04-py-ml-02-modeling.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "3b9af74e-3fd3-41a5-b2ea-bbc6879a113f",
   "metadata": {},
   "source": [
    "## Import packages"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "ebade665",
   "metadata": {},
   "outputs": [],
   "source": [
    "#!pip3 install numpy=1.20.2\n",
    "#!pip3 install pandas=1.2.4\n",
    "#!pip3 install pillow=8.1.2\n",
    "#!pip3 install scipy=1.6.3\n",
    "#!pip3 install rdkit=2022.3.4\n",
    "#!pip3 install scikit-learn=0.24.2\n",
    "#!pip3 install matplotlib=3.4.1\n",
    "#!pip3 install seaborn=0.11.1\n",
    "#!pip3 install shap=0.41.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "7d909141-9b45-4941-848e-c9693623a8bb",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.tree import DecisionTreeRegressor \n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn.model_selection import KFold\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.model_selection import RandomizedSearchCV\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import cross_validate\n",
    "import scipy\n",
    "import multiprocessing\n",
    "import pickle\n",
    "from sklearn import metrics\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "ac2278c5-032e-40e0-ba76-671eefa40cec",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "12"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "multiprocessing.cpu_count()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "11b5cd4b-1432-42e2-862e-6dd8de898aaf",
   "metadata": {},
   "source": [
    "## Define function to prepare data, train/test the ML models and evaluate them"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "0ea6f548-2d25-45ce-b9a6-677d004cbd21",
   "metadata": {},
   "outputs": [],
   "source": [
    "def prepareDataFrame(df, categorical_features):\n",
    "    \n",
    "    for ft in categorical_features:\n",
    "        \n",
    "        cat_ft = pd.get_dummies(df[ft], prefix='cat')\n",
    "        df = pd.concat([df,cat_ft], axis=1)\n",
    "        df = df.drop(ft, 1)\n",
    "    \n",
    "        \n",
    "    print(\"Dataframe after dummy variables:\")\n",
    "    print(df.shape)\n",
    "\n",
    "    Y = df[['ba']] \n",
    "    Y = Y.values.ravel()\n",
    "    \n",
    "    X = df.drop('ba', axis=1)\n",
    "    \n",
    "    return X, Y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "1e51c21e-b00d-4ebc-9567-a5afac37c6dd",
   "metadata": {},
   "outputs": [],
   "source": [
    "def evaluate(model, test_features, test_labels, model_type):\n",
    "    \n",
    "    predictions = model.predict(test_features)\n",
    "    \n",
    "    pred_dict = {'observed':test_labels,'predicted':predictions}\n",
    "    \n",
    "    pred_df = pd.DataFrame(pred_dict)\n",
    "    pred_df.to_csv(\"work/predictions/\"+model_type+\".csv\", encoding=\"utf-8\", index=False)\n",
    "    \n",
    "    pearson_r = scipy.stats.pearsonr(test_labels, predictions)\n",
    "    print('Pearson R: ', pearson_r)\n",
    "    print('Pearson R squared: ', pearson_r[0]**2)\n",
    "\n",
    "    r2 = metrics.r2_score(test_labels, predictions)\n",
    "    mae = metrics.mean_absolute_error(test_labels, predictions)\n",
    "    mse = metrics.mean_squared_error(test_labels, predictions)\n",
    "    rmse = np.sqrt(metrics.mean_squared_error(test_labels, predictions))\n",
    "    rae = metrics.mean_absolute_error(test_labels, predictions) / metrics.mean_absolute_error(test_labels, [test_labels.mean()] * len(test_labels))\n",
    "    rrse = np.sqrt(metrics.mean_squared_error(test_labels, predictions)) / np.sqrt(metrics.mean_squared_error(test_labels, [test_labels.mean()] * len(test_labels)))\n",
    "    \n",
    "    print('R Squared Error:', r2)\n",
    "    print('Mean Absolute Error:', mae)\n",
    "    print('Mean Squared Error:', mse)\n",
    "    print('Root Mean Squared Error:', rmse)\n",
    "    print('Relative absolute Error:', rae)\n",
    "    print('Root relateive squared Error:', rrse)\n",
    "    \n",
    "    return [r2, mae, mse, rmse, rae, rrse]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "f74173b8-004e-4bfb-b409-e2be3bddf567",
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_test(model_type):\n",
    "    \n",
    "    categorical_features = []\n",
    "    \n",
    "    if(model_type == \"dual_model_pl_split_random\"):\n",
    "     \n",
    "        train = pd.read_csv(\"work/training/dual-model/train-random.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "        test = pd.read_csv(\"work/training/dual-model/test-random.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "        \n",
    "    elif(model_type == \"dual_model_pl_split_protein\"):\n",
    "     \n",
    "        train = pd.read_csv(\"work/training/dual-model/train-protein.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "        test = pd.read_csv(\"work/training/dual-model/test-protein.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "        \n",
    "    elif(model_type == \"dual_model_pl_split_pocket\"):\n",
    "     \n",
    "        train = pd.read_csv(\"work/training/dual-model/train-pocket.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "        test = pd.read_csv(\"work/training/dual-model/test-pocket.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "                \n",
    "    elif(model_type == \"dual_model_pl_split_ligand_weight\"):\n",
    "     \n",
    "        train = pd.read_csv(\"work/training/dual-model/train-ligand-weight.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "        test = pd.read_csv(\"work/training/dual-model/test-ligand-weight.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "        \n",
    "    elif(model_type == \"dual_model_pl_split_ligand_diversity\"):\n",
    "     \n",
    "        train = pd.read_csv(\"work/training/dual-model/train-ligand-diversity.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "        test = pd.read_csv(\"work/training/dual-model/test-ligand-diversity.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "                \n",
    "    elif(model_type == \"dual_model_pl_split_ligand_tpsa\"):\n",
    "     \n",
    "        train = pd.read_csv(\"work/training/dual-model/train-ligand-tpsa.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "        test = pd.read_csv(\"work/training/dual-model/test-ligand-tpsa.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "                \n",
    "    elif(model_type == \"dual_model_pl_split_ligand_volume\"):\n",
    "     \n",
    "        train = pd.read_csv(\"work/training/dual-model/train-ligand-volume.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "        test = pd.read_csv(\"work/training/dual-model/test-ligand-volume.csv\", sep=\",\", encoding=\"utf-8\")    \n",
    "        \n",
    "    print(train.shape)\n",
    "    print(test.shape)\n",
    "    \n",
    "    train_X, train_Y = prepareDataFrame(train, categorical_features)\n",
    "    test_X, test_Y = prepareDataFrame(test, categorical_features)\n",
    "    \n",
    "    train_X = np.array(train_X)\n",
    "\n",
    "    num_of_features = round(train_X.shape[1]/2)\n",
    "    \n",
    "    regressor = RandomForestRegressor(n_jobs=-1, max_features=num_of_features, n_estimators=500, random_state = 123, criterion= 'mse') \n",
    "\n",
    "    kf = KFold(n_splits=5, random_state=123, shuffle=True)\n",
    "    \n",
    "    RMSE_sum=0\n",
    "    RMSE_length=5\n",
    "    MSE_sum=0\n",
    "    MSE_length=5\n",
    "    MAE_sum=0\n",
    "    MAE_length=5\n",
    "    R2_sum=0\n",
    "    R2_length=5\n",
    "    \n",
    "    for loop_number, (train_k, test_k) in enumerate(kf.split(train_X)):\n",
    "\n",
    "        ## Get Training Matrix and Y Vector\n",
    "\n",
    "        training_X_array = train_X[train_k]\n",
    "        training_y_array = train_Y[train_k].reshape(-1, 1)\n",
    "\n",
    "        training_y_array = training_y_array.ravel()\n",
    "        \n",
    "        ## Get Testing Matrix and Y vector\n",
    "\n",
    "        X_test_array = train_X[test_k]\n",
    "        y_actual_values = train_Y[test_k]\n",
    "\n",
    "        ## Fit the Random Forest Regression Model\n",
    "\n",
    "        lr_model = regressor.fit(training_X_array, training_y_array)\n",
    "\n",
    "        ## Compute the predictions for the test data\n",
    "\n",
    "        prediction = lr_model.predict(X_test_array)      \n",
    "        predicted_Y = np.array(prediction)   \n",
    "\n",
    "        ## Calculate the RMSE\n",
    "\n",
    "        RMSE_cross_fold = np.sqrt(metrics.mean_squared_error(y_actual_values, predicted_Y))\n",
