pss_prediction_notebook.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "82fb793b-de91-4aab-92ce-476493c83902",
   "metadata": {},
   "source": [
    "# Summary"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2b4eaf13-6f0e-4e2d-9693-399fa060da1d",
   "metadata": {},
   "source": [
    "The notebook consists of the following parts.\n",
    "\n",
    "1.   Importing Modules: Imports all the necessary modules.\n",
    "2.   Data Preprocessing: Parses and preprocesses the data. Creates the cross-validation sets and the blind test stes.\n",
    "3.   Model Development: Contains some utility functions, 2 suboptimal models, the final model trained/tested on the cross-validation sets, the final model tested on the blind test set, and several metrics plots for the predicted classes.\n",
    "4.   User Interface: Contains some utility functions and the user interface when the user inserts one sequence PSSM file and gets the resulted predicted secondary structure together with the respective probabilities of prediction."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7314fd00-7cd0-4495-9136-45fd161a9344",
   "metadata": {},
   "source": [
    "# Import Modules"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "850635fb-f60f-4d52-8807-1c9fbdd6632d",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\Users\\kanto\\anaconda3\\Lib\\site-packages\\keras\\src\\losses.py:2976: The name tf.losses.sparse_softmax_cross_entropy is deprecated. Please use tf.compat.v1.losses.sparse_softmax_cross_entropy instead.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras import layers, Model\n",
    "from keras_tuner.tuners import BayesianOptimization\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from matplotlib import pyplot as plt\n",
    "from matplotlib.font_manager import FontProperties\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "66c73d02-6326-4ffb-9e2e-38c8dbb6bec7",
   "metadata": {},
   "source": [
    "# Data Preprocessing"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6058f177-4317-4b35-a0c8-3d0eb0a32e9f",
   "metadata": {},
   "source": [
    "## Parsing data functions - PSSM, DSSP, sequences"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3befdc55-4322-4523-918b-604af3921c61",
   "metadata": {},
   "source": [
    "Utility functions to parse and preprocess data. DSSP labels, PSSM and amino acid sequence features.\n",
    "\n",
    "Preprocessing steps:\n",
    "- Parse PSSM, DSSP and sequence files.\n",
    "- Concatenate vertically all examples.\n",
    "- Use sliding window size method to improve localization and extract features from PSSM. \n",
    "- Encode with numbers and one-hot-encoding for DSSP.\n",
    "- Use one-hot-encoding and sliding window size method to improve localization and extract features from amino acid sequences.\n",
    "\n",
    "Get data ready by using all predefined functions. Output data: \n",
    "- a 2D numpy array \"profile_x\" with PSSM matrix extracted features.\n",
    "- a 2D numpy array \"seq_x\" with one-hot-encoded amino acid sequences extracted features.\n",
    "- a 2D numpy array \"ss_y\" with one-hot-encoded DSSP."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "9ee0c679-80d0-4b93-982f-c15ef00293ec",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def parse_pssm(pssm_filename):\n",
    "    \"\"\"\n",
    "    Parses a PSSM (Position Specific Scoring Matrix) file.\n",
    "\n",
    "    Args:\n",
    "        pssm_filename (str): The filename of the PSSM file to parse.\n",
    "\n",
    "    Returns:\n",
    "        list: A 2D list representing the parsed profile.\n",
    "\n",
    "    \"\"\"\n",
    "    profile = []\n",
    "    with open(pssm_filename) as pssm:\n",
    "        pssm_lines = pssm.readlines()\n",
    "        \n",
    "        for line in pssm_lines[3:-6]:\n",
    "            profile_line = []\n",
    "            \n",
    "            for n in line.rstrip().split()[22:-2]:\n",
    "                profile_line.append(float(n)/100)\n",
    "                \n",
    "            profile.append(profile_line)\n",
    "            \n",
    "    return profile\n",
    "\n",
    "\n",
    "def parse_dssp(dssp_filename):\n",
    "    \"\"\"\n",
    "    Parses a DSSP (Dictionary of Secondary Structure of Proteins) file.\n",
    "\n",
    "    Args:\n",
    "        dssp_filename (str): The filename of the DSSP file to parse.\n",
    "\n",
    "    Returns:\n",
    "        list: A list representing the secondary structure.\n",
    "\n",
    "    \"\"\"\n",
    "    ss = ''\n",
    "    with open(dssp_filename) as dssp:\n",
    "        dssp.readline()\n",
    "        ss = dssp.readline().rstrip()\n",
    "        ss = ss.replace('C','-')\n",
    "        ss = list(ss)\n",
    "        \n",
    "    return ss\n",
    "\n",
    "\n",
    "def extract_sequence_from_pssm(pssm_filename):\n",
    "    \"\"\"\n",
    "    Extracts amino acid sequence from a PSSM file.\n",
    "\n",
    "    Args:\n",
    "        pssm_filename (str): The filename of the PSSM file.\n",
    "\n",
    "    Returns:\n",
    "        list: A list representing the amino acid sequence.\n",
    "\n",
    "    \"\"\"\n",
    "    with open(pssm_filename) as pssm:\n",
    "        pssm_lines = pssm.readlines()\n",
    "        seq = []\n",
    "        \n",
    "        for line in pssm_lines[3:-6]:\n",
    "            line = line.rstrip().split()\n",
    "            seq.append(line[1])\n",
    "            \n",
    "    return seq\n",
    "\n",
    "\n",
    "def sequence_to_one_hot(seq):\n",
    "    \"\"\"\n",
    "    Converts amino acid sequence to one-hot encoding.\n",
    "\n",
    "    Args:\n",
    "        seq (list): A list representing the amino acid sequence.\n",
    "\n",
    "    Returns:\n",
    "        numpy.ndarray: A 2D numpy array representing the one-hot encoded sequence.\n",
    "\n",
    "    \"\"\"\n",
    "    amino_acids = ['A', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'K', 'L', 'M', 'N', 'P', 'Q', 'R', 'S', 'T', 'V', 'W', 'Y']\n",
    "    one_hot_seq = np.zeros((len(seq), len(amino_acids)))\n",
    "    for i, aa in enumerate(seq):\n",
    "        if aa in amino_acids:\n",
    "            one_hot_seq[i, amino_acids.index(aa)] = 1\n",
    "    return one_hot_seq\n",
    "\n",
    "\n",
    "def create_profiles(profile, ss, seq, profile_x, ss_y, seq_x, window_size):\n",
    "    \"\"\"\n",
    "    Creates profiles for deep learning models.\n",
    "\n",
    "    Args:\n",
    "        profile (list): A 2D list representing the profile.\n",
    "        ss (list): A list representing the secondary structure.\n",
    "        seq (list): A list representing the amino acid sequence.\n",
    "        profile_x (list): A list to store profile features.\n",
    "        ss_y (list): A list to store secondary structure labels.\n",
    "        seq_x (list): A list to store sequence features.\n",
    "        window_size (int): The window size for creating profiles.\n",
    "\n",
    "    Returns:\n",
    "        tuple: A tuple containing the updated profile_x, ss_y, and seq_x lists.\n",
    "\n",
    "    \"\"\"\n",
    "    ss_map = {'H': 0, 'E': 1, 'C': 2, '-': 2}\n",
    "    \n",
    "    for i in range(0, len(profile)):\n",
    "        half_window_size = int((window_size-1)/2)\n",
    "        # Apply the sliding window method to PSSM profiles\n",
    "        part1 = np.zeros(20*max(0, half_window_size-i))\n",
    "        part2 = np.ndarray.flatten(profile[max(0,i-half_window_size):min(i+half_window_size+1,len(profile))])\n",
    "        part3 = np.zeros(20*max(0, half_window_size-(len(profile)-i-1)))\n",
    "        vec = np.concatenate((part1, part2, part3))\n",
    "        profile_x.append(vec.tolist())\n",
    "                    \n",
    "        ss_y.append(ss_map[ss[i]])\n",
    "                    \n",
    "        one_hot_seq = sequence_to_one_hot(seq)\n",
    "        # Apply the sliding window method to one-hot-encoded sequences\n",
    "        part1 = np.zeros(20*max(0, half_window_size-i))\n",
    "        part2 = np.ndarray.flatten(one_hot_seq[max(0,i-half_window_size):min(i+half_window_size+1,len(one_hot_seq))])\n",
    "        part3 = np.zeros(20*max(0, half_window_size-(len(one_hot_seq)-i-1)))\n",
    "        vec = np.concatenate((part1, part2, part3))\n",
    "        seq_x.append(vec.tolist())\n",
    "                    \n",
    "    return profile_x, ss_y, seq_x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "6f04bfb1-935f-4467-a06f-ea505592ffd9",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def preprocess_data(list_file_path, folder_file_path, window_size = 15, str_transformation = False):\n",
    "    \"\"\"\n",
    "    Preprocesses data for deep learning models.\n",
    "\n",
    "    Args:\n",
    "        list_file_path (str): The path to the file containing a list of IDs.\n",
    "        folder_file_path (str): The path to the folder containing PSSM and DSSP files.\n",
    "        window_size (int, optional): The window size for creating profiles. Defaults to 15.\n",
    "        str_transformation (bool, optional): Whether to perform string transformation. Defaults to False.\n",
    "\n",
    "    Returns:\n",
    "        tuple: A tuple containing the profile features, sequence features, and secondary structure labels.\n",
    "\n",
    "    \"\"\"\n",
    "    with open(list_file_path) as id_list:\n",
    "        profile_x = []\n",
    "        ss_y = []\n",
    "        seq_x = []\n",
    "\n",
    "        for line in id_list:\n",
    "            line = line.rstrip()\n",
    "            \n",
    "            # Solve problem with file names in blindTest folder, they have ':' instead of '_'\n",
    "            if str_transformation == True:\n",
    "                line = line.replace(\":\", \"_\")\n",
    "                \n",
    "            profile = np.array(parse_pssm(folder_file_path+'pssm/'+line+'.pssm'))\n",
    "            \n",
    "            if np.sum(profile) != 0:\n",
    "                ss = parse_dssp(folder_file_path+'dssp/'+line+'.dssp')\n",
    "                seq = extract_sequence_from_pssm(folder_file_path+'pssm/'+line+'.pssm')\n",
    "                profile_x, ss_y, seq_x = create_profiles(profile, ss, seq, profile_x, ss_y, seq_x, window_size)\n",
    "            \n",
    "        profile_x = np.array(profile_x)\n",
    "        seq_x = np.array(seq_x)\n",
    "        ss_y = np.array(ss_y)\n",
    "        ss_y = keras.utils.to_categorical(ss_y, 3)\n",
    "        \n",
    "    return profile_x, seq_x, ss_y "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "53cde423-7301-4993-a6ad-dd6e3cf8d6d3",
   "metadata": {},
   "source": [
    "## Apply functions to data - training, test, cross-validation sets"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bc073ff7-792c-405c-a5c6-fbd68a0f2ccc",
   "metadata": {},
   "source": [
    "Creating cross-validation features and labels."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "aead004e-6ddf-4944-bf54-9606875947d9",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def create_validation_sets(window_size):\n",
    "    \"\"\"\n",
    "    Creates validation sets for cross-validation.\n",
    "\n",
    "    Args:\n",
    "        window_size (int): The window size for creating profiles.\n",
    "\n",
    "    Returns:\n",
    "        tuple: A tuple containing lists of profile features, sequence features, and secondary structure labels for each \n",
    "        validation set.\n",
    "\n",
    "    \"\"\"\n",
    "    names_list = [\n",
    "        ['train1', 'test1'], \n",
    "        ['train2', 'test2'], \n",
    "        ['train3', 'test3'], \n",
    "        ['train4', 'test4'], \n",
    "        ['train5', 'test5']\n",
    "    ]\n",
    "\n",
    "    profile_x_list = []\n",
    "    seq_x_list = []\n",
    "    ss_y_list = []\n",
    "\n",
    "    for train_name, test_name in names_list:\n",
    "        profile_x_sublist = []\n",
    "        seq_x_sublist = []\n",
    "        ss_y_sublist = []\n",
    "\n",
    "        train_path = 'Data/training/cv/' + train_name + '.txt'\n",
    "        test_path = 'Data/training/cv/' + test_name + '.txt'\n",
    "        print(f'{train_name} is preprocessed.')\n",
    "\n",
    "        profile_x, seq_x, ss_y  = preprocess_data(train_path,\n",
    "                                                  'Data/training/',\n",
    "                                                  window_size = window_size,\n",
    "                                                  str_transformation = True\n",
    "                                                 )\n",
    "        profile_x_sublist.append(profile_x)\n",
    "        seq_x_sublist.append(seq_x)\n",
    "        ss_y_sublist.append(ss_y)\n",
    "\n",
    "        print(f'{test_name} is preprocessed.')\n",
    "\n",
    "        profile_x_test, seq_x_test, ss_y_test  = preprocess_data(test_path,\n",
    "                                                                 'Data/training/',\n",
    "                                                                 window_size = window_size,\n",
    "                                                                 str_transformation = True\n",
    "                                                                )\n",
    "        profile_x_sublist.append(profile_x_test)\n",
    "        seq_x_sublist.append(seq_x_test)\n",
    "        ss_y_sublist.append(ss_y_test)\n",
    "\n",
    "        profile_x_list.append(profile_x_sublist)\n",
    "        seq_x_list.append(seq_x_sublist)\n",
    "        ss_y_list.append(ss_y_sublist)\n",
    "        \n",
    "    return profile_x_list, seq_x_list, ss_y_list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "a5b189b3-6e10-497f-a9d1-8fa254e28a21",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train1 is preprocessed.\n",
      "test1 is preprocessed.\n",
      "train2 is preprocessed.\n",
      "test2 is preprocessed.\n",
      "train3 is preprocessed.\n",
      "test3 is preprocessed.\n",
      "train4 is preprocessed.\n",
      "test4 is preprocessed.\n",
      "train5 is preprocessed.\n",
      "test5 is preprocessed.\n"
     ]
    }
   ],
   "source": [
    "profile_x_list, seq_x_list, ss_y_list = create_validation_sets(window_size = 17)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "890ad03a-7591-493e-a4fe-345b8d9481c9",
   "metadata": {},
   "source": [
    "Creating blind test features and labels."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "04d1434b-2df7-4130-bba5-ed15f212fd34",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "profile_x_blindtest, seq_x_blindtest, ss_y_blindtest  = preprocess_data(\n",
    "    'Data/blindTest/list.txt',\n",
    "    'Data/blindTest/',\n",
    "    window_size = 17,\n",
    "    str_transformation = True\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2e9c02e6-4c6b-4dc9-99c4-481fe444280b",
   "metadata": {},
   "source": [
    "# Model Development"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "810b2e05-0551-47d1-b95b-d333bd51e73b",
   "metadata": {},
   "source": [
    "## Utility functions"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d8a91443-6ade-4cb4-b4a2-96ffb51b42a8",
   "metadata": {},
   "source": [
    "Calculate mean and standard deviation of the accuracy of the five cross-validation sets. Print them as table."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "d4c45361-1a68-4fd8-85bf-ec2c368ac38c",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def print_mean_std_table(mean_accuracy, std_accuracy, mean_loss, std_loss):\n",
    "    \"\"\"\n",
    "    Prints a table of mean accuracy, standard deviation of accuracy, mean loss, and standard deviation of loss.\n",
    "\n",
    "    Args:\n",
    "        mean_accuracy (float): The mean accuracy value.\n",
    "        std_accuracy (float): The standard deviation of accuracy.\n",
    "        mean_loss (float): The mean loss value.\n",
    "        std_loss (float): The standard deviation of loss.\n",
    "\n",
    "    Returns:\n",
    "        None\n",
    "\n",
    "    \"\"\"\n",
    "    print(f'Mean Accuracy:     {mean_accuracy}')\n",
    "    print(f'Std Dev Accuracy:  {std_accuracy}')\n",
    "    print('-----------------------------')\n",
    "    print(f'Mean Loss:         {mean_loss}')\n",
    "    print(f'Std Dev Loss:      {std_loss}')\n",
    "\n",
    "\n",
    "def calculate_mean_std(accuracy_cv, loss_cv):\n",
    "    \"\"\"\n",
    "    Calculates the mean and standard deviation of accuracy and loss.\n",
    "\n",
    "    Args:\n",
    "        accuracy_cv (list): A list of accuracy values from cross-validation sets.\n",
