notebooks/01_Benchmark_DUDE.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "1f4ebd59-178e-436d-bee0-f3411320c255",
   "metadata": {},
   "source": [
    "# Benchmark DUD-E\n",
    "This notebook performs a performance comparison between different models using the DUD-E dataset.\n",
    "\n",
    "The models in the benchmark are:\n",
    "* Our atom/surface-level ensemble (DENVIS) trained on PDBbind general set (DENVIS-G).\n",
    "* Our atom/surface-level ensemble (DENVIS) trained on PDBbind refined set (DENVIS-R).\n",
    "* GNINA\n",
    "* Vina\n",
    "* RF-score\n",
    "* NN-score\n",
    "* DeepDTA\n",
    "* Gold\n",
    "* Glide\n",
    "* Surflex\n",
    "* FlexX\n",
    "* Our ligand baseline trained on PDBbind refined set.\n",
    "\n",
    "We compare AUROC, EF1 and BedROC metrics. For each model we have one group comprising multiple targets. For micro-average scores we use median (as we make no assumptions regarding distributions of metrics).\n",
    "\n",
    "For statistical analysis, we perform group comparisons using the Friedman test and post-hoc comparisons using Wilcoxon signed-rank tests with Sidak correction (non-parametric)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "0231770e-d369-44d9-9ef7-f11df639cac3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "env: OUTDATED_IGNORE=1  # suppress pingouing out-of-date warnings\n"
     ]
    }
   ],
   "source": [
    "%set_env OUTDATED_IGNORE=1  # suppress pingouing out-of-date warnings\n",
    "\n",
    "import os\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import pingouin as pg\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns; sns.set()\n",
    "\n",
    "import ensembling\n",
    "import filtering\n",
    "import metrics\n",
    "import parsing\n",
    "import plotting_utils\n",
    "import postprocessing"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eacf05df-feb3-4fc1-a155-98625d5bd4df",
   "metadata": {},
   "source": [
    "# 1. Configuration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "7d053b44-f2e8-45b9-8cb0-64f45ba73eda",
   "metadata": {},
   "outputs": [],
   "source": [
    "AVG_FUN = np.nanmedian  # Reduce function to compute averages across targets. Using nanmedian because there is one target missing from RF/NN-score\n",
    "EF_ALPHA = 0.01 # alpha for EF metric (i.e., EF1)\n",
    "BEDROC_ALPHA = 80.5 # alpha for BedROC metric (i.e, BedROC80.5)\n",
    "\n",
    "PERFORMANCE_METRICS = ['AUROC', 'EF1', 'BEDROC80.5']\n",
    "DATASET = 'DUD-E'\n",
    "\n",
    "MAIN_PATH = '../data/outputs'\n",
    "PROTEIN_LIGAND_PAIR_REDUCE = 'max'  # Reduce function for protein-ligand pairs multiple entries (e.g. Vina, docking, RF/NN-score, DeepDTA)\n",
    "PAIR_ID_COLS = ['target_id', 'ligand_id']  # Specifies a unique protein-ligand pair\n",
    "\n",
    "PATH_RESULTS = {\n",
    "    'DENVIS-G': {\n",
    "        'atom': os.path.join(MAIN_PATH, 'denvis_outputs', 'dude_main_general_atom.parquet'),\n",
    "        'surface': os.path.join(MAIN_PATH, 'denvis_outputs', 'dude_main_general_surface.parquet'),},\n",
    "    'DENVIS-R': {\n",
    "        'atom': os.path.join(MAIN_PATH, 'denvis_outputs', 'dude_main_refined_atom.parquet'),\n",
    "        'surface': os.path.join(MAIN_PATH, 'denvis_outputs', 'dude_main_refined_surface.parquet')},\n",
    "    'Ligand bsl.': os.path.join(MAIN_PATH, 'denvis_outputs', 'dude_main_refined_ligand_bsl.parquet'),\n",
    "    'GNINA': os.path.join(MAIN_PATH, 'gnina_outputs/dude', 'newdefault_CNNaffinity-max.summary'),\n",
    "    'Vina': os.path.join(MAIN_PATH, 'vina_outputs', 'dude.csv'),\n",
    "    'RF-score': os.path.join(MAIN_PATH, 'rfnn_dude_scores'),\n",
    "    'NN-score': os.path.join(MAIN_PATH, 'rfnn_dude_scores'),\n",
    "    'Deep DTA': os.path.join(MAIN_PATH, 'deepdta_outputs', 'dude_trained_pdbbind.json'),\n",
    "    'Gold': os.path.join(MAIN_PATH, 'docking_performance_scores', 'gold.csv'),\n",
    "    'Glide': os.path.join(MAIN_PATH, 'docking_performance_scores', 'glide.csv'),\n",
    "    'Surflex': os.path.join(MAIN_PATH, 'docking_performance_scores', 'surflex.csv'),\n",
    "    'Flexx': os.path.join(MAIN_PATH, 'docking_performance_scores', 'flexx.csv')\n",
    "}\n",
    "\n",
    "# Network output aggregation kwargs for refined/general models\n",
    "# General models have three outpus (Kd, Ki, IC50), refined models have two (Ki, Kd)\n",
    "OUTPUT_COMBINATION_KWS = {\n",
    "    'DENVIS-G': {'y_aff_weight': 0., 'y_kd_weight': 0.5, 'y_ki_weight': 0.5, 'y_ic50_weight': 0.5, 'use_clf': False, 'clf_strategy': None},\n",
    "    'DENVIS-R': {'y_aff_weight': 0., 'y_kd_weight': 0.5, 'y_ki_weight': 0.5, 'y_ic50_weight': 0.0, 'use_clf': False, 'clf_strategy': None},\n",
    "    'Ligand bsl.': {'y_aff_weight': 0., 'y_kd_weight': 0.5, 'y_ki_weight': 0.5, 'y_ic50_weight': 0.0, 'use_clf': False, 'clf_strategy': None},\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e62d99aa-43ad-40be-b881-c32a8f2bebd5",
   "metadata": {},
   "source": [
    "# 2. Parse results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "ff4f2d97-4295-4f5f-8103-53c7ce04b46d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Parsing DENVIS-G results...\n",
      "Parsing DENVIS-R results...\n",
      "Parsing ligand baseline results (refined)...\n",
      "Parsing DeepDTA results...\n",
      "Parsing Vina results...\n",
      "Parsing GNINA results...\n",
      "Parsing NN-score results...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 101/101 [10:16<00:00,  6.10s/it]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Parsing RF-score results...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 101/101 [10:15<00:00,  6.09s/it]\n"
     ]
    }
   ],
   "source": [
    "results = dict()  # key-value pairs that will store the outputs DataFrame instances\n",
    "\n",
    "# For DENVIS, we have two models, one trained on PDBBind general and one trained on PDBbind refined sets.\n",
    "for model in ['DENVIS-G', 'DENVIS-R']:\n",
    "    results_cache = dict()\n",
    "    print(f\"Parsing {model} results...\")\n",
    "    # For each case we have two different types of model, depending on protein representation used: atom-level and surface-level\n",
    "    for protein_model in ['atom', 'surface']:\n",
    "        results_tmp, _ = parsing.parse_results_denvis(PATH_RESULTS[model][protein_model])\n",
    "        results_cache[protein_model] = ensembling.compute_ensemble_scores(results_tmp, version=True, pair_id_cols=PAIR_ID_COLS)  # Multi-run ensembles\n",
    "    \n",
    "    \n",
    "    results_ens = ensembling.compute_level_ensemble_scores(\n",
    "        results_cache['atom'], results_cache['surface'], atom_weight=0.5, use_target_intersection=False, pair_id_cols=PAIR_ID_COLS)  # Multi-level (atom/surface) ensemble\n",
    "    results_ens = postprocessing.combine_outputs(results_ens, **OUTPUT_COMBINATION_KWS[model])  # Aggregate model outputs\n",
    "    results[model] = results_ens\n",
    "\n",
    "print(f\"Parsing ligand baseline results (refined)...\")\n",
    "model = 'Ligand bsl.'\n",
    "results_tmp, _ = parsing.parse_results_denvis(PATH_RESULTS[model])  \n",
    "results_tmp = ensembling.compute_ensemble_scores(results_tmp, version=True, pair_id_cols=PAIR_ID_COLS)\n",
    "results_tmp = postprocessing.combine_outputs(results_tmp, **OUTPUT_COMBINATION_KWS[model])\n",
    "results[model] = results_tmp\n",
    "\n",
    "print(f\"Parsing DeepDTA results...\")\n",
    "model = 'Deep DTA'\n",
    "results[model] = parsing.parse_results_deeppurpose(PATH_RESULTS[model], reduce=PROTEIN_LIGAND_PAIR_REDUCE)\n",
    "\n",
    "print(f\"Parsing Vina results...\")\n",
    "model = 'Vina'\n",