    "        RMSE_sum = RMSE_cross_fold + RMSE_sum\n",
    "        \n",
    "        MSE_cross_fold = metrics.mean_squared_error(y_actual_values, predicted_Y)\n",
    "        MSE_sum = MSE_cross_fold + MSE_sum\n",
    "        \n",
    "        MAE_cross_fold = metrics.mean_absolute_error(y_actual_values, predicted_Y)\n",
    "        MAE_sum = MAE_cross_fold + MAE_sum\n",
    "        \n",
    "        R2_cross_fold = metrics.r2_score(y_actual_values, predicted_Y)\n",
    "        R2_sum = R2_cross_fold + R2_sum\n",
    "\n",
    "    ## Calculate the average and print    \n",
    "\n",
    "    RMSE_cross_fold_avg = RMSE_sum / RMSE_length\n",
    "    MSE_cross_fold_avg = MSE_sum / MSE_length\n",
    "    MAE_cross_fold_avg = MAE_sum / MAE_length\n",
    "    R2_cross_fold_avg = R2_sum / R2_length\n",
    "\n",
    "    print('The Mean R2 across all folds is: ', R2_cross_fold_avg)\n",
    "    print('The Mean MAE across all folds is: ', MAE_cross_fold_avg)\n",
    "    print('The Mean MSE across all folds is: ', MSE_cross_fold_avg)\n",
    "    print('The Mean RMSE across all folds is: ', RMSE_cross_fold_avg)\n",
    "    \n",
    "    # Now, lets train the model on the whole dataset\n",
    "    \n",
    "    regressor.fit(train_X, train_Y)\n",
    "    \n",
    "    print(\"Evaluation with the test set: \")\n",
    "    final_metrics = evaluate(regressor, test_X, test_Y, model_type)\n",
    "    \n",
    "    final_metrics = [R2_cross_fold_avg, MAE_cross_fold_avg, MSE_cross_fold_avg, RMSE_cross_fold_avg] + final_metrics\n",
    "    \n",
    "    pickle.dump(regressor, open(\"work/models/\"+model_type+\".mdl\", 'wb'))\n",
    "    \n",
    "    return final_metrics, regressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "ad647e9a-181d-4dee-bb9f-d972899cef67",
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_test_mutation_model(model_type):\n",
    "\n",
    "    rf_model_ba = pickle.load(open(\"work/models/dual_model_pl_split_ligand_weight.mdl\", 'rb'))\n",
    "\n",
    "    train_pl = pd.read_csv(\"work/training/dual-model/train-ligand-weight.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "    test_pl = pd.read_csv(\"work/training/dual-model/test-ligand-weight.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "\n",
    "    train_meta_pl = pd.read_csv(\"work/training/dual-model/train-meta-ligand-weight.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "    test_meta_pl = pd.read_csv(\"work/training/dual-model/test-meta-ligand-weight.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "        \n",
    "        \n",
    "    train_pl = pd.concat([train_meta_pl, train_pl], axis=1)\n",
    "    test_pl = pd.concat([test_meta_pl, test_pl], axis=1)\n",
    "\n",
    "    train_test_pl = pd.concat([train_pl, test_pl])\n",
    "    \n",
    "    \n",
    "    pl_Y = train_test_pl[['ba']] \n",
    "    pl_Y = pl_Y.values.ravel()\n",
    "    \n",
    "    pl_X = train_test_pl.drop(['pdb', 'ligand_file', 'chembl_id', 'tanimoto_index', 'ba'], axis=1)\n",
    "    \n",
    "    ba_predictions = rf_model_ba.predict(pl_X)\n",
    "    \n",
    "    train_test_pl['predicted_wt_ba'] = ba_predictions\n",
    "    \n",
    "    train_test_pl.rename(columns = {'ba':'wt_pl_ba'}, inplace = True)\n",
    "    \n",
    "    \n",
    "    train_mut = pd.read_csv(\"work/training/dual-model/train-mutations-random.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "    test_mut = pd.read_csv(\"work/training/dual-model/test-mutations-random.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "    \n",
    "    train_meta_mut = pd.read_csv(\"work/training/dual-model/train-mutations-meta-random.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "    test_meta_mut = pd.read_csv(\"work/training/dual-modelf/test-mutations-meta-random.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "    \n",
    "    train_mut = pd.concat([train_meta_mut, train_mut], axis=1)\n",
    "    test_mut = pd.concat([test_meta_mut, test_mut], axis=1)\n",
    "\n",
    "    # real wt_ba training\n",
    "    train_mut_real_ba = train_mut.copy()\n",
    "    \n",
    "    train_mut_real_ba_cols = list(train_mut_real_ba.columns.values)\n",
    "    train_mut_real_ba_cols.pop(train_mut_real_ba_cols.index('FoldXname.x'))\n",
    "    train_mut_real_ba_cols.pop(train_mut_real_ba_cols.index('ligand_file'))\n",
    "    train_mut_real_ba_cols.pop(train_mut_real_ba_cols.index('FoldXmutation'))\n",
    "    train_mut_real_ba = train_mut_real_ba[train_mut_real_ba_cols] \n",
    "\n",
    "    # predicted wt_ba train set\n",
    "    train_mut_predicted_ba = train_mut.merge(train_test_pl[['ligand_file', 'predicted_wt_ba']], on='ligand_file', how='left')\n",
    "    \n",
    "    train_mut_predicted_ba_cols = list(train_mut_predicted_ba.columns.values)\n",
    "    train_mut_predicted_ba_cols.pop(train_mut_predicted_ba_cols.index('FoldXname.x'))\n",
    "    train_mut_predicted_ba_cols.pop(train_mut_predicted_ba_cols.index('ligand_file'))\n",
    "    train_mut_predicted_ba_cols.pop(train_mut_predicted_ba_cols.index('FoldXmutation'))\n",
    "    train_mut_predicted_ba_cols.pop(train_mut_predicted_ba_cols.index('wt_ba'))\n",
    "    train_mut_predicted_ba_cols.pop(train_mut_predicted_ba_cols.index('ba'))\n",
    "    train_mut_predicted_ba = train_mut_predicted_ba[train_mut_predicted_ba_cols + ['ba']] \n",
    "     \n",
    "    # real wt_ba test set\n",
    "    test_mut_real_ba = test_mut.copy()\n",
    "    \n",
    "    test_mut_real_ba_cols = list(test_mut_real_ba.columns.values)\n",
    "    test_mut_real_ba_cols.pop(test_mut_real_ba_cols.index('FoldXname.x'))\n",
    "    test_mut_real_ba_cols.pop(test_mut_real_ba_cols.index('ligand_file'))\n",
    "    test_mut_real_ba_cols.pop(test_mut_real_ba_cols.index('FoldXmutation'))\n",
    "    test_mut_real_ba = test_mut_real_ba[test_mut_real_ba_cols] \n",
    "    \n",
    "    # predicted wt_ba test set\n",
    "    test_mut_predicted_ba = test_mut.merge(train_test_pl[['ligand_file', 'predicted_wt_ba']], on='ligand_file', how='left')\n",
    "    \n",
    "    test_mut_predicted_ba_cols = list(test_mut_predicted_ba.columns.values)\n",
    "    test_mut_predicted_ba_cols.pop(test_mut_predicted_ba_cols.index('FoldXname.x'))\n",
    "    test_mut_predicted_ba_cols.pop(test_mut_predicted_ba_cols.index('ligand_file'))\n",
    "    test_mut_predicted_ba_cols.pop(test_mut_predicted_ba_cols.index('FoldXmutation'))\n",
    "    test_mut_predicted_ba_cols.pop(test_mut_predicted_ba_cols.index('wt_ba'))\n",
    "    test_mut_predicted_ba_cols.pop(test_mut_predicted_ba_cols.index('ba'))\n",
    "    test_mut_predicted_ba = test_mut_predicted_ba[test_mut_predicted_ba_cols + ['ba']] \n",
    "    \n",
    "\n",
    "    categorical_features = []\n",
    "    \n",
    "    print(train_mut_real_ba.shape)\n",
    "    print(test_mut_real_ba.shape)\n",
    "    \n",
    "    train_real_X, train_real_Y = prepareDataFrame(train_mut_real_ba, categorical_features)\n",
    "    train_pred_X, train_pred_Y = prepareDataFrame(train_mut_predicted_ba, categorical_features)\n",
    "\n",
    "    test_real_X, test_real_Y = prepareDataFrame(test_mut_real_ba, categorical_features)\n",
    "    test_pred_X, test_pred_Y = prepareDataFrame(test_mut_predicted_ba, categorical_features)\n",
    "\n",
    "    num_of_features = round(train_real_X.shape[1]/2)\n",
    "    \n",
    "    regressor = RandomForestRegressor(n_jobs=-1, max_features=num_of_features, n_estimators=500, random_state = 123, criterion= 'mse') \n",
    "\n",
    "    kf = KFold(n_splits=5, random_state=123, shuffle=True)\n",
    "    \n",
    "    RMSE_sum=0\n",
    "    RMSE_length=5\n",
    "    MSE_sum=0\n",
    "    MSE_length=5\n",
    "    MAE_sum=0\n",
    "    MAE_length=5\n",
    "    R2_sum=0\n",
    "    R2_length=5\n",
    "    \n",
    "    if(model_type == \"mutation_model_real_wt_ba\"):\n",
    "        train_X = np.array(train_real_X)\n",
    "        train_Y = train_real_Y\n",
    "    elif(model_type == \"mutation_model_pred_wt_ba\"):\n",
    "        train_X = np.array(train_pred_X)\n",
    "        train_Y = train_pred_Y\n",
    "    \n",
    "    for loop_number, (train_k, test_k) in enumerate(kf.split(train_X)):\n",
    "\n",
    "        ## Get Training Matrix and Y Vector\n",
    "\n",
    "        training_X_array = train_X[train_k]\n",
    "        training_y_array = train_Y[train_k].reshape(-1, 1)\n",
    "\n",
    "        training_y_array = training_y_array.ravel()\n",
    "        \n",
