBitVect(y, 2) for y in mol_real]\n",
    "fp_real = [Chem.RDKFingerprint(y) for y in mol_real]\n",
    "fp_real = [MACCSkeys.GenMACCSKeys(y) for y in mol_real]\n",
    "\n",
    "mol_gen = [Chem.MolFromSmiles(x) for x in df_generated_list if x is not None]\n",
    "#fp_gen = [AllChem.GetMorganFingerprintAsBitVect(y,2) for y in mol_gen]\n",
    "fp_gen = [Chem.RDKFingerprint(y) for y in mol_gen]\n",
    "fp_gen = [MACCSkeys.GenMACCSKeys(y) for y in mol_gen]\n",
    "\n",
    "similarities_ij= []\n",
    "\n",
    "for i in fp_real:\n",
    "    for j in fp_gen:\n",
    "        similarities_ij.append(DataStructs.FingerprintSimilarity(i, j))\n",
    "\n",
    "similarities_ii= []        \n",
    "for i in fp_real:\n",
    "    for j in fp_real:\n",
    "        similarities_ii.append(DataStructs.FingerprintSimilarity(i, j))\n",
    "        \n",
    "similarities_jj= []        \n",
    "for i in fp_gen:\n",
    "    for j in fp_gen:\n",
    "        similarities_jj.append(DataStructs.FingerprintSimilarity(i, j))\n",
    "        \n",
    "\n",
    "\n",
    "similarities_ij = np.asarray(similarities_ij, dtype=np.float32)\n",
    "similarities_ij = np.sort(similarities_ij)\n",
    "\n",
    "similarities_ii = np.asarray(similarities_ii, dtype=np.float32)\n",
    "similarities_ii = np.sort(similarities_ii)\n",
    "\n",
    "similarities_jj = np.asarray(similarities_jj, dtype=np.float32)\n",
    "similarities_jj = np.sort(similarities_jj)\n",
    "\n",
    "# Plotting with shading\n",
    "plt.figure(figsize=(6, 4))\n",
    "\n",
    "\n",
    "p = sns.kdeplot(similarities_ij,  fill=True, color='green',\n",
    "                label='Mean Tanimoto index = ' + str(np.round(np.mean(similarities_ij), 2)), linewidth=3,\n",
    "                shade=True, alpha=.3)\n",
    "\n",
    "p.legend(bbox_to_anchor=(1.00, 1.00), loc='best', fontsize=12)\n",
    "#p.set(xlabel='Tanimoto coefficient', title='Distribution of Tanimoto coefficient with mean value')\n",
    "#plt.tight_layout()\n",
    "p.set_xlabel('Tanimoto coefficient', fontsize=14)  # Adjust the font size as needed\n",
    "p.set_ylabel('Density', fontsize=14)  # Adjust the font size as needed\n",
    "plt.tight_layout()\n",
    "plt.savefig('./results/TANIMOTO_experimental_vs_generated.png',dpi = 600)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c0e29124-3d99-45b3-a125-76e1ba8c261f",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "# Diversity \n",
    "mol_real = [Chem.MolFromSmiles(x) for x in exp_smiles_list]\n",
    "fp_real = [AllChem.GetMorganFingerprintAsBitVect(y, 2) for y in mol_real]\n",
    "#fp_real = [Chem.RDKFingerprint(y) for y in mol_real]\n",
    "\n",
    "mol_gen = [Chem.MolFromSmiles(x) for x in df_generated_list if x is not None]\n",
    "fp_gen = [AllChem.GetMorganFingerprintAsBitVect(y,2) for y in mol_gen]\n",
    "#fp_gen = [Chem.RDKFingerprint(y) for y in mol_gen]\n",
    "\n",
    "similarities_ij= []\n",
    "\n",
    "for i in fp_real:\n",
    "    for j in fp_gen:\n",
    "        similarities_ij.append(1-DataStructs.FingerprintSimilarity(i, j))\n",
    "\n",
    "similarities_ii= []        \n",
    "for i in fp_real:\n",
    "    for j in fp_real:\n",
    "        similarities_ii.append(1-DataStructs.FingerprintSimilarity(i, j))\n",
    "        \n",
    "similarities_jj= []        \n",
    "for i in fp_gen:\n",
    "    for j in fp_gen:\n",
    "        similarities_jj.append(1-DataStructs.FingerprintSimilarity(i, j))\n",
    "\n",
    "similarities_ij = np.asarray(similarities_ij, dtype=np.float32)\n",
    "similarities_ij = np.sort(similarities_ij)\n",
    "\n",
    "similarities_ii = np.asarray(similarities_ii, dtype=np.float32)\n",
    "similarities_ii = np.sort(similarities_ii)\n",
    "\n",