    "results[model] = parsing.parse_results_vina(PATH_RESULTS[model], reduce=PROTEIN_LIGAND_PAIR_REDUCE)  \n",
    "\n",
    "print(f\"Parsing GNINA results...\")\n",
    "model = 'GNINA'\n",
    "results[model] = parsing.parse_results_gnina(PATH_RESULTS[model])\n",
    "\n",
    "print(f\"Parsing NN-score results...\")\n",
    "model = 'NN-score'\n",
    "results[model] = parsing.parse_results_nn_score(PATH_RESULTS[model], reduce=PROTEIN_LIGAND_PAIR_REDUCE, prog_bar=True)\n",
    "\n",
    "print(f\"Parsing RF-score results...\")\n",
    "model = 'RF-score'\n",
    "results[model] = parsing.parse_results_rf_score(PATH_RESULTS[model], reduce=PROTEIN_LIGAND_PAIR_REDUCE, prog_bar=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a9eb888e-deb4-4c86-9b41-e9db9d0a8abb",
   "metadata": {},
   "source": [
    "# 3. Identify missing targets\n",
    "A few methods have missing targets. We will identify them here so that we can add `nan` entries in the respective DataFrames later to enable fair comparisons between the algorithms."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "88caca7c-1bec-409c-a676-fd9879c6e47e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Missing targets, DENVIS-G: set()\n",
      "Missing targets, DENVIS-R: set()\n",
      "Missing targets, Ligand bsl.: set()\n",
      "Missing targets, Deep DTA: set()\n",
      "Missing targets, Vina: set()\n",
      "Missing targets, GNINA: set()\n",
      "Missing targets, NN-score: {'csf1r'}\n",
      "Missing targets, RF-score: {'csf1r'}\n"
     ]
    }
   ],
   "source": [
    "filtering_list = []\n",
    "for _, result in results.items():\n",
    "    filtering_list.append(list(result.target_id.unique()))\n",
    "target_union = filtering.target_union(filtering_list)\n",
    "\n",
    "missing_targets = dict()  # One entry per model\n",
    "for model, result in results.items():\n",
    "    taget_diff = (set(target_union)) - set(result.target_id.unique())\n",
    "    missing_targets[model] = taget_diff\n",
    "    print(f'Missing targets, {model}: {missing_targets[model]}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2db63892-d6ec-47f5-89a1-11ee91abd262",
   "metadata": {},
   "source": [
    "# 4. Benchmark"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fafb5498-9a9e-4c43-aeba-e1b1026a18e8",
   "metadata": {},
   "source": [
    "## 4.1 Compute performance scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "59e85d29-023a-4a85-8864-d0b2fa5b0f14",
   "metadata": {},
   "outputs": [],
   "source": [
    "results_tf = pd.DataFrame(columns=['Dataset', 'Model'] + PERFORMANCE_METRICS)  # micro-average DataFrame (tidy-format)\n",
    "results_pt_tf = pd.DataFrame(columns=['Dataset', 'Model', 'Target'] + PERFORMANCE_METRICS)  # Per-targety DataFrame (tidy-format)\n",
    "\n",
    "for model, result in results.items():\n",
    "    # Compute per-target and micro-average scores\n",
    "    auroc_pt, auroc_micro = metrics.compute_auroc_scores(result, avg_fun=AVG_FUN)\n",
    "    ef1_pt, ef1_micro = metrics.compute_ef_scores(result, alpha=EF_ALPHA, avg_fun=AVG_FUN)\n",
    "    bedroc805_pt, bedroc805_micro = metrics.compute_bedroc_scores(result, alpha=BEDROC_ALPHA, avg_fun=AVG_FUN)\n",
    "\n",
    "    # Update micro DataFrame\n",
    "    results_tf = pd.concat((results_tf, pd.DataFrame.from_records({\n",
    "        'Dataset': DATASET,\n",
    "        'Model': model,\n",
    "        'AUROC': auroc_micro,\n",
    "        'EF1': ef1_micro,\n",
    "        'BEDROC80.5': bedroc805_micro}, \n",
    "        index=[len(results_tf)])))\n",
    "\n",
    "    new_df_pt = pd.DataFrame.from_records(\n",
    "        [auroc_pt, ef1_pt, bedroc805_pt],\n",
    "        index=['AUROC', 'EF1', 'BEDROC80.5']).T.reset_index(level=0, inplace=False).rename({'index': 'Target'}, axis=1, inplace=False)\n",
    "    # If there are missing targets for this model, add entries with nan values (to make statistical comparisons feasible)\n",
    "    if len(missing_targets[model]):\n",
    "        for missing_target in missing_targets[model]:\n",
    "            new_df_pt = pd.concat(\n",
    "                (new_df_pt, pd.DataFrame({\n",
    "                    'Target': missing_target,\n",
    "                    'AUROC': np.nan,\n",
    "                    'EF1': np.nan,\n",
    "                    'BEDROC80.5': np.nan},\n",
    "                    index=[len(new_df_pt)])), axis='index')\n",
    "    new_df_pt.insert(0, 'Dataset', DATASET)\n",
    "    new_df_pt.insert(1, 'Model', model)\n",
    "    results_pt_tf = pd.concat((results_pt_tf, new_df_pt))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "911968ca-eb6f-4256-8c05-47cf458330dd",
   "metadata": {},
   "source": [
    "## 4.2 Append performance scores for docking methods\n",
    "For the four docking methods (Gold, Glide, Surflex, Flexx), the raw outputs (i.e., scores) are not available. For this reason, we provide the computed performance scores for each method per-target. We append these results in the DataFrames computed above."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "739ca5b0-ab79-4e47-a409-bdeefa891e13",
   "metadata": {},
   "outputs": [],
   "source": [
    "for model in ['Gold', 'Glide', 'Surflex', 'Flexx']:\n",
    "    # Update per-target DataFrame\n",
    "    new_df_pt = pd.read_csv(PATH_RESULTS[model])\n",
    "    new_df_pt.insert(0, 'Dataset', DATASET)\n",
    "    new_df_pt.insert(1, 'Model', model)\n",
    "    results_pt_tf = pd.concat((results_pt_tf, new_df_pt))\n",
    "    \n",
    "    # Update micro-average DataFrame\n",
    "    results_tf = pd.concat((results_tf, pd.DataFrame.from_records({\n",
    "        'Dataset': DATASET,\n",
    "        'Model': model,\n",
    "        'AUROC': AVG_FUN(new_df_pt['AUROC']),\n",
    "        'EF1': AVG_FUN(new_df_pt['EF1']),\n",
    "        'BEDROC80.5': AVG_FUN(new_df_pt['BEDROC80.5'])},\n",
    "        index=[len(results_tf)])))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f1cda039-6da0-40e6-b6e0-623d8dd4187b",
   "metadata": {},
   "source": [
    "## 4.3 Micro-average score benchmark (Table 3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "65be38ba-9f1d-4b51-9531-4e95deb1347a",
   "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>AUROC</th>\n",
       "      <th>EF1</th>\n",
       "      <th>BEDROC80.5</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Model</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>DENVIS-G</th>\n",
       "      <td>0.92</td>\n",
       "      <td>38.92</td>\n",
       "      <td>0.66</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DENVIS-R</th>\n",
       "      <td>0.86</td>\n",
       "      <td>18.37</td>\n",
       "      <td>0.33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Deep DTA</th>\n",
       "      <td>0.58</td>\n",
       "      <td>1.62</td>\n",
       "      <td>0.04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Gold</th>\n",
       "      <td>0.86</td>\n",
       "      <td>21.08</td>\n",
       "      <td>0.39</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Glide</th>\n",
       "      <td>0.82</td>\n",
       "      <td>21.96</td>\n",
       "      <td>0.35</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Surflex</th>\n",
       "      <td>0.75</td>\n",
       "      <td>9.36</td>\n",
       "      <td>0.19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Flexx</th>\n",
       "      <td>0.76</td>\n",
       "      <td>10.10</td>\n",
       "      <td>0.20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Vina</th>\n",
       "      <td>0.75</td>\n",
       "      <td>6.84</td>\n",
       "      <td>0.14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GNINA</th>\n",
       "      <td>0.79</td>\n",
       "      <td>15.48</td>\n",