    "        ## Get Testing Matrix and Y vector\n",
    "\n",
    "        X_test_array = train_X[test_k]\n",
    "        y_actual_values = train_Y[test_k]\n",
    "\n",
    "        ## Fit the Random Forest Regression Model\n",
    "\n",
    "        lr_model = regressor.fit(training_X_array, training_y_array)\n",
    "\n",
    "        ## Compute the predictions for the test data\n",
    "\n",
    "        prediction = lr_model.predict(X_test_array)      \n",
    "        predicted_Y = np.array(prediction)   \n",
    "\n",
    "        ## Calculate the RMSE\n",
    "\n",
    "        RMSE_cross_fold = np.sqrt(metrics.mean_squared_error(y_actual_values, predicted_Y))\n",
    "        RMSE_sum = RMSE_cross_fold + RMSE_sum\n",
    "        \n",
    "        MSE_cross_fold = metrics.mean_squared_error(y_actual_values, predicted_Y)\n",
    "        MSE_sum = MSE_cross_fold + MSE_sum\n",
    "        \n",
    "        MAE_cross_fold = metrics.mean_absolute_error(y_actual_values, predicted_Y)\n",
    "        MAE_sum = MAE_cross_fold + MAE_sum\n",
    "        \n",
    "        R2_cross_fold = metrics.r2_score(y_actual_values, predicted_Y)\n",
    "        R2_sum = R2_cross_fold + R2_sum\n",
    "\n",
    "    ## Calculate the average and print    \n",
    "\n",
    "    RMSE_cross_fold_avg = RMSE_sum / RMSE_length\n",
    "    MSE_cross_fold_avg = MSE_sum / MSE_length\n",
    "    MAE_cross_fold_avg = MAE_sum / MAE_length\n",
    "    R2_cross_fold_avg = R2_sum / R2_length\n",
    "\n",
    "    print('The Mean R2 across all folds is: ', R2_cross_fold_avg)\n",
    "    print('The Mean MAE across all folds is: ', MAE_cross_fold_avg)\n",
    "    print('The Mean MSE across all folds is: ', MSE_cross_fold_avg)\n",
    "    print('The Mean RMSE across all folds is: ', RMSE_cross_fold_avg)\n",
    "    print('\\n')\n",
    "    \n",
    "    # Now, lets train the model on the whole dataset\n",
    "    \n",
    "    regressor.fit(train_X, train_Y)\n",
    "    \n",
    "    print(\"Evaluation with the test set (real wt_bt): \")\n",
    "    final_metrics_real = evaluate(regressor, test_real_X, test_real_Y, model_type+\"_test_real_wt_ba\")\n",
    "    \n",
    "    final_metrics_real = [R2_cross_fold_avg, MAE_cross_fold_avg, MSE_cross_fold_avg, RMSE_cross_fold_avg] + final_metrics_real\n",
    "    \n",
    "    print('\\n')\n",
    "    print(\"Evaluation with the test set (predicted wt_bt): \")\n",
    "    final_metrics_pred = evaluate(regressor, test_pred_X, test_pred_Y, model_type+\"_test_pred_wt_ba\")\n",
    "    \n",
    "    final_metrics_pred = [R2_cross_fold_avg, MAE_cross_fold_avg, MSE_cross_fold_avg, RMSE_cross_fold_avg] + final_metrics_pred\n",
    "    \n",
    "    pickle.dump(regressor, open(\"work/models/\"+model_type+\".mdl\", 'wb'))\n",
    "    \n",
    "    return final_metrics_real, final_metrics_pred, regressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "041be870-a16f-420b-9623-53df68ec3afb",
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_test_y_randomization():\n",
    "    \n",
    "    for i in range(1,11):\n",
    "        \n",
    "        train = pd.read_csv(\"work/training/randomization/randomization_y\"+str(i)+\".csv\", sep=\",\", encoding=\"utf-8\")\n",
    "        test = pd.read_csv(\"work/training/randomization/randomization_y\"+str(i)+\".csv\", sep=\",\", encoding=\"utf-8\") \n",
    "        \n",
    "        print(train.shape)\n",
    "        print(test.shape)\n",
    "\n",
    "        categorical_features = []\n",
    "\n",
    "        train_Y = train[['y_scramble_'+str(i)]] \n",
    "        train_Y = train_Y.values.ravel()\n",
    "        train_X = train.drop('y_scramble_'+str(i), axis=1)\n",
    "        \n",
    "        test_Y = test[['y_scramble_'+str(i)]] \n",
    "        test_Y = test_Y.values.ravel()\n",
    "        test_X = test.drop('y_scramble_'+str(i), axis=1)\n",
    "        \n",
    "\n",
    "        train_X = np.array(train_X)\n",
    "\n",
    "        num_of_features = round(train_X.shape[1]/2)\n",
    "\n",
    "        regressor = RandomForestRegressor(n_jobs=-1, max_features=num_of_features, n_estimators=500, random_state = 123, criterion= 'squared_error') \n",
    "\n",
    "        kf = KFold(n_splits=5, random_state=123, shuffle=True)\n",
    "\n",
    "        RMSE_sum=0\n",
    "        RMSE_length=5\n",
    "        MSE_sum=0\n",
    "        MSE_length=5\n",
    "        MAE_sum=0\n",
    "        MAE_length=5\n",
    "        R2_sum=0\n",
    "        R2_length=5\n",
    "\n",
    "        for loop_number, (train_k, test_k) in enumerate(kf.split(train_X)):\n",
    "\n",
    "            ## Get Training Matrix and Y Vector\n",
    "\n",
    "            training_X_array = train_X[train_k]\n",
    "            training_y_array = train_Y[train_k].reshape(-1, 1)\n",
    "\n",
    "            training_y_array = training_y_array.ravel()\n",
    "\n",
    "            ## Get Testing Matrix and Y vector\n",
    "\n",
    "            X_test_array = train_X[test_k]\n",
    "            y_actual_values = train_Y[test_k]\n",
    "\n",
    "            ## Fit the Random Forest Regression Model\n",
    "\n",
    "            lr_model = regressor.fit(training_X_array, training_y_array)\n",
    "\n",
    "            ## Compute the predictions for the test data\n",
    "\n",
    "            prediction = lr_model.predict(X_test_array)      \n",
    "            predicted_Y = np.array(prediction)   \n",
    "\n",
    "            ## Calculate the RMSE\n",
    "\n",
    "            RMSE_cross_fold = np.sqrt(metrics.mean_squared_error(y_actual_values, predicted_Y))\n",
    "            RMSE_sum = RMSE_cross_fold + RMSE_sum\n",
    "\n",
    "            MSE_cross_fold = metrics.mean_squared_error(y_actual_values, predicted_Y)\n",
    "            MSE_sum = MSE_cross_fold + MSE_sum\n",
    "\n",
    "            MAE_cross_fold = metrics.mean_absolute_error(y_actual_values, predicted_Y)\n",
    "            MAE_sum = MAE_cross_fold + MAE_sum\n",
    "\n",
    "            R2_cross_fold = metrics.r2_score(y_actual_values, predicted_Y)\n",
    "            R2_sum = R2_cross_fold + R2_sum\n",
    "\n",
    "        ## Calculate the average and print    \n",
    "\n",
    "        RMSE_cross_fold_avg = RMSE_sum / RMSE_length\n",
    "        MSE_cross_fold_avg = MSE_sum / MSE_length\n",
    "        MAE_cross_fold_avg = MAE_sum / MAE_length\n",
    "        R2_cross_fold_avg = R2_sum / R2_length\n",
    "\n",
    "        print('The Mean R2 across all folds is: ', R2_cross_fold_avg)\n",
    "        print('The Mean MAE across all folds is: ', MAE_cross_fold_avg)\n",
    "        print('The Mean MSE across all folds is: ', MSE_cross_fold_avg)\n",
    "        print('The Mean RMSE across all folds is: ', RMSE_cross_fold_avg)\n",
    "        print('\\n')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "4351ef7c-cd46-43ca-a20e-e2848b0a2c67",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_df_feature_importance(importance,names,model_type, top=20, execlude=None):\n",
    "\n",
    "    feature_importance = np.array(importance)\n",
    "    feature_names = np.array(names)\n",
    "\n",
    "    data={'feature_names':feature_names,'feature_importance':feature_importance}\n",
    "    ft_df = pd.DataFrame(data)\n",
    "\n",
    "    ft_df.sort_values(by=['feature_importance'], ascending=False,inplace=True)\n",
    "    \n",
    "    ft_df = ft_df.iloc[0:top]\n",
    "    \n",
    "    if(execlude is not None):\n",
    "        ft_df = ft_df[ft_df.feature_names != execlude]\n",
    "\n",
    "    return ft_df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "825ebcdf-6087-4b61-94c0-02d0851ca21f",
   "metadata": {},
   "source": [
    "## Train/Test ML models to predict binding affinity from protein, binding-site and ligand features"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3ef9e751-5b7a-45f2-9d2f-06f45031af95",
   "metadata": {},
   "source": [
    "In the dual model approach in the paper, a separate model was proposed to predict the binding affinity of wild-type (WT) protein-ligand pairs. Then, another model will be developed to use the predicted WT protein-ligand binding affinity besides mutation features to predict the binding affinity of mutated protein-ligand pair.\n",
    "The below is to train/test a model to predict WT protein-ligand binding affinity using protein, binding-site and ligand featues. Different splitting scenarios were tested for this model (split by protein similarity, binding pocket similarity, ligand similarity and stratified random splitting by binding affinity, ligand molecular weight, ligand surface area and ligand volume)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "id": "af8f4c34-3c4b-4263-bf6a-91d217e55903",