    "similarities_jj = np.asarray(similarities_jj, dtype=np.float32)\n",
    "similarities_jj = np.sort(similarities_jj)\n",
    "\n",
    "\n",
    "# Plotting with shading\n",
    "plt.figure(figsize=(6, 4))\n",
    "\n",
    "\n",
    "p = sns.kdeplot(similarities_jj,  fill=True, #color='green',\n",
    "                label='Internal diversity (Gen.) = ' + str(np.round(np.sum(similarities_jj) / (len(fp_gen)**2), 2)), linewidth=3,\n",
    "                shade=True, alpha=.3)\n",
    "p = sns.kdeplot(similarities_ii,  fill=True, #color='red',\n",
    "                label='Internal diversity (Exp.) = ' + str(np.round(np.sum(similarities_ii) / (len(fp_real)**2), 2)), linewidth=3,\n",
    "                shade=True, alpha=.3)\n",
    "p = sns.kdeplot(similarities_ij,  fill=True, #color='blue',\n",
    "                label='External diversity = ' + str(np.round(np.sum(similarities_ij) / (len(fp_real) * len(fp_gen)), 2)), linewidth=3,\n",
    "                shade=True, alpha=.3)\n",
    "\n",
    "p.legend(bbox_to_anchor=(0.00, 1.00), loc='upper left', fontsize=12)\n",
    "#p.set(xlabel='Tanimoto coefficient', title='Distribution of Tanimoto coefficient with mean value')\n",
    "#plt.tight_layout()\n",
    "p.set_xlabel('Diversity', fontsize=14)  # Adjust the font size as needed\n",
    "p.set_ylabel('Density', fontsize=14)  # Adjust the font size as needed\n",
    "plt.tight_layout()\n",
    "plt.savefig('./results/Diversity_experimental_vs_generated.png',dpi = 600)\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "db0c0deb-ef10-4416-8114-301e88ddc7ba",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "# Tanimoto Similarity\n",
    "mol_real = [Chem.MolFromSmiles(x) for x in exp_smiles_list]\n",
    "#fp_real = [AllChem.GetMorganFingerprintAsBitVect(y, 2) for y in mol_real]\n",
    "fp_real = [Chem.RDKFingerprint(y) for y in mol_real]\n",
    "fp_real = [MACCSkeys.GenMACCSKeys(y) for y in mol_real]\n",
    "\n",
    "mol_gen = [Chem.MolFromSmiles(x) for x in df_generated_list if x is not None]\n",
    "#fp_gen = [AllChem.GetMorganFingerprintAsBitVect(y,2) for y in mol_gen]\n",
    "fp_gen = [Chem.RDKFingerprint(y) for y in mol_gen]\n",
    "fp_gen = [MACCSkeys.GenMACCSKeys(y) for y in mol_gen]\n",
    "\n",
    "similarities_ij= []\n",
    "\n",
    "for i in fp_real:\n",
    "    for j in fp_gen:\n",
    "        similarities_ij.append(1-DataStructs.FingerprintSimilarity(i, j))\n",
    "\n",
    "similarities_ii= []        \n",
    "for i in fp_real:\n",
    "    for j in fp_real:\n",
    "        similarities_ii.append(1-DataStructs.FingerprintSimilarity(i, j))\n",
    "        \n",
    "similarities_jj= []        \n",
    "for i in fp_gen:\n",
    "    for j in fp_gen:\n",
    "        similarities_jj.append(1-DataStructs.FingerprintSimilarity(i, j))\n",
    "\n",
    "\n",
    "        \n",
    "similarities_ij = np.asarray(similarities_ij, dtype=np.float32)\n",
    "similarities_ij = np.sort(similarities_ij)\n",
    "\n",
    "similarities_ii = np.asarray(similarities_ii, dtype=np.float32)\n",
    "similarities_ii = np.sort(similarities_ii)\n",
    "\n",
    "similarities_jj = np.asarray(similarities_jj, dtype=np.float32)\n",
    "similarities_jj = np.sort(similarities_jj)\n",
    "\n",
    "\n",
    "# Plotting with shading\n",
    "plt.figure(figsize=(6, 4))\n",
    "\n",
    "\n",
    "p = sns.kdeplot(similarities_jj,  fill=True, color='green',\n",
    "                label='Mean diversity = ' + str(np.round(np.sum(similarities_jj) / (len(fp_gen)**2), 2)), linewidth=3,\n",
    "                shade=True, alpha=.3)\n",
    "\n",
    "p.legend(bbox_to_anchor=(0.00, 1.00), loc='upper left', fontsize=12)\n",
    "#p.set(xlabel='Tanimoto coefficient', title='Distribution of Tanimoto coefficient with mean value')\n",