       "      <td>0.29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RF-score</th>\n",
       "      <td>0.62</td>\n",
       "      <td>2.00</td>\n",
       "      <td>0.05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NN-score</th>\n",
       "      <td>0.58</td>\n",
       "      <td>1.38</td>\n",
       "      <td>0.03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Ligand bsl.</th>\n",
       "      <td>0.66</td>\n",
       "      <td>1.33</td>\n",
       "      <td>0.04</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             AUROC    EF1  BEDROC80.5\n",
       "Model                                \n",
       "DENVIS-G      0.92  38.92        0.66\n",
       "DENVIS-R      0.86  18.37        0.33\n",
       "Deep DTA      0.58   1.62        0.04\n",
       "Gold          0.86  21.08        0.39\n",
       "Glide         0.82  21.96        0.35\n",
       "Surflex       0.75   9.36        0.19\n",
       "Flexx         0.76  10.10        0.20\n",
       "Vina          0.75   6.84        0.14\n",
       "GNINA         0.79  15.48        0.29\n",
       "RF-score      0.62   2.00        0.05\n",
       "NN-score      0.58   1.38        0.03\n",
       "Ligand bsl.   0.66   1.33        0.04"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print_order = ['DENVIS-G', 'DENVIS-R', 'Deep DTA', 'Gold', 'Glide', 'Surflex', 'Flexx', 'Vina', 'GNINA', 'RF-score', 'NN-score', 'Ligand bsl.']\n",
    "cols = ['AUROC', 'EF1', 'BEDROC80.5']\n",
    "print_df = results_tf.set_index('Model').loc[print_order][cols].round(decimals=2)\n",
    "print_df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2f5afd6d-c4f1-4643-8119-9db2c748d622",
   "metadata": {},
   "source": [
    "## 4.4 Per-target benchmark (Figure 3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "2ae7f180-13ff-4e69-9e07-e61173a5eaf0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 201.6x381.6 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set_context(\"paper\", font_scale=.8, rc={\"lines.linewidth\": 1.5})\n",
    "hue_order = ['DENVIS-G', 'DENVIS-R', 'Deep DTA', 'Gold', 'Glide', 'Surflex', 'Flexx', 'Vina', 'GNINA', 'RF-score', 'NN-score', 'Ligand bsl.']\n",
    "color_palette = sns.color_palette(\"Blues\", 3) + sns.color_palette(\"Oranges\", 5) + sns.color_palette(\"Greens\", 3) + sns.color_palette(\"Greys\", 1)\n",
    "metrics_ = ['AUROC', 'EF1', 'BEDROC80.5']\n",
    "\n",
    "showfliers = True\n",
    "fliersize = 1\n",
    "linewidth = 1.2\n",
    "figsize = (2.8,5.3)\n",
    "leg_pos = (-0.01, 1.9)\n",
    "\n",
    "label_alias = {\n",
    "    'DENVIS-G': 'DENVIS-G (ours) ',\n",
    "    'DENVIS-R': 'DENVIS-R (ours)',\n",
    "    'GNINA': 'GNINA',\n",
    "    'Vina': 'Vina',\n",
    "    'RF-score': 'RF-score',\n",
    "    'NN-score': 'NN-score',\n",
    "    'Deep DTA': 'DeepDTA',\n",
    "    'Ligand bsl.': 'Ligand bsl.',\n",
    "    'Gold': 'Gold',\n",
    "    'Glide': 'Glide',\n",
    "    'Surflex': 'Surflex',\n",
    "    'Flexx': 'FlexX'\n",
    "}\n",
    "\n",
    "fig, ax = plt.subplots(figsize=figsize, nrows=3, ncols=1)\n",
    "for i, metric in enumerate(metrics_):\n",
    "    sns.boxenplot(\n",
    "        data=results_pt_tf,\n",
    "        x='Dataset',\n",
    "        order=['DUD-E'],\n",
    "        hue='Model',\n",
    "        y=metric,\n",
    "        hue_order=hue_order,\n",
    "        palette=color_palette,\n",
    "        showfliers=showfliers,\n",
    "        linewidth=linewidth,\n",
    "        ax=ax[i])\n",
    "\n",
    "\n",
    "# Legend options\n",
    "handles, labels = ax[0].get_legend_handles_labels()\n",
    "ax[0].legend(handles=handles, labels=[label_alias[l] for l in labels], bbox_to_anchor=leg_pos, loc='upper left', ncol=2)\n",
    "\n",
    "for ax_ in ax:\n",
    "    ax_.set_xticklabels([])\n",
    "    ax_.set_xlabel('')\n",
    "    # Temporary workaround for setting fliersize for boxenplot\n",
    "    # see: https://github.com/mwaskom/seaborn/issues/2701\n",
    "    plt.setp(ax_.collections[::2], sizes=[2 ** fliersize])\n",
    "\n",
    "for ax_ in ax[1:]:\n",
    "    ax_.legend().remove()\n",
    "\n",
    "ax[1].set_ylabel(r'EF$_{1}$')\n",
    "ax[2].set_ylabel(r'BEDROC$_{80.5}$')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8bd3072d-2e04-42f7-911f-679b1afbe938",
   "metadata": {},
   "source": [
    "## 4.5 One-to-one comparisons (Figure 4)\n",
    "* DENVIS-R vs. GNINA\n",
    "* DENVIS-G vs. Gold"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "abf977f2-04cf-4adc-bdc0-2608be2d99b7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 396x288 with 6 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set_context(\"paper\", font_scale=.8, rc={\"lines.linewidth\": 1.5})\n",
    "figsize = (5.5,4.)\n",
    "size = 10\n",
    "aspect_ratio = 0.7\n",
    "aspect_ratio = 1.0\n",
    "\n",
    "# Colors are defined in the cell above, we just index them here\n",
    "denvis_g_col = color_palette[0]\n",
    "denvis_r_col = color_palette[1]\n",
    "gnina_col = color_palette[9]\n",
    "gold_col = color_palette[4]\n",
    "col_1, col_2 = sns.color_palette('Paired')[8:10]\n",
    "col_1, col_2 = sns.color_palette(n_colors=4)[0], sns.color_palette(n_colors=4)[3]\n",
    "\n",
    "\n",
    "fig, ax = plt.subplots(figsize=figsize, nrows=2, ncols=3)\n",
    "comparisons = [\n",
    "    ('DENVIS-R', 'GNINA'),\n",
    "    ('DENVIS-G', 'Gold')]\n",
    "\n",
    "for row_id, comparison in enumerate(comparisons):\n",
    "    for col_id, metric in enumerate(PERFORMANCE_METRICS):\n",
    "        our_scores = results_pt_tf[(results_pt_tf['Model']==comparison[0])][metric].values\n",
    "        other_scores = results_pt_tf[(results_pt_tf['Model']==comparison[1])][metric].values\n",
    "        \n",
    "        sns.scatterplot(\n",
    "            x=other_scores[other_scores > our_scores],\n",
    "            y=our_scores[other_scores > our_scores],\n",
    "            color=col_1,\n",
    "            s=size,\n",
    "            ax=ax[row_id, col_id])\n",
    "        sns.scatterplot(\n",
    "            x=other_scores[other_scores < our_scores],\n",
    "            y=our_scores[other_scores < our_scores],\n",
    "            color=col_2,\n",
    "            s=size,\n",
    "            ax=ax[row_id, col_id])\n",
    "        plotting_utils.plot_diagonal_line(ax[row_id, col_id])\n",
    "        ax[row_id, col_id].set_ylabel(comparison[0])\n",
    "        ax[row_id, col_id].set_xlabel(comparison[1])\n",
    "        ax[row_id, col_id].set_title(metric)\n",
    "\n",
    "# Tidy-up\n",
    "for ax_ in ax.reshape(-1,):\n",
    "    ax_.set_xticks(ax_.get_yticks()[1:-1])\n",
    "    ax_.set_yticks(ax_.get_yticks()[1:-1])\n",
    "          \n",
    "fig.text(0, 1.0, '(a)')\n",
    "fig.text(0, 0.5, '(b)')\n",
    "fig.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9af98580-2f2c-49a0-86b0-1c1307aa56bf",
   "metadata": {},
   "source": [
    "## 4.6 Statitical analysis (Table 4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "dbc9b6ad-7b2f-4866-aef7-679b0621c688",
   "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></th>\n",
       "      <th>AUROC</th>\n",
       "      <th>EF1</th>\n",
       "      <th>BEDROC80.5</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>DENVIS-G</th>\n",
       "      <th>DENVIS-R</th>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DENVIS-R</th>\n",
       "      <th>GNINA</th>\n",
       "      <td>n.s.</td>\n",
       "      <td>n.s.</td>\n",
       "      <td>n.s.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">DENVIS-G</th>\n",
       "      <th>Gold</th>\n",
       "      <td>n.s.</td>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Glide</th>\n",