    "        loss_cv (list): A list of loss values from cross-validation sets.\n",
    "\n",
    "    Returns:\n",
    "        tuple: A tuple containing the mean accuracy, standard deviation of accuracy, mean loss, and standard deviation of loss.\n",
    "\n",
    "    \"\"\"\n",
    "    mean_accuracy = np.mean(accuracy_cv)\n",
    "    std_accuracy = np.std(accuracy_cv)\n",
    "    mean_loss = np.mean(loss_cv)\n",
    "    std_loss = np.std(loss_cv)\n",
    "    \n",
    "    print_mean_std_table(mean_accuracy, std_accuracy, mean_loss, std_loss)\n",
    "    \n",
    "    return mean_accuracy, std_accuracy, mean_loss, std_loss"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "92fce594-6360-4c75-b069-3f92e22b57b7",
   "metadata": {},
   "source": [
    "Create plots for model accuracy or loss vs number of epoch for each of the five cross-validation sets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "8839b77e-d6d6-4cc5-996d-9084e4907bd7",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def plot_accuracy_epoch(history_cv):\n",
    "    \"\"\"\n",
    "    Plots the accuracy versus epoch for cross-validation sets.\n",
    "\n",
    "    Args:\n",
    "        history_cv (list): A list of Keras History objects, each corresponding to a cross-validation set.\n",
    "\n",
    "    Returns:\n",
    "        None\n",
    "\n",
    "    \"\"\"\n",
    "    sns.set_style('ticks')\n",
    "    palette = sns.color_palette('deep')\n",
    "    plt.figure(figsize=(4, 4))\n",
    "    for i, history in enumerate(history_cv, start=1):\n",
    "        plt.plot(history.history['accuracy'], label=f'{i} cv set', color=palette[i-1])\n",
    "    plt.xlabel('Epoch', fontname='Arial', fontsize=11)\n",
    "    plt.ylabel('Accuracy', fontname='Arial', fontsize=11)\n",
    "    plt.legend(fontsize=11)\n",
    "    plt.show()\n",
    "\n",
    "    \n",
    "def plot_loss_epoch(history_cv):\n",
    "    \"\"\"\n",
    "    Plots the loss versus epoch for cross-validation sets.\n",
    "\n",
    "    Args:\n",
    "        history_cv (list): A list of Keras History objects, each corresponding to a cross-validation set.\n",
    "\n",
    "    Returns:\n",
    "        None\n",
    "\n",
    "    \"\"\"\n",
    "    sns.set_style('ticks')\n",
    "    palette = sns.color_palette('deep')\n",
    "\n",
    "    plt.figure(figsize=(4, 4))\n",
    "    for i, history in enumerate(history_cv, start=1):\n",
    "        plt.plot(history.history['loss'], label=f'{i} cv set', color=palette[i-1])\n",
    "    plt.xlabel('Epoch', fontname='Arial', fontsize=11)\n",
    "    plt.ylabel('Loss', fontname='Arial', fontsize=11)\n",
    "    plt.legend(fontsize=11)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "286817e0-b028-4251-b314-61f4f6742dea",
   "metadata": {},
   "source": [
    "## Model Screening & Optimization"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d11c3199-3f9d-418d-b48f-703234314ffc",
   "metadata": {},
   "source": [
    "### Select best optimizer and hidden layer activation function\n",
    "\n",
    "Steps:\n",
    "- Develop simple model (baseline)\n",
    "- Try different combination of optimizers and activation functions.\n",
    "- Evaluate the model with the respective cross-validation test sets.\n",
    "- Collect and store accuracy metrics in list.\n",
    "\n",
    "Selected:\n",
    "- Optimizer: Adam.\n",
    "- Hidden Layers Activation Function: ReLu"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "f1802108-d340-4fed-8927-f24f59e0cf78",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [],
   "source": [
    "def develop_simple_model(profile_x, ss_y, seq_x, profile_x_test, ss_y_test, seq_x_test, accuracy_cv, loss_cv, history_cv, ss_pred_cv):\n",
    "    \"\"\"\n",
    "    Develops a simple neural network model for secondary structure prediction.\n",
    "\n",
    "    Args:\n",
    "        profile_x (numpy.ndarray): Training profile features.\n",
    "        ss_y (numpy.ndarray): Training secondary structure labels.\n",
    "        seq_x (numpy.ndarray): Training sequence features.\n",
    "        profile_x_test (numpy.ndarray): Test profile features.\n",
    "        ss_y_test (numpy.ndarray): Test secondary structure labels.\n",
    "        seq_x_test (numpy.ndarray): Test sequence features.\n",
    "        accuracy_cv (list): List to store accuracy values from cross-validation sets.\n",
    "        loss_cv (list): List to store loss values from cross-validation sets.\n",
    "        history_cv (list): List to store Keras History objects from cross-validation sets.\n",
    "        ss_pred_cv (list): List to store predicted secondary structures from cross-validation sets.\n",
    "\n",
    "    Returns:\n",
    "        tuple: A tuple containing updated accuracy_cv, loss_cv, history_cv, and ss_pred_cv lists.\n",
    "\n",
    "    \"\"\"\n",
    "    model = keras.Sequential([\n",
    "        layers.Input(shape=(profile_x.shape[1])),\n",
    "        layers.Flatten(),\n",
    "        layers.Dense(64, activation='relu'),\n",
    "        layers.Dense(3, activation='softmax')\n",
    "    ])\n",
    "\n",
    "    model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "        \n",
    "    history = model.fit(profile_x, ss_y, epochs=10, batch_size=128)\n",
    "    history_cv.append(history)\n",
    "        \n",
    "    test_loss, test_accuracy = model.evaluate(profile_x_test, ss_y_test)\n",
    "    accuracy_cv.append(test_accuracy)\n",
    "    loss_cv.append(test_loss)\n",
    "        \n",
    "    ss_y_pred = model.predict(profile_x_test)\n",
    "    ss_pred_cv.append(ss_y_pred)\n",
    "    \n",
    "    return accuracy_cv, loss_cv, history_cv, ss_pred_cv\n",
    "\n",
    "\n",
    "def simple_model_cross_validation(profile_x_list, seq_x_list, ss_y_list):\n",
    "    \"\"\"\n",
    "    Performs cross-validation for a simple neural network model for secondary structure prediction.\n",
    "\n",
    "    Args:\n",
    "        profile_x_list (list): A list of lists containing profile features for each cross-validation set.\n",
    "        seq_x_list (list): A list of lists containing sequence features for each cross-validation set.\n",
    "        ss_y_list (list): A list of lists containing secondary structure labels for each cross-validation set.\n",
    "\n",
    "    Returns:\n",
    "        tuple: A tuple containing accuracy_cv, loss_cv, history_cv, ss_pred_cv, and ss_true_cv lists.\n",
    "\n",
    "    \"\"\"\n",
    "    accuracy_cv = []\n",
    "    loss_cv = []\n",
    "    history_cv = []\n",
    "    ss_pred_cv = []\n",
    "    ss_true_cv = []\n",
    "    \n",
    "    for s in range(len(profile_x_list)):       \n",
    "        print(f'Train/test cross-validation set {s+1}.\\n-------------------------------')\n",
    "        profile_x = profile_x_list[s][0]\n",
    "        seq_x = seq_x_list[s][0]\n",
    "        ss_y = ss_y_list[s][0]\n",
    "        profile_x_test = profile_x_list[s][1]\n",
    "        seq_x_test = seq_x_list[s][1]\n",
    "        ss_y_test = ss_y_list[s][1]\n",
    "        ss_true_cv.append(ss_y_test)\n",
    "        \n",
    "        accuracy_cv, loss_cv, history_cv, ss_pred_cv = develop_simple_model(\n",
    "            profile_x, \n",
    "            ss_y, \n",
    "            seq_x, \n",
    "            profile_x_test, \n",
    "            ss_y_test, \n",
    "            seq_x_test, \n",
    "            accuracy_cv, \n",
    "            loss_cv, \n",
    "            history_cv, \n",
    "            ss_pred_cv\n",
    "        )\n",
    "        \n",
    "    return accuracy_cv, loss_cv, history_cv, ss_pred_cv, ss_true_cv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "74a6b98a-fb7e-46d2-81b1-bdc2d9fa0e70",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train/test cross-validation set 1.\n",
      "-------------------------------\n",
      "WARNING:tensorflow:From C:\\Users\\kanto\\anaconda3\\Lib\\site-packages\\keras\\src\\backend.py:1398: The name tf.executing_eagerly_outside_functions is deprecated. Please use tf.compat.v1.executing_eagerly_outside_functions instead.\n",
      "\n",
      "WARNING:tensorflow:From C:\\Users\\kanto\\anaconda3\\Lib\\site-packages\\keras\\src\\optimizers\\__init__.py:309: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
      "\n",
      "Epoch 1/10\n",
      "WARNING:tensorflow:From C:\\Users\\kanto\\anaconda3\\Lib\\site-packages\\keras\\src\\utils\\tf_utils.py:492: The name tf.ragged.RaggedTensorValue is deprecated. Please use tf.compat.v1.ragged.RaggedTensorValue instead.\n",
      "\n",
      "WARNING:tensorflow:From C:\\Users\\kanto\\anaconda3\\Lib\\site-packages\\keras\\src\\engine\\base_layer_utils.py:384: The name tf.executing_eagerly_outside_functions is deprecated. Please use tf.compat.v1.executing_eagerly_outside_functions instead.\n",
      "\n",
      "1230/1230 [==============================] - 4s 2ms/step - loss: 0.7231 - accuracy: 0.6968\n",
      "Epoch 2/10\n",
      "1230/1230 [==============================] - 3s 3ms/step - loss: 0.6565 - accuracy: 0.7293\n",
      "Epoch 3/10\n",
      "1230/1230 [==============================] - 3s 2ms/step - loss: 0.6052 - accuracy: 0.7513\n",
      "Epoch 4/10\n",
      "1230/1230 [==============================] - 3s 2ms/step - loss: 0.5774 - accuracy: 0.7641\n",
      "Epoch 5/10\n",
      "1230/1230 [==============================] - 3s 2ms/step - loss: 0.5599 - accuracy: 0.7712\n",
      "Epoch 6/10\n",
      "1230/1230 [==============================] - 2s 2ms/step - loss: 0.5464 - accuracy: 0.7777\n",
      "Epoch 7/10\n",
      "1230/1230 [==============================] - 2s 2ms/step - loss: 0.5359 - accuracy: 0.7820\n",
      "Epoch 8/10\n",
      "1230/1230 [==============================] - 3s 2ms/step - loss: 0.5270 - accuracy: 0.7862\n",
      "Epoch 9/10\n",
      "1230/1230 [==============================] - 3s 2ms/step - loss: 0.5191 - accuracy: 0.7889\n",
      "Epoch 10/10\n",
      "1230/1230 [==============================] - 3s 2ms/step - loss: 0.5120 - accuracy: 0.7921\n",
      "1273/1273 [==============================] - 4s 3ms/step - loss: 0.5888 - accuracy: 0.7605\n",
      "1273/1273 [==============================] - 3s 3ms/step\n",
      "Train/test cross-validation set 2.\n",
      "-------------------------------\n",
      "Epoch 1/10\n",
      "1241/1241 [==============================] - 5s 4ms/step - loss: 0.7255 - accuracy: 0.6952\n",
      "Epoch 2/10\n",
      "1241/1241 [==============================] - 4s 3ms/step - loss: 0.6486 - accuracy: 0.7338\n",
      "Epoch 3/10\n",
      "1241/1241 [==============================] - 3s 3ms/step - loss: 0.5996 - accuracy: 0.7546\n",
      "Epoch 4/10\n",
      "1241/1241 [==============================] - 3s 2ms/step - loss: 0.5733 - accuracy: 0.7651\n",
      "Epoch 5/10\n",
      "1241/1241 [==============================] - 3s 2ms/step - loss: 0.5564 - accuracy: 0.7729\n",
      "Epoch 6/10\n",
      "1241/1241 [==============================] - 3s 2ms/step - loss: 0.5434 - accuracy: 0.7785\n",
      "Epoch 7/10\n",
      "1241/1241 [==============================] - 3s 2ms/step - loss: 0.5335 - accuracy: 0.7833\n",
      "Epoch 8/10\n",
      "1241/1241 [==============================] - 3s 2ms/step - loss: 0.5251 - accuracy: 0.7868\n",
      "Epoch 9/10\n",
      "1241/1241 [==============================] - 3s 3ms/step - loss: 0.5176 - accuracy: 0.7892\n",
      "Epoch 10/10\n",
      "1241/1241 [==============================] - 3s 2ms/step - loss: 0.5115 - accuracy: 0.7925\n",
      "1229/1229 [==============================] - 3s 2ms/step - loss: 0.5836 - accuracy: 0.7628\n",
      "1229/1229 [==============================] - 3s 3ms/step\n",
      "Train/test cross-validation set 3.\n",
      "-------------------------------\n",
      "Epoch 1/10\n",
      "1238/1238 [==============================] - 5s 3ms/step - loss: 0.7196 - accuracy: 0.6987\n",
      "Epoch 2/10\n",
      "1238/1238 [==============================] - 2s 2ms/step - loss: 0.6362 - accuracy: 0.7391\n",
      "Epoch 3/10\n",
      "1238/1238 [==============================] - 2s 2ms/step - loss: 0.5880 - accuracy: 0.7603\n",
      "Epoch 4/10\n",
      "1238/1238 [==============================] - 3s 2ms/step - loss: 0.5644 - accuracy: 0.7692\n",
      "Epoch 5/10\n",
      "1238/1238 [==============================] - 3s 2ms/step - loss: 0.5490 - accuracy: 0.7769\n",
      "Epoch 6/10\n",
      "1238/1238 [==============================] - 3s 2ms/step - loss: 0.5367 - accuracy: 0.7822\n",
      "Epoch 7/10\n",
      "1238/1238 [==============================] - 3s 3ms/step - loss: 0.5267 - accuracy: 0.7873\n",
      "Epoch 8/10\n",
      "1238/1238 [==============================] - 3s 2ms/step - loss: 0.5184 - accuracy: 0.7903\n",
      "Epoch 9/10\n",
      "1238/1238 [==============================] - 5s 4ms/step - loss: 0.5116 - accuracy: 0.7935\n",
      "Epoch 10/10\n",
      "1238/1238 [==============================] - 4s 3ms/step - loss: 0.5048 - accuracy: 0.7966\n",
      "1241/1241 [==============================] - 3s 2ms/step - loss: 0.6197 - accuracy: 0.7478\n",
      "1241/1241 [==============================] - 2s 2ms/step\n",
      "Train/test cross-validation set 4.\n",
      "-------------------------------\n",
      "Epoch 1/10\n",
      "1245/1245 [==============================] - 4s 2ms/step - loss: 0.7260 - accuracy: 0.6949\n",
      "Epoch 2/10\n",
      "1245/1245 [==============================] - 3s 3ms/step - loss: 0.6535 - accuracy: 0.7293\n",
      "Epoch 3/10\n",
      "1245/1245 [==============================] - 3s 2ms/step - loss: 0.6066 - accuracy: 0.7510\n",
      "Epoch 4/10\n",
      "1245/1245 [==============================] - 3s 2ms/step - loss: 0.5783 - accuracy: 0.7632\n",
      "Epoch 5/10\n",
      "1245/1245 [==============================] - 3s 2ms/step - loss: 0.5595 - accuracy: 0.7709\n",
      "Epoch 6/10\n",
      "1245/1245 [==============================] - 4s 3ms/step - loss: 0.5460 - accuracy: 0.7766\n",
      "Epoch 7/10\n",
      "1245/1245 [==============================] - 3s 2ms/step - loss: 0.5350 - accuracy: 0.7816\n",
      "Epoch 8/10\n",
      "1245/1245 [==============================] - 3s 2ms/step - loss: 0.5257 - accuracy: 0.7859\n",
      "Epoch 9/10\n",
      "1245/1245 [==============================] - 3s 2ms/step - loss: 0.5180 - accuracy: 0.7887\n",
      "Epoch 10/10\n",
      "1245/1245 [==============================] - 3s 2ms/step - loss: 0.5118 - accuracy: 0.7913\n",
      "1213/1213 [==============================] - 2s 2ms/step - loss: 0.5909 - accuracy: 0.7571\n",
      "1213/1213 [==============================] - 2s 2ms/step\n",
      "Train/test cross-validation set 5.\n",
      "-------------------------------\n",
      "Epoch 1/10\n",
      "1239/1239 [==============================] - 3s 2ms/step - loss: 0.7239 - accuracy: 0.6966\n",
      "Epoch 2/10\n",
      "1239/1239 [==============================] - 3s 2ms/step - loss: 0.6518 - accuracy: 0.7320\n",
      "Epoch 3/10\n",
      "1239/1239 [==============================] - 3s 2ms/step - loss: 0.5969 - accuracy: 0.7560\n",
      "Epoch 4/10\n",
      "1239/1239 [==============================] - 3s 2ms/step - loss: 0.5698 - accuracy: 0.7662\n",
      "Epoch 5/10\n",
      "1239/1239 [==============================] - 2s 2ms/step - loss: 0.5535 - accuracy: 0.7734\n",
      "Epoch 6/10\n",
      "1239/1239 [==============================] - 3s 2ms/step - loss: 0.5422 - accuracy: 0.7783\n",
      "Epoch 7/10\n",
      "1239/1239 [==============================] - 3s 2ms/step - loss: 0.5323 - accuracy: 0.7828\n",
      "Epoch 8/10\n",
      "1239/1239 [==============================] - 3s 2ms/step - loss: 0.5240 - accuracy: 0.7862\n",
      "Epoch 9/10\n",
      "1239/1239 [==============================] - 2s 2ms/step - loss: 0.5168 - accuracy: 0.7898\n",
      "Epoch 10/10\n",
      "1239/1239 [==============================] - 2s 2ms/step - loss: 0.5110 - accuracy: 0.7925\n",
      "1239/1239 [==============================] - 2s 2ms/step - loss: 0.5946 - accuracy: 0.7598\n",
      "1239/1239 [==============================] - 2s 2ms/step\n",
      "Mean Accuracy:     0.7576032757759095\n",
      "Std Dev Accuracy:  0.005239895648719019\n",
      "-----------------------------\n",
      "Mean Loss:         0.5955265760421753\n",
      "Std Dev Loss:      0.012617303702922546\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "None\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "None\n"
     ]
    }
   ],
   "source": [
    "accuracy_cv, loss_cv, history_cv, ss_pred_cv, ss_true_cv = simple_model_cross_validation(profile_x_list, seq_x_list, ss_y_list)\n",
