classification.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "8396cc39",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X_protein_esm1.npy\n",
      "X_drug.npy\n",
      "yreg.npy\n",
      "X_protein_esm1v.npy\n",
      "X_protein_Prot_T5_XL_BFD.npy\n",
      "X_protein_ProtBert_BFD.npy\n",
      "y.npy\n",
      "X_protein_ProtAlbert.npy\n",
      "X_protein_esm1b.npy\n",
      "train_dnndr2_v2.ipynb\n",
      "X_protein_Prot_XLNet.npy\n",
      "X_protein_Prot_T5_XL.npy\n",
      "X_protein_ProtBert.npy\n"
     ]
    }
   ],
   "source": [
    "# Import packages\n",
    "import sys\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import os\n",
    "from tqdm import tqdm\n",
    "from natsort import natsorted\n",
    "import matplotlib.pyplot as plt\n",
    "plt.style.use('classic')\n",
    "%matplotlib inline\n",
    "from tensorflow.keras.models import Sequential, Model\n",
    "from tensorflow.keras.layers import Layer,Input, Dense, Dropout, Activation, Concatenate, Flatten, BatchNormalization\n",
    "from tensorflow.keras.regularizers import l2,l1\n",
    "from tensorflow.keras.optimizers import SGD,Adam,RMSprop\n",
    "from tensorflow.compat.v1 import InteractiveSession\n",
    "import tensorflow.keras.backend as K\n",
    "from tensorflow.keras.preprocessing.image import array_to_img, img_to_array, load_img\n",
    "from tensorflow.keras.callbacks import EarlyStopping, ModelCheckpoint,ReduceLROnPlateau\n",
    "from tensorflow.keras.models import load_model\n",
    "import tensorflow.keras.metrics\n",
    "####import tensorflow_addons as tfa not required\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "import tensorflow as tf\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from tensorflow.keras.utils import to_categorical, plot_model\n",
    "# from pycm import *\n",
    "#import matplotlib.pyplot as plt\n",
    "\n",
    "#from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.metrics import precision_recall_curve,roc_auc_score\n",
    "from sklearn.metrics import average_precision_score\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "from sklearn.metrics import balanced_accuracy_score\n",
    "from sklearn.metrics import matthews_corrcoef, cohen_kappa_score\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.metrics import average_precision_score\n",
    "from sklearn import metrics\n",
    "from sklearn.metrics import r2_score\n",
    "\n",
    "### \n",
    "\n",
    "ROOT = '/mnt/media/dnndr2/finalEmbs/'\n",
    "\n",
    "# View all available embeddings\n",
    "for f in os.listdir(ROOT):\n",
    "    print(f)\n",
    "    \n",
    "# Chosoe emb from \n",
    "# ['esm1', 'esm1b', 'esm1v','ProtBert','ProtBert_BFD', 'Prot_T5_XL', 'Prot_T5_XL_BFD',\n",
    "#                  'Prot_XLNet', 'ProtAlbert', 'alphafold']\n",
    "emb = 'ProtAlbert'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "8832ddec",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(41288, 4096) (41288, 1024) (41288, 3) (20336, 4096) (20336, 1024) (20336, 3)\n"
     ]
    }
   ],
   "source": [
    "X_protein = np.load(ROOT+'X_protein_%s.npy'%(emb))\n",
    "X_drug = np.load(ROOT+'X_drug.npy')\n",
    "y = np.load(ROOT+'y.npy')\n",
    "\n",
    "X1_train, X1_test, X2_train, X2_test, y_train, y_test = train_test_split(X_protein, X_drug, y, test_size=0.33, random_state=42)\n",
    "print(X1_train.shape, X2_train.shape, y_train.shape, X1_test.shape, X2_test.shape, y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "141b0856",
   "metadata": {},
   "outputs": [],
   "source": [
    "#@title Classification model\n",
    "####model MLP  Classification\n",
    "\n",
    "##################\n",
    "# 1: it works good\n",
    "\n",
    "#input_1 protien\n",
    "if emb in ['esm1', 'esm1b', 'esm1v']:\n",
    "    input_1 = Input(shape = (1280,), name='i_1')  ### ESM, ESM1b, ESM1v\n",
    "\n",
    "elif emb in ['ProtBert','ProtBert_BFD', 'Prot_T5_XL', 'Prot_T5_XL_BFD', 'Prot_XLNet']:\n",
    "    input_1 = Input(shape = (1024,), name='i_1')  ### ProtBert, ProtBert-BFD, ProtT5-XL, ProtT5-XL-BFD, ProtXLNet\n",
    "\n",
    "elif emb in ['ProtAlbert']:\n",