    "#plt.tight_layout()\n",
    "p.set_xlabel('Diversity', fontsize=14)  # Adjust the font size as needed\n",
    "p.set_ylabel('Density', fontsize=14)  # Adjust the font size as needed\n",
    "plt.tight_layout()\n",
    "plt.savefig('./results/Diversity_experimental_vs_generated.png',dpi = 600)\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7f78c71d-c084-4512-a0a0-42f3fa42fa6b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# LogP\n",
    "\n",
    "gen_logp = [Descriptors.MolLogP(mol) for mol in mol_gen]\n",
    "real_logp = [Descriptors.MolLogP(mol) for mol in mol_real]\n",
    "\n",
    "# Plotting with shading\n",
    "plt.figure(figsize=(6, 4))\n",
    "p = sns.kdeplot(gen_logp,  fill=True, #color='green',\n",
    "                label='Generated', linewidth=3,\n",
    "                shade=True, alpha=.5)\n",
    "p = sns.kdeplot(real_logp, fill=True, #color='red'\n",
    "                label='Experimental',  linewidth=3,\n",
    "                shade=True, alpha=.5)\n",
    "\n",
    "#p.legend(bbox_to_anchor=(1.00, 1.00), loc='best', fontsize=12)\n",
    "#p.set(xlabel='LogP', title='Distribution of LogP')\n",
    "p.set_xlabel('LogP', fontsize=14)  # Adjust the font size as needed\n",
    "p.set_ylabel('Density', fontsize=14)  # Adjust the font size as needed\n",
    "plt.tight_layout()\n",
    "plt.savefig('./results/LOGP_experimental_vs_generated.png',dpi = 600)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "04a03bcb-ef80-4adf-8108-8feaa8ca7eb1",
   "metadata": {},
   "outputs": [],
   "source": [
    "# QED\n",
    "\n",
    "gen_qed = [Descriptors.qed(mol) for mol in mol_gen]\n",
    "real_qed = [Descriptors.qed(mol) for mol in mol_real]\n",
    "\n",
    "# Plotting with shading\n",
    "plt.figure(figsize=(6, 4))\n",
    "\n",
    "p = sns.kdeplot(gen_qed,  fill=True, #color='green',\n",
    "                label='Generated',  linewidth=3,\n",
    "                shade=True, alpha=.5)\n",
    "p = sns.kdeplot(real_qed,  fill=True, #color='red',\n",
    "                label='Experimental', linewidth=3,\n",
    "                shade=True, alpha=.5)\n",
    "\n",
    "#p.legend(bbox_to_anchor=(1.00, 1.00), loc='best')\n",
    "#p.set(xlabel='QED', title='Distribution of QED')\n",
    "p.set_xlabel('QED', fontsize=14)  # Adjust the font size as needed\n",
    "p.set_ylabel('Density', fontsize=14)  # Adjust the font size as needed\n",
    "plt.tight_layout()\n",
    "plt.savefig('./results/QED_experimental_vs_generated.png',dpi = 600)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "db8d6f6e-4c50-4e14-ad6b-a295141cdbad",
   "metadata": {},
   "outputs": [],
   "source": [
    "# np\n",
    "\n",
    "fscore = npscorer.readNPModel()\n",
    "\n",
    "gen_npscore = [npscorer.scoreMol(mol, fscore) for mol in mol_gen]\n",
    "real_npscore = [npscorer.scoreMol(mol, fscore) for mol in mol_real]\n",
    "\n",
    "# Plotting with shading\n",
    "plt.figure(figsize=(6, 4))\n",
    "\n",
    "p = sns.kdeplot(gen_npscore, fill=True, # color='green',\n",
    "                label='Generated', linewidth=3,\n",
    "                shade=True, alpha=.5)\n",
    "\n",
    "p = sns.kdeplot(real_npscore,  fill=True, #color='red',\n",
    "                label='Experimental', linewidth=3,\n",
    "                shade=True, alpha=.5)\n",
    "\n",
    "#p.legend(bbox_to_anchor=(1.00, 1.00), loc='best', fontsize=12)\n",
    "p.set(xlabel='NP−likeness score') #, title='Distribution of NP−likeness'\n",
    "p.set_xlabel('NP−likeness score', fontsize=14)  # Adjust the font size as needed\n",
    "p.set_ylabel('Density', fontsize=14)  # Adjust the font size as needed\n",
    "plt.tight_layout()\n",