   "metadata": {},
   "outputs": [],
   "source": [
    "final_metrics_df = pd.DataFrame()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "id": "03169c5b-ffd8-435d-bc39-3e3c394f4047",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0, 0)"
      ]
     },
     "execution_count": 147,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_metrics_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "id": "687f1493-d10f-4693-b917-6ea1e43275a3",
   "metadata": {},
   "outputs": [],
   "source": [
    "final_metrics_df['metric'] = [\"R2 (5-fold cross-validation)\", \"MAE (5-fold cross-validation)\", \"MSE (5-fold cross-validation)\", \"RMSE (5-fold cross-validation)\",\n",
    "                              \"R2 (test set)\", \"MAE (test set)\", \"MSE (test set)\", \"RMSE (test set)\", \n",
    "                              \"Relative absolute error (test set)\", \"Root relative squared error (test set)\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "id": "1b2c850d-91fe-4cf1-8659-1b24fc1edf82",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10, 1)"
      ]
     },
     "execution_count": 149,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_metrics_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "id": "b64918b1-0d29-4cec-ae72-8a6ca619040e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(6738, 126)\n",
      "(1683, 126)\n",
      "Dataframe after dummy variables:\n",
      "(6738, 126)\n",
      "Dataframe after dummy variables:\n",
      "(1683, 126)\n",
      "The Mean R2 across all folds is:  0.8509576779588084\n",
      "The Mean MAE across all folds is:  0.4197126871011304\n",
      "The Mean MSE across all folds is:  0.33699007424893235\n",
      "The Mean RMSE across all folds is:  0.580365995865205\n",
      "Evaluation with the test set: \n",
      "Pearson R:  (0.9321436427463227, 0.0)\n",
      "Pearson R squared:  0.868891770712384\n",
      "R Squared Error: 0.8671290619537905\n",
      "Mean Absolute Error: 0.40136244800950693\n",
      "Mean Squared Error: 0.30246556922162887\n",
      "Root Mean Squared Error: 0.5499686984016716\n",
      "Relative absolute Error: 0.326883818580622\n",
      "Root relateive squared Error: 0.36451466094823887\n"
     ]
    }
   ],
   "source": [
    "final_metrics, reg_model = train_test(\"dual_model_pl_split_random\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "id": "292ab66e-c65d-4666-be68-37acead43af4",
   "metadata": {},
   "outputs": [],
   "source": [
    "final_metrics_df['random_split'] = pd.Series(final_metrics)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "id": "504d2ca3-192e-46b5-b456-014834ff8f55",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(6969, 126)\n",
      "(1452, 126)\n",
      "Dataframe after dummy variables:\n",
      "(6969, 126)\n",
      "Dataframe after dummy variables:\n",
      "(1452, 126)\n",
      "The Mean R2 across all folds is:  0.8692430070804479\n",
      "The Mean MAE across all folds is:  0.4063900384892287\n",
      "The Mean MSE across all folds is:  0.3172436039732673\n",
      "The Mean RMSE across all folds is:  0.5631445867101779\n",
      "Evaluation with the test set: \n",
      "Pearson R:  (0.5318876117837603, 7.726084104267749e-107)\n",
      "Pearson R squared:  0.2829044315690321\n",
      "R Squared Error: 0.2485707648145905\n",
      "Mean Absolute Error: 0.8591214876033066\n",
      "Mean Squared Error: 1.1148497339393952\n",
      "Root Mean Squared Error: 1.0558644486577788\n",
      "Relative absolute Error: 0.8500956445633928\n",
      "Root relateive squared Error: 0.8668501803572572\n"
     ]
    }
   ],
   "source": [
    "final_metrics, reg_model = train_test(\"dual_model_pl_split_pocket\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "id": "4f797fbd-c5f2-4418-a815-4f59e494f6fb",
   "metadata": {},
   "outputs": [],
   "source": [
    "final_metrics_df['pocket_split'] = pd.Series(final_metrics)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "id": "ee8f1370-4524-450c-8b0f-ac0e0907d087",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(6731, 126)\n",
      "(1690, 126)\n",
      "Dataframe after dummy variables:\n",
      "(6731, 126)\n",
      "Dataframe after dummy variables:\n",
      "(1690, 126)\n",
      "The Mean R2 across all folds is:  0.8610701844281682\n",
      "The Mean MAE across all folds is:  0.4143660251221414\n",
      "The Mean MSE across all folds is:  0.3267372232559859\n",
      "The Mean RMSE across all folds is:  0.5715057123852781\n",
      "Evaluation with the test set: \n",
      "Pearson R:  (0.7200352715281191, 3.795990719058383e-270)\n",
      "Pearson R squared:  0.5184507922445721\n",
      "R Squared Error: 0.4152196084128633\n",
      "Mean Absolute Error: 0.7777157396449698\n",
      "Mean Squared Error: 1.0045361886863893\n",
      "Root Mean Squared Error: 1.002265528034557\n",
      "Relative absolute Error: 0.7472472252495457\n",
      "Root relateive squared Error: 0.7647093510525007\n"
     ]
    }
   ],
   "source": [
    "final_metrics, reg_model = train_test(\"dual_model_pl_split_protein\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "id": "94a0a407-e173-41aa-879e-50706d3717fe",
   "metadata": {},
   "outputs": [],
   "source": [
    "final_metrics_df['protein_split'] = pd.Series(final_metrics)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "id": "2eb1e34c-78ff-4b3b-9f58-4c2716f1b9d8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(6737, 126)\n",
      "(1684, 126)\n",
      "Dataframe after dummy variables:\n",
      "(6737, 126)\n",
      "Dataframe after dummy variables:\n",
      "(1684, 126)\n",
      "The Mean R2 across all folds is:  0.8652961606230015\n",
      "The Mean MAE across all folds is:  0.39896914799124994\n",
      "The Mean MSE across all folds is:  0.30062893179847083\n",
      "The Mean RMSE across all folds is:  0.5480166140975424\n",
      "Evaluation with the test set: \n",
      "Pearson R:  (0.8973566286085314, 0.0)\n",
      "Pearson R squared:  0.8052489189076697\n",
      "R Squared Error: 0.8045853343982321\n",
      "Mean Absolute Error: 0.47772161520190015\n",
      "Mean Squared Error: 0.4598789959619957\n",
      "Root Mean Squared Error: 0.6781437870850073\n",
      "Relative absolute Error: 0.38495807949821514\n",
      "Root relateive squared Error: 0.44205731031368317\n"
     ]
    }
   ],
   "source": [
    "final_metrics, reg_model = train_test(\"dual_model_pl_split_ligand_diversity\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "id": "75a143e7-ed3a-4f44-818f-76d75c89e105",
   "metadata": {},
   "outputs": [],
   "source": [
    "final_metrics_df['ligand_diversity_split'] = pd.Series(final_metrics)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "id": "8c49d428-73cf-45bb-a3f8-40755d9455c7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(6770, 126)\n",
      "(1651, 126)\n",
      "Dataframe after dummy variables:\n",
      "(6770, 126)\n",
      "Dataframe after dummy variables:\n",
      "(1651, 126)\n",
      "The Mean R2 across all folds is:  0.8520771745184981\n",
      "The Mean MAE across all folds is:  0.41659497784342686\n",
      "The Mean MSE across all folds is:  0.33222098626883306\n",
      "The Mean RMSE across all folds is:  0.5761344038141452\n",
      "Evaluation with the test set: \n",
      "Pearson R:  (0.9311295929536413, 0.0)\n",
      "Pearson R squared:  0.8670023188740137\n",
      "R Squared Error: 0.8664236759625299\n",
      "Mean Absolute Error: 0.4067153240460329\n",
      "Mean Squared Error: 0.31206524140520975\n",
      "Root Mean Squared Error: 0.5586279991239338\n",
      "Relative absolute Error: 0.3264477856115103\n",
      "Root relateive squared Error: 0.3654809489391618\n"
     ]
    }
   ],
   "source": [
    "final_metrics, reg_model = train_test(\"dual_model_pl_split_ligand_weight\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "id": "d525fc24-63e7-4913-b010-cdf758e2a366",
   "metadata": {},
   "outputs": [],
   "source": [
    "final_metrics_df['ligand_weight_split'] = pd.Series(final_metrics)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "id": "67666dc8-5363-411e-b8cf-c0a9c3bf6cde",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(6765, 126)\n",
      "(1656, 126)\n",
      "Dataframe after dummy variables:\n",
      "(6765, 126)\n",
      "Dataframe after dummy variables:\n",
      "(1656, 126)\n",