       "      <td>**</td>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DENVIS-R</th>\n",
       "      <th>DeepDTA</th>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DENVIS-G</th>\n",
       "      <th>Ligand baseline</th>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DENVIS-R</th>\n",
       "      <th>Ligand baseline</th>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>GNINA</th>\n",
       "      <th>Ligand baseline</th>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DeepDTA</th>\n",
       "      <th>Ligand baseline</th>\n",
       "      <td>**</td>\n",
       "      <td>n.s.</td>\n",
       "      <td>n.s.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Vina</th>\n",
       "      <th>Ligand baseline</th>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RF-score</th>\n",
       "      <th>Ligand baseline</th>\n",
       "      <td>n.s.</td>\n",
       "      <td>n.s.</td>\n",
       "      <td>n.s.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>NN-score</th>\n",
       "      <th>Ligand baseline</th>\n",
       "      <td>*</td>\n",
       "      <td>n.s.</td>\n",
       "      <td>n.s.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Gold</th>\n",
       "      <th>Ligand baseline</th>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Surflex</th>\n",
       "      <th>Ligand baseline</th>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>FlexX</th>\n",
       "      <th>Ligand baseline</th>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Glide</th>\n",
       "      <th>Ligand baseline</th>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "      <td>***</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         AUROC   EF1 BEDROC80.5\n",
       "DENVIS-G DENVIS-R          ***   ***        ***\n",
       "DENVIS-R GNINA            n.s.  n.s.       n.s.\n",
       "DENVIS-G Gold             n.s.   ***        ***\n",
       "         Glide              **   ***        ***\n",
       "DENVIS-R DeepDTA           ***   ***        ***\n",
       "DENVIS-G Ligand baseline   ***   ***        ***\n",
       "DENVIS-R Ligand baseline   ***   ***        ***\n",
       "GNINA    Ligand baseline   ***   ***        ***\n",
       "DeepDTA  Ligand baseline    **  n.s.       n.s.\n",
       "Vina     Ligand baseline   ***   ***        ***\n",
       "RF-score Ligand baseline  n.s.  n.s.       n.s.\n",
       "NN-score Ligand baseline     *  n.s.       n.s.\n",
       "Gold     Ligand baseline   ***   ***        ***\n",
       "Surflex  Ligand baseline   ***   ***        ***\n",
       "FlexX    Ligand baseline   ***   ***        ***\n",
       "Glide    Ligand baseline   ***   ***        ***"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.options.display.float_format = '{:.2e}'.format\n",
    "\n",
    "def statistical_notation(p_val):\n",
    "    if p_val < 1e-3:\n",
    "        res = '***'\n",
    "    elif p_val < 1e-2:\n",
    "        res = '**'\n",
    "    elif p_val < 0.05:\n",
    "        res = '*'\n",
    "    else:\n",
    "        res = 'n.s.'\n",
    "    \n",
    "    return res\n",
    "\n",
    "comparisons = [\n",
    "    ('DENVIS-G', 'DENVIS-R'),\n",
    "    ('DENVIS-R', 'GNINA'),\n",
    "    ('DENVIS-G', 'Gold'),\n",
    "    ('DENVIS-G', 'Glide'),\n",
    "    ('DENVIS-R', 'Deep DTA'),\n",
    "    ('DENVIS-G', 'Ligand bsl.'),\n",
    "    ('DENVIS-R', 'Ligand bsl.'),\n",
    "    ('GNINA', 'Ligand bsl.'),\n",
    "    ('Deep DTA', 'Ligand bsl.'),\n",
    "    ('Vina', 'Ligand bsl.'),\n",
    "    ('RF-score', 'Ligand bsl.'),\n",
    "    ('NN-score', 'Ligand bsl.'),\n",
    "    ('Gold', 'Ligand bsl.'),\n",
    "    ('Surflex', 'Ligand bsl.'),\n",
    "    ('Flexx', 'Ligand bsl.'),\n",
    "    ('Glide', 'Ligand bsl.'),\n",
    "]\n",
    "perf_metrics = ['AUROC', 'EF1', 'BEDROC80.5']\n",
    "label_alias = {\n",
    "    'DENVIS-G': 'DENVIS-G',\n",
    "    'DENVIS-R': 'DENVIS-R',\n",
    "    'GNINA': 'GNINA',\n",
    "    'Vina': 'Vina',\n",
    "    'RF-score': 'RF-score',\n",
    "    'NN-score': 'NN-score',\n",
    "    'Deep DTA': 'DeepDTA',\n",
    "    'Ligand bsl.': 'Ligand baseline',\n",
    "    'Gold': 'Gold',\n",
    "    'Glide': 'Glide',\n",
    "    'Surflex': 'Surflex',\n",
    "    'Flexx': 'FlexX'}\n",
    "\n",
    "df_comparison = pd.DataFrame(columns=['Model 1', 'Model 2'] + PERFORMANCE_METRICS)\n",
    "df_comparison['Model 1'] = [model[0] for model in comparisons]\n",
    "df_comparison['Model 2'] = [model[1] for model in comparisons]\n",
    "\n",
    "multcompare_method = 'sidak'\n",
    "\n",
    "for perf_metric in PERFORMANCE_METRICS:\n",
    "    friedman_pval = pg.friedman(data=results_pt_tf, dv=perf_metric, within='Model', subject='Target')['p-unc']['Friedman']\n",
    "    \n",
    "    p_val_uncorr = []\n",
    "    n_comparisons = len(comparisons)\n",
    "    for comparison in comparisons:\n",
    "        model_x, model_y = comparison\n",
    "        x = results_pt_tf[results_pt_tf['Model']==model_x].set_index('Target').sort_index()[perf_metric].values\n",
    "        y = results_pt_tf[results_pt_tf['Model']==model_y].set_index('Target').sort_index()[perf_metric].values\n",
    "        p_val_uncorr.append(pg.wilcoxon(x=x, y=y, alternative='two-sided')['p-val']['Wilcoxon'])\n",
    "\n",
    "    # Perform multiple comparisons\n",
    "    nh_reject, p_val_corr = pg.multicomp(pvals=p_val_uncorr, alpha=0.05, method=multcompare_method)\n",
    "    df_comparison[perf_metric] = p_val_corr\n",
    "\n",
    "for perf_metric in perf_metrics:\n",
    "    df_comparison[perf_metric] = df_comparison.apply(lambda row: statistical_notation(row[perf_metric]), axis=1)\n",
    "\n",
    "print_df = df_comparison.replace(to_replace=label_alias).set_index(['Model 1', 'Model 2'])\n",
    "print_df.index.names = None, None\n",
    "print_df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fb838c78-fef2-4390-9cee-d249383f9187",
   "metadata": {},
   "source": [
    "## 4.7 Number of targets that one model is better than the other"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "4d65f747-ac37-40c9-b11e-1e4afa316e4c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Models: DENVIS-G vs. GNINA\n",
      "----------------------------------------------\n",
      "AUROC: DENVIS-G > GNINA: 83\n",
      "AUROC: GNINA > DENVIS-G: 19\n",
      "EF1: DENVIS-G > GNINA: 81\n",
      "EF1: GNINA > DENVIS-G: 20\n",
      "BEDROC80.5: DENVIS-G > GNINA: 83\n",
      "BEDROC80.5: GNINA > DENVIS-G: 19\n",
      "----------------------------------------------\n",
      "Models: DENVIS-G vs. Gold\n",
      "----------------------------------------------\n",
      "AUROC: DENVIS-G > Gold: 66\n",
      "AUROC: Gold > DENVIS-G: 36\n",
      "EF1: DENVIS-G > Gold: 73\n",
      "EF1: Gold > DENVIS-G: 28\n",
      "BEDROC80.5: DENVIS-G > Gold: 76\n",
      "BEDROC80.5: Gold > DENVIS-G: 26\n",
      "----------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "targets = results_pt_tf['Target'].unique()\n",
    "for (model_x, model_y) in [('DENVIS-G', 'GNINA'), ('DENVIS-G', 'Gold')]:\n",
    "    print(f\"Models: {model_x} vs. {model_y}\")\n",
    "    print(\"----------------------------------------------\")\n",
    "    for metric in ['AUROC', 'EF1', 'BEDROC80.5']:\n",
    "        x = results_pt_tf[results_pt_tf['Model']==model_x].set_index('Target').reindex(targets)[metric]\n",
    "        y = results_pt_tf[results_pt_tf['Model']==model_y].set_index('Target').reindex(targets)[metric]\n",
    "        print(f\"{metric}: {model_x} > {model_y}: {(x > y).sum()}\")\n",
    "        print(f\"{metric}: {model_y} > {model_x}: {(y > x).sum()}\")\n",