    "plt.savefig('./results/NP_experimental_vs_generated.png',dpi = 600)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "26ab1d97-73e6-46dd-af1b-cca95d512f23",
   "metadata": {},
   "outputs": [],
   "source": [
    "# PSA\n",
    "from rdkit import Chem\n",
    "from rdkit.Chem import rdMolDescriptors\n",
    "import rdkit.Chem as Chem \n",
    "from rdkit.Chem import Descriptors3D\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from rdkit.Chem.Draw import SimilarityMaps\n",
    "import rdkit.Chem as Chem\n",
    "from rdkit.Chem import MolSurf\n",
    "\n",
    "gen_psa = [MolSurf.pyLabuteASA(mol) for mol in mol_gen]\n",
    "real_psa = [MolSurf.pyLabuteASA(mol) for mol in mol_real]\n",
    "\n",
    "#MolSurf.pyLabuteASA\n",
    "#MolSurf.TPSA\n",
    "\n",
    "# Plotting with shading\n",
    "plt.figure(figsize=(6, 4))\n",
    "p = sns.kdeplot(gen_psa,  fill=True, #color='green',\n",
    "                label='Generated', linewidth=3,\n",
    "                shade=True, alpha=.5)\n",
    "p = sns.kdeplot(real_psa, fill=True, #color='red'\n",
    "                label='Experimental',  linewidth=3,\n",
    "                shade=True, alpha=.5)\n",
    "\n",
    "#p.legend(bbox_to_anchor=(1.00, 1.00), loc='best', fontsize=12)\n",
    "#p.set(xlabel='LogP', title='Distribution of LogP')\n",
    "p.set_xlabel('PSA', fontsize=14)  # Adjust the font size as needed\n",
    "p.set_ylabel('Density', fontsize=14)  # Adjust the font size as needed\n",
    "plt.tight_layout()\n",
    "plt.savefig('./results/PSA_experimental_vs_generated.png',dpi = 600)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cb476fce-35e0-4d24-b2b0-f9850ca6f25f",
   "metadata": {},
   "outputs": [],
   "source": [
    "#HBA\n",
    "gen_HBA = [rdMolDescriptors.CalcNumLipinskiHBA(mol) for mol in mol_gen]\n",
    "real_HBA = [rdMolDescriptors.CalcNumLipinskiHBA(mol) for mol in mol_real]\n",
    "\n",
    "# Plotting with shading\n",
    "plt.figure(figsize=(6, 4))\n",
    "p = sns.kdeplot(gen_HBA,  fill=True, #color='green',\n",
    "                label='Generated', linewidth=3,\n",
    "                shade=True, alpha=.5)\n",
    "p = sns.kdeplot(real_HBA, fill=True, #color='red'\n",
    "                label='Experimental',  linewidth=3,\n",
    "                shade=True, alpha=.5)\n",
    "\n",
    "#p.legend(bbox_to_anchor=(1.00, 1.00), loc='best', fontsize=12)\n",
    "#p.set(xlabel='LogP', title='Distribution of LogP')\n",
    "p.set_xlabel('HBA', fontsize=14)  # Adjust the font size as needed\n",
    "p.set_ylabel('Density', fontsize=14)  # Adjust the font size as needed\n",
    "plt.tight_layout()\n",
    "plt.savefig('./results/HBA_experimental_vs_generated.png',dpi = 600)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6adf7348-4bf0-4079-8e87-daab6a95e2c5",
   "metadata": {},
   "outputs": [],
   "source": [
    "#HBD\n",
    "gen_HBD = [rdMolDescriptors.CalcNumLipinskiHBD(mol) for mol in mol_gen]\n",
    "real_HBD = [rdMolDescriptors.CalcNumLipinskiHBD(mol) for mol in mol_real]\n",
    "\n",
    "# Plotting with shading\n",
    "plt.figure(figsize=(6, 4))\n",
    "p = sns.kdeplot(gen_HBD,  fill=True, #color='green',\n",
    "                label='Generated', linewidth=3,\n",
    "                shade=True, alpha=.5)\n",
    "p = sns.kdeplot(real_HBD, fill=True, #color='red'\n",
    "                label='Experimental',  linewidth=3,\n",
    "                shade=True, alpha=.5)\n",
    "\n",
    "p.legend(bbox_to_anchor=(1.00, 1.00), loc='best', fontsize=12)\n",
    "#p.set(xlabel='LogP', title='Distribution of LogP')\n",
    "p.set_xlabel('HBD', fontsize=14)  # Adjust the font size as needed\n",
    "p.set_ylabel('Density', fontsize=14)  # Adjust the font size as needed\n",
    "plt.tight_layout()\n",