    "mean_accuracy, std_accuracy, mean_loss, std_loss = calculate_mean_std(accuracy_cv, loss_cv)\n",
    "print(plot_accuracy_epoch(history_cv))\n",
    "print(plot_loss_epoch(history_cv))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c40cb062-0c3a-47d6-ad63-bee2b1143d73",
   "metadata": {},
   "source": [
    "### Optimize window-size in preprocessing step on simple model\n",
    "\n",
    "- Window size: 7-17 with step 1\n",
    "\n",
    "*Ref: Optimization of the Sliding Window Size for Protein Structure Prediction: (https://ieeexplore.ieee.org/document/4133195)*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "933ee531-8a88-48e1-9954-571ef02b67ec",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [],
   "source": [
    "window_size_list = list(range(7, 18, 2))\n",
    "mean_accuracy_list = []\n",
    "std_accuracy_list = []\n",
    "mean_loss_list = []\n",
    "std_loss_list = []\n",
    "\n",
    "for window_size in window_size_list:\n",
    "    print(f'Window size: {window_size}.\\n-------------------------------\\n-------------------------------')\n",
    "    profile_x_list, seq_x_list, ss_y_list = create_validation_sets(window_size)\n",
    "    accuracy_cv, loss_cv, history_cv, ss_pred_cv, ss_true_cv = simple_model_cross_validation(\n",
    "        profile_x_list, \n",
    "        seq_x_list, \n",
    "        ss_y_list)\n",
    "    mean_accuracy, std_accuracy, mean_loss, std_loss = calculate_mean_std(accuracy_cv, loss_cv)\n",
    "    \n",
    "    mean_accuracy_list.append(mean_accuracy)\n",
    "    std_accuracy_list.append(std_accuracy)\n",
    "    mean_loss_list.append(mean_loss)\n",
    "    std_loss_list.append(std_loss)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2f7585be-1c72-4bfb-a3f0-ea15c17d74b3",
   "metadata": {},
   "source": [
    "Plot accuracy vs window size. The accuracy increases by increasing the window size. After window size = 13, it seems that it reaches a plateau."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "18733202-dcba-48d2-b570-bac767a32449",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "data = {\n",
    "    'Window Size': window_size_list,\n",
    "    'Mean Accuracy': mean_accuracy_list,\n",
    "    'Std Deviation': std_accuracy_list\n",
    "}\n",
    "df = pd.DataFrame(data)\n",
    "\n",
    "sns.set_style('ticks')\n",
    "palette = sns.color_palette('deep')\n",
    "plt.figure(figsize=(8, 4))\n",
    "sns.lineplot(data=df, x='Window Size', y='Mean Accuracy', errorbar='sd', color=palette[0])\n",
    "plt.errorbar(df['Window Size'], df['Mean Accuracy'], yerr=df['Std Deviation'], fmt='o', color=palette[0])\n",
    "plt.xlim(6, 18)\n",
    "plt.ylim(0.7, 0.8)\n",
    "plt.xticks(range(7, 18, 2))\n",
    "plt.xlabel('Window Size', fontname='Arial', fontsize=11)\n",
    "plt.ylabel('Mean Validation Accuracy', fontname='Arial', fontsize=11)\n",
    "plt.savefig('Results/windowsize_vs_accuracy.png', dpi=300)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2a1ec772-d7d7-4651-8d4e-0429f9b44e82",
   "metadata": {
    "tags": []
   },
   "source": [
    "### Model with PSSM input\n",
    "\n",
    "Optimize this model with trial and error. Try different:\n",
    "- Number of hidden layers.\n",
    "- Regularization methods\n",
    "- Activation functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "id": "f47cd221-4cce-4169-a301-d4d4f49d5426",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [],
   "source": [
    "def develop_model(profile_x, ss_y, seq_x, profile_x_test, ss_y_test, seq_x_test, accuracy_cv, loss_cv, history_cv, ss_pred_cv):\n",
    "    \"\"\"\n",
    "    Develops a neural network model for secondary structure prediction.\n",
    "\n",
    "    Args:\n",
    "        profile_x (numpy.ndarray): Training profile features.\n",
    "        ss_y (numpy.ndarray): Training secondary structure labels.\n",
    "        seq_x (numpy.ndarray): Training sequence features.\n",
    "        profile_x_test (numpy.ndarray): Test profile features.\n",
    "        ss_y_test (numpy.ndarray): Test secondary structure labels.\n",
    "        seq_x_test (numpy.ndarray): Test sequence features.\n",
    "        accuracy_cv (list): List to store accuracy values from cross-validation sets.\n",
    "        loss_cv (list): List to store loss values from cross-validation sets.\n",
    "        history_cv (list): List to store Keras History objects from cross-validation sets.\n",
    "        ss_pred_cv (list): List to store predicted secondary structures from cross-validation sets.\n",
    "\n",
    "    Returns:\n",
    "        tuple: A tuple containing updated accuracy_cv, loss_cv, history_cv, and ss_pred_cv lists.\n",
    "\n",
    "    \"\"\"\n",
    "    model = keras.Sequential([\n",
    "        layers.Input(shape=(profile_x.shape[1])),\n",
    "        layers.Dense(1024, activation='relu'),\n",
    "        layers.Dropout(0.5),\n",
    "        layers.Dense(512, activation='relu'),\n",
    "        layers.Dropout(0.5),\n",
    "        layers.Dense(256, activation='relu'),\n",
    "        layers.Dropout(0.5),\n",
    "        layers.Dense(128, activation='relu'),\n",
    "        layers.Dropout(0.5),\n",
    "        layers.Dense(64, activation='relu'),\n",
    "        layers.Dropout(0.5),\n",
    "        layers.ActivityRegularization(l1=0.06, l2=0.06),\n",
    "        layers.Dense(3, activation='softmax')\n",
    "    ])\n",
    "\n",
    "    model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "        \n",
    "    history = model.fit(profile_x, ss_y, epochs=3, batch_size=128)\n",
    "    history_cv.append(history)\n",
    "        \n",
    "    test_loss, test_accuracy = model.evaluate(profile_x_test, ss_y_test)\n",
    "    accuracy_cv.append(test_accuracy)\n",
    "    loss_cv.append(test_loss)\n",
    "        \n",
    "    ss_y_pred = model.predict(profile_x_test)\n",
    "    ss_pred_cv.append(ss_y_pred)\n",
    "    \n",
    "    return accuracy_cv, loss_cv, history_cv, ss_pred_cv\n",
    "\n",
    "\n",
    "def model_cross_validation(profile_x_list, seq_x_list, ss_y_list):\n",
    "    \"\"\"\n",
    "    Performs cross-validation for a neural network model for secondary structure prediction.\n",
    "\n",
    "    Args:\n",
    "        profile_x_list (list): A list of lists containing profile features for each cross-validation set.\n",
    "        seq_x_list (list): A list of lists containing sequence features for each cross-validation set.\n",
    "        ss_y_list (list): A list of lists containing secondary structure labels for each cross-validation set.\n",
    "\n",
    "    Returns:\n",
    "        tuple: A tuple containing accuracy_cv, loss_cv, history_cv, ss_pred_cv, and ss_true_cv lists.\n",
    "\n",
    "    \"\"\"\n",
    "    accuracy_cv = []\n",
    "    loss_cv = []\n",
    "    history_cv = []\n",
    "    ss_pred_cv = []\n",
    "    ss_true_cv = []\n",
    "    \n",
    "    for s in range(len(profile_x_list)):       \n",
    "        print(f'Train/test cross-validation set {s+1}.\\n-------------------------------')\n",
    "        profile_x = profile_x_list[s][0]\n",
    "        seq_x = seq_x_list[s][0]\n",
    "        ss_y = ss_y_list[s][0]\n",
    "        profile_x_test = profile_x_list[s][1]\n",
    "        seq_x_test = seq_x_list[s][1]\n",
    "        ss_y_test = ss_y_list[s][1]\n",
    "        ss_true_cv.append(ss_y_test)\n",
    "        \n",
    "        accuracy_cv, loss_cv, history_cv, ss_pred_cv = develop_model(\n",
    "            profile_x, \n",
    "            ss_y, \n",
    "            seq_x, \n",
    "            profile_x_test, \n",
    "            ss_y_test, \n",
    "            seq_x_test, \n",
    "            accuracy_cv, \n",
    "            loss_cv, \n",
    "            history_cv, \n",
    "            ss_pred_cv\n",
    "        )\n",
    "        \n",
    "    return accuracy_cv, loss_cv, history_cv, ss_pred_cv, ss_true_cv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "id": "c650acc7-d826-4af4-af05-d42d5f3486a9",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train/test cross-validation set 1.\n",
      "-------------------------------\n",
      "Epoch 1/3\n",
      "1230/1230 [==============================] - 12s 9ms/step - loss: 0.8186 - accuracy: 0.7074\n",
      "Epoch 2/3\n",
      "1230/1230 [==============================] - 11s 9ms/step - loss: 0.6339 - accuracy: 0.7805\n",
      "Epoch 3/3\n",
      "1230/1230 [==============================] - 11s 9ms/step - loss: 0.5397 - accuracy: 0.8174\n",
      "1273/1273 [==============================] - 2s 2ms/step - loss: 0.6499 - accuracy: 0.7612\n",
      "1273/1273 [==============================] - 2s 2ms/step\n",
      "Train/test cross-validation set 2.\n",
      "-------------------------------\n",
      "Epoch 1/3\n",
      "1241/1241 [==============================] - 12s 9ms/step - loss: 0.8103 - accuracy: 0.7149\n",
      "Epoch 2/3\n",
      "1241/1241 [==============================] - 11s 9ms/step - loss: 0.6375 - accuracy: 0.7790\n",
      "Epoch 3/3\n",
      "1241/1241 [==============================] - 12s 9ms/step - loss: 0.5445 - accuracy: 0.8158\n",
      "1229/1229 [==============================] - 2s 2ms/step - loss: 0.6343 - accuracy: 0.7631\n",
      "1229/1229 [==============================] - 3s 2ms/step\n",
      "Train/test cross-validation set 3.\n",
      "-------------------------------\n",
      "Epoch 1/3\n",
      "1238/1238 [==============================] - 12s 9ms/step - loss: 0.8644 - accuracy: 0.6730\n",
      "Epoch 2/3\n",
      "1238/1238 [==============================] - 12s 9ms/step - loss: 0.6393 - accuracy: 0.7810\n",
      "Epoch 3/3\n",
      "1238/1238 [==============================] - 12s 10ms/step - loss: 0.5446 - accuracy: 0.8159\n",
      "1241/1241 [==============================] - 2s 2ms/step - loss: 0.6562 - accuracy: 0.7514\n",
      "1241/1241 [==============================] - 2s 2ms/step\n",
      "Train/test cross-validation set 4.\n",
      "-------------------------------\n",
      "Epoch 1/3\n",
      "1245/1245 [==============================] - 12s 9ms/step - loss: 0.8172 - accuracy: 0.7121\n",
      "Epoch 2/3\n",
      "1245/1245 [==============================] - 12s 10ms/step - loss: 0.6348 - accuracy: 0.7801\n",
      "Epoch 3/3\n",
      "1245/1245 [==============================] - 12s 10ms/step - loss: 0.5403 - accuracy: 0.8182\n",
      "1213/1213 [==============================] - 3s 2ms/step - loss: 0.6572 - accuracy: 0.7581\n",
      "1213/1213 [==============================] - 2s 2ms/step\n",
      "Train/test cross-validation set 5.\n",
      "-------------------------------\n",
      "Epoch 1/3\n",
      "1239/1239 [==============================] - 12s 9ms/step - loss: 0.8114 - accuracy: 0.7132\n",
      "Epoch 2/3\n",
      "1239/1239 [==============================] - 12s 9ms/step - loss: 0.6335 - accuracy: 0.7790\n",
      "Epoch 3/3\n",
      "1239/1239 [==============================] - 12s 10ms/step - loss: 0.5424 - accuracy: 0.8143\n",
      "1239/1239 [==============================] - 2s 2ms/step - loss: 0.6493 - accuracy: 0.7606\n",
      "1239/1239 [==============================] - 2s 2ms/step\n",
      "Mean Accuracy:     0.7588649988174438\n",
      "Std Dev Accuracy:  0.004075966918367876\n",
      "-----------------------------\n",
      "Mean Loss:         0.6493692398071289\n",
      "Std Dev Loss:      0.008175182943752357\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "None\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAF2CAYAAAB02w9PAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAAB1/ElEQVR4nO3deVxU9frA8c+s7KugqLiCgCgioALlkpqZejW11SXXrHszUkxbLbtlei1zT3NBs4xyy10rs7JcwH0XXBJBBVlc2BmYOb8/+MmNC+4MI/K8X695verM93zPc4bjPHPO95znq1IURUEIIYT4H2pLByCEEOLBJAlCCCFEuSRBCCGEKJckCCGEEOWSBCGEEKJckiCEEEKUSxKEEEKIcmktHcCDrlWrVhgMBtzd3S0dihBCVIi0tDT0ej379u27ZTtJELdRUFCA0Wi0dBhCCFFhioqKuJNnpCVB3EbNmjUB2LZtm4UjEUKIitG5c+c7aidjEEIIIcolCUIIIUS5JEEIIYQolyQIIYQQ5ZIEIYQQolxyF5MQ4q4YjUYKCwstHYa4CZ1Oh0ajqZC+JEEIIe6IoiikpKRw7do1S4cibsPZ2RkPDw9UKtV99SMJQghxR24kh5o1a2Jra3vfXz6i4imKQm5uLqmpqQDUrl37vvqTBCGEuC2j0ViSHGrUqGHpcMQt2NjYAJCamkrNmjXv63KTDFILIW7rxpiDra2thSMRd+LG3+l+x4okQQgh7phcVqoaKurvJAnCDNJyMpjy51yOpJy0dChCCHHPJEGYwemMBPZfOsq0XQtJzcmwdDhCCHFPJEGYQRvPljSp0Yjcwjxm7V6M0STlwoV4EM2dO5cXX3zR0mHcFUVRWLNmDRkZ5v/xKQnCDLRqDaPChmGjs+ZUxl+sPL7J0iEJIf7HV199xaxZsywdxl3bu3cvb7/9Nnl5eWbfliQIM6lp78bLrfoDsObEjxxPPWXhiIQQAJcvX+all15i5syZNGrUyNLh3LU7meinokiCMKNH67fmsUbhKCjMjllCVkG2pUMSokIpikJ+QZHFXvfyZXn8+HGcnJxYv349gYGBd7ROUlISI0eOJCQkhNDQUCIjI0lPTyc2NhZfX1/Onj1bqv3QoUMZN25cuX0lJCQwfPhwQkJCCAoKYvjw4cTHx5e8n5WVxfvvv09YWBghISEMGjSIo0ePAhAbG8ugQYOA4kl/fvjhh7ve/7shD8qZ2bCg54hPP0tyVirz9i5j3KOvyK2C4qGgKApvzdnByYQrFouhaUNXprzW9q7+TXXq1IlOnTrdcfusrCz69++Pt7c3X331FVqtlgkTJhAREUF0dDSenp6sX7+eyMhIoPgBtZiYGKKiosrtb8yYMfj6+rJ69WqKioqYMmUKr732Glu3bkVRFEaMGIFOp2P+/PnY29uzbt06+vXrx4oVKwgKCmL27NlERESwcuVKfHx87ng/7oWcQZiZtc6a0eEvoVVr2XfxMD+f+cPSIQkh7sLmzZvJyspi+vTpBAQE0LRpUz755BNCQkIwGAz06dOHjRs3lrTfuHEjtWrVIiwsrNz+EhMTcXNzw9PTE29vbyZNmsTEiRMxmUzExMRw8OBBZs6cSWBgIF5eXowZM4aWLVvy9ddfo9frcXJyAsDV1RVra2uz7rucQVSCRi71GNCiN0sPreLrQ6to6u5Nfee6lg5LiPuiUqmY8lpbCgyWu0vPSq8x+xl5fHw8DRs2xNnZuWRZkyZNGDt2LAC9e/dmzpw5HDhwgODgYNatW8dTTz2FWl3+7+/IyEgmTZrEd999R1hYGO3ataNbt26o1WqOHz8OlJ0z2mAwUFBQYJ4dvAWLJgiTycScOXNYuXIlmZmZhISEMGHCBBo0aFBu+7S0NCZPnszOnTsBCAsL45133sHDw6OkTadOnbh48WKp9Xr27MnUqVPNtyN3oLtPJ45cPsnB5OPM2B3F5C5vY6XVWzQmIe6XSqXC2urh/p2p1WpvmYQ8PT1p06YNGzZswMHBgbi4OGbMmHHT9gMGDODJJ59k+/bt7N69m2nTpjF79mzWrl2LyWTC3t6+3LEFvb7yvy8seolp7ty5fP/990ycOJHly5ejUqkYMWIEBoOh3PaRkZEkJyezZMkSlixZQkpKCq+++mrJ+9nZ2Vy6dIn58+ezY8eOkteECRMqa5duSqVS8WqbQThZO3IhM5mvD62ydEhCiDvg7e1NQkICWVlZJctOnDhBaGhoyY/Rvn37snXrVjZs2EBQUNBN745KT0/no48+orCwkL59+/LZZ5+xfv160tLS2LNnDz4+PmRnZ2MwGGjQoEHJa+HChWzbtg2o3HInFksQBoOBxYsXExERQYcOHfDz82P69OlcvnyZrVu3lmmfmZnJ3r17GTFiBP7+/vj7+/Pyyy9z/Phxrl69CsCpU6dQFIXg4GDc3d1LXg4ODpW9e+VysnYkInQIAFvP/knshYOWDUgIcVs9e/bEycmJcePGERcXx7Fjx/jwww/x8fGhbt3iS8Vdu3YlNzeXZcuW0adPn5v25ezszO+//8748eM5efIkSUlJREdHo9PpaN68Oe3ataNp06aMHj2a3bt3c/78eaZMmcLq1avx8vIC/luILy4ujpycHLPuu8USxI2d+/tAjqOjI/7+/uzdu7dMeysrK2xtbVm7di3Z2dlkZ2ezbt06GjZsWDJoEx8fj7u7O46OjpW2H3erhUdTevk9AcCXe74hPcdyd4AIIW7PxsaGqKgojEYj/fr1Y/jw4Xh5eZV6yM7GxoZu3bpRVFREt27dbtqXVqtl4cKFqNVqhgwZQo8ePYiJiWHBggXUr18fjUbD4sWLadGiBZGRkfTq1YvY2Fhmz55NeHg4AD4+PnTo0IHRo0ezfPlys+67SqnMpy7+5ueffyYiIoLDhw+XGokfNWoU+fn5zJ8/v8w6mzdv5qOPPuLatWuoVCrc3d1ZtmwZ9evXB+Df//43O3fuxNfXl4MHD+Lq6krfvn0ZNGjQTQeMoOyA0N8lJydTu3btktO7ilBkLOL9X6dy9sp5mrp788Fjo9GoK2aKQCHMIT8/n3PnztGoUSOz3zkj7t/t/l43vvNu971msTOIG4+J/+/Ai5WVVbmj9YqiEB8fT1BQEN9++y1Lly6lbt26jBw5kuzs4gfQTp8+TVZWFt27dycqKornn3+emTNnMnv2bPPv0F3QarSMCh+OtdaKk2ln+OHEFkuHJIQQZVjs9oMbWc1gMJTKcAUFBSUzIv3dpk2biI6O5rfffsPe3h6AL7/8ko4dO7J69WoGDx7MkiVLKCgoKHnf19eXnJwc5s2bR0RExE3PIm6VRW91dnE/POzdeSmkH3Niv2LVic0E1PLDz93bLNsSQoh7YbEziBtzpd6YO/WG1NTUUret3rB//34aNWpU8uUP4OTkRKNGjUhISABAp9OVeh+Kr9fl5uZy/fr1Ct6D+9e+YSjtG4SiKAozYxaTbTDvgJMQQtwNiyUIPz8/7O3tiY2NLVmWmZnJiRMnaNWqVZn2tWvX5vz586UuP+Xl5XHhwgUaNGiAyWSiU6dOzJs3r9R6R48exc3NDRcXF/PtzH0YHvICHvbuZOReZf7ebyu1EJcQQtyKxRKEXq9n4MCBTJ06lW3bthEXF0dkZCQeHh506dIFo9FIWloa+fn5QPHTigCjR48mLi6upL1er6dv376o1Wq6du3KokWL2LJlC4mJiSxfvpxFixYxatQoS+3mbdnorBkVPhyNSk3shYNs+2uHpUMSQgjAwk9Sv/766xQVFTF+/Hjy8/Np3bo1UVFR6PV6Lly4QOfOnZk8eTJ9+/alZs2aREdH89lnnzF48GDUajWtWrXiu+++K7mt9Y033sDR0ZHPP/+clJQUPD09ee+993juuecsuZu35eXagH4terPs8A98dXAlfm7eeDrVtnRYQohqzmK3uVYVd3o72P0yKSYm/zGHwyknqe9Ul0ld3kKv0Zl1m0LcKbnNtWqp8re5itLUKjUj2wzG0cqexOsXWXbIvHXehRDidiRBPECcbZwYGToYgB/P/M6+i4ctHJEQojqTBPGACardnH/4FJ/+zd3zDVdyr1k2ICEeMteuXeODDz6gffv2BAcH069fP/bt22fpsO6YoiisWbOGjIwMs29LEsQDqF+Lp2jkUo9sQw6zY5dgMpksHZIQD40xY8Zw+PBhpk2bxqpVq2jWrBnDhw8vM23og2rv3r28/fbbJdUozEkSxANIp9ExKnw4VlorjqeeYm3cT5YOSYiHwvnz59m5cycTJkygVatWNG7cmPfee49atWqVmhXuQVaZ9xVJgnhA1XGoxfDg5wFYcWwjp9L/snBEQpSlKAomQ77FXnf7Zeni4sKCBQto3rx5yTKVSoWiKLestpCUlMTIkSMJCQkhNDSUyMhI0tPTiY2NxdfXt8zZx9ChQxk3bly5fSUkJDB8+HBCQkIICgpi+PDhxMfHl7yflZXF+++/T1hYGCEhIQwaNIijR48CEBsby6BBg4DiO5HKm1ioIj3cU0FVcR0ahnE45QQ7E/cxc3cUn3Z9Dzu9raXDEgIoTg6Xvn6Pggvxt29sJlaeftQZNPGOJ9FxdHSkQ4cOpZbdeLC2bdu25a6TlZVF//798fb25quvvkKr1TJhwgQiIiKIjo7G09OT9evXExkZCRSXC4qJiSEqKqrc/saMGYOvry+rV6+mqKiIKVOm8Nprr7F161YURWHEiBHodDrmz5+Pvb0969ato1+/fqxYsYKgoCBmz55NREQEK1euxMfH5y4+rbsnZxAPMJVKxYiQ/tS0q0Fa7hUW7ouWUhziAVN5s5uZw/79+3n33Xfp3LkznTp1KrfN5s2bycrKYvr06QQEBNC0aVM++eQTQkJCMBgM9OnTp9TlqY0bN1KrVq1Sc938XWJiIm5ubnh6euLt7c2kSZOYOHEiJpOJmJgYDh48yMyZMwkMDMTLy4sxY8bQsmVLvv76a/R6fcn8N66urmZ/JkXOIB5wtnobRoUP54NtU9mVtJ8WHv50avyIpcMSApVKRZ1BE1EKy5bnr7QYdFb3PAXnL7/8wtixYwkMDGTatGk3bRcfH0/Dhg1xdnYuWdakSRPGjh0LFJcBmjNnDgcOHCA4OJh169bx1FNP3bR6dGRkJJMmTeK7774jLCyMdu3a0a1bN9RqNcePHwfKVpE2GAzlToNgbpIgqoAmNRrxXPOefHd0HUsOLMfXrTF1HctWvBWisqlUKlT6qvdk9bJly/jkk0/o0qULU6dOLTMvzd9ptdpbJiFPT0