    "    input_1 = Input(shape = (4096,), name='i_1')  ### ProtAlbert\n",
    "\n",
    "elif emb in ['alphafold']:\n",
    "    input_1 = Input(shape = (384,), name='i_1')  ### Alphafold\n",
    "\n",
    "#input_1 = Input(shape = (15599,), name='i_1')\n",
    "#bn1_0 = BatchNormalization()(input_1)\n",
    "dense1_1 = Dense(1024, activation = 'elu')(input_1)\n",
    "bn1_1 = BatchNormalization()(dense1_1)\n",
    "#dense1_2 = Dense(1024, activation = 'relu')(bn1_1)\n",
    "#bn1_2 = BatchNormalization()(dense1_2)\n",
    "\n",
    "#input_2 drug\n",
    "input_2 = Input(shape = (1024,), name='i_2')\n",
    "#bn2_0 = BatchNormalization()(input_2)\n",
    "dense2_1 = Dense(1024, activation = 'elu')(input_2)\n",
    "bn2_1 = BatchNormalization()(dense2_1)\n",
    "#dense2_2 = Dense(1024, activation = 'relu')(bn2_1)\n",
    "#bn2_2 = BatchNormalization()(dense2_2)\n",
    " \n",
    "# concatenate\n",
    "##concat   = Concatenate()([dense1_1, dense2_1])\n",
    "#concat   = Concatenate()([bn1_2, bn2_2])\n",
    "concat   = Concatenate()([bn1_1, bn2_1])\n",
    "fc_1   = Dense(1024, activation = 'elu')(concat)\n",
    "drop_1 = Dropout(0.5)(fc_1)\n",
    "fc_2   = Dense(512, activation = 'elu')(drop_1)\n",
    "#drop_2 = Dropout(0.5)(fc_1)\n",
    "#fc_3   = Dense(1024, activation = 'relu')(drop_2)\n",
    "#classification output\n",
    "output  = Dense(3, activation='softmax')(fc_2)\n",
    " \n",
    "# create model with two inputs\n",
    "model = Model(inputs=[input_1,input_2], outputs=output)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "a1e52dc5",
   "metadata": {},
   "outputs": [],
   "source": [
    "#@title classfication metrics\n",
    "metrics_c = ['accuracy',\n",
    "    #keras.metrics.FalseNegatives(name=\"fn\"),\n",
    "    #keras.metrics.FalsePositives(name=\"fp\"),\n",
    "    #keras.metrics.TrueNegatives(name=\"tn\"),\n",
    "    #keras.metrics.TruePositives(name=\"tp\"),\n",
    "    #keras.metrics.BinaryAccuracy(name='accuracy'),\n",
    "    tensorflow.keras.metrics.Precision(name=\"precision\"),\n",
    "    tensorflow.keras.metrics.Recall(name=\"recall\"),\n",
    "    tensorflow.keras.metrics.AUC(name=\"auc_pr\",curve=\"PR\"),\n",
    "    tensorflow.keras.metrics.AUC(name=\"auc_roc\",curve=\"ROC\")\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "90e79282",
   "metadata": {},
   "outputs": [],
   "source": [
    "## classi\n",
    "model.compile(loss='categorical_crossentropy', \n",
    "              optimizer=tensorflow.keras.optimizers.Adam(learning_rate=0.001), \n",
    "              metrics=metrics_c)\n",
    "##class\n",
    "reduce_lr = ReduceLROnPlateau(monitor='val_auc_pr', factor=0.5,patience=20, min_lr=0.00000001, verbose=1)\n",
    "\n",
    "### class setup\n",
    "###checkpoint_filepath = 'weights-improvement-{epoch:03d}-{val_auc_pr:.4f}.hdf5'\n",
    "checkpoint_filepath = 'checkpoints/weights-improvement-{val_auc_pr:.3f}.hdf5'\n",
    "model_checkpoint_callback = ModelCheckpoint(filepath=checkpoint_filepath,save_weights_only=False,\n",
    "                                            monitor='val_auc_pr',\n",
    "                                            mode='max',\n",
    "                                            save_best_only=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "44465c41",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 00021: ReduceLROnPlateau reducing learning rate to 0.0005000000237487257.\n",
      "\n",
      "Epoch 00041: ReduceLROnPlateau reducing learning rate to 0.0002500000118743628.\n",
      "\n",
      "Epoch 00061: ReduceLROnPlateau reducing learning rate to 0.0001250000059371814.\n",
      "\n",
      "Epoch 00081: ReduceLROnPlateau reducing learning rate to 6.25000029685907e-05.\n"
     ]
    }
   ],
   "source": [
    "# classification fit the keras model on the dataset\n",
    "history=model.fit([X1_train,X2_train],y_train,\n",
    "                  batch_size=2048, epochs=100,\n",
    "                  validation_data=([X1_test,X2_test],y_test),\n",
    "                  #validation_split=0.1,\n",
    "                  #class_weight=class_weights,\n",
