Project_Code3_BiLSTM.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Bioinformatics Project - BiLSTM @TarekHasan.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "JjHo5wIlI8ZA",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "9d29ff7e-ac3d-40b5-8ab3-cd7e3cf36e6f"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive/\n"
          ]
        }
      ],
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive/')"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "file = open(\"/content/drive/MyDrive/Introduction to Bioinformatics/Project Proposal/Dataset/histone/H3K4me2.txt\",\"r\")\n",
        "sequences = []\n",
        "class_label = []\n",
        "for line in file:\n",
        "    if '>' in line:\n",
        "        continue\n",
        "    line = line[:-1]\n",
        "    if len(line)>1:\n",
        "        sequences.append(line)\n",
        "    else:\n",
        "        class_label.append(int(line)) \n",
        "file.close()\n",
        "removable = []\n",
        "for i in range(len(sequences)):\n",
        "    if len(sequences[i]) < 500:\n",
        "        print(len(sequences[i]))\n",
        "        removable.append(i)\n",
        "\n",
        "iter = 0\n",
        "for x in removable:\n",
        "    sequences.pop(x-iter)\n",
        "    class_label.pop(x-iter)\n",
        "    iter+=1\n",
        "\n",
        "one = 0\n",
        "zero = 0\n",
        "for c in class_label:\n",
        "    if c == 0:\n",
        "        zero+=1\n",
        "    else:\n",
        "        one+=1\n",
        "print(zero, one)\n",
        "print(zero/(zero+one))\n",
        "print(one/(zero+one))\n",
        "class_label = np.array(class_label)"
      ],
      "metadata": {
        "id": "lgXv-KHTKKZ6",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "89cd28b9-2f0d-4e9d-f91d-f28a3431c26d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "310\n",
            "290\n",
            "12539 18142\n",
            "0.4086894169029693\n",
            "0.5913105830970308\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from numpy import array\n",
        "from keras.preprocessing.text import one_hot\n",
        "from keras.preprocessing.sequence import pad_sequences\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense\n",
        "from keras.layers import Flatten\n",
        "from keras.layers.embeddings import Embedding\n",
        "from keras.layers import Dense, Activation, Dropout, SpatialDropout1D, Bidirectional, Flatten\n",
        "from keras.layers.recurrent import LSTM"
      ],
      "metadata": {
        "id": "gOsSVCkgK0e5"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def SeqtoSen(sequences):\n",
        "    sentences = []\n",
        "    for sequence in sequences:\n",
        "        sentence = \"\"\n",
        "        for i in range(len(sequence)-2):\n",
        "            if (len(sentence)>0):\n",
        "                sentence = sentence + \" \";\n",
        "            sentence = sentence + sequence[i:i+3]\n",
        "        sentences.append(sentence)\n",
        "    return sentences"
      ],
      "metadata": {
        "id": "DlRdEPyEKVC5"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "sentences = ['CATGTAGTATTGGGC']\n",
        "print(SeqtoSen(sentences))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Vs6utlkuqeOd",
        "outputId": "ed12c9f6-da59-4428-ceda-d4ad2e6f9b5d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "['CAT ATG TGT GTA TAG AGT GTA TAT ATT TTG TGG GGG GGC']\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "sequences = SeqtoSen(sequences)"
      ],
      "metadata": {
        "id": "4j4Av9nULJGz"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "\n",
        "vocab_size = 64\n",
        "encoded_docs = [one_hot(sequence, vocab_size) for sequence in sequences]\n",
        "for i in range(len(encoded_docs)):\n",
        "    encoded_docs[i] = np.array(encoded_docs[i])\n",
        "encoded_docs = np.array(encoded_docs)"
      ],
      "metadata": {
        "id": "16N2QTKLLfB5"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def get_model():\n",
        "    model = Sequential()\n",
        "    model.add(Embedding(input_dim = vocab_size, output_dim = 256, input_length=498))\n",
        "    model.add(SpatialDropout1D(0.2))\n",
        "    model.add(Bidirectional(LSTM(units=256,return_sequences = True)))\n",
        "    model.add(Flatten())\n",
        "    model.add(Dense(128,activation='relu'))\n",
        "    model.add(Dropout(0.5))\n",
        "    model.add(Dense(2, activation='softmax'))\n",
        "    model.compile(optimizer='adam',loss='sparse_categorical_crossentropy',metrics=['accuracy'])\n",
        "    print(model.summary())\n",
        "    return model"
      ],
      "metadata": {
        "id": "sqLJq-SrNxYM"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import tensorflow as tf\n",
        "import keras\n",
        "from keras.callbacks import EarlyStopping, ModelCheckpoint\n",
        "from sklearn.metrics import accuracy_score, f1_score, precision_score, recall_score, matthews_corrcoef\n",
        "\n",
        "es = EarlyStopping(monitor='val_accuracy',mode='max',verbose=1,patience=200)\n",
        "mc = ModelCheckpoint('best_model.h5',monitor='val_accuracy',mode='max',verbose=1,save_best_only=True, save_weights_only=True)"
      ],
      "metadata": {
        "id": "aLP3JqZdO_yR"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from keras.utils.vis_utils import plot_model\n",
        "model = get_model()\n",
        "plot_model(model,to_file='model.png',show_shapes=True,show_layer_names=True)"
      ],
      "metadata": {
        "id": "fDvT7c_22gf-",
        "outputId": "c2ab6c00-3850-4c3d-ecba-37dae8d79a9c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " embedding (Embedding)       (None, 498, 256)          16384     \n",
            "                                                                 \n",
            " spatial_dropout1d (SpatialD  (None, 498, 256)         0         \n",