/atGnDhg0bcHBwIC4ujhkzZty0/YABA3jyySfZvn07u3fvZtq0acyePZu1a9diMpmwt7cvd2zhVjGai1xiqiKe8nuC5jV9KTAamLk7ikJjoaVDEqJKio6O5uOPP2bAgAHMmDHjtl+83t7eJCQkkJWVVbLsxIkThIaGcvHiRQD69u3L1q1b2bBhA0FBQTRq1KjcvtLT0/noo48oLCykb9++fPbZZ6xfv560tDT27NmDj48P2dnZGAwGGjRoUPJauHBhSVmMez1juheSIKoItVrNa2FDcNDbkXDtAt8eWWvpkISocs6dO8ekSZPo0qULr7zyChkZGaSlpZGWllYqAfxdz549cXJyYty4ccTFxXHs2DE+/PBDfHx8qFu3LgBdu3YlNzeXZcuW0adPn5tu39nZmd9//53x48dz8uRJkpKSiI6ORqfT0bx5c9q1a0fTpk0ZPXo0u3fv5vz580yZMoXVq1fj5eUFgK1t8Y0qcXFx5OSYdw4ZSRBViKuNM6/+fymOzad+5cCloxaOSIiq5aeffqKwsJCtW7fStm3bUq9PPvmk3HVsbGyIiorCaDTSr18/hg8fjpeXF7NmzSrVplu3bhQVFdGtW7ebbl+r1bJw4ULUajVDhgyhR48exMTEsGDBAurXr49Go2Hx4sW0aNGCyMhIevXqRWxsLLNnzyY8PBwongStQ4cOjB49muXLl1fsB/Q/pJrrbVRWNde7seTACrac/g0HK3umdh2Pi42TpUMSDzmp5lq1SDXXamxAYB8aOHuSVZDNnNglmBQpxSGEqHiSIKogvUbHqPBh6DU6jl6OZ33cVkuHJIR4CEmCqKI8HWszNKh4przlR9dzJiPBsgEJIR46kiCqsE6NHyWsXjBGxcTM3VHkFpq/uqMQovqQBGEGiqJwKT0bk8m84/8qlYpXWg3A3daVyznpLNr3nZTiEEJUGEkQZrDneAr/mrSViUtiMRrNO4Bsp7fl9fBhqFVqdiTu5Y+EWLNuTwhRfUiCMAPXKxcY89d3OPyxiXmrD5v9V72vmxfPNOsBwKID35OclWrW7QkhqgdJEGbgUcsJLSbaXDtB5paNrPjllNm32bfpkzR1b0JBUQEzd0dRZCwy+zaFEA83SRBm4ODrQ6OXhgHQ4cohDq/YyC97Es26TbVazethQ7HX2/HX1US+O7rOrNsTQjz8JEGYSZ1/dKfu08U1Wbql7mbTko0ciDPvpZ8ati78s/VAADbE/8Kh5BNm3Z4Q4uEmCcJMFJNCgxcH4N6xA2oUeiVvZ8mXmzhz4ZpZt9vGsyVPeLcH4IvYr7iWn2nW7QlR1WRkZDBu3DjCwsIICgri5Zdf5syZM5YO644pisKaNWvIyMgw+7YkQZhB4rkrfDr+R37dEo/XyH/h1DIQvVJEr/NbmTnnRy5fyTXr9gcFPk09pzpcL8jii9ilUopDiL/517/+RVJSEgsXLmTVqlVYW1szZMgQ8vKqxnNEe/fu5e23366UeCVBmEHBxVMUFBjZue0Mu34/R9O3x2HTuDG2pgKePL2FyXO2kZVrMNv29Vo9o8KGodPoOJxygs2nfjXbtoSoSq5evYqnpycff/wxAQEBeHl58eqrr5KWlsbp06ctHd4dqcxnnSRBmEH9Jh6E2O0D4LcfT3HoUCrNJ4xHV7MmzkXZtDu2nsnz/6Cg0Gi+GJzrMrjlMwB8e2Qtf105b7ZtiepLURTyiwos9rrbL0sXFxemTZtGkyZNgOIJfKKiovDw8MDb2/um6yUlJTFy5EhCQkIIDQ0lMjKS9PR0YmNj8fX15ezZs6XaDx06lHHjxpXbV0JCAsOHDyckJISgoCCGDx9OfHx8yftZWVm8//77hIWFERISwqBBgzh6tLi0f2xsLIMGDQKKK7KWN/NcRZIpR83AqlZDOvTvSsGy3ziW14LNq45gMyiEgI8mcGjcO3hkXcE/di3TnOx4c0gYGrV5Zojq4tWOIykn2XPxEDN2RzHliXex0UmpZlExFEXhg21Tic/4y2Ix+Lp58VGnN+5plrX333+fFStWoNfrmTdvXslEPP8rKyuL/v374+3tzVdffYVWq2XChAlEREQQHR2Np6cn69evJzIyEoDU1FRiYmKIiooqt78xY8bg6+vL6tWrKSoqYsqUKbz22mts3boVRVEYMWIEOp2O+fPnY29vz7p16+jXrx8rVqwgKCiI2bNnExERwcqVK/Hx8bnr/b4bcgZhJvZ+4Tz+bBjeVvEoqPhh2X6Ss7Q0n/Ae6K1olJdMjW2riVp3xGynjCqVin+2HkgNGxdSstNYfMC8k4uIaqgSp7+saIMHD2b16tX06tWLkSNHcvz48XLbbd68maysLKZPn05AQABNmzblk08+ISQkBIPBQJ8+fdi4cWNJ+40bN1KrVi3CwsLK7S8xMRE3Nzc8PT3x9vZm0qRJTJw4EZPJRExMDAcPHmTmzJkEBgbi5eXFmDFjaNmyJV9//TV6vR4np+L5X1xdXc0+N4ecQZiRY2AnuuXnsW7NORILG7I8ajeDX2uH/zvjOP7xJJplnyN27SrWONvRt+PNT2/vh72VHRFhQ/n379PZnhBDi1pNadewjVm2JaoXlUrFR53eoMBovvG027HS6O95juYbl5Q+/vhjDh06xLJly5g8eXKZdvHx8TRs2BBnZ+eSZU2aNGHs2LEA9O7dmzlz5nDgwAGCg4NZt24dTz31FGp1+b+/IyMjmTRpEt999x1hYWG0a9eObt26oVarS5LUjQl9bjAYDBQUFNzTft4Pi55BmEwmZs2aRbt27QgMDGTYsGGcP3/za+VpaWmMGTOG0NBQQkNDGTVqFCkpKaXabNmyhe7duxMQEEDPnj35448/zL0bt+QS2oN/dK+Dh/YShUUqvp23A2PdJvi8PhKA0GsnOLZsJX8cvGC2GPxrNuFp/+JpEBft/47L2Wlm25aoXlQqFdZaK4u97jY5ZGRksHHjRozG/47/qdVqvLy8SE0t/zklrVZ7y+14enrSpk0bNmzYwOnTp4mLi6N37943bT9gwAD++OMP3n33XWxsbJg2bRrdu3cnPT0dk8mEvb09a9euLfXavHlzqSlOK4tFE8TcuXP5/vvvmThxIsuXL0elUjFixAgMhvJ/kURGRpKcnMySJUtYsmQJKSkpvPrqqyXvx8TEMG7cOPr378/atWtp27YtI0eOLDOAVNnc2vel1+O2uGrSySuAb77YjnVQKA0GFT/U1jl9H5vnr+bomXSzxfC0f3d83bzIK8pn5u7FFJnMN0AuxIMqNTWVN954gz179pQsKyws5MSJE3h5eZW7jre3NwkJCWRlZZUsO3HiBKGhoVy8eBGAvn37snXrVjZs2EBQUBCNGjUqt6/09HQ++ugjCgsL6du3L5999hnr168nLS2NPXv24OPjQ3Z2NgaDgQYNGpS8Fi5cWDI96L2eMd0LiyUIg8HA4sWLiYiIoEOHDvj5+TF9+nQuX77M1q1lZ0jLzMxk7969jBgxAn9/f/z9/Xn55Zc5fvw4V69eBWDhwoV06dKFgQMH4uXlxVtvvUWzZs1YunRpZe9eGbUeH8BTbY04qK+TlaPw9ezfcenaHY8e3QHonryDpbPXcD7ZPA+2adQaXg8bip3OhjNXElhxbINZtiPEg8zPz4+2bdvy73//m3379nHq1CneeustMjMzGTJkSLnr9OzZEycnJ8aNG0dcXBzHjh3jww8/xMfHh7p16wLQtWtXcnNzWbZsGX369Lnp9p2dnfn9998ZP348J0+eJCkpiejoaHQ6Hc2bN6ddu3Y0bdqU0aNHs3v3bs6fP8+UKVNYvXp1SQK7MZgeFxdHTk5OxX5A/8NiCeLGzv19IMfR0RF/f3/27t1bpr2VlRW2trasXbuW7OxssrOzWbduHQ0bNsTJyQmTycSBAwfKDAyFhoayb98+s+/P7ahUKjx7DqVX62vYqnK4cs1I9NzteL74Ii7hYWgw0f38NmbN3ED6NfM8AONuV4NX/r8Ux7qTP3Mk5aRZtiPEg0qlUjFjxgzCwsIYPXo0zz77LNevX+fbb7+lTp065a5jY2NDVFQURqORfv36MXz4cLy8vEpd8rGxsaFbt24UFRXRrVu3m25fq9WycOFC1Go1Q4YMoUePHsTExLBgwQLq16+PRqNh8eLFtGjRgsjISHr16kVsbCyzZ88mPDwcAB8fHzp06MDo0aNZvty8N56oFAvNMPPzzz8TERHB4cOHS43Ejxo1ivz8fObPn19mnc2bN/PRRx9x7do1VCoV7u7uLFu2jPr163Pt2jVCQ0NZsGABHTp0KFnn22+/ZerUqRw8ePCmsfzvgNDfJScnU7t27ZLTu/ulmIzELZvNuiOeGBRrGjawpd9Lj3D8o4nknDxJtsaG34Kf5f03umNno6uQbf6vBXu/5Ze/duBs7cjUruNxtHYwy3bEwyM/P59z587RqFEjs985I+7f7f5eN77zbve9ZrEziBuPiev1+lLLraysyh2tVxSF+Ph4goKC+Pbbb1m6dCl169Zl5MiRZGdnk5+ff1f9WYpKrcGv/0ie9E1AQyEJ53NZ8+0e/N99G71nPeyNeTxyaC2fLdhOYZF5SmQMDnqWuo4eXMvPZO6er2UWOiFEuSx2m+uNrGYwGEpluIKCAmxsbMq037RpE9HR0fz222/Y29sD8OWXX9KxY8eSe5lv9Pd3N+vv726VRW91dnGvVFodLYZEYJg/i5/P+RAXl8nPG47x+L/Hc2jcu9S4koH/zpXMdrJn9KBQ1BX8IJ2VVs+osOG898sUDiQfY8vp3+ju06lCtyGEqPosdgZRu3ZtgDK3lqWmpuLh4VGm/f79+2nUqFFJcgBwcnKiUaNGJCQk4OzsjK2t7R33Z2lqnRWtXnqV9nVPAQoHDqSza2ciLT76AGxsqVuQTo2fvuPrjUfNsv2GLp682PJpAJYdXkPC1SSzbEcIUXVZLEH4+flhb29PbOx/51DOzMzkxIkTtGrVqkz72rVrc/78+VKXi/Ly8rhw4QINGjRApVIRHBxc6vY1KK5dEhISYr4duQ9qazse/ecrhNWMA2DHH5c4En+dgAnvoWh1eOdeIGflt2z60zy36Xb17kBInQCKTEXM2B1FftGDcylOCGF5FksQer2egQMHMnXqVLZt20ZcXByRkZF4eHjQpUsXjEYjaWlpJWMLNx48GT16NHFxcSXt9Xo9ffv2BYoLZG3atIklS5Zw9uxZPv30U06ePMngwYMttZu3pbF1pNOrw2npUjwt6c9bzpFwXUvTN8egqFS0zDzD8cXL2H30UoVvW6VS8a82g3CxceJS1mW+OrCiwrchhKi6LPqg3Ouvv84zzzzD+PHj6devHxqNhqioKPR6PcnJybRt25bNmzcDULNmTaKjo1EUhcGDBzN06FB0Oh3fffcdjo6OALRt27bkEfY+ffoQExPDl19+edMHYB4UWgdXnhzZHz+H4qJn61ae4Iptbbz+OQKAR68c4afZ0Zw8d6XCt+1oZU9E6FBUqPj13C52Je6v8G0IIaomi93mWlXc6e1gFaEgLYnlMzeQkFcfrdrIwFfCMMX+ysUVqzCh4scGnfnX+IF41qz421K/O7KONSd/xFZnw6dd36OmXY0K34aouuQ216qlyt/mKsqycq/HM//qSh2rZIpMGr5bsBubDl1x69QJNQpdEn9j7rQ1XM3Mr/BtP9v8HzSp0Yjcwjxm7V6MUUpxCFHtSYJ4wNjW9eK5EY/ipkunwKjlmy+24/7CQOyDgtApRjrFb2H6rM3kFRRV6Ha1ag2jwoZho7PmVMZfrDy+qUL7F0JUPZIgHkCOjfx5fnALnDTXyDXo+Gb2rzR8LQKrxl7YmAyEHVzD9Pm/UmSs2Afpatq78UqrAQCsOfEjx1NPVWj/QoiqRRLEA6pG02Cef8ELO3U21/N0RM/5Fd933kJTsxZORbk0/XM586NjK/wp6Efqt6Jjo0dQUJgVs5jMguwK7V+IB8m5c+cICgoy+9SdFUlRFNasWUNGRobZtyUJ4gHmERzOs71rYaXKIy1Lx4pFf9L8ww/AwZGahmu4bvyG7zcfq/DtDg1+jtoONbmad50v93wjpTjEQ6mwsJCxY8eSm5tr6VDuyt69e3n77bdLyhWZkySIB1z9RzvSt6sDOgxczNCy7vs9BHz4Poreivr5l8mJXsLPMecqdJvWWitGh7+EVq1l36Uj/HzGspMuCWEOs2fPxs7OztJh3LXK/MEmCaIKaNLlSf7RQYMaI2cvqdn600majX8bRa2hafZ5Ts6LYu+JlNt3dBcaudRjQIveAHx9aBWJ1y5WaP/i4aAoCsb8fIu97vXLcu/evSxfvpwpU6bcUfukpCRGjhxJSEgIoaGhREZGkp6eTmxsLL6+vmUmJRs6dCjjxo0rt6+EhASGDx9OSEgIQUFBDB8+nPj4+JL3s7KyeP/99wkLCyMkJIRBgwZx9GhxyZ3Y2FgGDRoEFN+qau5LYzIndRUR0Osp8nNW8uM+K47/pcbWNplWkRGc+nwGra6dZNv0JTh/8E+a1HOpsG129+nEkctxHEw+xozdUUzu8jZWWv3tVxTVgqIoHH37PbLi4m/f2EwcmvoRMHniXc2ylpmZyZtvvsn48eNLasLdSlZWFv3798fb25uvvvoKrVbLhAkTiIiIIDo6Gk9PT9avX09kZCRQXP8tJiaGqKiocvsbM2YMvr6+rF69mqKiIqZMmcJrr73G1q1bURSFESNGoNPpmD9/Pvb29qxbt45+/fqxYsUKgoKCmD17NhEREaxcuRIfH5873u97IWcQVUirF56hQ7PiQeO9xwo5kVJI/SHFZUTape7ju0+XkZJRcTNMqVQqXm3zIs7WjlzITGbpoVUV1rd4SFTi9JcV5cMPP6Rly5b07Nnzjtpv3ryZrKwspk+fTkBAAE2bNuWTTz4hJCQEg8FAnz592LhxY0n7jRs3UqtWrTKTl92QmJiIm5sbnp6eeHt7M2nSJCZOnIjJZCImJoaDBw8yc+ZMAgMD8fLyYsyYMbRs2ZKvv/4avV6Pk5MTAK6urmZ/aFHOIKoQlUpFu6EvkDf3W2L/cmF7bC62nWpQ8x89SN24iccStzP/UztGvz8AJ3urCtmmk7Ujr4UOYeL2Wfxy9k8CPZoS6hlUIX2Lqk2lUhEweSImC863orayuquzh7Vr17Jv3z42bLjzKXfj4+Np2LAhzs7OJcuaNGnC2LFjgeI6cXPmzOHAgQMEBwezbt06nnrqKdTq8n9/R0ZGlpQECgsLo127dnTr1g21Ws3x48eBstMMGAwGi8xrIwmiilGp1HT5Zz/yZkRz5JIrP/6aQe9/NMcx/CqZu3fRPv4n5sxw4I03+2Ctr5g/bwuPpvTye4L1cT/z5Z5v8HJpgJuda4X0Lao2lUqFpgqV3li9ejUZGRk89thjpZZPmDCBqKgoNm0q+4CoVqu9ZRLy9PSkTZs2bNiwAQcHB+Li4pgxY8ZN2w8YMIAnn3yS7du3s3v3bqZNm8bs2bNZu3YtJpMJe3v7cscW/ncytMogl5iqILVGS6/XX6BJjSsoqFm/8RJWHTph3bQZeqWI1gfW8sWCXzCaKu5uhxea98TLtQE5hXnMjl0ipThElTR16lQ2b97M2rVrS15QXDh0wYIF5a7j7e1NQkICWVlZJctOnDhBaGgoFy8W37zRt29ftm7dyoYNGwgKCqJRo0bl9pWens5HH31EYWEhffv25bPPPmP9+vWkpaWxZ88efHx8yM7OxmAw0KBBg5LXwoULS+om3c0Z0/2SBFFFqXV6nhnzLPUdr2JEy+rVCbg9+yyauvWwM+bj+/v3LP4upsJuidNqtIwKH46N1pqTaWf44cSWCulXiMpUq1atUl+8DRo0AKBGjRrUrVu33HV69uyJk5MT48aNIy4ujmPHjvHhhx/i4+NTsk7Xrl3Jzc1l2bJl9OnT56bbd3Z25vfff2f8+PGcPHmSpKQkoqOj0el0NG/enHbt2tG0aVNGjx7N7t27OX/+PFOmTGH16tUlValtbW0BiIuLIyen4sYcyyMJogrTWdvQ740+1LK9TqGiY/n38dR7eTi41MClMAuXdUtY89PxCtueh707L4X0A2DVic2cTDtdYX0L8aCysbEhKioKo9FIv379GD58OF5eXsyaNatUm27dulFUVES3bt1u2pdWq2XhwoWo1WqGDBlCjx49iImJYcGCBdSvXx+NRsPixYtp0aIFkZGR9OrVi9jYWGbPnk14eDgAPj4+dOjQgdGjR7N8+XKz7ruU+76Nyiz3fa9yrlxlydQfuVLggIM2hxdeaMqZKVNR5+Xwl20d6o0Zy2OtG1TY9ubEfMUf52OpYevCZ0+8h71V1XvYSNwdKfddtUi5b1HCztWFgRGdcNDmklVkx6oVx/F543VMWh2Ncy9xatYXHDqVevuO7tDwkBfwsHcnI/cqX+5bJqU4hHhISYJ4SDjXrsXAf4Zjrc7nqsGBtevi8Hr9NRSVimaZZ9k+ZR7nLl2vkG3Z6KwZFT4cjUrNnguH2PbXjgrpVwjxYJEE8RBxb1SP/sNaolMZuJznzE+//UX94cMAaJV+hJWTFpF2tWIKfHm5NqDf/5fiWHJwJUnXK37ObCGEZUmCeMh4NvXm2Rd80WAkKcuVnUdSqfn00wA8krSLqP98Q3ZeYYVs6x++nQn0aEqhsZCZuxdjKDJUSL9CiAeDJIiHkHerZjzVux4qTJy+4sbRFAOOj3VCBYTHbWXe56spLLr/5xjUKjUj2wzGycqBxOsX+eZw1ampL4S4PUkQD6nm7YJ5sktNAI5dduOixg6rwCC0mAjev5YFX/6EqQIepHO2cWJkaHE9qJ/ObGfvxcP33acQ4sEgCeIh1vrJMB4LdwBgz3k3suvWQ93QC2tTIV6/RrNs+a4K2U7L2s34h0/xbXPz9nzDldxrFdKvEMKyJEE85No/04E2gcX3Qf95ygVVywBMbrVwMObhtDqKDT8frZDt9GvxFI1c6pFtyGF27BJMpoqdL1sIUfkkQVQDXV/sTHNvDQpqfjnmiP1jj2K0c8St8Dq5UV+wc3/CfW9Dp9ExKnw4VlorjqeeYm3cT/cfuBDCoiRBVAMqlYrerzyJlyeY0LDlkC0uPR6nSG+NZ34ap6fN5NiZtPveTh2HWgwPfh6AFcc2Ep9+9jZrCFH5Ll68iK+vb5nXypUrLR3aHVEUhTVr1pCRkWH2bUmCqCbUahXPv9YNTzcjRejYfNAKt97dMKo1NMlOZMcnM0hMybzv7XRoGMaj9VthUkzM2r2YHEPVmhBePPzi4+OxsrLizz//ZMeOHSWvO51AyNL27t3L22+/TV5exTzTdCuSIKoRrU5D/9HdcHcsokCx5sdDWtx6d0MBAq7EsfajuVzJzL+vbahUKkaE9KemXQ3Scq+wcF+0lOIQD5RTp07RqFEjatasibu7e8mrqtSYqsx/T5IgqhlrGz0vjnkSJ9tCckz2bD2uw7lbFwBCLu7l64+jyM2/vwfpbPU2JaU4diXt57dzuysidPEAUhQFQ0GRxV738mUZHx+Pt7f3Xa2TlJTEyJEjCQkJITQ0lMjISNLT04mNjcXX15ezZ0tfTh06dCjjxo0rt6+EhASGDx9OSEgIQUFBDB8+nPj4/87rnZWVxfvvv09YWBghISEMGjSIo0eLbyaJjY1l0KBBQHHBvfImFqpIMqNcNWTvYM2g0U+weOpWrhuc+eNcIY+0fYScHbsIOfUbUZ858a93+6PV3PvvhyY1GvF8QC+ij6xlyYHl+Lo1pq6jRwXuhbA0RVFYMmcXFxKuWiyGeg1dGPLaI3c1ic6pU6dwd3enf//+JCQk0KBBA1599VXatWtXbvusrCz69++Pt7c3X331FVqtlgkTJhAREUF0dDSenp6sX7+eyMhIAFJTU4mJiSEqKqrc/saMGYOvry+rV6+mqKiIKVOm8Nprr7F161YURWHEiBHodDrmz5+Pvb0969ato1+/fqxYsYKgoCBmz55NREQEK1euxMfH5+4/tLsgZxDVlEsNOwa+1gErTRHpRe7sveKIrkUL1CgE7F/Pknmb7vtUtpdfFwJq+VJgNDBjdxQGY8WU+BAPjsqb26xiGAwGEhISyM7OZvTo0SxYsICAgABGjBjB7t3ln+lu3ryZrKwspk+fTkBAAE2bNuWTTz4hJCQEg8FAnz592LhxY0n7jRs3UqtWLcLCwsrtLzExETc3Nzw9PfH29mbSpElMnDgRk8lETEwMBw8eZObMmQQGBuLl5cWYMWNo2bIlX3/9NXq9HicnJwBcXV3NfllMziCqsVp1Xej3yiMs+3I3yYbaWKkL8W7she6vszTc9h0rajjxfL/299y/WqVmZOgQxv04kfPXLhB9eA1Dgp+rwD0QlqRSqRjy2iMUGiw3/axOr7mrswe9Xs/evXvRarUlczw3b96cs2fPEhUVVTIpz9/Fx8fTsGFDnJ2dS5Y1adKEsWPHAtC7d2/mzJnDgQMHCA4OZt26dTz11FOo1eX//o6MjGTSpEl89913hIWF0a5dO7p164Zareb48eIJvm7M13CDwWCgoKDgjvezolg0QZhMJubMmcPKlSvJzMwkJCSECRMmlEwD+HezZ89mzpw55fbTt29fJk+eDECnTp1K5om9oWfPnkydOrXid+AhUN+rJs8ObsXyr/aTkF8fK/tCatfMwTY1BcOqKH5yc6Zrlxb33L+rjTOvhg5myp9z2Xz6N1p4NCW4TkAF7oGwJJVKhd6qav3OvDFl59/5+PiwY0f5Zeu1Wu0tk5Cnpydt2rRhw4YNODg4EBcXx4wZM27afsCAATz55JNs376d3bt3M23aNGbPns3atWsxmUzY29uXO7ZwI6FVJoteYpo7dy7ff/89EydOZPny5cV3wIwYgcFQtirosGHDSt2StmPHDkaPHo21tTWDBxfXAsrOzubSpUvMnz+/VLsJEyZU9q5VKU2a16XXc80BhfhcL67UbUSRozPORdnkLJhF7IFz99V/SJ0AujXpCMAXe77mal7FzEshxN2Ki4sjKCiIffv2lVp+7Nixmw5ce3t7k5CQQFZWVsmyEydOEBoaWvJjtG/fvmzdupUNGzYQFBREo0aNyu0rPT2djz76iMLCQvr27ctnn33G+vXrSUtLY8+ePfj4+JCdnY3BYCg1b/bChQtLZn+7mzOm+2WxBGEwGFi8eDERERF06NABPz8/pk+fzuXLl9m6dWuZ9nZ2dqVuScvLy2P+/Pm8/fbb+Pn5AcWDT4qiEBwcXKqtg4NDZe9eldOiTWO69ij+B3Iky49cb28KrWypZbjKX1M/J/6v+3uQbkBgHxo4e5JVkM2c2CWYFCnFISqfj48PTZo04d///jf79u3j7NmzTJ48mUOHDvHPf/6z3HV69uyJk5MT48aNIy4ujmPHjvHhhx/i4+ND3bp1AejatSu5ubksW7aMPn363HT7zs7O/P7774wfP56TJ0+SlJREdHQ0Op2O5s2b065dO5o2bcro0aPZvXs358+fZ8qUKaxevRovLy/gv2dAcXFx5OTkVPAnVJrFEsSNnfv7QI6joyP+/v7s3bv3tuv/5z//oUmTJjz//PMly+Lj43F3d8fR0dEsMT/sQjs1pW0HTwD2ZjbH2NyHIo2O+jnJ7P73Z1xKzbpNDzen1+gYHT4cK42eo5fjWR9X9keAEOamVqv58ssvCQgIYPTo0fTp04fDhw+zZMkSfH19y13HxsaGqKgojEYj/fr1Y/jw4Xh5eTFr1qxSbbp160ZRURHdunW76fa1Wi0LFy5ErVYzZMgQevToQUxMDAsWLKB+/fpoNBoWL15MixYtiIyMpFevXsTGxjJ79uyS8REfHx86dOjA6NGjWb58ecV+QP9DpVjoKaaff/6ZiIgIDh8+XGokftSoUeTn5zN//vybrnv06FGeeeYZli5dWirB/Pvf/2bnzp34+vpy8OBBXF1d6du3L4MGDbrpgBGUHRD6u+TkZGrXrn3byb0fFoqisOn7PRzYl4YKE+0dD6A+eBy1YuJ47UCen/oWTvZW99z/r3/t5Mu9y9Co1HzUeSxNapR/Ki4eLPn5+Zw7d45GjRpVmQfKqrPb/b1ufOfd7nvNYmcQNx4T/9+BFysrq9uO1n/11VcEBgaWuY3s9OnTZGVl0b17d6Kionj++eeZOXMms2fPrtjgH2IqlYoez7ehaVMnFNTsyAxE1aL4l1Wz5MN8N2Ee+Yaie+6/Y6NHCKsXjFExMXN3FLmF5i8XIIS4Nxa7/eBGVjMYDKUyXEFBATY2NjddLzc3l61bt5Y78LxkyRIKCgqwt7cHwNfXl5ycHObNm0dERMRNzyJulUVvdXbxsFKpVfQd8ijR87ZzLiGHHbktaedfCCfOEHDmT77+jxPD3xuM5h4epFOpVLzSagBnMxJIzclg0b7viAgbWqkDb0KIO2OxM4jatWsDxU8d/l1qaioeHjd/4vbPP//EZDLRpUuXMu/pdLqS5HCDj48Pubm5XL8ud87cDY1WzfMvt6O2hxUGxZpdhcEo3sW3H/vt38y3c9fd84N0dnpbXg8fhlqlZkfiXv5IiK3I0IUQFcRiCcLPzw97e3tiY//75ZCZmcmJEydo1arVTdfbv38/zZo1KzMQbTKZ6NSpE/PmzSu1/OjRo7i5ueHi4lKxO1AN6K20DHi1A64uWvIUO/aqQ1A8PdBgwnPbctZ+v/2e+/Z18+LZZj0AWHTgey5lXa6osIUQFcRiCUKv1zNw4ECmTp3Ktm3biIuLIzIyEg8PD7p06YLRaCQtLY38/NLVRePi4sqtP6JWq+natSuLFi1iy5YtJCYmsnz5chYtWsSoUaMqa7ceOrZ2ega91gEHezVZJicO2bXG6FYDK6UQ25UL+XXroXvuu0/TJ/F3b0JBUQEzd0dRZLz3sQ0hRMWz6INyr7/+Os888wzjx4+nX79+aDQaoqKi0Ov1JCcn07ZtWzZv3lxqnfT09FKPvP/dG2+8wUsvvcTnn39O9+7dWbJkCe+99x7PPSflHe6Ho7MNg15rj40VXDG6cbJGKwodHLA35pMzfyYHDvx1T/2q1WoiwoZir7fj3NUkoo+uq+DIRUWT0u1VQ0X9nSx2m2tVcae3g1UHl5KusXTODgqLwFOTgNf53Wjz80m2caflpI/wblzznvrdc+EQU3cW39b8bvvXaFm7WUWGLSqA0Wjk1KlT1KxZkxo1alg6HHEbGRkZpKam4uPjg0ajKfP+nX6vVa0iKsKi6tRz5oWXwvh2wW4uGBuib1BAvbN7qJ2Xxp4Jk3CY/jG13O7+qfU2ni15wrs9P5/5gy9il/LZk+NxtpaHHR8kGo0GZ2fnkptKbG1t5c6zB5CiKOTm5pKamoqzs3O5yeFuSIIQd6VREzeefrEVq5bu4y+jL/rG+dQ+fYgGmUn8+M5/6DvjAxzs7v5BukGBT3My7QxJ1y/xRexS3mk/ErVKqtE/SG7cXfi/dx6KB4+zs/Mt7wa9U3KJ6TbkElP59u8+z6ZVxbNctVDtwe30cVTAqcahDPj0DfS6u//lknT9Em9v/Q+FxkJeDHyann6PV3DUoiIYjUYKC2VujweVTqe77ZmDXGISZhUS3oDcnAJ+23KKI0obQhoX4PzXGXz+iiX6kygGfTACtfruLkHUc6rD4JbPsGj/d0QfXYt/zSZ4uZYt/S4sS6PR3PelC1E1yDm8uGdtOzchtF1DAA6oHyWrfn0AvA7+zMrZq+6pzy5e7WhTtyVGk5GZu6PIK8y//UpCCLOQBCHumUql4olezQgIroOCmv36DuTWqY0aBY/fVrAx+pd76vOfrQdSw8aFlOw0Fh8wb7VKIcTNSYIQ90WlVtHrhZY08XPDiJZ9dh3Jd3NDq5iwWRnFjq37bt/J/7C3siupz7Q9IYY/E/aYIXIhxO1IghD3TaNR88yQ1tRr4EShYsV+l84UODtibTKQ9eUsjuw/fdd9+tdswtP+3QFYtP87UrLvb8IiIcTdkwQhKoROp6HfiDBq1rIjX7HlkHsXCm1tcSzK4dyUKST8lXLXfT7t3w0/Ny/yivKlFIcQFiAJQlQYaxsdA/4ZjrOLNdmKI0c8H6dQb4VbwVX2fTCRtLS7q6irUWt4PWwYdjobzl45z/JjG8wUuRCiPJIgRIVycLRm4D/DsbPTcc1Ug5MNOmHUaqmdlcy2tz8hJ/fWk0H9Lzc7V15pPRCA9XFbOZJy0hxhCyHKIQlCVDhXNzsG/DMcKysNaYoH8Y06YFSpqJd+lnVvf4qh0HhX/YXVC+bxxm1RUJgT+xWZ+fc+N7YQ4s5JghBm4VHHkRdeaoNWqyJZqc+5xo+iAA3OH2L1R3Pvutrk4KBnqevowbX8TL7Y87VUFRWiEkiCEGbToHENnh7UCpUKzquakNioNQpQ/8jvrJ0RfVd9WWn1jA4fjk6t5WDyMbac/s08QQshSkiCEGbl26wWvZ4PBOCMpjnJ9QIAcPt9DT8v23JXfTVw9uTFlk8DsOzwGs5dTarYYIUQpUiCEGYX2LoeXXr5A3DSqhWptZugRkG36itifoy5q766encgpE4ARaYiZu6OIr/o7ga9hRB3ThKEqBThHRrzaCcvAI7ZPcIV9wbolSKyFszm+L64O+5HpVLxrzaDcLFx4lLWZb46sMJcIQtR7UmCEJWmU3c/gkLroaDmsFMHrjt7YGvM5/yU/5B05uId9+NoZU9E6FBUqPj13C52Jd59OQ8hxO1JghCVRqVS0ePpAPwCPDCh4ZBbZ3Lsa+BkyOLABx+TkXrtjvtqXsuXPv5dAZi/71tSs9PNFLUQ1ZckCFGp1Bo1fQcE0dC7BkXoOejRhTxrR9xy0tj+5r/Jzbnz8t7PNPsHTWo0Iq8wn5kxiyky3d3zFUKIW5MEISqdVqfh+aGt8KjrSAE2HPLsSr7OllpXE9n85icUFd3ZF71WrWFU2DBsdNaczjjHquMbzRy5ENWLJAhhEVbWOgaMCMXV3Y5c7Dna4AkK1XpqXzjB+g9m3PGDcDXt3Xil1QAA1pz4ieOpp8wZthDVyj0niAMHDnDlyhUA1q5dyyuvvML8+fPlCVdxx+wcrBj4cigOjlZk4sKxRl0wqrTUOr6LzZ9/dcf9PFK/FR0bPYKCwqyYxWQWZJsvaCGqkXtKEN9//z0DBgwgPj6eU6dO8c4771BYWMiSJUv44osvKjpG8RBzdrVlwCthWNvouKKqyYmGHTGhxunPjfz21bo77mdo8HPUcajF1bzrfLnnG/mhIkQFuKcEsXTpUsaPH094eDhbtmyhSZMmLF68mE8//ZQffvihomMUD7maHg70e6kNOp2aVI0npxq0BUCzdhl7N/15R31Ya60YFT4crVrLvktH+OnMdnOGLES1cE8J4sKFC3Tq1AmAnTt30r59ewC8vb1JT5fbDcXdq9fQhWeHtEKtVnFR58XZuqGoFRPZi+ZwMvboHfXRyKUeA1r0BuCbQ6s5f+2CGSMW4uF3TwmiRo0apKamkp6ezrFjx3j00UcBiIuLw83NrUIDFNWHt19NevdrCSo4b+NPokcgelMRFz6bwoVTiXfUR3efTgTVbk6hqYiZuxdTUGQwb9BCPMTuKUH06NGDsWPHMmzYMDw8PGjTpg2bN2/mvffeo0ePHhUdo6hGmgfX5cnezQA4Yx/MJTc/bAvzOPrBh1y9nHHb9VUqFa+2eRFna0cuZCaz9NAqc4csxEPrnhLEG2+8waBBgwgLC2PJkiVoNBoyMjJ47rnnGD16dAWHKKqbNm0b0b5LEwBOOoeR6twIx7zr7Bz3PvnZebdd38nakddCh6BCxS9n/yQm6YC5QxbioaRS5HaPW+rcuTMA27Zts3Ak1YuiKGz54Rj7dp1HjYnAlK24Zl/iikdjus2ZjFanvW0f3x5ew7q4n7HT2fBZ1/G42blWQuRCPPju9Hvtns4gDAYDX375JefPnwfgvffeIygoiOHDh3P16tV76VKIUlQqFd36NKdZyzqYUHPEozPXrd1wTfmLH9/9DyaT6bZ9PB/QCy/XBuQU5jErZjFGKcUhxF25pwQxdepUlixZQnZ2Njt37mTNmjW88sorZGdn8+mnn95xPyaTiVmzZtGuXTsCAwMZNmxYSdL5X7Nnz8bX17fc1zvvvFPSbsuWLXTv3p2AgAB69uzJH3/8cS+7KB4AKrWK3v1a0tjHHSNajng+QbbOCZdTB9k65cvbrq9VaxgVPhwbrTVx6WdZfeLuJigSorq7pwTx448/Mm3aNJo1a8a2bdto06YN//znPxk/fjy///77Hfczd+5cvv/+eyZOnMjy5ctRqVSMGDECg6HsnSfDhg1jx44dpV6jR4/G2tqawYMHAxATE8O4cePo378/a9eupW3btowcOZKzZ8/ey26KB4BGq+a5ISHUre+MASsO1+9KntYO+5ht/Lno9nNBeNi781JIPwBWn9jMybTT5g5ZiIfGPSWIa9eu4eVVPPnLzp07S25zdXFxIT//zqpxGgwGFi9eTEREBB06dMDPz4/p06dz+fJltm7dWqa9nZ0d7u7uJa+8vDzmz5/P22+/jZ+fHwALFy6kS5cuDBw4EC8vL9566y2aNWvG0qVL72U3xQNCb6Wl30ttcKtlT77KjsP1u2JQW6HasJwDa28/NtSuYRvaNwxFURRmxSwhuyCnEqIWouq7pwRRv359jh49yokTJzh//jzt2rUD4JdffsHT0/OO+oiLiyMnJ4ewsLCSZY6Ojvj7+7N3797brv+f//yHJk2a8PzzzwPFl6sOHDhQqj+A0NBQ9u2TCWWqOls7PQNfDsXJxYYctRNH6hfXbcr5ah7xO/bfdv3hwS/gYe9