    "    print(\"----------------------------------------------\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "94f9f4c9-c984-434a-abef-2794de8f0b3d",
   "metadata": {},
   "source": [
    "## 4.8 Table per-target comparisons (Tables S1, S2, S3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "9ad8912e-9dd0-4b0f-8a94-4e24faf608d7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUROC\n",
      "--------------------------------------------------------------------------------------\n",
      "        D-G  D-R  DDTA  Gold  Glide  Surflex  FlexX  Vina  GNINA   RF   NN  Bsl.\n",
      "aa2ar  0.84 0.43  0.59  0.87   0.80     0.77   0.74  0.65   0.74 0.50 0.54  0.57\n",
      "abl1   0.89 0.90  0.65  0.90   0.74     0.76   0.85  0.78   0.79 0.68 0.62  0.81\n",
      "ace    0.90 0.96  0.41  0.74   0.72     0.77   0.61  0.56   0.61 0.67 0.75  0.40\n",
      "aces   0.81 0.82  0.62  0.75   0.67     0.73   0.57  0.78   0.62 0.66 0.65  0.67\n",
      "ada    0.99 0.99  0.61  0.91   0.68     0.83   0.55  0.57   0.76 0.61 0.58  0.61\n",
      "ada17  0.99 0.98  0.50  0.77   0.81     0.92   0.60  0.71   0.82 0.66 0.69  0.70\n",
      "adrb1  0.90 0.62  0.54  0.86   0.88     0.80   0.76  0.73   0.63 0.68 0.58  0.34\n",
      "adrb2  0.89 0.64  0.56  0.87   0.88     0.81   0.84  0.71   0.71 0.62 0.61  0.35\n",
      "akt1   0.88 0.96  0.79  0.91   0.78     0.63   0.73  0.76   0.76 0.63 0.49  0.71\n",
      "akt2   0.87 0.91  0.72  0.94   0.85     0.77   0.86  0.79   0.81 0.56 0.65  0.65\n",
      "aldr   0.56 0.60  0.45  0.79   0.79     0.65   0.74  0.74   0.56 0.59 0.50  0.55\n",
      "ampc   0.86 0.86  0.46  0.75   0.79     0.49   0.67  0.61   0.52 0.34 0.57  0.26\n",
      "andr   0.94 0.88  0.57  0.57   0.84     0.57   0.41  0.63   0.79 0.64 0.57  0.69\n",
      "aofb   0.58 0.54  0.47  0.84   0.76     0.68   0.71  0.78   0.51 0.53 0.56  0.54\n",
      "bace1  0.94 0.85  0.66  0.78   0.82     0.80   0.76  0.72   0.82 0.79 0.63  0.61\n",
      "braf   0.96 0.95  0.67  0.91   0.86     0.70   0.89  0.86   0.83 0.63 0.56  0.80\n",
      "cah2   0.97 0.94  0.75  0.82   0.77     0.64   0.73  0.59   0.93 0.82 0.68  0.76\n",
      "casp3  0.92 0.63  0.53  0.91   0.76     0.70   0.70  0.70   0.65 0.54 0.63  0.51\n",
      "cdk2   0.96 0.90  0.66  0.85   0.88     0.64   0.76  0.72   0.80 0.55 0.59  0.77\n",
      "comt   0.98 0.58  0.48  0.99   1.00     0.75   0.66  0.63   0.65 0.27 0.42  0.37\n",
      "cp2c9  0.56 0.55  0.57  0.82   0.64     0.59   0.70  0.62   0.62 0.59 0.53  0.58\n",
      "cp3a4  0.41 0.43  0.61  0.77   0.72     0.61   0.60  0.60   0.66 0.58 0.51  0.62\n",
      "csf1r  0.97 0.91  0.74  0.87   0.80     0.61   0.84  0.69   0.78  NaN  NaN  0.80\n",
      "cxcr4  0.78 0.77  0.52  0.92   0.72     0.76   0.47  0.60   0.85 0.68 0.80  0.71\n",
      "def    1.00 0.96  0.48  0.91   0.87     0.93   0.70  0.76   0.93 0.60 0.56  0.54\n",
      "dhi1   0.33 0.52  0.56  0.75   0.68     0.74   0.49  0.77   0.72 0.62 0.68  0.59\n",
      "dpp4   0.93 0.78  0.71  0.75   0.87     0.71   0.86  0.62   0.75 0.60 0.67  0.62\n",
      "drd3   0.56 0.55  0.55  0.73   0.72     0.77   0.67  0.75   0.71 0.66 0.60  0.66\n",
      "dyr    0.96 0.94  0.61  0.92   0.87     0.87   0.67  0.77   0.90 0.55 0.68  0.67\n",
      "egfr   0.93 0.91  0.58  0.84   0.84     0.67   0.83  0.64   0.77 0.54 0.55  0.69\n",
      "esr1   0.97 0.79  0.56  0.76   0.91     0.76   0.80  0.83   0.80 0.64 0.50  0.71\n",
      "esr2   0.93 0.75  0.55  0.79   0.94     0.76   0.84  0.80   0.79 0.65 0.50  0.71\n",
      "fa10   0.96 0.88  0.71  0.90   0.68     0.78   0.88  0.83   0.88 0.80 0.55  0.72\n",
      "fa7    0.99 0.99  0.67  0.97   0.91     0.98   0.97  0.91   0.91 0.85 0.70  0.55\n",
      "fabp4  0.95 0.90  0.53  0.87   0.85     0.76   0.87  0.78   0.53 0.66 0.30  0.44\n",
      "fak1   0.99 0.99  0.56  0.95   0.80     0.64   0.88  0.81   0.84 0.70 0.49  0.88\n",
      "fgfr1  0.47 0.56  0.53  0.70   0.48     0.58   0.59   NaN   0.94  NaN  NaN  0.49\n",
      "fkb1a  0.95 0.89  0.40  0.84   0.87     0.80   0.58  0.77   0.78 0.65 0.57  0.47\n",
      "fnta   0.90 0.74  0.58  0.78   0.75     0.69   0.76  0.65   0.78 0.74 0.57  0.73\n",
      "fpps   0.74 0.71  0.20  1.00   0.55     0.52   0.82  0.29   0.98 0.20 1.00  0.32\n",
      "gcr    0.69 0.80  0.64  0.54   0.72     0.67   0.54  0.64   0.75 0.60 0.64  0.75\n",
      "glcm   0.72 0.68  0.48  0.90   0.65     0.82   0.69  0.49   0.58 0.49 0.60  0.49\n",
      "gria2  0.67 0.72  0.67  0.83   0.77     0.75   0.76  0.75   0.83 0.56 0.62  0.60\n",
      "grik1  0.90 0.84  0.60  0.92   0.90     0.81   0.87  0.59   0.73 0.46 0.68  0.50\n",
      "hdac2  0.96 0.92  0.61  0.88   0.63     0.85   0.81  0.85   0.80 0.81 0.54  0.75\n",
      "hdac8  0.98 0.94  0.60  0.87   0.88     0.84   0.76  0.82   0.83 0.75 0.53  0.77\n",
      "hivint 0.47 0.44  0.44  0.78   0.72     0.67   0.67  0.71   0.50 0.48 0.47  0.52\n",
      "hivpr  0.94 0.92  0.51  0.64   0.72     0.82   0.55  0.72   0.85 0.77 0.58  0.65\n",
      "hivrt  0.84 0.58  0.45  0.80   0.83     0.67   0.61  0.68   0.82 0.50 0.56  0.56\n",
      "hmdh   0.87 0.85  0.57  0.87   0.92     0.83   0.67  0.79   0.92 0.71 0.69  0.66\n",
      "hs90a  0.99 0.89  0.57  0.85   0.65     0.49   0.34  0.26   0.73 0.39 0.44  0.52\n",
      "hxk4   0.84 0.78  0.54  0.88   0.57     0.56   0.67  0.57   0.82 0.62 0.63  0.74\n",
      "igf1r  0.94 0.93  0.68  0.90   0.86     0.70   0.89  0.83   0.85 0.57 0.62  0.76\n",
      "inha   0.86 0.62  0.38  0.96   0.56     0.84   0.69  0.72   0.68 0.66 0.54  0.57\n",
      "ital   0.89 0.65  0.47  0.80   0.68     0.50   0.61  0.60   0.66 0.61 0.48  0.76\n",
      "jak2   0.99 0.94  0.70  0.94   0.88     0.73   0.83  0.78   0.91 0.51 0.63  0.84\n",
      "kif11  0.81 0.73  0.54  0.89   0.90     0.70   0.69  0.85   0.81 0.72 0.54  0.70\n",
      "kit    0.96 0.96  0.67  0.83   0.65     0.60   0.79  0.78   0.79 0.65 0.56  0.74\n",
      "kith   0.28 0.38  0.36  0.87   0.72     0.83   0.82  0.74   0.73 0.57 0.67  0.30\n",
      "kpcb   0.82 0.78  0.59  0.90   0.86     0.84   0.80  0.77   0.71 0.69 0.49  0.60\n",
      "lck    0.91 0.94  0.64  0.80   0.83     0.72   0.84  0.80   0.76 0.49 0.53  0.79\n",
      "lkha4  0.99 0.95  0.52  0.85   0.94     0.90   0.83  0.89   0.88 0.63 0.35  0.58\n",
      "mapk2  0.95 0.92  0.75  0.91   0.92     0.68   0.81  0.89   0.90 0.64 0.54  0.80\n",
      "mcr    0.92 0.79  0.61  0.54   0.82     0.55   0.41  0.54   0.84 0.72 0.70  0.72\n",
      "met    0.92 0.94  0.65  0.93   0.83     0.72   0.82  0.81   0.88 0.74 0.59  0.78\n",
      "mk01   0.97 0.96  0.65  0.94   0.84     0.74   0.87  0.86   0.76 0.68 0.61  0.83\n",
      "mk10   0.95 0.89  0.66  0.91   0.85     0.61   0.78  0.75   0.81 0.56 0.57  0.81\n",
      "mk14   0.98 0.91  0.71  0.82   0.72     0.55   0.85  0.74   0.75 0.64 0.51  0.81\n",
      "mmp13  1.00 0.97  0.57  0.76   0.77     0.83   0.63  0.66   0.75 0.66 0.63  0.84\n",
      "mp2k1  0.81 0.73  0.55  0.86   0.75     0.59   0.81  0.54   0.70 0.56 0.68  0.67\n",
      "nos1   0.41 0.60  0.47  0.87   0.75     0.72   0.76  0.59   0.64 0.54 0.60  0.45\n",
      "nram   0.98 0.96  0.33  0.93   0.95     0.92   0.82  0.54   0.94 0.57 0.60  0.32\n",
      "pa2ga  0.91 0.84  0.53  0.82   0.79     0.72   0.72  0.62   0.81 0.62 0.41  0.49\n",
      "parp1  0.92 0.86  0.64  0.80   0.95     0.79   0.84  0.86   0.87 0.58 0.62  0.66\n",
      "pde5a  0.89 0.80  0.65  0.73   0.85     0.69   0.65  0.66   0.81 0.58 0.62  0.65\n",
      "pgh1   0.21 0.39  0.54  0.73   0.76     0.68   0.63  0.64   0.67 0.56 0.46  0.68\n",
      "pgh2   0.45 0.57  0.65  0.77   0.82     0.76   0.75  0.77   0.75 0.57 0.46  0.76\n",
      "plk1   0.93 0.94  0.60  0.90   0.91     0.62   0.81  0.64   0.91 0.60 0.47  0.63\n",
      "pnph   1.00 0.99  0.83  0.94   0.76     0.68   0.74  0.88   0.91 0.52 0.59  0.76\n",