    "                  #class_weight={0: 2.818, 1: 0.714, 2: 0.802},\n",
    "                  callbacks=[reduce_lr,model_checkpoint_callback], verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "71c56054",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9280688207041492,0.9504535317059217,0.9104399723667661,0.6950272963268356,0.945867316468167,0.8668837691992659,0.6917077594207605,0.6914253905912846\n"
     ]
    }
   ],
   "source": [
    "##### predictions\n",
    "y_pred = model.predict([X1_test,X2_test]) ###classification\n",
    "\n",
    "y_act = np.argmax(y_test, axis=1)\n",
    "y_mod = np.argmax(y_pred, axis=1)\n",
    "\n",
    "\n",
    "###\n",
    "res_list = roc_auc_score(y_test, y_pred, average=None, multi_class='ovr').tolist()+ \\\n",
    "      average_precision_score(y_test, y_pred, average=None).tolist()+ \\\n",
    "      [matthews_corrcoef(y_act, y_mod)]+[cohen_kappa_score(y_act, y_mod)]\n",
    "\n",
    "print(*res_list, sep=\",\")\n",
    "\n",
    "# Compute ROC/AUC\n",
    "fpr = dict()\n",
    "tpr = dict()\n",
    "roc_auc = dict()\n",
    "\n",
    "for i in range(3):\n",
    "    fpr[i], tpr[i], _ = roc_curve(y_test[:, i], y_pred[:, i])\n",
    "    roc_auc[i] = auc(fpr[i], tpr[i])\n",
    "    \n",
    "# compute PR/AUPR\n",
    "precision, recall, average_precision = dict(), dict(), dict()\n",
    "\n",
    "for i in range(3):\n",
    "    precision[i], recall[i], _ = precision_recall_curve(y_test[:, i], y_pred[:, i])\n",
    "    average_precision[i] = average_precision_score(y_test[:, i], y_pred[:, i])\n",
    "\n",
    "# A \"micro-average\": quantifying score on all classes jointly\n",
    "precision[\"micro\"], recall[\"micro\"], _ = precision_recall_curve(y_test.ravel(), y_pred.ravel())\n",
    "average_precision[\"micro\"] = average_precision_score(y_test, y_pred, average=\"micro\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "761245a4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "###plots\n",
    "fig, ax = plt.subplots(1,2,figsize=(9,5))\n",
    "\n",
    "for i, activity in zip(range(3),['Active', 'Intermediate','Inactive']):\n",
    "    ax[0].plot(fpr[i], tpr[i], label=activity+' (AUC: %0.2f)' % roc_auc[i], alpha=1)\n",
    "ax[0].plot([0, 1], [0, 1], 'k--')\n",
    "ax[0].set_ylim([0.0, 1.01])\n",
    "ax[0].set_xlim([0.0, 1.01])\n",
    "ax[0].set_yticks(np.arange(0, 1.1, 0.1))\n",
    "ax[0].set_xticks(np.arange(0, 1.1, 0.1))\n",
    "ax[0].set_title('ROC curves', fontsize=14)\n",
    "ax[0].set_xlabel('False Positive Rate', fontsize=14)\n",
    "ax[0].set_ylabel('True Positive Rate', fontsize=14)\n",
    "\n",
    "ax[0].tick_params(axis=\"x\", labelsize=12)\n",
    "ax[0].tick_params(axis=\"y\", labelsize=12) \n",
    "ax[0].grid(linestyle='-.', linewidth=0.7)\n",
    "ax[0].legend(fontsize=12, loc='lower right')\n",
    "\n",
    "# ############### auPR plot\n",
    "for i, activity in zip(range(3),['Active', 'Intermediate','Inactive']):\n",
    "    ax[1].plot(recall[i],precision[i], label=activity+' (AUC: %0.2f)' % average_precision[i], alpha=1)\n",
    "##################################\n",
    "#ax[1].step(recall['micro'], precision['micro'], where='post')\n",
    "ax[1].set_xlabel('Recall', fontsize=14)\n",
    "ax[1].set_ylabel('Precision', fontsize=14)\n",
    "ax[1].plot([0, 1], [1, 0], 'k--')\n",
    "ax[1].set_ylim([0.0, 1.01])\n",
    "ax[1].set_xlim([0.0, 1.00])\n",
    "ax[1].set_yticks(np.arange(0, 1.1, 0.1))\n",
    "ax[1].set_xticks(np.arange(0, 1.1, 0.1))\n",
    "ax[1].set_title('Precision-recall (micro avg :{0:0.2f})'.format(average_precision[\"micro\"]), fontsize=14)\n",
    "ax[1].tick_params(axis=\"x\", labelsize=12)\n",
    "ax[1].tick_params(axis=\"y\", labelsize=12) \n",
    "ax[1].grid(linestyle='-.', linewidth=0.7)\n",
    "ax[1].legend(fontsize=12, loc='lower right')\n",
    "fig.suptitle(\"Classification performance (%s)\"%(emb.upper()), fontsize=15)\n",
    "plt.tight_layout()\n",
    "plt.savefig('roc_auc_%s.png'%(emb), dpi=600, transparent=False, bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e0512c29",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}