ORu5Vvty3TEpxCHEHbn8rSDleeuklxowZg1qtJiwsDD8/P7744gu++OILJk2adEd9pKSkAFC7du1Sy2vWrElycvIt1z169Cjbtm1j6dKlqNXFOS4zM5Pc3Fw8PDzuur8bI/rlSU5OLhOjsAxHZxsGvBzKV3N2cT3HnWP1OtMicSuXp03BocZk6jT1uum6NjprRoUPZ/y2z9hz4RC/nN1BF+92lRi9EFXPPZ1B9O7dm5UrV/L555+zYMECAAICAli0aBG9e/e+oz7y8orvZ9fr9aWWW1lZUVBw64nov/rqKwIDA0udLdy4tHUv/Ymqw62mPf1HtEFvpSFDX4cTdTugNpo4NeF9rl28fMt1vVwb0C/gKQC+OrSSpOuXKiNkIaqsezqDAPDz88PT05O4uDh0Oh3BwcHY29vf8frW1tZA8VjEjf8GKCgowMbG5qbr5ebmsnXrViZMmFBquZWVVUl/f3e7/uDW9wLf6uxCWEades48N6Q13y3aw2Wbhug8wvFJ2cWBcW8SPm8ONk4ON133H76dOXr5JIdTTjJz92ImPf4meq3+pu2FqM7u6QxCURSmTJlCeHg4zz//PH369OGRRx7hk08+ueNruzcu26SmppZanpqaWuYy0d/9+eefmEwmunTpUmq5s7Mztra2d92fqJoa+7jRd2AQKhVcsPflnHswVjnZ7IocS1HBzesvqVVqRrYZjJOVA4nXL/LNYak+LMTN3FOCWLBgAatXr+att95izZo1/PDDD7zxxhusW7eOqKioO+rDz88Pe3t7YmNjS5ZlZmZy4sQJWrVqddP19u/fT7NmzXB0dCy1XKVSERwczJ49e0otj42NJSQk5C72TlQVTVvUpsczAQCccwokydkf64x0/njjLZRbPEjnbOPEyNDiW6N/OrOdvRcPV0q8QlQ195Qgli9fzoQJExg4cCB+fn74+/szePBg3n//fVasuP296VA8VjBw4ECmTp3Ktm3biIuLIzIyEg8PD7p06YLRaCQtLa3MbbNxcXH4+PiU2+fQoUPZtGkTS5Ys4ezZs3z66aecPHmy5DkJ8fAJDmtAx26+AJxyCyXZwQurpET+fP/ftzybbVm7Gf/wfRyAeXu+ISNXKgAI8b/uKUFkZGQQEBBQZnlgYOBt7xj6u9dff51nnnmG8ePH069fPzQaDVFRUej1epKTk2nbti2bN28utU56ejrOzs7l9te2bVsmTZrEd999R58+fYiJieHLL78seWZDPJzadvYmtH0jAE7Waku6rSeaY8eInTbrluv1D3iKRi71yDbkMDtmyR2V7xCiOrmnYn1PPfUUL7zwAv369Su1PDo6mqVLl/LTTz9VWICWJsX6qgbFpLD2+0Mc3X8RtWIk6OKPOOenYt+nN4FDXrzpepeyLvPWz5MpKCrg+eY9ebpZ90qMWgjLuNPvtXu6i2no0KF88MEHXLhwgeDgYFQqFfv27ePbb79l3Lhx99KlEPdFpVbR6/lA8nMLOX0ylcN1uxCctBnWruWUew18epT/xV/HoRbDg59n7p6vWXl8E81r+eLrJmecQsB9PAcxZswY1q1bx8iRI3n11VdZt24do0aNYuDAgRUdoxB3RKNR88ygEOo1cqVIpeeQZ1fyNPakLYriwu5dN12vQ8Mw2tZvjUkxMWv3YnIMuZUYtRAPrntKEABDhgxhx44d7Nq1i507d7Jz506aN2/OY489VoHhCXF3dHoN/Ya3pmZtBwxqGw56dqVAZcP5z6eRcfJEueuoVCpeatWPmnY1SMu9wvx930opDiG4jwRxg6urKzVq1ACKH0q7fPnWT7MKYW7WNjoGvByKs6steVpHDtV9gkKjjvgPJ5B9KancdWx1NowKH45GpSYm6QC/nbv5GYcQ1cV9JwghHkQOjtYMfCUUOwcrsvWuHKnTmaICFUfGjcNw7Uq56zSp0YjnA3oBsOTACi5mplRmyEI8cCRBiIeWq5sdA0a0wcpayzVrD455PIYxu4gDY16nKK/8kt+9/LoQUMuXAqOBGbujMBgLKzlqIR4ckiDEQ82jrhMvDGuNVqsm3a4+cbUepSgjjwNvRGAqKluSQ61SMzJ0CA5W9py/doHow2ssELUQD4Y7vs11zpw5t21zs+lChbCkBl41ePrFYFYs3U+yQxN0Rfk0ubiPQ++MJmjKbFRqTan2rjbOjGwziP/8OZfNp38jwKMpIXXKPhgqxMPujhPEDz/cWVEzmTtBPIh8m3vQ89kWrF9+mESXAPTGfBqcOsaxT96m+fgpqFSlT6aD6wTQrUlHtpz+jbl7vuazru/hauNsmeCFsJA7ThC//vqrOeMQwuxatqlHXq6BrRtOcsatNTpjAew7Tfysj/F9/QNUKlWp9gMC+3Ai7TTnr13gi9iveK/D66hVclVWVB9ytItqJfwxLx7pWPyk9Mmaj5JqV5+M345wbumMMm31Gh2jw4djpdFz9HI86+PKzpUuxMNMEoSodjr38COoTT1QqTjm8RhXrT1IWb+Di2vKlqqv6+jB0ODnAFh+dD2nM85VdrhCWIwkCFHtqFQqejwTgG/zWigqDYdqP851bQ0SozeT9svKMu07NnqE8HohGBUTM3dHkWvIs0DUQlQ+SRCiWlJr1Dw9MJgGXjUwqXUcqvME2Yojfy36nquxW0q1ValUvNyqP+62rqTmZLBwf7SU4hDVgiQIUW1pdRpeGNaK2p5OFGqsOVC3K9kGW07PXkTW0R2l2trpbXk9fBhqlZqdifvYnhBjoaiFqDySIES1ZmWto/9LbXB1s6NAa8+huk+Qm6Mn/vPp5JzeX6qtr5sXzzbrAUDUgeVcypK6Y+LhJglCVHt2DlYMfCUUB0crcvQuHKrThdxrWk599h/yEktXgO3T9En83ZtQUFTAzN1RFBmLLBS1EOYnCUIIwNnVlgEvh2JtoyPTuiZHPDqSfRlOf/ZvCpLPlrRTq9VEhA3FXm/HuatJRB9dZ8GohTAvSRBC/L+atR3p91IbdDoNV+w8OVGrLZmJRZyZ9j6GtP+WCa9h68K/2hRPY7ox/hcOJR+3VMhCmJUkCCH+pl5DF54dEoJareKygxen3EK5eqqAv2aNp/Dqf8t/t64byBPe7QH4InYp1/KuWypkIcxGEoQQ/8PbryZP9WsJwAVnfxJcArlyLJuEue9RlJlR0m5Q4NPUc6rD9YIsvtizFJNislDEQpiHJAghyhEQXJcnezcD4K8awSTZ+5J+4BqJC9/HmFN8tqDX6hkdPhydRsfhlJNsipd6ZeLhIglCiJto064R7bs0ASDePZxk64akxl7mwpIJGPOLJxyq51SHIS2fBSD66FrOXpGS9+LhIQlCiFvo0NWHVo80AJWK47Xak6quQ8rOJC598zGm/y+58bhXW9p4tsRoMjJzdxR5hfkWjlqIiiEJQohbUKlUPNmnOc1a1kFRaThSuxMZRjdS/jhNyveTMRUZUKlU/LPVQGrYuJCSnUbUge8tHbYQFUIShBC3oVar6N2vJY193ErqNl3NcyL59+Nc/uFzFGMR9lZ2RIQNRaVS8UdCLH8kxFo6bCHumyQIIe6ARqvmuSGtqFPPmSKNFQfrdOXadTsub9tH2oY5KIoJ/5pNeNq/OwCL9n9HSnaahaMW4v5IghDiDumttPR/qQ01atpRoLPjYJ0nuJpqxeVtf5K+ZSGKovC0fzf83LzIl1Ic4iEgCUKIu2Brr2fgy2HYO1mTq3fmUJ0uXEnSkvrLz1z59WvUKjWvhw3DTmfD2SvnWX5sg6VDFuKeSYIQ4i45udgw6JUwrGy0ZFm7c6R2Z678pSb15/Vc27kaNztXXmk9EIB1cT9zJOWkhSMW4t5IghDiHrjVsmfgy2FodWqu2tbheK32XDmlInXLd1zfs5GwesE83rgtAHNiv+J6fqaFIxbi7lk0QZhMJmbNmkW7du0IDAxk2LBhnD9/8weNCgsL+fzzz2nXrh0tW7Zk4MCBnDxZ+tdZp06d8PX1LfUaO3asuXdFVEN16zvzwrDWqNQqUu0bEVcjjIx4uLxxCZmHtjE46FnqOnpwLT+TuXu+kVnoRJVj0QQxd+5cvv/+eyZOnMjy5ctRqVSMGDECg8FQbvsPP/yQVatW8fHHH7N69WqcnZ0ZMWIEWVlZAGRnZ3Pp0iXmz5/Pjh07Sl4TJkyozN0S1UhjH3eeHhgEwEUnP844BHMlHlLXzaPw1N7iUhxqLQeTj7Hl9G8WjlaIu2OxBGEwGFi8eDERERF06NABPz8/pk+fzuXLl9m6dWuZ9klJSaxatYrJkyfz2GOP4eXlxaRJk9Dr9Rw7dgyAU6dOoSgKwcHBuLu7l7wcHBwqe/dENeIfWIcezwQAkOAayDkrf67EK1z+YQbu6Zd5seXTACw7vIZzV5Nu1ZUQDxSLJYi4uDhycnIICwsrWebo6Ii/vz979+4t037Hjh04OjrSvn37Uu1//fVXwsPDAYiPj8fd3R1HR0fz74AQfxMS3oDHnvQB4LR7KImqxlw9bSJl1We019WgVZ0WFJmKmLF7EflSikNUEVpLbTglpbi2fu3atUstr1mzJsnJyWXaJyQkUK9ePX7++WcWLFjA5cuX8ff35+2338bLywsoPoOwtbUlIiKCgwcP4urqSt++fRk0aBBq9c1zYefOnW/6XnJycpkYhShPu8ebkJNtYO+OBE7Waocu2YD6zAVYMZmhz73F2avnSc5KZcnBlSUTDgnxILPYGUReXnGhM71eX2q5lZUVBQUFZdpnZ2eTmJjI3LlzGTNmDPPmzUOr1dK/f38yMopr9J8+fZqsrCy6d+9OVFQUzz//PDNnzmT27Nnm3yFR7alUKp58qhn+LWujqNQc9ehIclZNss4VkLN6Gv/06YYKFb+d28WuxH2WDleI27LYGYS1tTVQPBZx478BCgoKsLGxKdNep9ORlZXF9OnTS84Ypk+fTocOHVizZg0vvfQSS5YsoaCgAHt7ewB8fX3Jyclh3rx5RERE3PQsYtu2bTeN81ZnF0L8L5VaRZ/+QeTkFHL+dDqHaz+O9uIWNPqruP74Nb3CO7EuYQfz932Lt2tDatq7WTpkIW7KYmcQNy7bpKamllqempqKh4dHmfYeHh5otdqS5ADFSaZevXpcuHABKE4iN5LDDT4+PuTm5nL9ukwJKSqHRqOm/7DW1Kzr+P91m54g5YI9OReuEb7nD5o41yOvMJ+ZMYspMhktHa4QN2WxBOHn54e9vT2xsf+tepmZmcmJEydo1apVmfatWrWiqKiIo0ePlizLz88nKSmJBg0aYDKZ6NSpE/PmzSu13tGjR3Fzc8PFxcV8OyPE/9DpNQz+ZxiONWwxaG05WKcrKQnWGC6m8+y5C9horTidcY5VxzdaOlQhbspiCUKv1zNw4ECmTp3Ktm3biIuLIzIyEg8PD7p06YLRaCQtLY38/OI7Plq1asUjjzzCW2+9xb59+zhz5gxvvvkmGo2Gp556CrVaTdeuXVm0aBFbtmwhMTGR5cuXs2jRIkaNGmWp3RTVmI2tnuEjH8HaTk+e3pFDHk+QekaPzYXLPJNZPH/1mhM/cexyvIUjFaJ8Fn1Q7vXXX+eZZ55h/Pjx9OvXD41GQ1RUFHq9nuTkZNq2bcvmzZtL2s+ePZs2bdrw2muv8cwzz5Cdnc3XX3+Nq6srAG+88QYvvfQSn3/+Od27d2fJkiW89957PPfcc5baRVHNOThZ89Lrj6LVa8i2qsGhmp1JO63FL+EioUU6FBRmxy4hsyDb0qEKUYZKkef/b+nGIPWtBrKFuJ2Ui9dZOHMHilHBLTuRoMzfcWpqZK5PbVJVRkLqBPBm23+hUqksHaqoBu70e02K9QlRCTzqOjFwRCioIN2+Psfsw8g+reb586logP2XjvLTme2WDlOIUiRBCFFJGjVx4+kXgwGFZEcfTuiCsY1X6JZaXOn1m0OrOXo5zrJBCvE3kiCEqETNAuvwRJ/iuk2JLgHEG/0JOFaAX3Y+haYiPv59JtN2LSQtJ8PCkQohCUKIShfWtiHhnb0BOOPWmjM53vQ+lEXo9VxUCsQkHWD0ln+z4tgG8ovKVhUQorJIghDCArp096NZa08ATtZ8hMSMenSNzyMi6QqN8gopNBay6vhmIjf/mx3n98pcEsIiJEEIYSF9nw+kvp87qNQcq9WBcxdrUrOwHiMuXGVA8nVcikxk5F1lVsxiPvj1c/66cvPJtIQwB0kQQliISqVi0LDW1PB0xKTWcqR2J04eLeTKhVo0M9RgzPl0umRko1MgPv0s72ydwpd7vuGaTF8qKokkCCEsSK1R88prj2LraoNRredwnS7sUQVyZm8ehiwvOmVqeeN8Oi2z8lFQ+PXcLkZtmsD6uK0UGYssHb54yEmCEMLCtDoNr4/tgG9wHRTgim1d9tR7iph0D84fULA1+vB8ah7/vHCVuvmF5BXls+zwD7zx48ccuHRUxieE2ciT1LchT1KLypSRnsPybw+QnlhcfVhtKqTBtWM0MibiEeiOkhfPAUdrfnKzJ1tT/PuupYc/g4Oepa5j2SrIQpTnTr/XJEHchiQIYQlnT6ezKvogBZnFt7nqi3LxyjhAY9d8XBpCfv4FfnW1ZaezLUaVCo1KTdcmj/Fssx7Y6W0tG7x44EmpDSGqMK8mbrz5weM83qcZ6NQYtLacrNWW7aZgju0xohj86J6lJzLxCn45BRgVE5tP/crrmyfwy9k/MZlMlt4F8RCQBCHEA0qlUvFI20a8O7ErzR9pgKJSyLZy5VDdrvye4snpgzpcNM0YkprH0EvXcDcUkVWQzYJ90by9dTInUk9behdEFSeXmG5DLjGJB0V2dgHffnOAlDPpqFCBYqJO5ml8SMAz0IWi68eJcbbhlxr25KuLq8KG1wthYGAf3O1qWDh68SCRS0xCPGTs7a145V/hDItsh3UNG1CpueTky58Ondi1X0vmtSa0U2oxNiGd0Ot5qBSF3Un7S8p2FBQZLL0LooqRBCFEFVPP05k33+1Mt36BYKXGqNbxV41gthmC2BerRUcgfbN1RCRdpVGeoaRsx+jNH7IzUcp2iDsnl5huQy4xiQeZYlLYsPkkB34/i1opvqzkkJ9Ok+wjNG3lDrlHOWoFm93suabTAODr5sXQoOdo7FrfkqELC5LbXCuIJAhRFeTnF/LNtwe4dPwyKlXxhQG3nESaqs7TqKU9BRmH+cPZht9d7ChUq1ChomPjR+gX0Asna0cLRy8qm4xBCFGNWFvrGDE8lJff6oh1LTsURSHdrj47bB7l1z3WZOW14Alq80ZiBoE3ynb8tZPXN09gQ9wvUrZDlEsShBAPkdq1HHjzzY48NbQVirUaRaXmgnNTfspqzo699tjq2zAgW//fsh2F+XxzeHVJ2Q4h/k4uMd2GXGISVZWiKGz6+RR7t8ahUYrHH6wLs/DOPUlQayeM1w+w3xp+qmFPtrb4t2JQ7WYMavmMlO14yMklJiGqOZVKxT+6+vLuJz2o3aIWJqWIfJ0Dx5zasO6IK+cu+RLuEMDY81dofzUHjaJwMPk4Y3/8mKUHV5FjyLX0LggLkwQhxEPOykrLiMGt+ee7j2NV2w4UI5nW7uzWBLMpxp5sYyi9NHUZ/beyHZtObWPU5gn8cnaHlO2oxiRBCFFNeLjZ89bYjjw1IgyjDaAopNk35KerTdi2xxUH23YMz7EuKduRWZDNgn3fStmOakzGIG5DxiDEw8hkUtjy22liNx9Hiw4ArbGAxgVnCG1jS1F6LLtt4BdXO/L/v6x4eL0QXgzsi5udqyVDFxVAxiCEEDelVqvo0dmHdyb1xKNlLUymAoo0VpyybcaqgzU4nRZER+eWjE28Spu/le0YtXkCK45tlLId1YQkCCGqMRsrLS+/2JpX3nsSbR1bMBnI0zmyz+jLD7tcyKIDL+jq/bdsh6mIVcc3SdmOakIShBCC2m52vPtGJ3q98ihF1iZUJiPXbWqxLa0BG/fUxsnxcV7NsaZ/8nWcC41k5F1l5u7FTPj1c/66kmjp8IWZyBjEbcgYhKhujCaFzdvPsGfDIXRYg0qFSjHSoCiRtqHWFF3+k9+tFbaXlO2Ajo0flbIdVYiMQQgh7olGraJnxya8/Ulv3IM9UIqyUVQaEnSNWL7XhRPp4XR3acUbiVf/v2wHxWU7Nk1gY7yU7XiYSIIQQpTLzkbHvwa25qX3eqCpbY26KJdCjTVHCurx3W53rqu7MljX4L9lO4ry+frQjbIdxywdvqgAFk0QJpOJWbNm0a5dOwIDAxk2bBjnz5+/afvCwkI+//xz2rVrR8uWLRk4cCAnT54s1WbLli10796dgIAAevbsyR9//GHu3RDioVa3pgPvjX2cJ1/pgMHKgNpYQJ7OkT9TavLDvga4OPZgVK41T1/OxL7IRHJ2Kv/58wsm/zGHi5kplg5f3AeLJoi5c+fy/fffM3HiRJYvX45KpWLEiBEYDOXfQvfhhx+yatUqPv74Y1avXo2zszMjRowgKysLgJiYGMaNG0f//v1Zu3Ytbdu2ZeTIkZw9e7Yyd0uIh1Ir/1p8+HFvmvcMoNCYicpk5JrenQ1xLvxyqiXBNZ9gXGp+mbIdX0vZjirLYoPUBoOBsLAwxo0bR79+/QDIzMykXbt2TJo0iR49epRqn5SUxOOPP878+fN57LHHStr37t2bTz75hPDwcIYPH46joyPTp08vWe+FF17Ax8eHjz766J7ilEFqIcrKzDGwbM0hMmLjUPTFD86pTUU0sb9GWEAeKYnb2eRqS5ydFQCOejteaNGbTo0eQa2WK9uW9sAPUsfFxZGTk0NYWFjJMkdHR/z9/dm7d2+Z9jt27MDR0ZH27duXav/rr78SHh6OyWTiwIEDpfoDCA0NZd++febbESGqIUc7Pa8ObMOL7/aGWlq0huuY1Fric92I3uXGNVUvXtY3/m/ZDkNOSdmOk2lStqOq0Fpqwykpxdcma9euXWp5zZo1SU5OLtM+ISGBevXq8fPPP7NgwQIuX76Mv78/b7/9Nl5eXmRmZpKbm4uHh8cd9fd3N7JpeZKTk8vEKIQo1rC2I++P60rMsWQ2LvsT+zyFQp09sZdsOJbkxaMBPryRs4M/r19hm6sdCdcuMOHXaTxSL4SBUrbjgWexM4i8vDwA9Hp9qeVWVlYUFBSUaZ+dnU1iYiJz585lzJgxzJs3D61WS//+/cnIyCA/P/+u+hNCVAyVSkV4QB0+/OQZmvQMwmhIQ2ssIEfjwM8nbPnxTBtau/fgzdSCkrIdu5L2M3rzBFZK2Y4HmsXOIKytrYHisYgb/w1QUFCAjY1NmfY6nY6srCymT5+Ol5cXANOnT6dDhw6sWbOGp59+uqS/v7tZf393q+twtzq7EEL8l06r4ZkuvnQObcDXPxwke89xTNbupOPMqliFRvZd6euZTej539hQw4YEG1h5fBO/nt3Ji0FPE14vBJVKZendEH9jsTOIG5dtUlNTSy1PTU0tc5kIwMPDA61WW5IcoDjJ1KtXjwsXLuDs7Iytre0d9yeEMA8XR2tGDQnnuXeepshdg3XeZVCpOJfjQPROd67yNCN13vRPuY5ToZGM/GvM2B3FhF8/59zVJEuHL/7GYgnCz88Pe3t7YmNjS5ZlZmZy4sQJWrVqVaZ9q1atKCoq4ujR/86bm5+fT1JSEg0aNEClUhEcHMyePXtKrRcbG0tISIj5dkQIUa4m9Vz48O0ehI3oSr7qKrYFGZjUWg5dsuK7fU1wtnuOt3JteTwjG51JIS79LG//PIn5e5dxPT/T0uELLHiJSa/XM3DgQKZOnYqrqyt169bls88+w8PDgy5dumA0Grly5QoODg5YW1vTqlUrHnnkEd566y0++ugjnJ2dmTVrFhqNhqeeegqAoUOH8vLLL+Pv70/79u1ZvXo1J0+e5JNPPrHUbgpRralUKtoHedKmWT9W/xLP6Y07sNE6UqBz4I+T4Kh+lPbBatqkbmSTHRx2sGbbXzvZdX4fzwb8gye9H0OrsdjXVLVn0WJ9RqORadOm8cMPP5Cfn0/r1q354IMP8PT05MKFC3Tu3JnJkyfTt29foHigeurUqfz444/k5+cTHBzMu+++i7e3d0mfa9euZe7cuaSkpODt7c24ceMIDw+/5xjlOQghKk7q1Vy+WbWfon2HMdrUpkhT/JxEbbsC2gVkc+nCz6x3teGSdfEkRnXs3Bkc8hxBtZtbMuyHzp1+r0k119uQBCFExTv+VwbfL9tOjYuXyLKrh6LSgKLgW9tEcO3THLhykJ9q2JOtLb4KHuThz+CgZ6njKOOJFeGBf1BOCFF9NWtcg3+P70Ozof8gX0nHOecCqFTEp2hYccAbJ+v+jMtxpN3VXNSKwsGUE7yx5SOWHlxFriHP0uFXG5IghBAWoVar6BLWkPcmD8KxWzjq7LPY56djVGnZE6+w8fQjBNs9zZjUQvxyCjCisOnUNiI2jmfb2R2YTCZL78JDTxKEEMKibK11DHkqkGGfDMfgXw/763FYF2aRZ9Kx7ZCaPam9eNr2cYamZOFuKCKrMJf5+77l7Z8+kbIdZiYJQgjxQKjjZs/bIzvT6c1BZDqrcbt2Eo3RwJU8Dev2OpKa9wL/ojk90rKwMppIyLzEhF+nMWPnQtJzrlg6/IeSJAghxAOlpU9N/j3hWRoM6UuBMQX3zNOoFBPn02DNocY4qgfyRrbrf8t2XDjA6E0fsELKdlQ4SRBCiAeORqOmRzsvxk15Cc0T7dFdP0mNnCQUVBw9W8SWU+G01D3La2kmGuYZMChGVh3fxOiN77MrcT9yc2bFkAQhhHhgOdjqefnZEPp98i+u+TXC9cpBHPLTKVTU7DqmsOdST3qrn6Df5Zzish0FmczYvYgJWz+Vsh0VQBKEEOKB18DDkXdHdaXN2BFkOKipm3EAq6Iccgwqfjlsx+XrL/BKUdB/y3ZcTeDtnyfxZezXZOZnWTr8KksShBCiSlCpVLRp5sGEiQNwefEF8g2XqHflMBpTIanXTGw+Wg+7goFEZNYkMCsfBfg1YTcRG95jY9wvFBmLLL0LVY4kCCFElaLTaujb2YfXP/0XuZ07Y331KLWvx4Ni4uyFQraeak2A8TleSlNRJ7+QPFMhXx9ezRubPuBQ8nFLh1+lSKmN25BSG0I82M5cuEb0t9upfSwGg10DMuw8AdBroVWTfNIKNvGzs56cG2U73H0Z3LofdRxqWTJsi5JSG0KIasHb05n33+yF3xuvcdlei9flP7EvuIKhCHadtOby5WcZkteKtlf+v2xHWjxjNv+bpQdWSNmO25AEIYSo8lQqFe1a1uW9yUPRDBxMriEZ77Td6Ityycwxsf1kbWyu92fEFU98cwowobDp9G9ErH+HX6Rsx03JJabbkEtMQlQ96dfy+Hb1XtTbf8RJY0Oic3NM6uIS4k0aaLFV/8hPTjmk64vnmmhoV5NhoS/i5+59q24fGlLuu4JIghCi6opLuMLyb36l8Ymd5NvXI9mhCahUqNUQ0NjAFeOP/OqqokBTfDElvHYAL7Z6ATdbVwtHbl4yBiGEqPb8Grry/ntPUydyNBdtrfG/9DOuuRcxmeDwGT1pl57ihWvhtL6Wj0pR2J18lFEb3mflkXVStgNJEEKIh5xaraJzmwa8OWUEef1fIrsgjebJ27AruEp+gcL+s25YXX6egWmNaZhroBATK0/+yKj177ArcW+1Ltshl5huQy4xCfFwScnI4dvlu7Hf+RPOah1/uQZRqLUBoG4tDXbW29nmeoXrOg0Avo51GR42mIYu9SwZdoWSMYgKIglCiIfTkTNprFm6laan/iTf1pNE52aY1MWD1t71jFzX/sSOGiYK1SpUQEfPEPqHPI+jtYNlA68AMgYhhBC30MLbnfEfvoDD62M5b2tH4IVNeGSeAeBMkoYriT3onfooAdeKist2XNhPxPp32HjiR4pMRssGX0kkQQghqi2NRk23R72I/OxVUp97mZzCK4QkbcA5LxmjUeFkogvWSX3om+xLnbwi8hQjXx9dxxvr3uHgpaOWDt/s5BLTbcglJiGqj6TLWURH/0nNPT/holJzpkYrcvVOALg4qbFz3MOfbsnk6Ip/Wwc6N2DoI0OrXNkOGYOoIJIghKh+9p28zKalmwg8s5N829qcc21JocYagDruRrLstxPrlotJpUIDPNnwEZ4NegZbvY1lA79DMgYhhBD3qFXTWrw7cQi8+hZ/2ToSnLSe+lePolKMXErTkHWuE08mtcf3mhojsClhFxFr3+KX+N8eqrIdkiCEEKIcOq2a3p18GDl1FOf6/IusokxCz6+hVtZfACSmOGLzV3eeTGqBe55CllLIgkMreHP9e5xMPWXh6CuGXGK6DbnEJIQA+Ovidb5f9isND27F2WTktFsbrtsUjz3Y2qiwdTnMbo/zFPx/WfGwGk0Y9MiQB7Jsh4xBVBBJEEKIGxRFYdeRS2z7eiPBCbsotKrBmRqtyNM7AuDsaCLPOYYDta6gqFToUNGrcXv6BPVFr9VbOPr/kjEIIYSoYCqVikcD6/LWf14iZ8Q4TtvWIDhpA03SYtEaC7iWqaYg8REeO9uJxletKURh9V/beX3NOHad3VXlynZIghBCiLtkpdPw/JPNGf75GE784xUyTfk8cn4V9a4dQ6UYSbtii92Zzjz2VwguuVqumAzM2PcN769/n3NXzls6/Dsml5huQy4xCSFuJ/78FVYv3UqTY9twNhk4U6MVafYNAdBpwdr1FPvqnqFQp6BSoENNfwY+MsRiZTtkDKKCSIIQQtwJk0nh9/2J7PpmPa0vxFKkc+C0W2syrWsCYGujkOd6kKN1kkEFNqh42udxugc+hVatqdRYZQxCCCEqkVqtolPrBrzx2b9IGzyWeNtatLywhWYpv2NdmE1ungrlYjDhJx+nXrozeSgsO7WVyNVjOXB+v6XDL5dFE4TJZGLWrFm0a9eOwMBAhg0bxvnzN78+t2bNGnx9fcu8/r5Op06dyrw/duzYytgdIYTAxkrLwKda0n/qOA49MYJMpYiwxB/wTt+LxmQgK9sKp78eJfxUOM7Ztlw25fOfmEVM3PAhl65dsnT4pVj0EtOcOXOIjo5m8uTJ1KpVi88++4ykpCQ2btyIXl/2lrD//Oc/HD9+nGnTppVa7urqikajITs7m1atWvHll1/SrFmzkvetra1xcLi3a31yiUkIcT+Onk1n/ZJNNIvfjlNRPudcW3LRyQ9FpUalAr1rIkfqxVGoL0SjwBMeLXj+0SHY6sxXtuOBv8RkMBhYvHgxERERdOjQAT8/P6ZPn87ly5fZunVrueucOnUKPz8/3N3dS700Gk3J+4qiEBwcXOr9e00OQghxvwK83Hj7o0FYj36PmNrB1L9ykNDENbhlJ6IoUJBRn2bHutA0sREmRc2Wy0d4bfVYth7ZiEmxbNkOraU2HBcXR05ODmFhYSXLHB0d8ff3Z+/evfTo0aPMOvHx8XTt2vWmfcbHx+Pu7o6jo+NdxXIjm5YnOTmZ2rVr31V/Qgjxdxq1iicf9aJt8ChWbjhI9sa1tEz5les2tTjl1oZsaqBJ8Sf4WhMu1zzChVopLDy5iZ9O/cawsCH4ewZYJG6LnUGkpKQAlPnyrVmzJsnJyWXaX7lyhfT0dPbu3cs//vEP2rZty8iRIzl37lxJm1OnTmFra0tERARt27alV69efPXVVw9V8SwhRNVlb6Nj6HNt6DP1HWI7DOY60CZpPf6X/8DKmEtBvg7nxBBCTrTH6ZozicZcPtw5l6kbPyY9K63S47XYGUReXh5AmbEGKysrrl+/Xqb9qVPFxa80Gg1TpkwhNzeXuXPn0r9/fzZs2ICbmxunT58mKyuL7t2789prr7Fv3z6mTp3K9evXGTVq1E1judV1uFudXQghxL3wrOnA2DE92XeyNT8t2UDLszsIT1hFonMzzrsGUpDjQL1Tj9LQKYW4+ifZwyUObfqAHnWC6Rs+CCudVaXEabEEYW1dXFvdYDCU/DdAQUEBNjZlB2fCwsLYs2cPTk5OJcu++OILOnbsyA8//MDLL7/MkiVLKCgowN7eHgBfX19ycnKYN28eERERqNVyV68Q4sHRqqkHLScPZ/Mf7Yj9bg2tUg9RJ/MU51yDuOjki/G6Bz7HalFQ4xxn651hTfIBtq8+zIDmPWnb7AlUKpVZ47PYN+aNS0upqamllqempuLh4VHuOn9PDgC2trZ4enpy+fJlAHQ6XUlyuMHHx4fc3Nxyz0qEEMLStBo1vTr68fK0MZx+OoLTtnXwS9tNaOJaXHMvoigq9OmNaX60M54XG3JVUZh9fC3jV7/JX8knzRqbxRKEn58f9vb2xMbGlizLzMzkxIkTtGrVqkz76OhoQkNDyc/PL1mWnZ1NQkIC3t7emEwmOnXqxLx580qtd/ToUdzc3HBxcTHfzgghxH1ysrfi5Rfb0vWz99kZ1o/rKi1Bl34m6OKP2Bddx1ikwfliM1oc7YhLugeni7J5Z/tMjp3eZbaYLJYg9Ho9AwcOZOrUqWzbto24uDgiIyPx8PCgS5cuGI1G0tLSShJCx44dURSFN998k9OnT3P06FEiIiJwdXWlT58+qNVqunbtyqJFi9iyZQuJiYksX76cRYsW3XL8QQghHiQNazsy7u2n8XzvA/5o3Amd4RptEtbQ9PKfWGHAWGBN3b9CaH7iEVyuO0NettliseiDckajkWnTpvHDDz+Qn59P69at+eCDD/D09OTChQt07tyZyZMn07dvXwBOnjzJ1KlTOXz4MIqi8Oijj/LOO++UXK4qKipi4cKFrF69mpSUFDw9PRk2bBjPPffcPccoD8oJISzFUGhk/S8nSFr1A8EZx1Ch5rxLcxJrBGJUin/fD/pXGA293e6qXynWV0EkQQghLO1KZj7LV+zE6teN+OYkUqCx4WzN1mTX8OL5f7WljqfzXfV3p99rFruLSQghxJ1xdbTmXy915tQTwaxZvBmf47/in/wHJP9BfpwjeLY3y3YlQQghRBXhU9+FNyf0Z/v+tuxc+gM1MhJR6V1pbKbtSYIQQogqRKVS8VirBoS1iOBSWg6N6txdaaG7IQlCCCGqIGu9lsZ1nW7f8D7Io8VCCCHKJQlCCCFEuSRBCCGEKJckCCGEEOWSBCGEEKJckiCEEEKUSxKEEEKIckmCEEIIUS5JEEIIIcolCUIIIUS5pNTGbaSmpmI0GkvK4wohRFWXnJyMRqO5bTs5g7gNKysrtNq7z6PJyckkJyebIaKqTT6Xm5PPpnzyuZTvfj4XrVaLlZXVbdvJhEFmIhMNlU8+l5uTz6Z88rmUrzI+FzmDEEIIUS5JEEIIIcolCUIIIUS5JEEIIYQolyQIIYQQ5ZIEIYQQolxym6sQQohyyRmEEEKIckmCEEIIUS5JEEIIIcolCUIIIUS5JEEIIYQolySIe2QymZg1axbt2rUjMDCQYcOGcf78+Zu2v3r1Km+88QatW7emdevWvP/+++Tm5lZixJXjbj+XNWvW4OvrW+Z1q3UeBnPnzuXFF1+8ZZvqcsz83Z18LtXlmLl27RoffPAB7du3Jzg4mH79+rFv376btjfL8aKIezJ79mwlPDxc+f3335WTJ08qw4YNU7p06aIUFBSU237gwIHKs88+qxw7dkzZtWuX0rFjR+XNN9+s5KjN724/l8mTJysDBw5UUlNTS72KiooqOfLKs2TJEsXX11cZOHDgLdtVl2Pmhjv9XKrLMTN06FClV69eyt69e5WzZ88qH3/8sdKiRQvlzJkz5bY3x/EiCeIeFBQUKEFBQUp0dHTJsuvXrystWrRQNm7cWKb9gQMHFB8fn1J/2D///FPx9fVVUlJSKiXmynC3n4uiFP8jmDhxYmWFaFEpKSnK8OHDlZYtWypPPvnkLb8Iq8sxoyh397koSvU4ZhISEhQfHx9l//79JctMJpPSpUsXZcaMGWXam+t4kUtM9yAuLo6cnBzCwsJKljk6OuLv78/evXvLtN+3bx/u7u54eXmVLGvTpg0qlYr9+/dXSsyV4W4/F4D4+Hi8vb0rK0SLOn78OE5OTqxfv57AwMBbtq0uxwzc3ecC1eOYcXFxYcGCBTRv3rxkmUqlQlEUrl+/Xqa9uY4XmXL0HqSkpABQu3btUstr1qxZ7gxPly9fLtNWr9fj7Oz8UM2Udbefy5UrV0hPT2fv3r188803XLt2jcDAQMaOHUujRo0qJebK1KlTJzp16nRHbavLMQN397lUl2PG0dGRDh06lFq2ZcsWEhMTadu2bZn25jpe5AziHuTl5QHFf4C/s7KyoqCgoNz2/9v2Vu2rqrv9XE6dOgWARqNhypQpTJ8+ndzcXPr37096err5A36AVZdj5m5V12Nm//79vPvuu3Tu3LncZGqu40XOIO6BtbU1AAaDoeS/AQoKCrCxsSm3vcFgKLO8oKAAW1tb8wVaye72cwkLC2PPnj04OTmVLPviiy/o2LEjP/zwAy+//LL5g35AVZdj5m5Vx2Pml19+YezYsQQGBjJt2rRy25jreJEziHtw41QuNTW11PLU1FQ8PDzKtPfw8CjT1mAwcO3aNWrVqmW+QCvZ3X4uQKl/6AC2trZ4enpy+fJl8wRZRVSXY+ZeVKdjZtmyZURERNC+fXsWLlxY6ofX35nreJEEcQ/8/Pywt7cnNja2ZFlmZiYnTpygVatWZdq3bt2alJSUUvdp31g3ODjY/AFXkrv9XKKjowkNDSU/P79kWXZ2NgkJCQ/9IOTtVJdj5m5Vp2MmOjqajz/+mAEDBjBjxoxyLyHdYK7jRRLEPdDr9QwcOJCpU6eybds24uLiiIyMxMPDgy5dumA0GklLSys5iAMDAwkODiYyMpIjR44QExPDhAkT6N2790P1a/BuP5eOHTuiKApvvvkmp0+f5ujRo0RERODq6kqfPn0svDeVq7oeM7dTXY+Zc+fOMWnSJLp06cIrr7xCRkYGaWlppKWlkZWVVXnHyz3fIFvNFRUVKZ9++qkSFhamtGzZUhkxYoSSlJSkKIqiJCUlKT4+Psrq1atL2qenpysRERFKy5YtldDQUGXChAlKfn6+pcI3m7v9XE6cOKEMGzZMCQkJUYKDg5WIiAjl0qVLlgq/0rz11lul7vevzsfM393J51Idjpl58+YpPj4+5b7eeuutSjteZMIgIYQQ5ZJLTEIIIcolCUIIIUS5JEEIIYQolyQIIYQQ5ZIEIYQQolySIIQQQpRLEoQQQohySYIQQghRLqnmKsQ9evHFF9mzZ89N39+xYwfu7u5mjSE2NpZBgwaxbds2PD09zbotUf1IghDiPnTr1o333nuv3Pdq1KhRydEIUbEkQQhxH6ytrc1+liCEpcgYhBBm1KlTJ7788kteeeUVWrRoQZcuXVi5cmWpNgcPHmTQoEGEhIQQGhrKu+++W2re4aKiImbPnk2nTp0IDAykb9++/PHHH6X62L59Oz179qR58+b06NGD33//vTJ2TzzkJEEIYWZffPEFAQEBrF27lgEDBvDBBx+wefNmAI4cOcKLL76It7c3y5cvZ9asWRw5coRhw4ZhMpkAmDRpEt9++y1jx45lw4YNdOjQgVdffZUzZ86UbOPrr79m/PjxbNiwgYYNGzJ69GhycnIssr/i4SGXmIS4Dxs2bOCnn34qs7xjx44l00M++uijvPbaawA0btyYw4cPs3TpUrp3787ixYvx9fXlgw8+AMDb25vPP/+cXr168eeffxISEsKKFSsYP3483bt3B2DUqFGYTKZSCeDdd98lNDQUgJEjR/LLL79w9uxZWrRoYdb9Fw83SRBC3IdOnToxduzYMsv/Pg/wjS/uG1q2bFlyCejUqVM8+uijpd739fXF0dGR+Ph4XF1dKSwspGXLlqXaREZGAv+dNaxRo0Yl7zk6OgKUmnVNiHshCUKI+2BnZ0eDBg1u2UarLf3PTFEU1Gp1yX+rVKoy65hMJnQ6HTqd7o7iuNHf/25HiPshYxBCmNnRo0dL/f+BAwfw9/cHwMfHh3379pV6Py4ujuzsbLy8vGjQoAE6na5MH8888wyLFi0yb+Ci2pMzCCHuQ35+PmlpaeW+d+NSz6ZNm2jRogVt27bll19+YevWrXz55ZcADBkyhAEDBvDRRx8xYMAAMjIy+Oijj/D39yc8PBydTsfAgQOZOXMmrq6uNGnShNWrV3PmzBk6duxIenp6pe2rqH4kQQhxH7Zs2cKWLVvKfe/GIHXv3r35+eefmTJlCg0bNmTGjBl06NABgKCgIBYuXMjMmTPp3bs39vb2PP7447zxxhsll5fGjBmDVqvlww8/JDMzE19fXxYsWICXl5ckCGFWMie1EGbUqVMn+vTpQ0REhKVDEeKuyRiEEEKIckmCEEIIUS65xCSEEKJccgYhhBCiXJIghBBClEsShBBCiHJJghBCCFEuSRBCCCHKJQlCCCFEuSRBCCGEKJckCCGEEOX6P1zmYw9xupH/AAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 400x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "None\n"