      "ppara  0.95 0.87  0.50  0.85   0.73     0.86   0.75  0.86   0.76 0.62 0.54  0.57\n",
      "ppard  0.96 0.93  0.52  0.84   0.66     0.85   0.68  0.76   0.74 0.54 0.51  0.58\n",
      "pparg  0.92 0.84  0.51  0.84   0.74     0.79   0.67  0.79   0.74 0.71 0.53  0.61\n",
      "prgr   0.90 0.82  0.62  0.64   0.84     0.63   0.68  0.67   0.80 0.67 0.65  0.79\n",
      "ptn1   0.76 0.62  0.58  0.91   0.84     0.85   0.82  0.83   0.81 0.62 0.60  0.65\n",
      "pur2   1.00 0.94  0.87  1.00   0.99     1.00   1.00  0.91   1.00 0.74 0.83  0.75\n",
      "pygm   0.62 0.43  0.42  0.80   0.56     0.55   0.74  0.60   0.51 0.62 0.67  0.55\n",
      "pyrd   0.59 0.67  0.53  0.85   0.88     0.77   0.83  0.83   0.73 0.78 0.49  0.73\n",
      "reni   0.83 0.74  0.62  0.85   0.84     0.90   0.73  0.66   0.75 0.71 0.65  0.48\n",
      "rock1  0.88 0.83  0.65  0.90   0.84     0.74   0.91  0.72   0.85 0.57 0.55  0.69\n",
      "rxra   0.95 0.94  0.65  0.82   0.97     0.84   0.80  0.81   0.93 0.47 0.52  0.93\n",
      "sahh   0.90 0.84  0.32  0.89   0.98     0.91   0.91  0.80   0.75 0.54 0.71  0.15\n",
      "src    0.93 0.93  0.72  0.77   0.74     0.62   0.80  0.65   0.84 0.53 0.55  0.78\n",
      "tgfr1  0.99 0.94  0.71  0.96   0.96     0.77   0.89  0.90   0.93 0.58 0.49  0.57\n",
      "thb    0.92 0.71  0.51  0.91   0.85     0.87   0.83  0.82   0.88 0.61 0.38  0.71\n",
      "thrb   0.99 0.96  0.66  0.85   0.92     0.88   0.88  0.77   0.92 0.81 0.79  0.58\n",
      "try1   0.98 0.95  0.74  0.88   0.91     0.94   0.90  0.80   0.89 0.78 0.72  0.61\n",
      "tryb1  0.73 0.87  0.67  0.90   0.88     0.91   0.86  0.71   0.66 0.72 0.66  0.49\n",
      "tysy   0.89 0.56  0.71  0.94   0.93     0.84   0.84  0.87   0.87 0.68 0.64  0.70\n",
      "urok   0.98 0.95  0.72  0.91   0.92     0.76   0.93  0.77   0.82 0.67 0.60  0.69\n",
      "vgfr2  0.82 0.94  0.70  0.82   0.68     0.69   0.85  0.77   0.83 0.71 0.50  0.73\n",
      "wee1   1.00 0.91  0.49  0.98   1.00     0.95   0.96  0.96   0.96 0.77 0.59  0.78\n",
      "xiap   0.91 0.89  0.58  0.94   0.82     0.90   0.91  0.73   0.89 0.74 0.80  0.41\n",
      "\n",
      "\n",
      "EF1\n",
      "--------------------------------------------------------------------------------------\n",
      "         D-G   D-R  DDTA  Gold  Glide  Surflex  FlexX  Vina  GNINA    RF    NN  Bsl.\n",
      "aa2ar  35.68  0.21  1.66 17.01   5.21    19.71   6.02  2.49   6.43  2.28  0.62  1.24\n",
      "abl1   51.65 41.21  8.24 26.92  17.42    17.03  27.78 13.26  20.44 11.05  2.21  4.40\n",
      "ace    29.08 32.27  0.00 14.54   5.42     2.48   1.42  2.86   4.29  5.71  4.64  0.00\n",
      "aces   20.97  3.31  8.83 22.52   8.13    20.31   1.35 14.57  15.45  3.97  1.55  3.31\n",
      "ada    59.14 40.86  0.00 18.28   5.38     9.68  10.75  1.08  13.98  4.30  0.00  0.00\n",
      "ada17  67.67 65.04  0.56 20.49  28.91    35.90   1.89 28.01  31.58  2.82  3.76  2.07\n",
      "adrb1  22.27  0.00  2.43 27.53  18.70    15.38   9.35  4.05   4.05  0.40  2.02  0.40\n",
      "adrb2  30.74  2.16  2.16 26.41  27.71    20.35  20.00  3.46   6.93  0.87  3.46  1.30\n",
      "akt1   17.75 31.74  4.44 20.48  10.92     4.44   5.17  5.46   4.10  1.37  0.00  1.37\n",
      "akt2   31.62 12.82  3.42 41.03  24.35    14.53  13.68 26.50  11.11  0.85  1.71  4.27\n",
      "aldr    5.66  1.89  0.00 23.90  27.67     7.55  11.95  9.43   8.81  1.26  0.00  1.89\n",
      "ampc   16.67 12.50  0.00  0.00   2.08     0.00   0.00  0.00   2.08  0.00  0.00  0.00\n",
      "andr   32.71 21.56  1.49  1.12  25.70     4.09   0.00 17.84   8.55  1.49  2.60  2.23\n",
      "aofb    2.46  0.00  0.82 16.39   6.67     4.92   5.26  6.56   1.64  0.82  0.00  4.92\n",
      "bace1  50.53 19.43  0.35 25.44  14.13    17.31  11.11  4.59  20.85 18.37  2.12  0.00\n",
      "braf   46.71 42.76  3.95 25.00  29.80     5.26  23.03 17.11  21.71 12.50  0.66  1.32\n",
      "cah2   63.62 63.21  1.42 16.70   3.89     3.05   7.39  0.00  26.22 18.50  3.66 14.02\n",
      "casp3  39.20  0.00  0.00 25.13  27.78    12.56  22.11  1.52  17.17  2.02  4.04  0.00\n",
      "cdk2   44.51 30.80  2.74 15.82  29.87     3.16  11.80  9.51  21.35  0.85  1.48  9.49\n",
      "comt   75.61  4.88  0.00 70.73  83.78     2.44   9.76  4.88   0.00  0.00  4.88  0.00\n",
      "cp2c9  10.00  5.83  1.67  5.83   0.88     3.33   9.09  3.33   5.00  5.00  0.00  0.00\n",
      "cp3a4   2.94  3.53  3.53 12.35  11.80     8.82   4.19  1.80   1.20  3.59  0.00  2.35\n",
      "csf1r  43.37 27.11  4.82 19.28  18.40     3.01  24.54  1.20  19.28   NaN   NaN  4.22\n",
      "cxcr4  30.00  2.50  2.50  2.50   0.00    15.00   0.00  0.00  15.00  0.00  5.00  0.00\n",
      "def    53.92 48.04  0.00 26.47   5.13    32.35   0.00  9.80  31.37  4.90  1.96  0.00\n",
      "dhi1    0.00  2.12  1.21  8.79   7.03     3.64   0.63  4.24  10.91  6.36  1.82  1.21\n",
      "dpp4   30.58  3.38  4.50 20.08  11.84     6.38  11.84  0.56  14.63  0.75  3.19  1.88\n",
      "drd3    7.92  1.04  1.04 11.04   1.69     7.08   2.51  5.02   8.16  3.77  1.26  4.17\n",
      "dyr    57.58 50.22  2.60 48.48  30.87    25.54  10.43  7.36  28.57  1.30  1.30  1.73\n",
      "egfr   62.36 46.49  2.58 25.09  24.95    13.65  19.29  5.72  22.69  2.58  2.77  8.86\n",
      "esr1   53.26  9.14  2.35 24.61  48.78    27.15  31.73 18.54  24.80  0.00  1.31  0.52\n",
      "esr2   50.68 11.72  1.91 17.49  44.26    21.80  31.02 14.99  29.16  0.00  0.27  0.82\n",
      "fa10   31.84 28.86  4.47 34.26  13.74     8.57  33.08 15.83  26.44 13.04  0.93  2.42\n",
      "fa7    56.14 41.23  3.51 45.61  44.04    42.11  55.26 11.40  30.70 18.42  0.00  0.00\n",
      "fabp4  48.94 17.02  0.00 31.91  25.58     2.13  17.02 29.79   0.00  0.00  0.00  0.00\n",
      "fak1   55.00 55.00  2.00 29.00  20.00     5.00  15.00 19.00  27.00  2.00  0.00  3.00\n",
      "fgfr1   1.44  2.16  2.16  3.60   0.00     0.72   0.76  1.44  42.45  1.44  1.44  1.44\n",
      "fkb1a  31.53 11.71  0.00 25.00  41.82     9.91   0.00  7.21   3.60  1.80  0.00  0.90\n",
      "fnta   53.04  2.20  0.84 21.62   6.85     4.94   6.26  2.70   8.45  8.11  1.52  1.18\n",
      "fpps    0.00  0.00  0.00 91.76   1.23     1.18   1.18  0.00  38.82  0.00 78.82  0.00\n",
      "gcr     9.30 15.50  0.78  6.98  16.56    15.12   7.63 15.12  15.50  2.33  2.33  3.88\n",
      "glcm    0.00  3.70  1.85 27.78  11.11    31.48   5.56  0.00   7.41  0.00  1.85  0.00\n",
      "gria2  13.29 16.46  1.90 14.56  33.33     8.23   7.59 10.76  18.35  4.43  8.86  2.53\n",
      "grik1  42.57 35.64  0.99 29.70  20.00     6.93  19.80  4.00  12.00  2.00  4.00  0.99\n",
      "hdac2  55.14 53.51  2.70 16.76  15.84    12.97  16.30 11.89   3.78  2.70  2.16  3.78\n",
      "hdac8  62.94 60.00  1.76 10.00   2.45     5.29  12.94 20.00  16.47  3.53  0.59  5.29\n",
      "hivint  0.00  0.00  0.00  9.00   2.06     3.00   4.00  2.00   0.00  0.00  0.00  0.00\n",
      "hivpr  50.93 18.28  1.49 17.91   9.14     5.04   2.08  4.10  28.36 23.51  0.19  7.09\n",
      "hivrt  14.20  2.37  0.59 12.72  22.94     4.73   2.08  4.15  11.28  0.30  2.08  6.21\n",
      "hmdh   45.29 47.06  0.00 24.12  39.41    17.06   1.76  4.12  36.47  4.71  1.76  0.00\n",
      "hs90a  38.64 15.91  2.27 13.64   1.16     1.14   2.35  0.00  11.36  0.00  2.27  0.00\n",
      "hxk4   19.57  2.17  1.09 14.13  22.83     0.00   4.35  3.26  20.65  0.00  0.00 14.13\n",
      "igf1r  51.35 46.62  3.38 28.38  32.65    10.81  27.70 16.22  30.41 12.84  2.70  1.35\n",
      "inha   20.93  0.00  0.00 25.58   9.30     4.65   9.30  6.98   2.33  0.00  0.00  0.00\n",
      "ital   44.20 10.87  0.00  6.52   8.91     2.90   1.45  0.00   6.57  0.73  0.73  2.90\n",
      "jak2   57.01 47.66  6.54 28.97  33.77    14.95   8.41 15.89  28.04  8.41  2.80  7.48\n",
