dockm8_ensemble.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Import required libraries and scripts**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Import required libraries and scripts\n",
    "from scripts.library_preparation import *\n",
    "from scripts.utilities import *\n",
    "from scripts.docking_functions import *\n",
    "from scripts.clustering_functions import *\n",
    "from scripts.rescoring_functions import *\n",
    "from scripts.consensus_methods import *\n",
    "from scripts.performance_calculation import *\n",
    "from scripts.dogsitescorer import *\n",
    "from scripts.get_pocket import *\n",
    "from scripts.postprocessing import *\n",
    "from scripts.protein_preparation import *"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Set up**\n",
    "- **software**: The path to the software folder. In most cases this is where the DockM8 repository was downloaded to.\n",
    "- **receptors**: The list of paths to the protein files (.pdb).\n",
    "- **prepare_proteins**: Whether or not protein files should be prepared using Protoss (True or False).\n",
    "- **pocket**: The method to use for pocket determination. Must be one of 'reference', 'RoG' or 'dogsitescorer'.\n",
    "- **ref_file**: The path to the reference ligand used to define the binding pocket (.sdf file).\n",
    "- **docking_library**: The path to the docking library file (.sdf).\n",
    "- **idcolumn**: The unique identifier column used in the docking library.\n",
    "- **conformers**: The method to use for conformer generation, must be one of 'GypsumDL', 'MMFF' or 'RDKit' (RDKit and MMFF are equivalent). \n",
    "- **protonation**: The method to use for compound protonation. Must be one of 'GypsumDL', 'None'.\n",
    "- **docking_programs**: The method(s) to use for docking. Must be one or more of 'GNINA', 'SMINA', 'QVINA2', 'QVINAW' or 'PLANTS'.\n",
    "- **nposes**: The number of poses to generate for each docking software. Default=10\n",
    "- **exhaustiveness**: The precision used if docking with SMINA/GNINA. Default=8\n",
    "- **pose_selection**: The method(s) to use for pose clustering. Must be one or more of 'RMSD', 'spyRMSD', 'espsim', 'USRCAT', '3DScore', 'bestpose', 'bestpose_GNINA', 'bestpose_SMINA', 'bestpose_QVINA2', 'bestpose_QVINAW' or 'bestpose_PLANTS'. You can also specify any of the scoring functions to select the poses.\n",
    "- **clustering_method**: Which algorithm to use for clustering. Must be one of 'KMedoids', 'Aff_prop'. Only valid for the descriptor based pose_selection methods (RMSD, spyRMSD, espsim, USRCAT)\n",
    "- **rescoring**: A list of scoring functions to use for rescoring. Must be one or more of 'GNINA-Affinity','CNN-Score','CNN-Affinity', 'AD4', 'CHEMPLP', 'RFScoreVS', 'LinF9', 'SCORCH', 'Vinardo', 'PLECScore', 'NNScore', 'KORP-PL', 'ConvexPLR', 'RTMScore', 'AAScore'.\n",
    "- **consensus**: Which consensus method to use. Must be one of :'ECR_best', 'ECR_avg', 'avg_ECR', 'RbR', 'RbV', 'Zscore_best', 'Zscore_avg'.\n",
    "We recommend to use the command line or GUI versions of DockM8 to generate decoys."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "CWD = os.getcwd()\n",
    "software = Path(CWD+'/software')\n",
    "receptors = [Path(CWD+'/dockm8_testing/1ffv_p.pdb'), Path(CWD+'/dockm8_testing/4kd1_p.pdb')]\n",
    "pocket = 'Reference'\n",
    "prepare_protein = True\n",
    "ref_files = [Path(CWD+'/dockm8_testing/1ffv_l.sdf'), Path(CWD+'/dockm8_testing/4kd1_l.sdf')]\n",
    "docking_library = Path(CWD+'/dockm8_testing/library.sdf')\n",
    "id_column = 'ID'\n",
    "conformers = 'GypsumDL'\n",
    "protonation = 'GypsumDL'\n",
    "docking_programs = ['PLANTS']\n",
    "n_poses = 10\n",
    "exhaustiveness = 8\n",
    "pose_selection = 'bestpose_PLANTS'\n",