     ]
    }
   ],
   "source": [
    "accuracy_cv, loss_cv, history_cv, ss_pred_cv, ss_true_cv = model_cross_validation(profile_x_list, seq_x_list, ss_y_list)\n",
    "mean_accuracy, std_accuracy, mean_loss, std_loss = calculate_mean_std(accuracy_cv, loss_cv)\n",
    "print(plot_accuracy_epoch(history_cv))\n",
    "print(plot_loss_epoch(history_cv))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "979862d4-aeb5-4e69-ba71-0a6793da391f",
   "metadata": {
    "tags": []
   },
   "source": [
    "### Ensemble Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "5c266021",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [],
   "source": [
    "def develop_ensemble_model(\n",
    "    profile_x, \n",
    "    ss_y, \n",
    "    seq_x, \n",
    "    profile_x_test, \n",
    "    ss_y_test, \n",
    "    seq_x_test, \n",
    "    accuracy_cv, \n",
    "    loss_cv, \n",
    "    history_cv, \n",
    "    ss_pred_cv\n",
    "):\n",
    "    \"\"\"\n",
    "    Develops an ensemble model for secondary structure prediction.\n",
    "\n",
    "    Args:\n",
    "        profile_x (numpy.ndarray): Training profile features.\n",
    "        ss_y (numpy.ndarray): Training secondary structure labels.\n",
    "        seq_x (numpy.ndarray): Training sequence features.\n",
    "        profile_x_test (numpy.ndarray): Test profile features.\n",
    "        ss_y_test (numpy.ndarray): Test secondary structure labels.\n",
    "        seq_x_test (numpy.ndarray): Test sequence features.\n",
    "        accuracy_cv (list): List to store accuracy values from cross-validation sets.\n",
    "        loss_cv (list): List to store loss values from cross-validation sets.\n",
    "        history_cv (list): List to store Keras History objects from cross-validation sets.\n",
    "        ss_pred_cv (list): List to store predicted secondary structures from cross-validation sets.\n",
    "\n",
    "    Returns:\n",
    "        tuple: A tuple containing updated accuracy_cv, loss_cv, history_cv, and ss_pred_cv lists.\n",
    "\n",
    "    \"\"\"\n",
    "    inputA = layers.Input(shape=(340))\n",
    "    inputB = layers.Input(shape=(340))\n",
    "    \n",
    "    # First branch operates on the first input\n",
    "    x = layers.Dense(1024, activation='leaky_relu')(inputA)\n",
    "    x = layers.Dropout(0.5)(x)\n",
    "    x = layers.Dense(512, activation='relu')(x)\n",
    "    x = layers.Dropout(0.5)(x)\n",
    "    x = layers.Dense(256, activation='relu')(x)\n",
    "    x = layers.Dropout(0.5)(x)\n",
    "    x = layers.Dense(128, activation=\"relu\")(x)\n",
    "    x = layers.Dropout(0.5)(x)\n",
    "    x = Model(inputs=inputA, outputs=x)\n",
    "    \n",
    "    # Second branch operates on the second input\n",
    "    y = Model(inputs=inputB, outputs=inputB)\n",
    "    \n",
    "    combined = layers.concatenate([x.output, y.output])  # combine the output of the two branches\n",
    "    z = layers.Dense(64, activation=\"relu\")(combined)\n",
    "    z = layers.Dropout(0.5)(z)\n",
    "    z = layers.Dense(3, activation='softmax')(z)\n",
    "    model = Model(inputs=[x.input, y.input], outputs=z)\n",
    "\n",
    "    model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "        \n",
    "    history = model.fit([profile_x, seq_x], ss_y, epochs=10, batch_size=128)\n",
    "    history_cv.append(history)\n",
    "        \n",
    "    test_loss, test_accuracy = model.evaluate([profile_x_test, seq_x_test], ss_y_test)\n",
    "    accuracy_cv.append(test_accuracy)\n",
    "    loss_cv.append(test_loss)\n",
    "        \n",
    "    ss_y_pred = model.predict([profile_x_test, seq_x_test])\n",
    "    ss_pred_cv.append(ss_y_pred)\n",
    "    \n",
    "    return accuracy_cv, loss_cv, history_cv, ss_pred_cv\n",
    "\n",
    "\n",
    "def ensemble_model_cross_validation(profile_x_list, seq_x_list, ss_y_list):\n",
    "    \"\"\"\n",
    "    Performs cross-validation for an ensemble model for secondary structure prediction.\n",
    "\n",
    "    Args:\n",
    "        profile_x_list (list): A list of lists containing profile features for each cross-validation set.\n",
    "        seq_x_list (list): A list of lists containing sequence features for each cross-validation set.\n",
    "        ss_y_list (list): A list of lists containing secondary structure labels for each cross-validation set.\n",
    "\n",
    "    Returns:\n",
    "        tuple: A tuple containing accuracy_cv, loss_cv, history_cv, ss_pred_cv, and ss_true_cv lists.\n",
    "\n",
    "    \"\"\"\n",
    "    accuracy_cv = []\n",
    "    loss_cv = []\n",
    "    history_cv = []\n",
    "    ss_pred_cv = []\n",
    "    ss_true_cv = []\n",
    "    \n",
    "    for s in range(len(profile_x_list)):       \n",
    "        print(f'Train/test cross-validation set {s+1}.\\n-------------------------------')\n",
    "        profile_x = profile_x_list[s][0]\n",
    "        seq_x = seq_x_list[s][0]\n",
    "        ss_y = ss_y_list[s][0]\n",
    "        profile_x_test = profile_x_list[s][1]\n",
    "        seq_x_test = seq_x_list[s][1]\n",
    "        ss_y_test = ss_y_list[s][1]\n",
    "        ss_true_cv.append(ss_y_test)\n",
    "        \n",
    "        accuracy_cv, loss_cv, history_cv, ss_pred_cv = develop_ensemble_model(\n",
    "            profile_x, \n",
    "            ss_y, \n",
    "            seq_x, \n",
    "            profile_x_test, \n",
    "            ss_y_test, \n",
    "            seq_x_test, \n",
    "            accuracy_cv, \n",
    "            loss_cv, \n",
    "            history_cv, \n",
    "            ss_pred_cv\n",
    "        )\n",
    "        \n",
    "    return accuracy_cv, loss_cv, history_cv, ss_pred_cv, ss_true_cv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "2f876d75-5734-45be-979d-a9ce07199e8c",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train/test cross-validation set 1.\n",
      "-------------------------------\n",
      "WARNING:tensorflow:From C:\\Users\\kanto\\anaconda3\\Lib\\site-packages\\keras\\src\\backend.py:1398: The name tf.executing_eagerly_outside_functions is deprecated. Please use tf.compat.v1.executing_eagerly_outside_functions instead.\n",
      "\n",
      "WARNING:tensorflow:From C:\\Users\\kanto\\anaconda3\\Lib\\site-packages\\keras\\src\\optimizers\\__init__.py:309: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
      "\n",
      "Epoch 1/10\n",
      "WARNING:tensorflow:From C:\\Users\\kanto\\anaconda3\\Lib\\site-packages\\keras\\src\\utils\\tf_utils.py:492: The name tf.ragged.RaggedTensorValue is deprecated. Please use tf.compat.v1.ragged.RaggedTensorValue instead.\n",
      "\n",
      "WARNING:tensorflow:From C:\\Users\\kanto\\anaconda3\\Lib\\site-packages\\keras\\src\\engine\\base_layer_utils.py:384: The name tf.executing_eagerly_outside_functions is deprecated. Please use tf.compat.v1.executing_eagerly_outside_functions instead.\n",
      "\n",
      "1230/1230 [==============================] - 25s 19ms/step - loss: 0.7650 - accuracy: 0.6755\n",
      "Epoch 2/10\n",
      "1230/1230 [==============================] - 25s 20ms/step - loss: 0.6539 - accuracy: 0.7341\n",
      "Epoch 3/10\n",
      "1230/1230 [==============================] - 24s 19ms/step - loss: 0.6232 - accuracy: 0.7470\n",
      "Epoch 4/10\n",
      "1230/1230 [==============================] - 24s 20ms/step - loss: 0.6034 - accuracy: 0.7566\n",
      "Epoch 5/10\n",
      "1230/1230 [==============================] - 24s 19ms/step - loss: 0.5909 - accuracy: 0.7615\n",
      "Epoch 6/10\n",
      "1230/1230 [==============================] - 25s 21ms/step - loss: 0.5822 - accuracy: 0.7641\n",
      "Epoch 7/10\n",
      "1230/1230 [==============================] - 25s 20ms/step - loss: 0.5745 - accuracy: 0.7680\n",
      "Epoch 8/10\n",
      "1230/1230 [==============================] - 24s 20ms/step - loss: 0.5673 - accuracy: 0.7724\n",
      "Epoch 9/10\n",
      "1230/1230 [==============================] - 26s 21ms/step - loss: 0.5605 - accuracy: 0.7738\n",
      "Epoch 10/10\n",
      "1230/1230 [==============================] - 24s 20ms/step - loss: 0.5558 - accuracy: 0.7763\n",
      "1273/1273 [==============================] - 6s 5ms/step - loss: 0.5617 - accuracy: 0.7705\n",
      "1273/1273 [==============================] - 6s 4ms/step\n",
      "Train/test cross-validation set 2.\n",
      "-------------------------------\n",
      "Epoch 1/10\n",
      "1241/1241 [==============================] - 28s 21ms/step - loss: 0.7664 - accuracy: 0.6764\n",
      "Epoch 2/10\n",
      "1241/1241 [==============================] - 24s 19ms/step - loss: 0.6559 - accuracy: 0.7345\n",
      "Epoch 3/10\n",
      "1241/1241 [==============================] - 25s 20ms/step - loss: 0.6263 - accuracy: 0.7468\n",
      "Epoch 4/10\n",
      "1241/1241 [==============================] - 26s 21ms/step - loss: 0.6067 - accuracy: 0.7557\n",
      "Epoch 5/10\n",
      "1241/1241 [==============================] - 25s 20ms/step - loss: 0.5932 - accuracy: 0.7605\n",
      "Epoch 6/10\n",
      "1241/1241 [==============================] - 27s 22ms/step - loss: 0.5845 - accuracy: 0.7655\n",
      "Epoch 7/10\n",
      "1241/1241 [==============================] - 26s 21ms/step - loss: 0.5757 - accuracy: 0.7682\n",
      "Epoch 8/10\n",
      "1241/1241 [==============================] - 25s 20ms/step - loss: 0.5668 - accuracy: 0.7718\n",
      "Epoch 9/10\n",
      "1241/1241 [==============================] - 26s 21ms/step - loss: 0.5625 - accuracy: 0.7730\n",
      "Epoch 10/10\n",
      "1241/1241 [==============================] - 23s 19ms/step - loss: 0.5588 - accuracy: 0.7748\n",
      "1229/1229 [==============================] - 6s 5ms/step - loss: 0.5484 - accuracy: 0.7730\n",
      "1229/1229 [==============================] - 5s 4ms/step\n",
      "Train/test cross-validation set 3.\n",
      "-------------------------------\n",
      "Epoch 1/10\n",
      "1238/1238 [==============================] - 28s 21ms/step - loss: 0.7631 - accuracy: 0.6778\n",
      "Epoch 2/10\n",
      "1238/1238 [==============================] - 22s 18ms/step - loss: 0.6505 - accuracy: 0.7361\n",
      "Epoch 3/10\n",
      "1238/1238 [==============================] - 25s 20ms/step - loss: 0.6196 - accuracy: 0.7504\n",
      "Epoch 4/10\n",
      "1238/1238 [==============================] - 22s 18ms/step - loss: 0.5995 - accuracy: 0.7585\n",
      "Epoch 5/10\n",
      "1238/1238 [==============================] - 25s 20ms/step - loss: 0.5885 - accuracy: 0.7634\n",
      "Epoch 6/10\n",
      "1238/1238 [==============================] - 23s 18ms/step - loss: 0.5768 - accuracy: 0.7686\n",
      "Epoch 7/10\n",
      "1238/1238 [==============================] - 24s 19ms/step - loss: 0.5693 - accuracy: 0.7715\n",
      "Epoch 8/10\n",
      "1238/1238 [==============================] - 23s 19ms/step - loss: 0.5614 - accuracy: 0.7740\n",
      "Epoch 9/10\n",
      "1238/1238 [==============================] - 25s 20ms/step - loss: 0.5572 - accuracy: 0.7771\n",
      "Epoch 10/10\n",
      "1238/1238 [==============================] - 24s 19ms/step - loss: 0.5506 - accuracy: 0.7794\n",
      "1241/1241 [==============================] - 6s 4ms/step - loss: 0.5766 - accuracy: 0.7614\n",
      "1241/1241 [==============================] - 5s 4ms/step\n",
      "Train/test cross-validation set 4.\n",
      "-------------------------------\n",
      "Epoch 1/10\n",
      "1245/1245 [==============================] - 25s 19ms/step - loss: 0.7641 - accuracy: 0.6777\n",
      "Epoch 2/10\n",
      "1245/1245 [==============================] - 24s 20ms/step - loss: 0.6556 - accuracy: 0.7346\n",
      "Epoch 3/10\n",
      "1245/1245 [==============================] - 25s 20ms/step - loss: 0.6233 - accuracy: 0.7478\n",
      "Epoch 4/10\n",
      "1245/1245 [==============================] - 24s 20ms/step - loss: 0.6038 - accuracy: 0.7557\n",
      "Epoch 5/10\n",
      "1245/1245 [==============================] - 24s 19ms/step - loss: 0.5931 - accuracy: 0.7612\n",
      "Epoch 6/10\n",
      "1245/1245 [==============================] - 23s 19ms/step - loss: 0.5806 - accuracy: 0.7659\n",
      "Epoch 7/10\n",
      "1245/1245 [==============================] - 23s 18ms/step - loss: 0.5737 - accuracy: 0.7686\n",
      "Epoch 8/10\n",
      "1245/1245 [==============================] - 25s 20ms/step - loss: 0.5655 - accuracy: 0.7723\n",
      "Epoch 9/10\n",
      "1245/1245 [==============================] - 21s 17ms/step - loss: 0.5611 - accuracy: 0.7746\n",
      "Epoch 10/10\n",
      "1245/1245 [==============================] - 24s 19ms/step - loss: 0.5542 - accuracy: 0.7771\n",
      "1213/1213 [==============================] - 5s 4ms/step - loss: 0.5581 - accuracy: 0.7714\n",
      "1213/1213 [==============================] - 5s 4ms/step\n",
      "Train/test cross-validation set 5.\n",
      "-------------------------------\n",
      "Epoch 1/10\n",
      "1239/1239 [==============================] - 25s 19ms/step - loss: 0.7650 - accuracy: 0.6747\n",
      "Epoch 2/10\n",
      "1239/1239 [==============================] - 23s 19ms/step - loss: 0.6551 - accuracy: 0.7331\n",
      "Epoch 3/10\n",
      "1239/1239 [==============================] - 23s 19ms/step - loss: 0.6201 - accuracy: 0.7487\n",
      "Epoch 4/10\n",
      "1239/1239 [==============================] - 21s 17ms/step - loss: 0.6039 - accuracy: 0.7556\n",
      "Epoch 5/10\n",
      "1239/1239 [==============================] - 24s 20ms/step - loss: 0.5914 - accuracy: 0.7613\n",
      "Epoch 6/10\n",
      "1239/1239 [==============================] - 22s 18ms/step - loss: 0.5822 - accuracy: 0.7644\n",
      "Epoch 7/10\n",
      "1239/1239 [==============================] - 23s 19ms/step - loss: 0.5740 - accuracy: 0.7681\n",
      "Epoch 8/10\n",
      "1239/1239 [==============================] - 22s 18ms/step - loss: 0.5664 - accuracy: 0.7712\n",
      "Epoch 9/10\n",
      "1239/1239 [==============================] - 22s 17ms/step - loss: 0.5615 - accuracy: 0.7749\n",
      "Epoch 10/10\n",
      "1239/1239 [==============================] - 22s 18ms/step - loss: 0.5553 - accuracy: 0.7754\n",
      "1239/1239 [==============================] - 5s 4ms/step - loss: 0.5621 - accuracy: 0.7708\n",
      "1239/1239 [==============================] - 5s 4ms/step\n",
      "Mean Accuracy:     0.7694002270698548\n",
      "Std Dev Accuracy:  0.004094429964340865\n",
      "-----------------------------\n",
      "Mean Loss:         0.5613723397254944\n",
      "Std Dev Loss:      0.009072772490825495\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "None\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "None\n"
     ]
    }
   ],
   "source": [
    "accuracy_cv, loss_cv, history_cv, ss_pred_cv, ss_true_cv = ensemble_model_cross_validation(profile_x_list, seq_x_list, ss_y_list)\n",
    "mean_accuracy, std_accuracy, mean_loss, std_loss = calculate_mean_std(accuracy_cv, loss_cv)\n",