      "kif11  27.59 16.38  0.00 33.62  38.79     6.03   2.59 24.14  37.93 14.66  0.00  4.31\n",
      "kit    35.54 30.12  4.22 14.46   6.10     3.01   7.88  4.82  13.86  6.63  2.41  4.22\n",
      "kith    0.00  0.00  0.00 15.79  30.77    42.11   0.00 33.33  38.60  1.75  0.00  0.00\n",
      "kpcb   58.52 31.11  0.00 28.89  52.46    27.41  27.48 28.15  23.70  0.00  2.22  6.67\n",
      "lck    50.24 42.14  3.81 15.48  25.12    13.57  18.12 10.00  15.24  3.57  0.71  2.14\n",
      "lkha4  52.63 33.92  0.00 13.45  28.24    18.71   7.60 14.71  15.29  0.00  0.00  1.17\n",
      "mapk2  43.56 25.74 14.85 21.78  27.72     2.97   6.93 13.86  16.83  1.98  1.98  3.96\n",
      "mcr    27.66 22.34  1.06  2.13  29.41     2.13   0.00  7.45  10.64  4.26  9.57  4.26\n",
      "met    68.07 58.43  1.20 39.76  27.71     9.04  18.07 12.05  37.95 19.28  1.20  1.20\n",
      "mk01   59.49 43.04  1.27 20.25  12.66     1.27  21.52  3.80  10.13  5.06  1.27  1.27\n",
      "mk10   44.23 26.92  3.85 23.08  12.50     3.85  14.42  7.69   7.69  4.81  0.96  4.81\n",
      "mk14   57.61 27.16  2.42 14.19  22.15     3.29  14.19  6.57  19.38  7.44  0.52  4.84\n",
      "mmp13  65.56 64.86  1.57 25.70  18.41    12.76   4.20  4.02   9.09  5.59  2.10  9.27\n",
      "mp2k1  28.10 32.23  0.00 23.14  13.22     0.00   8.33  0.00   5.83  0.83  5.00  6.61\n",
      "nos1    3.00  1.00  1.00 20.00  14.29    24.00   7.07  2.00   8.00  0.00  2.00  1.00\n",
      "nram   50.00 41.84  0.00 19.39  41.84    22.45  14.29  0.00  17.35  0.00  0.00  0.00\n",
      "pa2ga  37.37 20.20  2.02 20.20  35.35     2.02  17.53  1.02   9.18  0.00  0.00  0.00\n",
      "parp1  36.22 18.50  0.98 14.17  41.12    19.09  12.40 13.78  16.14  2.56  1.18  0.79\n",
      "pde5a  27.14 13.82  2.76 10.55  21.77     8.79   1.76 11.81  14.32  1.01  1.26  0.50\n",
      "pgh1    0.00  0.00  3.08 14.36  17.11     7.18   5.21  6.15   9.23  1.54  1.03  9.23\n",
      "pgh2    0.92  0.00  5.06 18.62  34.91    14.94  11.08 25.98  17.24  1.84  1.38 12.18\n",
      "plk1   31.78 38.32  0.00 30.84  43.81     3.74   8.41  0.00  38.68  5.66  5.66  0.93\n",
      "pnph   65.05 63.11 19.42 25.24   7.77    10.68  10.68 11.76  35.29  0.00  1.96  3.88\n",
      "ppara  50.13 22.25  1.34 17.43   8.94    12.60   3.77  6.70   4.83  0.80  2.14  0.80\n",
      "ppard  48.75 33.75  3.33 12.08   3.61    10.42   1.25  1.25   4.58  1.25  0.42  1.25\n",
      "pparg  50.00  6.61  1.24 17.36   7.53     4.96   1.66  6.20   5.99  3.10  1.24  1.03\n",
      "prgr   22.87  9.90  3.41  2.39  19.05     4.44   3.64 12.97  18.43  2.39  3.75  7.85\n",
      "ptn1   24.62 18.46  0.77 41.54  15.45    10.77  14.84 26.15  25.38 10.00  4.62  0.00\n",
      "pur2   56.00 12.00  6.00 56.00  54.00    56.00  56.00  4.00  52.00  0.00 12.00  0.00\n",
      "pygm    0.00  1.30  1.30  6.49   1.30     0.00   9.09  2.60   3.90  0.00  9.09  0.00\n",
      "pyrd    0.90  0.00  0.00 31.53  41.18    11.71  29.09 20.72  22.52 11.71  0.00  3.60\n",
      "reni   17.31  3.85  0.00 24.04  33.98    36.54  11.65  4.85   8.74  3.88  0.97  0.00\n",
      "rock1  24.00 25.00  4.00 12.00  25.25     4.00  23.00  7.00  19.00  1.00  1.00  2.00\n",
      "rxra   54.20 24.43  2.29 30.53  39.32    27.48   7.69 29.77  35.88  0.76  0.00 35.11\n",
      "sahh   12.70  0.00  0.00 42.86  43.33    55.56  30.16 22.22   0.00  0.00  0.00  0.00\n",
      "src    59.16 50.38  1.91 14.31  12.21     7.82  14.76  5.53  11.83  1.53  1.72 13.17\n",
      "tgfr1  54.89 31.58 10.53 38.35  44.36    18.05  25.56 10.53  28.57  0.75  0.00  1.50\n",
      "thb    58.25  0.00  0.00 31.07  35.79    16.50  15.53 26.21  34.95  1.94  0.00  1.94\n",
      "thrb   54.66 34.27  2.17 32.75  39.47    27.55  33.11  1.52  35.14  7.38  5.86  0.00\n",
      "try1   54.79 42.09  4.23 42.76  44.19    40.98  50.56  2.67  22.94  9.35  3.34  0.22\n",
      "tryb1  12.84 15.54  1.35 33.78  28.57    19.59  27.89  8.11   2.03  7.43  0.68  0.00\n",
      "tysy   48.62  0.00  3.67 45.87  38.53    23.85  18.35 22.94  30.28  3.67  2.75  0.00\n",
      "urok   53.70 48.77  2.47 43.21  53.70    12.96  47.53  8.02  12.35  2.47  0.00  3.09\n",
      "vgfr2  37.90 36.19  4.65 20.54  17.00     4.65  31.68 17.36  22.74  5.87  0.00  3.67\n",
      "wee1   55.88 16.67  0.00 61.76  59.80    52.94  40.20 54.46  59.41  1.98  0.00  0.98\n",
      "xiap   32.00 16.00  0.00 26.00  50.51    37.00  27.00 10.10  35.35  0.00  4.04  0.00\n",
      "\n",
      "\n",
      "BEDROC80.5\n",
      "--------------------------------------------------------------------------------------\n",
      "        D-G  D-R  DDTA  Gold  Glide  Surflex  FlexX  Vina  GNINA   RF   NN  Bsl.\n",
      "aa2ar  0.53 0.00  0.05  0.29   0.13     0.32   0.11  0.06   0.14 0.04 0.01  0.04\n",
      "abl1   0.73 0.63  0.12  0.46   0.33     0.30   0.48  0.25   0.34 0.19 0.03  0.14\n",
      "ace    0.48 0.52  0.00  0.27   0.09     0.07   0.03  0.04   0.08 0.10 0.10  0.00\n",
      "aces   0.38 0.09  0.14  0.36   0.16     0.34   0.03  0.25   0.27 0.09 0.03  0.07\n",
      "ada    0.85 0.67  0.03  0.32   0.15     0.18   0.20  0.02   0.23 0.10 0.00  0.00\n",
      "ada17  0.95 0.89  0.01  0.34   0.43     0.57   0.04  0.42   0.48 0.07 0.07  0.06\n",
      "adrb1  0.38 0.01  0.04  0.45   0.32     0.26   0.18  0.08   0.08 0.03 0.03  0.01\n",
      "adrb2  0.45 0.03  0.04  0.42   0.47     0.33   0.34  0.09   0.13 0.02 0.06  0.02\n",
      "akt1   0.36 0.58  0.10  0.38   0.21     0.08   0.10  0.13   0.11 0.04 0.01  0.03\n",
      "akt2   0.50 0.28  0.10  0.63   0.42     0.24   0.24  0.39   0.22 0.02 0.05  0.07\n",
      "aldr   0.11 0.04  0.01  0.40   0.43     0.15   0.22  0.16   0.17 0.03 0.01  0.04\n",
      "ampc   0.26 0.23  0.00  0.04   0.09     0.02   0.04  0.02   0.02 0.01 0.01  0.00\n",
      "andr   0.60 0.38  0.02  0.04   0.42     0.08   0.01  0.31   0.17 0.03 0.04  0.05\n",
      "aofb   0.05 0.02  0.02  0.27   0.14     0.10   0.11  0.16   0.04 0.02 0.02  0.07\n",
      "bace1  0.77 0.33  0.02  0.41   0.26     0.29   0.21  0.09   0.32 0.30 0.03  0.01\n",
      "braf   0.71 0.67  0.09  0.39   0.50     0.10   0.41  0.27   0.34 0.20 0.03  0.05\n",
      "cah2   0.92 0.89  0.05  0.29   0.08     0.06   0.13  0.01   0.44 0.30 0.06  0.24\n",
      "casp3  0.63 0.00  0.01  0.49   0.46     0.23   0.38  0.05   0.30 0.04 0.07  0.01\n",
      "cdk2   0.74 0.52  0.07  0.29   0.50     0.07   0.22  0.17   0.37 0.02 0.03  0.17\n",
      "comt   0.86 0.08  0.00  0.82   0.95     0.05   0.15  0.06   0.02 0.00 0.06  0.00\n",
      "cp2c9  0.17 0.09  0.03  0.12   0.04     0.05   0.17  0.05   0.08 0.09 0.01  0.01\n",
      "cp3a4  0.05 0.06  0.05  0.22   0.17     0.14   0.07  0.04   0.03 0.07 0.01  0.06\n",
      "csf1r  0.67 0.44  0.11  0.34   0.31     0.05   0.38  0.05   0.31  NaN  NaN  0.07\n",
      "cxcr4  0.43 0.14  0.03  0.11   0.01     0.26   0.01  0.00   0.33 0.00 0.10  0.01\n",
      "def    0.97 0.82  0.00  0.48   0.14     0.59   0.01  0.20   0.58 0.08 0.04  0.00\n",
      "dhi1   0.00 0.05  0.02  0.16   0.13     0.09   0.01  0.09   0.20 0.12 0.04  0.03\n",
      "dpp4   0.44 0.07  0.08  0.30   0.22     0.10   0.22  0.02   0.23 0.02 0.06  0.04\n",
      "drd3   0.12 0.02  0.02  0.18   0.04     0.13   0.05  0.10   0.14 0.08 0.03  0.08\n",
      "dyr    0.80 0.69  0.07  0.66   0.47     0.40   0.17  0.12   0.44 0.03 0.04  0.04\n",
      "egfr   0.87 0.70  0.05  0.40   0.41     0.22   0.31  0.10   0.35 0.06 0.06  0.14\n",
      "esr1   0.88 0.20  0.04  0.42   0.81     0.47   0.51  0.35   0.43 0.00 0.02  0.03\n",
      "esr2   0.80 0.21  0.04  0.31   0.75     0.39   0.51  0.28   0.46 0.00 0.01  0.03\n",
      "fa10   0.70 0.62  0.11  0.72   0.31     0.24   0.72  0.38   0.61 0.31 0.03  0.06\n",