      "The Mean R2 across all folds is:  0.8566582183015944\n",
      "The Mean MAE across all folds is:  0.41469626016260175\n",
      "The Mean MSE across all folds is:  0.3264438738861791\n",
      "The Mean RMSE across all folds is:  0.5711376396596861\n",
      "Evaluation with the test set: \n",
      "Pearson R:  (0.9177243839503039, 0.0)\n",
      "Pearson R squared:  0.8422180448969647\n",
      "R Squared Error: 0.8411442943636551\n",
      "Mean Absolute Error: 0.42402499999999976\n",
      "Mean Squared Error: 0.35141057876811554\n",
      "Root Mean Squared Error: 0.5927989362069702\n",
      "Relative absolute Error: 0.35405314875615723\n",
      "Root relateive squared Error: 0.3985670654185376\n"
     ]
    }
   ],
   "source": [
    "final_metrics, reg_model = train_test(\"dual_model_pl_split_ligand_tpsa\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "id": "2eb0e5ea-4543-4dec-8078-ad0226b74c54",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(6770, 126)\n",
      "(1651, 126)\n",
      "Dataframe after dummy variables:\n",
      "(6770, 126)\n",
      "Dataframe after dummy variables:\n",
      "(1651, 126)\n",
      "The Mean R2 across all folds is:  0.8505307591558356\n",
      "The Mean MAE across all folds is:  0.4184301033973414\n",
      "The Mean MSE across all folds is:  0.3347123998641065\n",
      "The Mean RMSE across all folds is:  0.5784778733404837\n",
      "Evaluation with the test set: \n",
      "Pearson R:  (0.9322320187915425, 0.0)\n",
      "Pearson R squared:  0.8690565368601548\n",
      "R Squared Error: 0.8675394216068888\n",
      "Mean Absolute Error: 0.40391338582677144\n",
      "Mean Squared Error: 0.3138169781950334\n",
      "Root Mean Squared Error: 0.560193697032583\n",
      "Relative absolute Error: 0.32776083315426835\n",
      "Root relateive squared Error: 0.36395134069420776\n"
     ]
    }
   ],
   "source": [
    "final_metrics, reg_model = train_test(\"dual_model_pl_split_ligand_volume\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "id": "585907f3-a678-4790-a1ac-f3712f3a06fc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10, 6)"
      ]
     },
     "execution_count": 160,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_metrics_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "id": "3e97a5cf-e384-4c6c-b63a-e7adf61a2097",
   "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>metric</th>\n",
       "      <th>random_split</th>\n",
       "      <th>pocket_split</th>\n",
       "      <th>protein_split</th>\n",
       "      <th>ligand_diversity_split</th>\n",
       "      <th>ligand_weight_split</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>R2 (5-fold cross-validation)</td>\n",
       "      <td>0.850958</td>\n",
       "      <td>0.869243</td>\n",
       "      <td>0.861070</td>\n",
       "      <td>0.865296</td>\n",
       "      <td>0.852077</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>MAE (5-fold cross-validation)</td>\n",
       "      <td>0.419713</td>\n",
       "      <td>0.406390</td>\n",
       "      <td>0.414366</td>\n",
       "      <td>0.398969</td>\n",
       "      <td>0.416595</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>MSE (5-fold cross-validation)</td>\n",
       "      <td>0.336990</td>\n",
       "      <td>0.317244</td>\n",
       "      <td>0.326737</td>\n",
       "      <td>0.300629</td>\n",
       "      <td>0.332221</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>RMSE (5-fold cross-validation)</td>\n",
       "      <td>0.580366</td>\n",
       "      <td>0.563145</td>\n",
       "      <td>0.571506</td>\n",
       "      <td>0.548017</td>\n",
       "      <td>0.576134</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>R2 (test set)</td>\n",
       "      <td>0.867129</td>\n",
       "      <td>0.248571</td>\n",
       "      <td>0.415220</td>\n",
       "      <td>0.804585</td>\n",
       "      <td>0.866424</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>MAE (test set)</td>\n",
       "      <td>0.401362</td>\n",
       "      <td>0.859121</td>\n",
       "      <td>0.777716</td>\n",
       "      <td>0.477722</td>\n",
       "      <td>0.406715</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>MSE (test set)</td>\n",
       "      <td>0.302466</td>\n",
       "      <td>1.114850</td>\n",
       "      <td>1.004536</td>\n",
       "      <td>0.459879</td>\n",
       "      <td>0.312065</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>RMSE (test set)</td>\n",
       "      <td>0.549969</td>\n",
       "      <td>1.055864</td>\n",
       "      <td>1.002266</td>\n",
       "      <td>0.678144</td>\n",
       "      <td>0.558628</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Relative absolute error (test set)</td>\n",
       "      <td>0.326884</td>\n",
       "      <td>0.850096</td>\n",
       "      <td>0.747247</td>\n",
       "      <td>0.384958</td>\n",
       "      <td>0.326448</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Root relative squared error (test set)</td>\n",
       "      <td>0.364515</td>\n",
       "      <td>0.866850</td>\n",
       "      <td>0.764709</td>\n",
       "      <td>0.442057</td>\n",
       "      <td>0.365481</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                   metric  random_split  pocket_split  \\\n",
       "0            R2 (5-fold cross-validation)      0.850958      0.869243   \n",
       "1           MAE (5-fold cross-validation)      0.419713      0.406390   \n",
       "2           MSE (5-fold cross-validation)      0.336990      0.317244   \n",
       "3          RMSE (5-fold cross-validation)      0.580366      0.563145   \n",
       "4                           R2 (test set)      0.867129      0.248571   \n",
       "5                          MAE (test set)      0.401362      0.859121   \n",
       "6                          MSE (test set)      0.302466      1.114850   \n",
       "7                         RMSE (test set)      0.549969      1.055864   \n",
       "8      Relative absolute error (test set)      0.326884      0.850096   \n",
       "9  Root relative squared error (test set)      0.364515      0.866850   \n",
       "\n",
       "   protein_split  ligand_diversity_split  ligand_weight_split  \n",
       "0       0.861070                0.865296             0.852077  \n",
       "1       0.414366                0.398969             0.416595  \n",
       "2       0.326737                0.300629             0.332221  \n",
       "3       0.571506                0.548017             0.576134  \n",
       "4       0.415220                0.804585             0.866424  \n",
       "5       0.777716                0.477722             0.406715  \n",
       "6       1.004536                0.459879             0.312065  \n",
       "7       1.002266                0.678144             0.558628  \n",
       "8       0.747247                0.384958             0.326448  \n",
       "9       0.764709                0.442057             0.365481  "
      ]
     },
     "execution_count": 162,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_metrics_df.head(n=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "id": "3b8a1135-f281-40a8-8137-3a87507a23bc",
   "metadata": {},
   "outputs": [],
   "source": [
    "final_metrics_df.to_csv(\"work/training/dual-model/final_metrics.csv\", sep=\",\", index=False, encoding=\"utf-8\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5284b9ce-7495-4e7d-8e22-92566b31e6af",
   "metadata": {},
   "source": [
    "## Train/Test a Random Forest model to predict binding affinity of mutated protein-ligand from the WT protein-ligand binding affinity and mutation features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "41e1d613-1264-4047-a4b8-1162c1154aa2",
   "metadata": {},
   "outputs": [],
   "source": [
    "final_metrics_df = pd.DataFrame()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "a0d696a8-fd2a-4900-a6f3-ca264a9ec77a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0, 0)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_metrics_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "c4463bee-df70-49a7-bbe6-659c59eef1dd",
   "metadata": {},
   "outputs": [],
   "source": [
    "final_metrics_df['metric'] = [\"R2 (5-fold cross-validation)\", \"MAE (5-fold cross-validation)\", \"MSE (5-fold cross-validation)\", \"RMSE (5-fold cross-validation)\",\n",
    "                              \"R2 (test set)\", \"MAE (test set)\", \"MSE (test set)\", \"RMSE (test set)\", \n",