    "plt.savefig('./results/HBD_experimental_vs_generated.png',dpi = 600)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dd30e92b-7482-4388-b21f-8e6df5b96443",
   "metadata": {},
   "outputs": [],
   "source": [
    "def calculate_molecular_volume(mol):\n",
    "    \n",
    "    # Generate 3D coordinates\n",
    "    AllChem.EmbedMolecule(mol)\n",
    "    \n",
    "    # Calculate volume using GetConformer and GetAtomPosition\n",
    "    conformer = mol.GetConformer()\n",
    "    volume = AllChem.ComputeMolVolume(mol, confId=conformer.GetId())\n",
    "    \n",
    "    return volume\n",
    "\n",
    "gen_vol = [calculate_molecular_volume(mol) for mol in mol_gen]\n",
    "real_vol = [calculate_molecular_volume(mol) for mol in mol_real]\n",
    "\n",
    "# Plotting with shading\n",
    "plt.figure(figsize=(6, 4))\n",
    "p = sns.kdeplot(gen_vol,  fill=True, #color='green',\n",
    "                label='Generated', linewidth=3,\n",
    "                shade=True, alpha=.5)\n",
    "p = sns.kdeplot(real_vol, fill=True, #color='red'\n",
    "                label='Experimental',  linewidth=3,\n",
    "                shade=True, alpha=.5)\n",
    "\n",
    "#p.legend(bbox_to_anchor=(1.00, 1.00), loc='best', fontsize=12)\n",
    "#p.set(xlabel='LogP', title='Distribution of LogP')\n",
    "p.set_xlabel('Molecular volume', fontsize=14)  # Adjust the font size as needed\n",
    "p.set_ylabel('Density', fontsize=14)  # Adjust the font size as needed\n",
    "plt.tight_layout()\n",
    "plt.savefig('./results/vol_experimental_vs_generated.png',dpi = 600)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b036983c-3816-4234-bfcd-70014057849a",
   "metadata": {
    "tags": []
   },
   "source": [
    "# tMAP plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "623c423c-9769-4339-9e31-5c0f885176ca",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import tmap\n",
    "from faerun import Faerun\n",
    "from mhfp.encoder import MHFPEncoder\n",
    "from rdkit.Chem import AllChem"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e77caa8e-ba47-42a2-9e72-912cca32815c",
   "metadata": {},
   "outputs": [],
   "source": [
    "exp_smiles = pd.read_csv('./data/Experimental_Ligands.csv')\n",
    "print('Dataset:', exp_smiles.shape)\n",
    "exp_smiles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1305b1a8-db34-4afc-aedb-19921870d257",
   "metadata": {},
   "outputs": [],
   "source": [
    "exp_smiles['smiles'] = exp_smiles['smiles'].apply(lambda x: Chem.MolToSmiles(Chem.MolFromSmiles(x)) if Chem.MolFromSmiles(x) is not None else None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "687f2fbd-c7a8-40c5-9933-22a0f8e1d811",
   "metadata": {},
   "outputs": [],
   "source": [
    "exp_smiles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b9cd9a3d-0331-46de-b5c4-b5c835dc2b4c",
   "metadata": {},
   "outputs": [],
   "source": [
    "exp_smiles['label'] = 'real'\n",
    "exp_smiles"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fb547625-6d6a-40fe-89e3-b7a9297f7587",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_generated = pd.read_csv('./results/generative_model/Generated_Practical_Set.csv')\n",
    "df_generated = df_generated[['smiles']]\n",
    "df_generated"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "96e566e1-ba72-4004-8a56-feb40711ac9e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Assuming your DataFrame is named df\n",
    "df_generated['smiles'] = df_generated['smiles'].apply(lambda x: Chem.MolToSmiles(Chem.MolFromSmiles(x)) if Chem.MolFromSmiles(x) is not None else None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6548b22b-cf81-462c-895f-727b75b22738",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_generated"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "93dd1ced-f544-4357-8a6c-c3b399d6ed5e",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_generated['label'] = 'generated'\n",