    "print(plot_accuracy_epoch(history_cv))\n",
    "print(plot_loss_epoch(history_cv))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4b5b52ec-8c31-4ab1-a148-912928150b9a",
   "metadata": {},
   "source": [
    "### Hyperparameter Optimization with Bayesian Tuner"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d9417f6e-1ebd-4e51-ad50-ce5e2f200ded",
   "metadata": {},
   "source": [
    "Optimize selected model hyperparameters with the Bayesian optimization method."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4ec9ba19-09da-4844-a745-26618bfb991f",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def build_model(hp):\n",
    "    inputA = layers.Input(shape=(340))\n",
    "    inputB = layers.Input(shape=(340))\n",
    "    \n",
    "    # First branch operates on the first input\n",
    "    x = layers.Dense(\n",
    "        hp.Int('units_hp_4', min_value=1024, max_value=1536, step=256),\n",
    "        activation=hp.Choice('act_fun_4', values=['relu','leaky_relu'])\n",
    "    )(inputA)\n",
    "    x = layers.Dropout(hp.Float('dropout_4', min_value=0.4, max_value=0.5, step=0.1))(x)\n",
    "    x = layers.Dense(hp.Int('units_hp_3', min_value=512, max_value=1024, step=128), activation='relu')(x)\n",
    "    x = layers.Dropout(hp.Float('dropout_3', min_value=0.4, max_value=0.5, step=0.1))(x)\n",
    "    x = layers.Dense(hp.Int('units_hp_2', min_value=256, max_value=512, step=64), activation='relu')(x)\n",
    "    x = layers.Dropout(hp.Float('dropout_2', min_value=0.3, max_value=0.5, step=0.1))(x)\n",
    "    x = layers.Dense(hp.Int('units_hp_1', min_value=128, max_value=256, step=32), activation='relu')(x)\n",
    "    x = layers.Dropout(hp.Float('dropout_1', min_value=0.3, max_value=0.5, step=0.1))(x)\n",
    "    x = Model(inputs=inputA, outputs=x)\n",
    "    \n",
    "    # Second branch operates on the second input\n",
    "    y = Model(inputs=inputB, outputs=inputB)\n",
    "    \n",
    "    combined = layers.concatenate([x.output, y.output])  # combine the output of the two branches\n",
    "    z = layers.Dense(hp.Int('units_hp_0', min_value=32, max_value=128, step=32), activation='relu')(combined)\n",
    "    z = layers.Dropout(hp.Float('dropout_0', min_value=0.3, max_value=0.5, step=0.1))(z)\n",
    "    z = layers.Dense(3, activation='softmax')(z)\n",
    "    model = Model(inputs=[x.input, y.input], outputs=z)\n",
    "\n",
    "    model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "        \n",
    "    return model\n",
    "    \n",
    "tuner = BayesianOptimization(build_model, objective = 'val_accuracy', max_trials = 50, overwrite = True)\n",
    "tuner.search(\n",
    "    [profile_x_list[0][0], seq_x_list[0][0]], \n",
    "    ss_y_list[0][0], \n",
    "    epochs = 10, \n",
    "    validation_data = ([profile_x_list[0][1], seq_x_list[0][1]], ss_y_list[0][1])\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "852d8fe6-13af-47a7-88b4-bad2df70d278",
   "metadata": {},
   "source": [
    "### Ablation Study"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7cd39621-fcfd-43a2-bfcb-3ebdbd1984fd",
   "metadata": {},
   "source": [
    "Visualization of the results of an ablation study conducted on deep learning models. The purpose of this visualization is to show the impact of individual components or changes in a model by systematically adding or altering them (from bottom to top). The study investigates various configurations and their corresponding validation mean accuracies. All data came from alternatives of model with PSSM input or ensemble model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "96787dea-1088-4797-a8f0-ef1ef99b3fad",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "labels = [\n",
    "    'Optimized Ensemble Model', \n",
    "    'Leaky-Relu instead of\\nReLu in 1st hidden layer',\n",
    "    '5 hidden layers,\\ndropout layers',\n",
    "    '5 hidden layers,\\nactivity regularization layer',\n",
    "    '5 hidden layers\\n(no regularization)',\n",
    "    'ReLu instead of tanh',\n",
    "    'Adam instead of SGD',\n",
    "    'SGD, tanh activation function,\\n1 hidden layer'\n",
    "]\n",
    "data = {\n",
    "    'Label': [label for label in labels for _ in range(5)],\n",
    "    'Values': [\n",
    "        0.7705, 0.7730, 0.7614, 0.7714, 0.7708, \n",
    "        0.7690, 0.7680, 0.7567, 0.7656, 0.7704, \n",
    "        0.7677, 0.7651, 0.7535, 0.7623, 0.7656,\n",
    "        0.7612, 0.7631, 0.7514, 0.7581, 0.7606,\n",
    "        0.7377, 0.7365, 0.7301, 0.7392, 0.7454,\n",
    "        0.7569, 0.7570, 0.7505, 0.7583, 0.7580,\n",
    "        0.7438, 0.7413, 0.7335, 0.7500, 0.7458,\n",
    "        0.7383, 0.7397, 0.7317, 0.7481, 0.7449\n",
    "    ]\n",
    "}\n",
    "sens_analysis_data = pd.DataFrame(data)\n",
    "\n",
    "sns.set(style='ticks')\n",
    "palette = sns.color_palette('deep')\n",
    "plt.figure(figsize=(8, 6))\n",
    "ax = sns.boxplot(x='Values', y='Label', data=sens_analysis_data, orient=\"h\", showfliers=False)\n",
    "for patch in ax.patches:\n",
    "    r, g, b, a = patch.get_facecolor()\n",
    "    patch.set_facecolor((r, g, b, .5))\n",
    "sns.stripplot(x='Values', y='Label', hue='Label', data=sens_analysis_data, orient='h', size=6)\n",
    "\n",
    "plt.xlim(0.7, 0.8)\n",
    "plt.xlabel('Validation Accuracy', fontname='Arial', fontsize=11)\n",
    "plt.ylabel('',  fontname='Arial', fontsize=11)\n",
    "plt.legend([],[], frameon=False)\n",
    "plt.savefig('Results/ablation_study.png', dpi=300, bbox_inches='tight')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d2115140-70f8-4265-95ec-e8abce4011d8",
   "metadata": {},
   "source": [
    "# Final Model Testing"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d6863684-602b-40ba-8944-4aab2bbbcc0a",
   "metadata": {},
   "source": [
    "Train with cv train sets and test with blind test sets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "240a910d-b413-407d-a0b2-e60badddb3e3",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def ensemble_model_blind_test(profile_x_list, seq_x_list, ss_y_list, profile_x_blindtest, seq_x_blindtest, ss_y_blindtest):\n",
    "    accuracy_cv = []\n",
    "    loss_cv = []\n",
    "    history_cv = []\n",
    "    ss_pred_cv = []\n",
    "    ss_true_cv = []\n",
    "    \n",
    "    for s in range(len(profile_x_list)):       \n",
    "        print(f'Train/test cross-validation set {s+1}.\\n-------------------------------')\n",
    "        profile_x = profile_x_list[s][0]\n",
    "        seq_x = seq_x_list[s][0]\n",
    "        ss_y = ss_y_list[s][0]\n",
    "        profile_x_test = profile_x_blindtest\n",
    "        seq_x_test = seq_x_blindtest\n",
    "        ss_y_test = ss_y_blindtest\n",
    "        ss_true_cv.append(ss_y_test)\n",
    "        \n",
    "        accuracy_cv, loss_cv, history_cv, ss_pred_cv = develop_ensemble_model(\n",
    "            profile_x, \n",
    "            ss_y, \n",
    "            seq_x, \n",
    "            profile_x_test, \n",
    "            ss_y_test, \n",
    "            seq_x_test, \n",
    "            accuracy_cv, \n",
    "            loss_cv, \n",
    "            history_cv, \n",
    "            ss_pred_cv\n",
    "        )\n",
    "        \n",
    "    return accuracy_cv, loss_cv, history_cv, ss_pred_cv, ss_true_cv"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fd0bbed6-8bc7-4f86-bd08-482df6633e2c",
   "metadata": {},
   "source": [
    "Use optimized model on the blind test set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "1f78b2c0-99bc-4336-be44-4aebe0b792ac",
   "metadata": {
    "scrolled": true,
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train/test cross-validation set 1.\n",
      "-------------------------------\n",
      "Epoch 1/10\n",
      "1230/1230 [==============================] - 20s 15ms/step - loss: 0.7648 - accuracy: 0.6772\n",
      "Epoch 2/10\n",
      "1230/1230 [==============================] - 20s 16ms/step - loss: 0.6558 - accuracy: 0.7335\n",
      "Epoch 3/10\n",
      "1230/1230 [==============================] - 19s 15ms/step - loss: 0.6230 - accuracy: 0.7487\n",
      "Epoch 4/10\n",
      "1230/1230 [==============================] - 18s 14ms/step - loss: 0.6050 - accuracy: 0.7566\n",
      "Epoch 5/10\n",
      "1230/1230 [==============================] - 17s 14ms/step - loss: 0.5921 - accuracy: 0.7613\n",
      "Epoch 6/10\n",
      "1230/1230 [==============================] - 18s 14ms/step - loss: 0.5840 - accuracy: 0.7655\n",
      "Epoch 7/10\n",
      "1230/1230 [==============================] - 17s 14ms/step - loss: 0.5755 - accuracy: 0.7677\n",
      "Epoch 8/10\n",
      "1230/1230 [==============================] - 17s 14ms/step - loss: 0.5685 - accuracy: 0.7715\n",
      "Epoch 9/10\n",
      "1230/1230 [==============================] - 18s 15ms/step - loss: 0.5637 - accuracy: 0.7740\n",
      "Epoch 10/10\n",
      "1230/1230 [==============================] - 18s 14ms/step - loss: 0.5556 - accuracy: 0.7761\n",
      "1942/1942 [==============================] - 6s 3ms/step - loss: 0.6382 - accuracy: 0.7486\n",
      "1942/1942 [==============================] - 7s 3ms/step\n",
      "Train/test cross-validation set 2.\n",
      "-------------------------------\n",
      "Epoch 1/10\n",
      "1241/1241 [==============================] - 19s 14ms/step - loss: 0.7648 - accuracy: 0.6788\n",
      "Epoch 2/10\n",
      "1241/1241 [==============================] - 18s 14ms/step - loss: 0.6553 - accuracy: 0.7342\n",
      "Epoch 3/10\n",
      "1241/1241 [==============================] - 19s 15ms/step - loss: 0.6252 - accuracy: 0.7469\n",
      "Epoch 4/10\n",
      "1241/1241 [==============================] - 22s 17ms/step - loss: 0.6064 - accuracy: 0.7560\n",
      "Epoch 5/10\n",
      "1241/1241 [==============================] - 22s 18ms/step - loss: 0.5955 - accuracy: 0.7596\n",
      "Epoch 6/10\n",
      "1241/1241 [==============================] - 22s 18ms/step - loss: 0.5849 - accuracy: 0.7640\n",
      "Epoch 7/10\n",
      "1241/1241 [==============================] - 23s 18ms/step - loss: 0.5760 - accuracy: 0.7672\n",
      "Epoch 8/10\n",
      "1241/1241 [==============================] - 18s 15ms/step - loss: 0.5693 - accuracy: 0.7705\n",
      "Epoch 9/10\n",
      "1241/1241 [==============================] - 17s 14ms/step - loss: 0.5623 - accuracy: 0.7739\n",
      "Epoch 10/10\n",
      "1241/1241 [==============================] - 18s 14ms/step - loss: 0.5589 - accuracy: 0.7747\n",
      "1942/1942 [==============================] - 7s 3ms/step - loss: 0.6028 - accuracy: 0.7543\n",
      "1942/1942 [==============================] - 6s 3ms/step\n",
      "Train/test cross-validation set 3.\n",
      "-------------------------------\n",
      "Epoch 1/10\n",
      "1238/1238 [==============================] - 18s 14ms/step - loss: 0.7598 - accuracy: 0.6800\n",
      "Epoch 2/10\n",
      "1238/1238 [==============================] - 18s 14ms/step - loss: 0.6487 - accuracy: 0.7379\n",
      "Epoch 3/10\n",
      "1238/1238 [==============================] - 18s 15ms/step - loss: 0.6169 - accuracy: 0.7503\n",
      "Epoch 4/10\n",
      "1238/1238 [==============================] - 16s 13ms/step - loss: 0.5986 - accuracy: 0.7587\n",
      "Epoch 5/10\n",
      "1238/1238 [==============================] - 17s 14ms/step - loss: 0.5877 - accuracy: 0.7641\n",
      "Epoch 6/10\n",
      "1238/1238 [==============================] - 17s 14ms/step - loss: 0.5764 - accuracy: 0.7689\n",
      "Epoch 7/10\n",
      "1238/1238 [==============================] - 17s 14ms/step - loss: 0.5684 - accuracy: 0.7709\n",
      "Epoch 8/10\n",
      "1238/1238 [==============================] - 17s 14ms/step - loss: 0.5619 - accuracy: 0.7735\n",
      "Epoch 9/10\n",
      "1238/1238 [==============================] - 17s 14ms/step - loss: 0.5552 - accuracy: 0.7773\n",
      "Epoch 10/10\n",
      "1238/1238 [==============================] - 18s 15ms/step - loss: 0.5494 - accuracy: 0.7792\n",
      "1942/1942 [==============================] - 7s 4ms/step - loss: 0.6059 - accuracy: 0.7559\n",
      "1942/1942 [==============================] - 6s 3ms/step\n",
      "Train/test cross-validation set 4.\n",
      "-------------------------------\n",
      "Epoch 1/10\n",
      "1245/1245 [==============================] - 18s 13ms/step - loss: 0.7615 - accuracy: 0.6782\n",
      "Epoch 2/10\n",
      "1245/1245 [==============================] - 20s 16ms/step - loss: 0.6545 - accuracy: 0.7352\n",
      "Epoch 3/10\n",
      "1245/1245 [==============================] - 17s 14ms/step - loss: 0.6230 - accuracy: 0.7489\n",
      "Epoch 4/10\n",
      "1245/1245 [==============================] - 18s 14ms/step - loss: 0.6041 - accuracy: 0.7561\n",
      "Epoch 5/10\n",
      "1245/1245 [==============================] - 18s 14ms/step - loss: 0.5933 - accuracy: 0.7612\n",
      "Epoch 6/10\n",
      "1245/1245 [==============================] - 18s 14ms/step - loss: 0.5806 - accuracy: 0.7661\n",
      "Epoch 7/10\n",
      "1245/1245 [==============================] - 20s 16ms/step - loss: 0.5748 - accuracy: 0.7683\n",
      "Epoch 8/10\n",
      "1245/1245 [==============================] - 22s 18ms/step - loss: 0.5662 - accuracy: 0.7715\n",
      "Epoch 9/10\n",
      "1245/1245 [==============================] - 16s 13ms/step - loss: 0.5614 - accuracy: 0.7744\n",
      "Epoch 10/10\n",
      "1245/1245 [==============================] - 19s 15ms/step - loss: 0.5545 - accuracy: 0.7774\n",
      "1942/1942 [==============================] - 7s 4ms/step - loss: 0.6048 - accuracy: 0.7575\n",
      "1942/1942 [==============================] - 6s 3ms/step\n",
      "Train/test cross-validation set 5.\n",
      "-------------------------------\n",
      "Epoch 1/10\n",
      "1239/1239 [==============================] - 19s 14ms/step - loss: 0.7648 - accuracy: 0.6782\n",
      "Epoch 2/10\n",
      "1239/1239 [==============================] - 17s 14ms/step - loss: 0.6540 - accuracy: 0.7349\n",
      "Epoch 3/10\n",
      "1239/1239 [==============================] - 17s 14ms/step - loss: 0.6219 - accuracy: 0.7474\n",
      "Epoch 4/10\n",
      "1239/1239 [==============================] - 18s 15ms/step - loss: 0.6047 - accuracy: 0.7558\n",
      "Epoch 5/10\n",
      "1239/1239 [==============================] - 18s 15ms/step - loss: 0.5909 - accuracy: 0.7606\n",
      "Epoch 6/10\n",
      "1239/1239 [==============================] - 18s 15ms/step - loss: 0.5819 - accuracy: 0.7647\n",
      "Epoch 7/10\n",
      "1239/1239 [==============================] - 20s 16ms/step - loss: 0.5753 - accuracy: 0.7671\n",
      "Epoch 8/10\n",
      "1239/1239 [==============================] - 18s 15ms/step - loss: 0.5668 - accuracy: 0.7708\n",
      "Epoch 9/10\n",
      "1239/1239 [==============================] - 18s 15ms/step - loss: 0.5618 - accuracy: 0.7727\n",
      "Epoch 10/10\n",
      "1239/1239 [==============================] - 20s 16ms/step - loss: 0.5558 - accuracy: 0.7763\n",
      "1942/1942 [==============================] - 7s 4ms/step - loss: 0.6098 - accuracy: 0.7526\n",
      "1942/1942 [==============================] - 7s 3ms/step\n",
      "Mean Accuracy:     0.7538013339042664\n",
      "Std Dev Accuracy:  0.003039067837588215\n",
      "-----------------------------\n",
      "Mean Loss:         0.6122960567474365\n",
      "Std Dev Loss:      0.013167133550483743\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "None\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "None\n"
     ]
    }
   ],
   "source": [
    "accuracy_test, loss_test, history_test, ss_pred_test, ss_true_test = ensemble_model_blind_test(\n",
    "    profile_x_list, \n",
    "    seq_x_list, \n",
    "    ss_y_list, \n",
    "    profile_x_blindtest, \n",
    "    seq_x_blindtest, \n",
    "    ss_y_blindtest\n",
    ")\n",
    "mean_accuracy, std_accuracy, mean_loss, std_loss = calculate_mean_std(accuracy_test, loss_test)\n",
    "print(plot_accuracy_epoch(history_test))\n",
    "print(plot_loss_epoch(history_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0ea970d7-f2aa-4b93-95b2-8d0765ff1a0d",
   "metadata": {},
   "source": [
    "## Calculate Metrics for Final Model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4cd635f0-bbf2-4f95-b2d2-91e585557dba",
   "metadata": {},
   "source": [
    "Classes:\n",
    "- H-helix\n",
    "- E-sheet\n",
    "- C-coil\n",
    "\n",
    "Calculate confusion matrix, accuracy, sensitivity and specificity for each class using the \"one-vs-all\" approach."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5ebbd7be-a28b-4299-9166-2f1aab6e95fa",