    "                              \"Relative absolute error (test set)\", \"Root relative squared error (test set)\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "552bfd34-91d1-473c-bb7d-1b5f411a8b7d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10, 1)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_metrics_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "1fb5d461-0d37-4cca-844c-73977aedba36",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(158341, 190)\n",
      "(39584, 190)\n",
      "Dataframe after dummy variables:\n",
      "(158341, 190)\n",
      "Dataframe after dummy variables:\n",
      "(158341, 190)\n",
      "Dataframe after dummy variables:\n",
      "(39584, 190)\n",
      "Dataframe after dummy variables:\n",
      "(39584, 190)\n",
      "The Mean R2 across all folds is:  0.8963458018916505\n",
      "The Mean MAE across all folds is:  0.352762619785875\n",
      "The Mean MSE across all folds is:  0.2465786485365496\n",
      "The Mean RMSE across all folds is:  0.4965565693027725\n",
      "\n",
      "\n",
      "Evaluation with the test set (real wt_bt): \n",
      "Pearson R:  (0.9469612802175152, 0.0)\n",
      "Pearson R squared:  0.8967356662311954\n",
      "R Squared Error: 0.896511085074408\n",
      "Mean Absolute Error: 0.34941937098092185\n",
      "Mean Squared Error: 0.24506832485284885\n",
      "Root Mean Squared Error: 0.4950437605432967\n",
      "Relative absolute Error: 0.27593320775644575\n",
      "Root relateive squared Error: 0.32169693023961543\n",
      "\n",
      "\n",
      "Evaluation with the test set (predicted wt_bt): \n",
      "Pearson R:  (0.9409324405702544, 0.0)\n",
      "Pearson R squared:  0.8853538577174953\n",
      "R Squared Error: 0.8850130867815884\n",
      "Mean Absolute Error: 0.37820656925033547\n",
      "Mean Squared Error: 0.27229631524011116\n",
      "Root Mean Squared Error: 0.5218201943582781\n",
      "Relative absolute Error: 0.2986661888688003\n",
      "Root relateive squared Error: 0.33909720320051534\n"
     ]
    }
   ],
   "source": [
    "final_metrics_real, final_metrics_pred, reg_model = train_test_mutation_model(\"mutation_model_real_wt_ba\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "5a8fb1f0-267b-4423-87c4-84b0953370dc",
   "metadata": {},
   "outputs": [],
   "source": [
    "final_metrics_df['mutation_model_real_wt_ba_test_real_wt_ba'] = pd.Series(final_metrics_real)\n",
    "final_metrics_df['mutation_model_real_wt_ba_test_pred_wt_ba'] = pd.Series(final_metrics_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "c1026722-4d71-4639-9f54-da31a138feea",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(158341, 190)\n",
      "(39584, 190)\n",
      "Dataframe after dummy variables:\n",
      "(158341, 190)\n",
      "Dataframe after dummy variables:\n",
      "(158341, 190)\n",
      "Dataframe after dummy variables:\n",
      "(39584, 190)\n",
      "Dataframe after dummy variables:\n",
      "(39584, 190)\n",
      "The Mean R2 across all folds is:  0.8824172767872792\n",
      "The Mean MAE across all folds is:  0.38214091719493015\n",
      "The Mean MSE across all folds is:  0.27971359342825824\n",
      "The Mean RMSE across all folds is:  0.5288727839178008\n",
      "\n",
      "\n",
      "Evaluation with the test set (real wt_bt): \n",
      "Pearson R:  (0.9341108179070892, 0.0)\n",
      "Pearson R squared:  0.8725630201310511\n",
      "R Squared Error: 0.8655399916211797\n",
      "Mean Absolute Error: 0.39980744746354147\n",
      "Mean Squared Error: 0.3184098416413948\n",
      "Root Mean Squared Error: 0.564278159812512\n",
      "Relative absolute Error: 0.31572420027496156\n",
      "Root relateive squared Error: 0.3666878895993435\n",
      "\n",
      "\n",
      "Evaluation with the test set (predicted wt_bt): \n",
      "Pearson R:  (0.9399625736993191, 0.0)\n",
      "Pearson R squared:  0.883529639955448\n",
      "R Squared Error: 0.8828514632313114\n",
      "Mean Absolute Error: 0.3810855430760208\n",
      "Mean Squared Error: 0.27741517712797353\n",
      "Root Mean Squared Error: 0.5267021711821336\n",
      "Relative absolute Error: 0.30093968756046724\n",
      "Root relateive squared Error: 0.3422696842676672\n"
     ]
    }
   ],
   "source": [
    "final_metrics_real, final_metrics_pred, reg_model = train_test_mutation_model(\"mutation_model_pred_wt_ba\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "2888e8c7-6a5d-47fb-8841-6f6ab0e92587",
   "metadata": {},
   "outputs": [],
   "source": [
    "final_metrics_df['mutation_model_pred_wt_ba_test_real_wt_ba'] = pd.Series(final_metrics_real)\n",
    "final_metrics_df['mutation_model_pred_wt_ba_test_pred_wt_ba'] = pd.Series(final_metrics_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "3ecc9f26-c9a9-4053-bb7e-b30f9e04ae0b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10, 5)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_metrics_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "b7d391b9-bea9-43f9-9214-eed70036cd2c",
   "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>metric</th>\n",
       "      <th>mutation_model_real_wt_ba_test_real_wt_ba</th>\n",
       "      <th>mutation_model_real_wt_ba_test_pred_wt_ba</th>\n",
       "      <th>mutation_model_pred_wt_ba_test_real_wt_ba</th>\n",
       "      <th>mutation_model_pred_wt_ba_test_pred_wt_ba</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>R2 (5-fold cross-validation)</td>\n",
       "      <td>0.896346</td>\n",
       "      <td>0.896346</td>\n",
       "      <td>0.882417</td>\n",
       "      <td>0.882417</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>MAE (5-fold cross-validation)</td>\n",
       "      <td>0.352763</td>\n",
       "      <td>0.352763</td>\n",
       "      <td>0.382141</td>\n",
       "      <td>0.382141</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>MSE (5-fold cross-validation)</td>\n",
       "      <td>0.246579</td>\n",
       "      <td>0.246579</td>\n",
       "      <td>0.279714</td>\n",
       "      <td>0.279714</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>RMSE (5-fold cross-validation)</td>\n",
       "      <td>0.496557</td>\n",
       "      <td>0.496557</td>\n",
       "      <td>0.528873</td>\n",
       "      <td>0.528873</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>R2 (test set)</td>\n",
       "      <td>0.896511</td>\n",
       "      <td>0.885013</td>\n",
       "      <td>0.865540</td>\n",
       "      <td>0.882851</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>MAE (test set)</td>\n",
       "      <td>0.349419</td>\n",
       "      <td>0.378207</td>\n",
       "      <td>0.399807</td>\n",
       "      <td>0.381086</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>MSE (test set)</td>\n",
       "      <td>0.245068</td>\n",
       "      <td>0.272296</td>\n",
       "      <td>0.318410</td>\n",
       "      <td>0.277415</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>RMSE (test set)</td>\n",
       "      <td>0.495044</td>\n",
       "      <td>0.521820</td>\n",
       "      <td>0.564278</td>\n",
       "      <td>0.526702</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>Relative absolute error (test set)</td>\n",
       "      <td>0.275933</td>\n",
       "      <td>0.298666</td>\n",
       "      <td>0.315724</td>\n",
       "      <td>0.300940</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>Root relative squared error (test set)</td>\n",
       "      <td>0.321697</td>\n",
       "      <td>0.339097</td>\n",
       "      <td>0.366688</td>\n",
       "      <td>0.342270</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                   metric  \\\n",
       "0            R2 (5-fold cross-validation)   \n",
       "1           MAE (5-fold cross-validation)   \n",
       "2           MSE (5-fold cross-validation)   \n",
       "3          RMSE (5-fold cross-validation)   \n",
       "4                           R2 (test set)   \n",
       "5                          MAE (test set)   \n",
       "6                          MSE (test set)   \n",
       "7                         RMSE (test set)   \n",
       "8      Relative absolute error (test set)   \n",
       "9  Root relative squared error (test set)   \n",
       "\n",
       "   mutation_model_real_wt_ba_test_real_wt_ba  \\\n",
       "0                                   0.896346   \n",
       "1                                   0.352763   \n",
       "2                                   0.246579   \n",
       "3                                   0.496557   \n",
       "4                                   0.896511   \n",