    "clustering_method = 'KMedoids'\n",
    "rescoring_functions = ['GNINA_Affinity','CNN-Score','CNN-Affinity', 'AD4', 'CHEMPLP', 'RFScoreVS']\n",
    "consensus_method = 'ECR_best'\n",
    "ncpus = int(os.cpu_count()*0.9)\n",
    "open('log.txt', 'w').close()\n",
    "\n",
    "print('DockM8 is running in ensemble mode...')\n",
    "        \n",
    "receptor_dict = {}\n",
    "for i, receptor in enumerate(receptors):\n",
    "    receptor_dict[receptor] = ref_files[i]\n",
    "    \n",
    "for receptor, ref_file in receptor_dict.items():\n",
    "\n",
    "    w_dir = Path(receptor).parent / Path(receptor).stem\n",
    "    print('The working directory has been set to:', w_dir)\n",
    "    (w_dir).mkdir(exist_ok=True)\n",
    "    \n",
    "    if prepare_protein == 1:\n",
    "        #Prepare the protein\n",
    "        prepared_receptor = prepare_protein_protoss(receptor)\n",
    "    else:\n",
    "        prepared_receptor = receptor\n",
    "    \n",
    "    #if os.path.isfile(str(receptor).replace('.pdb', '_pocket.pdb')) == False:\n",
    "    if pocket == 'Reference':\n",
    "        pocket_definition = get_pocket(ref_file, prepared_receptor, 8)\n",
    "        print(pocket_definition)\n",
    "    if pocket == 'RoG':\n",
    "        pocket_definition = get_pocket_RoG(ref_file, prepared_receptor)\n",
    "        print(pocket_definition)\n",
    "    elif pocket == 'Dogsitescorer':\n",
    "        pocket_definition = binding_site_coordinates_dogsitescorer(prepared_receptor, w_dir, method='Volume')\n",
    "        print(pocket_definition)\n",
    "    #else:\n",
    "        #pocket_definition = calculate_pocket_coordinates_from_pocket_pdb_file((str(prepared_receptor).replace('.pdb', '_pocket.pdb')))\n",
    "        \n",
    "    if (w_dir / 'final_library.sdf').is_file() == False:\n",
    "        prepare_library(docking_library, w_dir, id_column, conformers, protonation, software, ncpus)\n",
    "        \n",
    "    docking(w_dir, prepared_receptor, pocket_definition, software, docking_programs, exhaustiveness, n_poses, ncpus, 'concurrent_process')\n",
    "    concat_all_poses(w_dir, docking_programs, prepared_receptor, ncpus, bust_poses=False)\n",
    "\n",
    "    print('Loading all poses SDF file...')\n",
    "    tic = time.perf_counter()\n",
    "    all_poses = PandasTools.LoadSDF(str(w_dir / 'allposes.sdf'), idName='Pose ID', molColName='Molecule', includeFingerprints=False, strictParsing=True)\n",
    "    print(f'Loaded {len(all_poses)} poses.')\n",
    "    toc = time.perf_counter()\n",
    "    print(f'Finished loading all poses SDF in {toc-tic:0.4f}!...')\n",
    "\n",
    "    for method in pose_selection:\n",
    "        if os.path.isfile(w_dir / 'clustering/'f'{method}_clustered.sdf') == False:\n",
    "            select_poses(method, 'KMedoids', w_dir, prepared_receptor, pocket_definition, software, all_poses, ncpus)\n",
    "    for method in pose_selection:\n",
    "        rescore_poses(w_dir, prepared_receptor, pocket_definition, software, w_dir / 'clustering' / f'{method}_clustered.sdf', rescoring_functions , ncpus)\n",
    "    for method in pose_selection:\n",
    "        apply_consensus_methods(w_dir, method, consensus_method, rescoring_functions, 'min_max')\n",
    "\n",
    "for method in pose_selection:\n",
    "    ensemble_results = ensemble_consensus(receptors, method, consensus_method, 1)\n",
    "\n",
    "print(ensemble_results)\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.8.13 ('wocondock')",
   "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": "3.10.13"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "ac9acbc93af693fb4e01b586b9d883cf48eab2850c268069ebbf85c5f9fbe2b9"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}