            " ropout1D)                                                       \n",
            "                                                                 \n",
            " bidirectional (Bidirectiona  (None, 498, 512)         1050624   \n",
            " l)                                                              \n",
            "                                                                 \n",
            " flatten (Flatten)           (None, 254976)            0         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 128)               32637056  \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 128)               0         \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 2)                 258       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 33,704,322\n",
            "Trainable params: 33,704,322\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "X_train_val, X_test, y_train_val, y_test = train_test_split(encoded_docs, class_label, train_size = 0.8, random_state=42, stratify=class_label)\n",
        "train_X, val_X, train_y, val_y = train_test_split(X_train_val, y_train_val, train_size = 0.9, random_state = 42, stratify=y_train_val)\n",
        "model.fit(train_X,train_y, validation_data=(val_X,val_y), epochs=300,batch_size=256,callbacks=[es,mc])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GRuEzPAAQbg8",
        "outputId": "85d888fa-e938-404e-d8e8-d77c5e8fa3b0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_10\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " embedding_10 (Embedding)    (None, 498, 256)          16384     \n",
            "                                                                 \n",
            " spatial_dropout1d_10 (Spati  (None, 498, 256)         0         \n",
            " alDropout1D)                                                    \n",
            "                                                                 \n",
            " bidirectional_8 (Bidirectio  (None, 498, 512)         1050624   \n",
            " nal)                                                            \n",
            "                                                                 \n",
            " flatten_7 (Flatten)         (None, 254976)            0         \n",
            "                                                                 \n",
            " dense_13 (Dense)            (None, 128)               32637056  \n",
            "                                                                 \n",
            " dropout_6 (Dropout)         (None, 128)               0         \n",
            "                                                                 \n",
            " dense_14 (Dense)            (None, 2)                 258       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 33,704,322\n",
            "Trainable params: 33,704,322\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n",
            "Epoch 1/300\n",
            "87/87 [==============================] - ETA: 0s - loss: 0.6731 - accuracy: 0.6079\n",
            "Epoch 00001: val_accuracy improved from -inf to 0.63014, saving model to best_model.h5\n",
            "87/87 [==============================] - 360s 4s/step - loss: 0.6731 - accuracy: 0.6079 - val_loss: 0.6435 - val_accuracy: 0.6301\n",
            "Epoch 2/300\n",
            "87/87 [==============================] - ETA: 0s - loss: 0.6344 - accuracy: 0.6525\n",
            "Epoch 00002: val_accuracy improved from 0.63014 to 0.63910, saving model to best_model.h5\n",
            "87/87 [==============================] - 354s 4s/step - loss: 0.6344 - accuracy: 0.6525 - val_loss: 0.6355 - val_accuracy: 0.6391\n",
            "Epoch 3/300\n",
            "87/87 [==============================] - ETA: 0s - loss: 0.6062 - accuracy: 0.6768\n",
            "Epoch 00003: val_accuracy did not improve from 0.63910\n",
            "87/87 [==============================] - 356s 4s/step - loss: 0.6062 - accuracy: 0.6768 - val_loss: 0.6467 - val_accuracy: 0.6367\n",
            "Epoch 4/300\n",
            "87/87 [==============================] - ETA: 0s - loss: 0.5567 - accuracy: 0.7173\n",
            "Epoch 00004: val_accuracy did not improve from 0.63910\n",
            "87/87 [==============================] - 358s 4s/step - loss: 0.5567 - accuracy: 0.7173 - val_loss: 0.6692 - val_accuracy: 0.6375\n",
            "Epoch 5/300\n",
            "87/87 [==============================] - ETA: 0s - loss: 0.4847 - accuracy: 0.7700\n",
            "Epoch 00005: val_accuracy did not improve from 0.63910\n",
            "87/87 [==============================] - 354s 4s/step - loss: 0.4847 - accuracy: 0.7700 - val_loss: 0.6929 - val_accuracy: 0.6387\n",
            "Epoch 6/300\n",
            "87/87 [==============================] - ETA: 0s - loss: 0.4003 - accuracy: 0.8151\n",
            "Epoch 00006: val_accuracy did not improve from 0.63910\n",
            "87/87 [==============================] - 354s 4s/step - loss: 0.4003 - accuracy: 0.8151 - val_loss: 0.7502 - val_accuracy: 0.6151\n",
            "Epoch 7/300\n",
            "17/87 [====>.........................] - ETA: 4:40 - loss: 0.3315 - accuracy: 0.8497"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.load_weights('best_model.h5')\n",
        "p = model.predict(X_test)\n",
        "prediction = []\n",
        "for x in p:\n",
        "    if(x[0]>x[1]):\n",
        "        prediction.append(0)\n",
        "    else:\n",
        "        prediction.append(1)\n",
        "\n",
        "f = open(output_file, \"w\")\n",
        "f.write(f'Accuracy: {accuracy_score(y_test,prediction)}\\n')\n",
        "f.write(f'Precision: {precision_score(y_test,prediction)}\\n')\n",
        "f.write(f'Recall: {recall_score(y_test,prediction)}\\n')\n",
        "f.write(f'F1 Score: {f1_score(y_test,prediction)}\\n')\n",
        "f.write(f'MCC Score: {matthews_corrcoef(y_test,prediction)}\\n')\n",
        "f.close()"
      ],
      "metadata": {
        "id": "opUY6VowR_w_"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}