       "5                                   0.349419   \n",
       "6                                   0.245068   \n",
       "7                                   0.495044   \n",
       "8                                   0.275933   \n",
       "9                                   0.321697   \n",
       "\n",
       "   mutation_model_real_wt_ba_test_pred_wt_ba  \\\n",
       "0                                   0.896346   \n",
       "1                                   0.352763   \n",
       "2                                   0.246579   \n",
       "3                                   0.496557   \n",
       "4                                   0.885013   \n",
       "5                                   0.378207   \n",
       "6                                   0.272296   \n",
       "7                                   0.521820   \n",
       "8                                   0.298666   \n",
       "9                                   0.339097   \n",
       "\n",
       "   mutation_model_pred_wt_ba_test_real_wt_ba  \\\n",
       "0                                   0.882417   \n",
       "1                                   0.382141   \n",
       "2                                   0.279714   \n",
       "3                                   0.528873   \n",
       "4                                   0.865540   \n",
       "5                                   0.399807   \n",
       "6                                   0.318410   \n",
       "7                                   0.564278   \n",
       "8                                   0.315724   \n",
       "9                                   0.366688   \n",
       "\n",
       "   mutation_model_pred_wt_ba_test_pred_wt_ba  \n",
       "0                                   0.882417  \n",
       "1                                   0.382141  \n",
       "2                                   0.279714  \n",
       "3                                   0.528873  \n",
       "4                                   0.882851  \n",
       "5                                   0.381086  \n",
       "6                                   0.277415  \n",
       "7                                   0.526702  \n",
       "8                                   0.300940  \n",
       "9                                   0.342270  "
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_metrics_df.head(n=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "8426f8de-f2ee-41e7-88da-8ea1ab9b7f7b",
   "metadata": {},
   "outputs": [],
   "source": [
    "final_metrics_df.to_csv(\"work/training/dual-model/final_metrics_mutation.csv\", sep=\",\", index=False, encoding=\"utf-8\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ad20e2b0-4ca8-42e9-977c-b2ed610f4857",
   "metadata": {},
   "source": [
    "# Y-randomization test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "936ed2cf-3fec-4a9b-9eb6-27faba84d5da",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(6770, 126)\n",
      "(6770, 126)\n",
      "The Mean R2 across all folds is:  -0.07256870182450789\n",
      "The Mean MAE across all folds is:  1.2487744756277694\n",
      "The Mean MSE across all folds is:  2.4124579176720826\n",
      "The Mean RMSE across all folds is:  1.553005686232836\n",
      "\n",
      "\n",
      "(6770, 126)\n",
      "(6770, 126)\n",
      "The Mean R2 across all folds is:  -0.07905331690022481\n",
      "The Mean MAE across all folds is:  1.2574497193500742\n",
      "The Mean MSE across all folds is:  2.427963054871492\n",
      "The Mean RMSE across all folds is:  1.5578052826916586\n",
      "\n",
      "\n",
      "(6770, 126)\n",
      "(6770, 126)\n",
      "The Mean R2 across all folds is:  -0.07719008547442359\n",
      "The Mean MAE across all folds is:  1.2503636632200885\n",
      "The Mean MSE across all folds is:  2.4218693326203837\n",
      "The Mean RMSE across all folds is:  1.5556610343343893\n",
      "\n",
      "\n",
      "(6770, 126)\n",
      "(6770, 126)\n",
      "The Mean R2 across all folds is:  -0.06532529466173091\n",
      "The Mean MAE across all folds is:  1.2405533825701627\n",
      "The Mean MSE across all folds is:  2.3934215508774006\n",
      "The Mean RMSE across all folds is:  1.5470127139273615\n",
      "\n",
      "\n",
      "(6770, 126)\n",
      "(6770, 126)\n",
      "The Mean R2 across all folds is:  -0.06681282550174013\n",
      "The Mean MAE across all folds is:  1.2462793500738552\n",
      "The Mean MSE across all folds is:  2.3989873448862626\n",
      "The Mean RMSE across all folds is:  1.5486167884001951\n",
      "\n",
      "\n",
      "(6770, 126)\n",
      "(6770, 126)\n",
      "The Mean R2 across all folds is:  -0.08217779464389574\n",
      "The Mean MAE across all folds is:  1.2556472082717876\n",
      "The Mean MSE across all folds is:  2.4306626242363367\n",
      "The Mean RMSE across all folds is:  1.558776684339813\n",
      "\n",
      "\n",
      "(6770, 126)\n",
      "(6770, 126)\n",
      "The Mean R2 across all folds is:  -0.06199286138228555\n",
      "The Mean MAE across all folds is:  1.2434163367799116\n",
      "The Mean MSE across all folds is:  2.3878258206617433\n",
      "The Mean RMSE across all folds is:  1.5450559768009893\n",
      "\n",
      "\n",
      "(6770, 126)\n",
      "(6770, 126)\n",
      "The Mean R2 across all folds is:  -0.0598257425904289\n",
      "The Mean MAE across all folds is:  1.2418920531757753\n",
      "The Mean MSE across all folds is:  2.382575078558344\n",
      "The Mean RMSE across all folds is:  1.543283006468551\n",
      "\n",
      "\n",
      "(6770, 126)\n",
      "(6770, 126)\n",
      "The Mean R2 across all folds is:  -0.07173742280223512\n",
      "The Mean MAE across all folds is:  1.2497821565731164\n",
      "The Mean MSE across all folds is:  2.4110745963604123\n",
      "The Mean RMSE across all folds is:  1.5526132725739212\n",
      "\n",
      "\n",
      "(6770, 126)\n",
      "(6770, 126)\n",
      "The Mean R2 across all folds is:  -0.06694977923149277\n",
      "The Mean MAE across all folds is:  1.2441208862629247\n",
      "The Mean MSE across all folds is:  2.3978920649217126\n",
      "The Mean RMSE across all folds is:  1.5483236758708148\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "train_test_y_randomization()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "588d9c21",
   "metadata": {},
   "source": [
    "# Feature Importance"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "46475cd6",
   "metadata": {},
   "source": [
    "### Wild-type protein-ligand binding affinity prediction random forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "f859f7eb",
   "metadata": {},
   "outputs": [],
   "source": [
    "rf = pickle.load(open(\"work/models/dual_model_pl_split_ligand_weight.mdl\", 'rb'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "ee2286fd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(6770, 126)\n",
      "(1651, 126)\n"
     ]
    }
   ],
   "source": [
    "train = pd.read_csv(\"work/training/dual-model/train-ligand-weight.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "test = pd.read_csv(\"work/training/dual-model/test-ligand-weight.csv\", sep=\",\", encoding=\"utf-8\") \n",
    "\n",
    "print(train.shape)\n",
    "print(test.shape)\n",
    "\n",
    "train_X = train.drop('ba', axis=1)\n",
    "wt_pl_train_X = train_X\n",
    "\n",
    "test_X = test.drop('ba', axis=1)\n",
    "wt_pl_test_X = test_X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "d03e9a13-5172-443a-955c-19d5d7ff9f4a",
   "metadata": {},
   "outputs": [],
   "source": [
    "wt_pl_ft_df = get_df_feature_importance(rf.feature_importances_,train_X.columns,'WT protein-ligand binding affinity RF model ', top=20)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8c38fb96-e005-4728-a98d-3a43d94d7c93",
   "metadata": {},
   "source": [
    "### Mutated protein-ligand binding affinity prediction random forest (real WT BA)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "e7de40ed-1a3c-4c2c-a9d9-9bada6954073",
   "metadata": {},
   "outputs": [],
   "source": [
    "rf = pickle.load(open(\"work/models/mutation_model_real_wt_ba.mdl\", 'rb'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "a0fbc946-f424-4e40-8ebf-8a984e405005",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(158341, 190)\n",
      "(39584, 190)\n"
     ]
    }
   ],
   "source": [
    "train = pd.read_csv(\"work/training/dual-model/train-mutations-random.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "test = pd.read_csv(\"work/training/dual-model/test-mutations-random.csv\", sep=\",\", encoding=\"utf-8\") \n",