    "df_generated"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d6051938-3b66-45c7-9056-43cbba1ec5b8",
   "metadata": {},
   "outputs": [],
   "source": [
    "mixed_df = pd.concat([exp_smiles,df_generated], ignore_index=True)\n",
    "mixed_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "be7e8ff9-e9a7-40ad-bc01-c5c1f19e8809",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Set a fixed random seed (choose any integer value)\n",
    "import numpy as np\n",
    "np.random.seed(42)  # Example seed value (you can change this)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4a119d89-9988-495f-a7ec-4b62f84452e9",
   "metadata": {},
   "outputs": [],
   "source": [
    "from collections import Counter\n",
    "from matplotlib.colors import ListedColormap\n",
    "from matplotlib import pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7643aa03-34a9-477a-bc8e-8feee7096eb6",
   "metadata": {},
   "outputs": [],
   "source": [
    "type_labels, type_data = Faerun.create_categories(mixed_df[\"label\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3c408069-5693-4a7e-b093-9edca02f0684",
   "metadata": {},
   "outputs": [],
   "source": [
    "type_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ec6635d2-e34e-4a31-82be-2c8a7333e90b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# The number of permutations used by the MinHashing algorithm\n",
    "perm = 512\n",
    "\n",
    "# Initializing the MHFP encoder with 512 permutations\n",
    "enc = MHFPEncoder(perm)\n",
    "\n",
    "# Create MHFP fingerprints from SMILES\n",
    "# The fingerprint vectors have to be of the tm.VectorUint data type\n",
    "fingerprints = [tmap.VectorUint(enc.encode(s)) for s in mixed_df[\"smiles\"]]\n",
    "\n",
    "\n",
    "#df['yield'] = df['yield'].astype(str)\n",
    "# Initialize the LSH Forest\n",
    "lf = tmap.LSHForest(perm)\n",
    "\n",
    "# Add the Fingerprints to the LSH Forest and index\n",
    "lf.batch_add(fingerprints)\n",
    "lf.index()\n",
    "\n",
    "# Get the coordinates\n",
    "x, y, s, t, _ = tmap.layout_from_lsh_forest(lf)\n",
    "\n",
    "\n",
    "# Create a Faerun plot for the ESOL dataset\n",
    "faerun = Faerun(view=\"front\", coords=False, clear_color='#FFFFFF') #,clear_color='#FFFFFF'\n",
    "\n",
    "#Hex code for white: #FFFFFF\n",
    "#Hex code for black: #000000\n",
    "#Hex code for red: #FF0000\n",
    "#Hex code for green: #00FF00\n",
    "#Hex code for blue: #0000FF\n",
    "\n",
    "\n",
    "# Add scatter points with colors based on \"label\" values\n",
    "faerun.add_scatter(\n",
    "    \"tmap_real_vs_generated_512\",\n",
    "    {\n",
    "        \"x\": x,\n",
    "        \"y\": y,\n",
    "        \"c\": type_data,  # Use df['label'] directly\n",
    "        \"labels\": mixed_df[\"smiles\"]\n",
    "    },\n",
    "    point_scale=15,\n",
    "    colormap=['tab10'],\n",
    "    has_legend=True,\n",
    "    legend_title=[''],\n",
    "    legend_labels=[type_labels],\n",
    "    categorical=[True],\n",
    "    shader='sphere' #circle, sphere\n",
    ")\n",
    "\n",
    "# Add a tree structure connecting related points\n",
    "faerun.add_tree(\"tmap_real_vs_generated_512_tree\", {\"from\": s, \"to\": t}, point_helper=\"tmap_real_vs_generated_512\")\n",
    "\n",
    "# Choose the \"smiles\" template to display molecular structure on hover\n",
    "faerun.plot('tmap_real_vs_generated_512', template=\"smiles\", notebook_height=900)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "re-explore",
   "language": "python",
   "name": "re-explore"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}