      "fa7    0.92 0.72  0.07  0.77   0.71     0.73   0.90  0.28   0.52 0.35 0.02  0.00\n",
      "fabp4  0.77 0.33  0.00  0.50   0.48     0.10   0.33  0.43   0.01 0.05 0.00  0.00\n",
      "fak1   0.98 0.93  0.03  0.52   0.33     0.09   0.31  0.33   0.43 0.05 0.00  0.14\n",
      "fgfr1  0.38 0.36  0.44  0.83   0.04     0.28   0.30  1.00   0.66 1.00 1.00  0.30\n",
      "fkb1a  0.55 0.26  0.00  0.39   0.71     0.19   0.01  0.13   0.09 0.03 0.01  0.03\n",
      "fnta   0.67 0.05  0.02  0.32   0.12     0.13   0.11  0.05   0.13 0.11 0.03  0.03\n",
      "fpps   0.01 0.02  0.00  0.97   0.01     0.03   0.07  0.00   0.52 0.00 0.92  0.00\n",
      "gcr    0.16 0.27  0.04  0.13   0.25     0.26   0.13  0.24   0.29 0.05 0.06  0.07\n",
      "glcm   0.01 0.07  0.03  0.40   0.18     0.49   0.16  0.00   0.10 0.00 0.05  0.00\n",
      "gria2  0.23 0.23  0.03  0.23   0.48     0.14   0.13  0.17   0.30 0.07 0.14  0.05\n",
      "grik1  0.63 0.48  0.03  0.47   0.33     0.18   0.38  0.08   0.20 0.03 0.06  0.03\n",
      "hdac2  0.87 0.78  0.06  0.32   0.27     0.26   0.29  0.24   0.09 0.08 0.05  0.09\n",
      "hdac8  0.93 0.86  0.04  0.21   0.10     0.09   0.23  0.34   0.31 0.07 0.02  0.12\n",
      "hivint 0.01 0.01  0.01  0.17   0.04     0.07   0.07  0.05   0.01 0.00 0.00  0.02\n",
      "hivpr  0.73 0.41  0.03  0.28   0.16     0.11   0.04  0.09   0.45 0.36 0.01  0.11\n",
      "hivrt  0.31 0.05  0.01  0.23   0.41     0.09   0.05  0.09   0.22 0.02 0.05  0.10\n",
      "hmdh   0.71 0.70  0.01  0.42   0.69     0.34   0.05  0.10   0.65 0.10 0.04  0.01\n",
      "hs90a  0.69 0.27  0.05  0.26   0.03     0.02   0.03  0.00   0.17 0.00 0.04  0.03\n",
      "hxk4   0.32 0.07  0.05  0.31   0.38     0.03   0.09  0.05   0.31 0.01 0.00  0.27\n",
      "igf1r  0.78 0.69  0.06  0.46   0.50     0.19   0.47  0.25   0.48 0.20 0.04  0.05\n",
      "inha   0.38 0.00  0.00  0.55   0.15     0.19   0.12  0.20   0.04 0.03 0.00  0.01\n",
      "ital   0.69 0.20  0.00  0.12   0.13     0.04   0.03  0.01   0.11 0.05 0.01  0.06\n",
      "jak2   0.90 0.75  0.12  0.49   0.51     0.21   0.16  0.26   0.47 0.13 0.06  0.15\n",
      "kif11  0.46 0.27  0.01  0.54   0.59     0.11   0.09  0.44   0.57 0.29 0.01  0.07\n",
      "kit    0.60 0.52  0.07  0.25   0.11     0.06   0.16  0.10   0.22 0.12 0.03  0.09\n",
      "kith   0.00 0.01  0.00  0.30   0.63     0.70   0.08  0.51   0.61 0.06 0.04  0.00\n",
      "kpcb   0.77 0.50  0.01  0.50   0.69     0.44   0.44  0.46   0.36 0.03 0.04  0.13\n",
      "lck    0.74 0.67  0.07  0.26   0.39     0.22   0.31  0.16   0.26 0.07 0.02  0.06\n",
      "lkha4  0.85 0.57  0.00  0.28   0.57     0.34   0.15  0.29   0.30 0.02 0.00  0.04\n",
      "mapk2  0.70 0.51  0.26  0.43   0.49     0.05   0.20  0.27   0.31 0.04 0.03  0.12\n",
      "mcr    0.51 0.38  0.02  0.05   0.39     0.06   0.01  0.14   0.23 0.11 0.20  0.11\n",
      "met    0.84 0.80  0.03  0.61   0.44     0.15   0.35  0.22   0.56 0.32 0.03  0.05\n",
      "mk01   0.87 0.74  0.05  0.44   0.28     0.04   0.38  0.11   0.19 0.11 0.06  0.02\n",
      "mk10   0.68 0.46  0.10  0.41   0.26     0.07   0.25  0.14   0.13 0.07 0.03  0.09\n",
      "mk14   0.87 0.45  0.06  0.26   0.36     0.06   0.26  0.13   0.32 0.13 0.02  0.12\n",
      "mmp13  0.95 0.88  0.03  0.40   0.30     0.24   0.08  0.08   0.17 0.10 0.04  0.17\n",
      "mp2k1  0.41 0.44  0.01  0.37   0.22     0.02   0.16  0.01   0.13 0.02 0.07  0.14\n",
      "nos1   0.05 0.02  0.04  0.30   0.20     0.33   0.14  0.03   0.13 0.01 0.04  0.02\n",
      "nram   0.80 0.67  0.00  0.39   0.68     0.40   0.21  0.00   0.30 0.01 0.03  0.00\n",
      "pa2ga  0.66 0.43  0.04  0.34   0.59     0.06   0.32  0.02   0.16 0.03 0.02  0.01\n",
      "parp1  0.58 0.33  0.03  0.24   0.67     0.32   0.22  0.26   0.31 0.06 0.02  0.03\n",
      "pde5a  0.42 0.22  0.06  0.18   0.34     0.15   0.04  0.19   0.24 0.02 0.02  0.03\n",
      "pgh1   0.00 0.00  0.07  0.25   0.30     0.16   0.09  0.11   0.17 0.04 0.01  0.19\n",
      "pgh2   0.02 0.01  0.11  0.33   0.60     0.28   0.22  0.45   0.29 0.04 0.02  0.24\n",
      "plk1   0.51 0.60  0.00  0.49   0.62     0.07   0.19  0.03   0.61 0.11 0.09  0.03\n",
      "pnph   0.97 0.88  0.30  0.47   0.16     0.18   0.17  0.20   0.57 0.01 0.04  0.06\n",
      "ppara  0.84 0.41  0.04  0.34   0.21     0.26   0.10  0.14   0.11 0.02 0.04  0.02\n",
      "ppard  0.86 0.62  0.07  0.24   0.11     0.24   0.04  0.05   0.08 0.04 0.01  0.04\n",
      "pparg  0.78 0.18  0.03  0.34   0.17     0.11   0.04  0.13   0.12 0.08 0.03  0.03\n",
      "prgr   0.40 0.19  0.07  0.07   0.22     0.08   0.09  0.22   0.32 0.05 0.09  0.16\n",
      "ptn1   0.42 0.28  0.03  0.66   0.34     0.23   0.28  0.45   0.45 0.20 0.08  0.04\n",
      "pur2   1.00 0.29  0.16  0.95   0.97     0.98   0.98  0.15   0.92 0.00 0.29  0.01\n",
      "pygm   0.00 0.02  0.03  0.19   0.02     0.01   0.13  0.04   0.08 0.00 0.20  0.00\n",
      "pyrd   0.02 0.02  0.00  0.52   0.62     0.25   0.49  0.34   0.34 0.18 0.00  0.07\n",
      "reni   0.26 0.07  0.01  0.44   0.48     0.56   0.21  0.11   0.15 0.06 0.02  0.00\n",
      "rock1  0.35 0.40  0.06  0.24   0.41     0.06   0.39  0.10   0.29 0.02 0.02  0.04\n",
      "rxra   0.83 0.44  0.05  0.52   0.84     0.53   0.24  0.51   0.58 0.02 0.01  0.59\n",
      "sahh   0.23 0.01  0.00  0.69   0.96     0.87   0.58  0.41   0.04 0.00 0.08  0.00\n",
      "src    0.82 0.76  0.06  0.24   0.23     0.14   0.24  0.09   0.27 0.04 0.04  0.21\n",
      "tgfr1  0.82 0.50  0.14  0.60   0.67     0.29   0.43  0.23   0.47 0.02 0.00  0.03\n",
      "thb    0.76 0.00  0.02  0.46   0.61     0.31   0.28  0.43   0.54 0.03 0.02  0.04\n",
      "thrb   0.88 0.59  0.06  0.53   0.65     0.48   0.55  0.07   0.57 0.13 0.12  0.01\n",
      "try1   0.89 0.70  0.09  0.67   0.69     0.68   0.78  0.07   0.43 0.18 0.08  0.01\n",
      "tryb1  0.28 0.29  0.03  0.60   0.47     0.38   0.45  0.16   0.07 0.16 0.03  0.01\n",
      "tysy   0.73 0.00  0.07  0.70   0.64     0.42   0.32  0.36   0.49 0.08 0.05  0.03\n",
      "urok   0.80 0.73  0.06  0.66   0.79     0.23   0.70  0.15   0.25 0.05 0.02  0.08\n",
      "vgfr2  0.59 0.60  0.08  0.35   0.30     0.08   0.50  0.28   0.38 0.12 0.01  0.08\n",
      "wee1   0.93 0.25  0.00  0.89   0.99     0.78   0.65  0.81   0.84 0.08 0.01  0.02\n",
      "xiap   0.56 0.29  0.00  0.50   0.77     0.64   0.58  0.16   0.59 0.01 0.09  0.00\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "pd.options.display.float_format = '{:,.2f}'.format\n",
    "pd.options.display.max_rows = 120\n",
    "pd.set_option('display.expand_frame_repr', False) \n",
    "\n",
    "perf_metrics = ['AUROC', 'EF1', 'BEDROC80.5']\n",
    "columns_order = ['DENVIS-G', 'DENVIS-R', 'Deep DTA', 'Gold', 'Glide', 'Surflex', 'Flexx', 'Vina', 'GNINA', 'RF-score', 'NN-score', 'Ligand bsl.']\n",
    "columns_alias = {\n",
    "    'DENVIS-G': 'D-G',\n",
    "    'DENVIS-R': 'D-R',\n",
    "    'GNINA': 'GNINA',\n",
    "    'Vina': 'Vina',\n",
    "    'RF-score': 'RF',\n",
    "    'NN-score': 'NN',\n",
    "    'Deep DTA': 'DDTA',\n",
    "    'Ligand bsl.': 'Bsl.',\n",
    "    'Gold': 'Gold',\n",
    "    'Glide': 'Glide',\n",
    "    'Surflex': 'Surflex',\n",
    "    'Flexx': 'FlexX'}\n",
    "\n",
    "for metric in perf_metrics:\n",
    "    print_df = pd.pivot(results_pt_tf, index='Target', columns='Model', values=metric)\n",
    "    print_df.columns.name, print_df.index.name = None, None\n",
    "    print_df = print_df[columns_order].rename(columns=columns_alias)\n",
    "    print(metric)\n",
    "    print('--------------------------------------------------------------------------------------')\n",
    "    print(print_df)\n",
    "    print('\\n')"
   ]
  }
 ],
 "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": "3.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}