    "\n",
    "print(train.shape)\n",
    "print(test.shape)\n",
    "\n",
    "train_X = train.drop('ba', axis=1)\n",
    "\n",
    "test_X = test.drop('ba', axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "11bc4909-623c-45ac-b75d-9671b1eb9cb9",
   "metadata": {},
   "outputs": [],
   "source": [
    "mut_real_wt_ba_ft_df = get_df_feature_importance(rf.feature_importances_,train_X.columns,'Mutation protein-ligand binding affinity RF model (real WT BA) ', top=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "f592142c-05e4-4369-a9b0-b4adf996d1cc",
   "metadata": {},
   "outputs": [],
   "source": [
    "mut_real_no_wt_ba_ft_df = get_df_feature_importance(rf.feature_importances_,train_X.columns,'Mutation protein-ligand binding affinity RF model (real WT BA) ', top=10, execlude=\"wt_ba\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "96408451-7c3c-4f63-a1fa-eb676cae68e6",
   "metadata": {},
   "source": [
    "### Mutated protein-ligand binding affinity prediction random forest (predicted WT BA)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "e0d0f5de-0c5b-484a-b7d8-2821e1be7650",
   "metadata": {},
   "outputs": [],
   "source": [
    "rf = pickle.load(open(\"work/models/mutation_model_pred_wt_ba.mdl\", 'rb'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "f2bf1655-0d9d-4eab-8112-cc9610a1f1c0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(158341, 190)\n",
      "(39584, 190)\n"
     ]
    }
   ],
   "source": [
    "train = pd.read_csv(\"work/training/dual-model/train-mutations-random.csv\", sep=\",\", encoding=\"utf-8\")\n",
    "test = pd.read_csv(\"work/training/dual-model/test-mutations-random.csv\", sep=\",\", encoding=\"utf-8\") \n",
    "\n",
    "print(train.shape)\n",
    "print(test.shape)\n",
    "\n",
    "train_X = train.drop('ba', axis=1)\n",
    "\n",
    "test_X = test.drop('ba', axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "acd22c2b-80c8-457e-9b15-ba6058d62207",
   "metadata": {},
   "outputs": [],
   "source": [
    "mut_pred_wt_ba_ft_df = get_df_feature_importance(rf.feature_importances_,train_X.columns,'Mutation protein-ligand binding affinity RF model (predicted WT BA) ', top=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "3263897c-9bc0-4c5a-a2bc-0db56441551e",
   "metadata": {},
   "outputs": [],
   "source": [
    "mut_pred_no_wt_ba_ft_df = get_df_feature_importance(rf.feature_importances_,train_X.columns,'Mutation protein-ligand binding affinity RF model (predicted WT BA) ', top=10, execlude=\"wt_ba\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aeb34287-881c-4a1f-9e2d-c62654422c6e",
   "metadata": {},
   "source": [
    "### Making the feature importance plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "44d3a248-233d-4d4b-9517-4b265dc205dd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x576 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axes = plt.subplot_mosaic([['left', 'right1'], ['left', 'right2'], ['bottom1', 'bottom2']],\n",
    "                              constrained_layout=True, figsize=(16,8))\n",
    "fig.suptitle('Random forest models feature importance')\n",
    "\n",
    "axes['left'].set_title('WT protein-ligand binding affinity RF model')\n",
    "axes['right1'].set_title('Mutation binding affinity RF model (with real WT BA)')\n",
    "axes['right2'].set_title('Mutation binding affinity RF model (without real WT BA)')\n",
    "axes['bottom1'].set_title('Mutation affinity RF model (with predicted WT BA)')\n",
    "axes['bottom2'].set_title('Mutation affinity RF model (without predicted WT BA)')\n",
    "\n",
    "sns.barplot(ax=axes['left'], x=wt_pl_ft_df['feature_importance'], y=wt_pl_ft_df['feature_names'], color=\"#008bfb\").set(ylabel=None)\n",
    "sns.barplot(ax=axes['right1'], x=mut_real_wt_ba_ft_df['feature_importance'], y=mut_real_wt_ba_ft_df['feature_names'], color=\"#008bfb\").set(ylabel=None)\n",
    "sns.barplot(ax=axes['right2'], x=mut_real_no_wt_ba_ft_df['feature_importance'], y=mut_real_no_wt_ba_ft_df['feature_names'], color=\"#008bfb\").set(ylabel=None)\n",
    "sns.barplot(ax=axes['bottom1'], x=mut_pred_wt_ba_ft_df['feature_importance'], y=mut_pred_wt_ba_ft_df['feature_names'], color=\"#008bfb\").set(ylabel=None)\n",
    "sns.barplot(ax=axes['bottom2'], x=mut_pred_no_wt_ba_ft_df['feature_importance'], y=mut_pred_no_wt_ba_ft_df['feature_names'], color=\"#008bfb\").set(ylabel=None)\n",
    "\n",
    "plt.savefig('images/importance/feature_importance_all.png', dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "id": "a0c5332c-200d-4d32-bfbc-33fbbdd66302",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "dfs = [mut_real_wt_ba_ft_df, mut_real_no_wt_ba_ft_df, mut_pred_wt_ba_ft_df, mut_pred_no_wt_ba_ft_df]\n",
    "titles = ['Mutation binding affinity RF model (with real WT BA)',\n",
    "          'Mutation binding affinity RF model (without real WT BA)',\n",
    "          'Mutation affinity RF model (with predicted WT BA)',\n",
    "          'Mutation affinity RF model (without predicted WT BA)']\n",
    "filenames = ['mut_model_real_wt_ba_feature_importance',\n",
    "             'mut_model_real_wt_ba_feature_importance_no_wtba',\n",
    "             'mut_model_pred_wt_ba_feature_importance',\n",
    "             'mut_model_pred_wt_ba_feature_importance_no_wtba']\n",
    "\n",
    "for i in range(4):\n",
    "    plt.figure(figsize=(10,8))\n",
    "    fig = sns.barplot(x=dfs[i]['feature_importance'], y=dfs[i]['feature_names'], color=\"#008bfb\")\n",
    "    plt.title(titles[i])\n",
    "    plt.xlabel('Feature Importance')\n",
    "    plt.ylabel('Feature Name')\n",
    "    plt.savefig('images/importance/'+filenames[i]+'.png', dpi=300, bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "27e7c917-02be-4c79-897f-a774345b4d54",
   "metadata": {},
   "source": [
    "# Evaluate prediction speed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "id": "ee4b664a-42b7-472a-8caa-b11c728bd158",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1651, 126)\n",
      "0.17403916601324454 seconds to predict 1651 instances\n",
      "0.00010541439492019657 seconds to predict one instance\n",
      "0.19761443103197962 seconds to predict 1651 instances\n",
      "0.00011969378015262243 seconds to predict one instance\n",
      "0.19888473598985001 seconds to predict 1651 instances\n",
      "0.00012046319563285889 seconds to predict one instance\n",
      "0.20000486000208184 seconds to predict 1651 instances\n",
      "0.00012114164748763286 seconds to predict one instance\n",
      "0.19828114699339494 seconds to predict 1651 instances\n",
      "0.00012009760568951844 seconds to predict one instance\n",
      "0.19632747501600534 seconds to predict 1651 instances\n",
      "0.00011891427923440663 seconds to predict one instance\n",
      "0.4413825879455544 seconds to predict 1651 instances\n",
      "0.0002673425729530917 seconds to predict one instance\n",
      "0.20655659801559523 seconds to predict 1651 instances\n",
      "0.0001251099927411237 seconds to predict one instance\n",
      "0.21580177202122286 seconds to predict 1651 instances\n",
      "0.0001307097347190932 seconds to predict one instance\n",
      "0.22281083604320884 seconds to predict 1651 instances\n",
      "0.0001349550793720223 seconds to predict one instance\n",
      "Average speed of prediction for one instance: 0.00013638422829025668 seconds\n"
     ]
    }
   ],
   "source": [
    "from timeit import default_timer as timer\n",
    "\n",
    "rf = pickle.load(open(\"work/models/dual_model_pl_split_ligand_weight.mdl\", 'rb'))\n",
    "\n",
    "test = pd.read_csv(\"work/training/dual-model/test-ligand-weight.csv\", sep=\",\", encoding=\"utf-8\") \n",
    "\n",
    "test_X = test.drop('ba', axis=1)\n",
    "    \n",
    "print(test.shape)\n",
    "\n",
    "per_instance_sum = 0\n",
    "\n",
    "for i in range(10):\n",
    "\n",
    "    start = timer()\n",
    "\n",
    "    predictions = rf.predict(test_X)\n",
    "\n",
    "    end = timer()\n",
    "\n",
    "    time_spent = end - start\n",
    "\n",
    "    per_instance = time_spent / test_X.shape[0]\n",
    "    \n",
    "    per_instance_sum += per_instance\n",
    "\n",
    "    print(str(time_spent) + \" seconds to predict \" + str(test_X.shape[0]) + \" instances\" )\n",
    "    print(str(per_instance) + \" seconds to predict one instance\")\n",
    "    \n",
    "\n",
    "print(\"Average speed of prediction for one instance: \" + str(per_instance_sum / 10) + \" seconds\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}