   "metadata": {},
   "source": [
    "### Utility function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "b8cda5c3-b3c9-4906-8b0e-fa1c3e765351",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def make_one_hot(matrix):\n",
    "    max_indices = np.argmax(matrix, axis=-1)  # Find the index of the maximum value along the last axis\n",
    "    one_hot_matrix = np.zeros_like(matrix)  # Create a new array with the same shape as the original matrix\n",
    "    one_hot_matrix[np.arange(len(matrix)), max_indices] = 1  # Set the maximum value to 1\n",
    "    return one_hot_matrix\n",
    "\n",
    "\n",
    "def transform_output(sets, decode_dict):\n",
    "    decoded_sets = []\n",
    "    \n",
    "    for cv_set in sets:\n",
    "        decoded_set = []\n",
    "        cv_set = make_one_hot(cv_set)\n",
    "        \n",
    "        for row in cv_set:\n",
    "            index = np.argmax(row)\n",
    "            decoded_value = decode_dict[tuple(row)]\n",
    "            decoded_set.append(decoded_value)\n",
    "            \n",
    "        decoded_set = np.array(decoded_set)\n",
    "        decoded_sets.append(decoded_set)\n",
    "        \n",
    "    return decoded_sets\n",
    "\n",
    "\n",
    "def create_cm(ss_true_test, ss_pred_test, decode_dict):\n",
    "    ss_pred = transform_output(ss_pred_test, decode_dict)\n",
    "    ss_true = transform_output(ss_true_test, decode_dict)\n",
    "    cm_list = []\n",
    "    \n",
    "    for i in range(len(ss_true)):\n",
    "        cm = confusion_matrix(ss_true[i], ss_pred[i])\n",
    "        cm_list.append(cm)\n",
    "        \n",
    "    return cm_list, ss_true\n",
    "\n",
    "\n",
    "def calculate_class_metrics(confusion_matrix):\n",
    "    sensitivity = np.zeros(confusion_matrix.shape[0])\n",
    "    specificity = np.zeros(confusion_matrix.shape[0])\n",
    "    accuracy = np.zeros(confusion_matrix.shape[0])\n",
    "    \n",
    "    for i in range(confusion_matrix.shape[0]):\n",
    "        TP = confusion_matrix[i, i]\n",
    "        FN = confusion_matrix[i, :].sum() - TP\n",
    "        FP = np.sum(confusion_matrix[:, i]) - TP\n",
    "        TN = np.sum(confusion_matrix) - TP - FP - FN\n",
    "\n",
    "        sensitivity[i] = TP / (TP + FN)\n",
    "        specificity[i] = TN / (TN + FP)\n",
    "        accuracy[i] = (TP + TN) / (TP + FN + TN + FP)\n",
    "        \n",
    "    return sensitivity, specificity, accuracy\n",
    "\n",
    "\n",
    "def print_mean_std_metrics_table(\n",
    "    mean_accuracy, \n",
    "    std_accuracy, \n",
    "    mean_sensitivity, \n",
    "    std_sensitivity, \n",
    "    mean_specificity, \n",
    "    std_specificity\n",
    "):\n",
    "    print(f'-----------------------------\\nHelix:\\n-----------------------------')\n",
    "    print(f'Mean Accuracy:     {mean_accuracy[0]}')\n",
    "    print(f'Std Dev Accuracy:  {std_accuracy[0]}')\n",
    "    print('-----------------------------')\n",
    "    print(f'Mean Sensitivity:     {mean_sensitivity[0]}')\n",
    "    print(f'Std Dev Sensitivity:  {std_sensitivity[0]}')\n",
    "    print('-----------------------------')\n",
    "    print(f'Mean Specificity:     {mean_specificity[0]}')\n",
    "    print(f'Std Dev Specificity:  {std_specificity[0]}')\n",
    "    print(f'-----------------------------\\nSheet:\\n-----------------------------')\n",
    "    print(f'Mean Accuracy:     {mean_accuracy[1]}')\n",
    "    print(f'Std Dev Accuracy:  {std_accuracy[1]}')\n",
    "    print('-----------------------------')\n",
    "    print(f'Mean Sensitivity:     {mean_sensitivity[1]}')\n",
    "    print(f'Std Dev Sensitivity:  {std_sensitivity[1]}')\n",
    "    print('-----------------------------')\n",
    "    print(f'Mean Specificity:     {mean_specificity[1]}')\n",
    "    print(f'Std Dev Specificity:  {std_specificity[1]}')\n",
    "    print(f'-----------------------------\\nCoil:\\n-----------------------------')\n",
    "    print(f'Mean Accuracy:     {mean_accuracy[2]}')\n",
    "    print(f'Std Dev Accuracy:  {std_accuracy[2]}')\n",
    "    print('-----------------------------')\n",
    "    print(f'Mean Sensitivity:     {mean_sensitivity[2]}')\n",
    "    print(f'Std Dev Sensitivity:  {std_sensitivity[2]}')\n",
    "    print('-----------------------------')\n",
    "    print(f'Mean Specificity:     {mean_specificity[2]}')\n",
    "    print(f'Std Dev Specificity:  {std_specificity[2]}')\n",
    "\n",
    "\n",
    "def calculate_mean_std_metrics(sensitivity_list, specificity_list, accuracy_list):\n",
    "    accuracy_array = np.array(accuracy_list)\n",
    "    sensitivity_array = np.array(sensitivity_list)\n",
    "    specificity_array = np.array(specificity_list)\n",
    "    \n",
    "    mean_accuracy = np.mean(accuracy_array, axis=0)\n",
    "    std_accuracy = np.std(accuracy_array, axis=0)\n",
    "    mean_sensitivity = np.mean(sensitivity_array, axis=0)\n",
    "    std_sensitivity = np.std(sensitivity_array, axis=0)\n",
    "    mean_specificity = np.mean(specificity_array, axis=0)\n",
    "    std_specificity = np.std(specificity_array, axis=0)\n",
    "    \n",
    "    print_mean_std_metrics_table(\n",
    "    mean_accuracy, \n",
    "    std_accuracy, \n",
    "    mean_sensitivity, \n",
    "    std_sensitivity, \n",
    "    mean_specificity, \n",
    "    std_specificity\n",
    ")\n",
    "    \n",
    "    return mean_accuracy, std_accuracy, mean_sensitivity, std_sensitivity, mean_specificity, std_specificity"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "26f76f8d-2d15-41e1-a2ae-58abf0fa54b4",
   "metadata": {},
   "source": [
    "- Decode the data. Softmax output to one-hot-encoding to class (0, 1, 2).\n",
    "- Confusion matrix for each validation set.\n",
    "- Accuracy, sensitivity and specificity calculation for each class and each valdiation set.\n",
    "- Mean and standard deviation of accuracy, sensitivity and specificity calculation for each class."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "d0b0024f-69c5-4462-9095-cb026bf4fb40",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-----------------------------\n",
      "Helix:\n",
      "-----------------------------\n",
      "Mean Accuracy:     0.8511426548216814\n",
      "Std Dev Accuracy:  0.0013259748967228936\n",
      "-----------------------------\n",
      "Mean Sensitivity:     0.741452708298824\n",
      "Std Dev Sensitivity:  0.01467929955044216\n",
      "-----------------------------\n",
      "Mean Specificity:     0.9208697483415816\n",
      "Std Dev Specificity:  0.008887915170538203\n",
      "-----------------------------\n",
      "Sheet:\n",
      "-----------------------------\n",
      "Mean Accuracy:     0.8724346594566758\n",
      "Std Dev Accuracy:  0.0009079260276016023\n",
      "-----------------------------\n",
      "Mean Sensitivity:     0.6436036036036036\n",
      "Std Dev Sensitivity:  0.012066136910812151\n",
      "-----------------------------\n",
      "Mean Specificity:     0.9451591694767874\n",
      "Std Dev Specificity:  0.003814585104733233\n",
      "-----------------------------\n",
      "Coil:\n",
      "-----------------------------\n",
      "Mean Accuracy:     0.7831305523368096\n",
      "Std Dev Accuracy:  0.0035253218878091735\n",
      "-----------------------------\n",
      "Mean Sensitivity:     0.8373429552667044\n",
      "Std Dev Sensitivity:  0.013742031745546212\n",
      "-----------------------------\n",
      "Mean Specificity:     0.7512636393836404\n",
      "Std Dev Specificity:  0.013401984885840719\n"
     ]
    }
   ],
   "source": [
    "decode_dict = {\n",
    "    (1, 0, 0): 0,\n",
    "    (0, 1, 0): 1,\n",
    "    (0, 0, 1): 2\n",
    "}\n",
    "\n",
    "cm_list, ss_true = create_cm(ss_true_test, ss_pred_test, decode_dict)\n",
    "\n",
    "sensitivity_list = []\n",
    "specificity_list = []\n",
    "accuracy_list = []\n",
    "\n",
    "for cm in cm_list:\n",
    "    sensitivity, specificity, accuracy = calculate_class_metrics(cm)\n",
    "    sensitivity_list.append(sensitivity)\n",
    "    specificity_list.append(specificity)\n",
    "    accuracy_list.append(accuracy)\n",
    "    \n",
    "mean_accuracy, std_accuracy, mean_sensitivity, std_sensitivity, mean_specificity, std_specificity = calculate_mean_std_metrics(\n",
    "    sensitivity_list, \n",
    "    specificity_list, \n",
    "    accuracy_list\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "206aba9a-605b-4dcd-a112-e92fb669206c",
   "metadata": {},
   "source": [
    "### Plot results for each class"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c8c74517-4af0-411a-9c1e-6b5a9c7fb74b",
   "metadata": {},
   "source": [
    "Plot confusion matrix for the best set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "3cee15b8-7cd5-459b-9738-e0985c52aa16",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "class_labels = ['α-Helix', 'β-Sheet', 'Coil']\n",
    "sns.set(style='ticks')\n",
    "palette = sns.color_palette('deep')\n",
    "sns.set(font_scale = 1)\n",
    "sns.heatmap(cm_list[3], annot = True, fmt='d', cmap='Reds', xticklabels=class_labels, yticklabels=class_labels)\n",
    "plt.xlabel('Predicted Labels', fontname='Arial', fontsize=11)\n",
    "plt.ylabel('True Labels', fontname='Arial', fontsize=11)\n",
    "plt.title('')\n",
    "plt.savefig('Results/confmatrix_blindtest.png', dpi=300, bbox_inches='tight')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e8b5694a-7113-4380-896f-754c3f34a37d",
   "metadata": {},
   "source": [
    "Plot mean accuracy, sensitivity and specificity for each class."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "6f94ce20-21f9-4c8b-8483-95913ec9825b",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "class_labels = ['α-Helix', 'β-sheet', 'Coil']\n",
    "\n",
    "# Process data to plot\n",
    "metrics_data = {\n",
    "    'class': [],\n",
    "    'metric': [],\n",
    "    'value': []\n",
    "}\n",
    "\n",
    "for label in class_labels:\n",
    "    metrics_data['class'].extend([label for label in class_labels])\n",
    "metrics_data['metric'].extend(3*['Accuracy'])\n",
    "metrics_data['metric'].extend(3*['Sensitivity'])\n",
    "metrics_data['metric'].extend(3*['Specificity'])\n",
    "metrics_data['value'].extend(mean_accuracy)\n",
    "metrics_data['value'].extend(mean_sensitivity)\n",
    "metrics_data['value'].extend(mean_specificity)\n",
    "\n",
    "metrics_data = pd.DataFrame(metrics_data)\n",
    "\n",
    "sns.set(style='ticks')\n",
    "palette = sns.color_palette('deep')\n",
    "plt.figure(figsize=(4, 4))\n",
    "ax = sns.barplot(data=metrics_data, x='class', y='value', hue='metric', edgecolor='black', linewidth=1)\n",
    "for patch in ax.patches:\n",
    "    r, g, b, a = patch.get_facecolor()\n",
    "    patch.set_facecolor((r, g, b, 0.5))\n",
    "plt.xlabel('Class', fontname='Arial', fontsize=11)\n",
    "plt.ylabel('', fontname='Arial', fontsize=11)\n",
    "font = FontProperties()\n",
    "font.set_family('Arial')\n",
    "font.set_size('11')\n",
    "plt.legend(loc='center left', bbox_to_anchor=(1, 0.5), title='', prop=font)\n",
    "plt.savefig('Results/classmetrics_blindtest.png', dpi=300, bbox_inches='tight')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7c6211fa-1b2e-41bb-b0e7-c3753ed91d59",
   "metadata": {},
   "source": [
    "Plot percentage of individualized classes in blind test set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "id": "8d7e3a18-fdb0-4ba1-a89e-f01f03462d9a",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "counts = {}\n",
    "\n",
    "for val in ss_true_test[0]:\n",
    "    arr_tuple = tuple(val)\n",
    "    counts[arr_tuple] = counts.get(arr_tuple, 0) + 1\n",
    "\n",
    "class_labels = ['Coil', 'β-sheet', 'α-Helix']\n",
    "\n",
    "sns.set(style='ticks')\n",
    "palette = sns.color_palette('deep')\n",
    "plt.figure(figsize=(4, 4))\n",
    "plt.pie(counts.values(), labels=class_labels, autopct='%1.1f%%', startangle=140, wedgeprops={'alpha':0.6})\n",
    "plt.axis('equal')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "072b3b38-3767-4748-b68c-c1953ab3ccd9",
   "metadata": {},
   "source": [
    "# User Interface"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3ea2bf6e-0f9d-4e69-ab8e-2a7ec552a38e",
   "metadata": {},
   "source": [
    "The user interface is designed in a way that it allows the user to insert a single pssm file as input and get a prediction of the protein's secondary structure. Some functions are similar with the functions of preprocessing section, but have minor changes that allow the user to run a single sequence.\n",
    "\n",
    "To use it just import the packages (first code block in this notebook), then run the parsing data pipeline (second code block in this notebook), and then run the following code blocks. You do not have to run the whole notebook to test this.\n",
    "\n",
    "The model is trained on cross-validation set 3, as it performed best on the blind set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "de6a8a6f-5de5-4827-ba53-410ebf2a526b",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def create_user_profiles(profile, seq, window_size):\n",
    "    \"\"\"\n",
    "    Create user profiles from PSSM profiles and amino acid sequences using a sliding window method.\n",
    "\n",
    "    Args:\n",
    "        profile (numpy.ndarray): Position Specific Scoring Matrix (PSSM) profiles.\n",
    "        seq (str): Amino acid sequence.\n",
    "        window_size (int): The size of the sliding window.\n",
    "\n",
    "    Returns:\n",
    "        tuple: A tuple containing the profile features and sequence features.\n",
    "\n",
    "    \"\"\"\n",
    "    ss_map = {'H': 0, 'E': 1, 'C': 2, '-': 2}\n",
    "    profile_x = []\n",
    "    seq_x = []\n",
    "    for i in range(0, len(profile)):\n",
    "        half_window_size = int((window_size-1)/2)\n",
    "        # Apply the sliding window method to PSSM profiles\n",
    "        part1 = np.zeros(20*max(0, half_window_size-i))\n",
    "        part2 = np.ndarray.flatten(profile[max(0,i-half_window_size):min(i+half_window_size+1,len(profile))])\n",
    "        part3 = np.zeros(20*max(0, half_window_size-(len(profile)-i-1)))\n",
    "        vec = np.concatenate((part1, part2, part3))\n",
    "        profile_x.append(vec.tolist())\n",
    "\n",
    "        one_hot_seq = sequence_to_one_hot(seq)\n",
    "        # Apply the sliding window method to one-hot-encoded sequences\n",
    "        part1 = np.zeros(20*max(0, half_window_size-i))\n",
    "        part2 = np.ndarray.flatten(one_hot_seq[max(0,i-half_window_size):min(i+half_window_size+1,len(one_hot_seq))])\n",
    "        part3 = np.zeros(20*max(0, half_window_size-(len(one_hot_seq)-i-1)))\n",
    "        vec = np.concatenate((part1, part2, part3))\n",
    "        seq_x.append(vec.tolist())\n",
    "\n",
    "    return profile_x, seq_x\n",
    "\n",
    "\n",
    "def preprocess_user_data(file_path, window_size = 17, str_transformation = True):\n",
    "    \"\"\"\n",
    "    Preprocesses user data for machine learning models.\n",
    "\n",
    "    Args:\n",
    "        file_path (str): The path to the PSSM file.\n",
    "        window_size (int, optional): The window size for creating profiles. Defaults to 17.\n",
    "        str_transformation (bool, optional): Whether to perform string transformation. Defaults to True.\n",
    "\n",
    "    Returns:\n",
    "        tuple: A tuple containing the profile features and sequence features.\n",
    "\n",
    "    \"\"\"\n",
    "    profile = np.array(parse_pssm(file_path))\n",
    "\n",
    "    if np.sum(profile) != 0:\n",
    "        seq = extract_sequence_from_pssm(file_path)\n",
    "        profile_x, seq_x = create_user_profiles(profile, seq, window_size)\n",
    "        profile_x = np.array(profile_x)\n",
    "        seq_x = np.array(seq_x)\n",
    "\n",
    "    return profile_x, seq_x\n",
    "\n",
    "\n",
    "def create_user_validation_sets(window_size):\n",
    "    \"\"\"\n",
    "    Creates validation sets for cross-validation.\n",
    "\n",
    "    Args:\n",
    "        window_size (int): The window size for creating profiles.\n",
    "\n",
    "    Returns:\n",
    "        tuple: A tuple containing lists of profile features, sequence features, and secondary structure labels for each\n",
    "        validation set.\n",
    "\n",
    "    \"\"\"\n",
    "    names_list = [\n",
    "        ['train3', 'test3']\n",
    "    ]\n",
    "\n",
    "    profile_x_list = []\n",
    "    seq_x_list = []\n",
    "    ss_y_list = []\n",
    "\n",
    "    for train_name, test_name in names_list:\n",
    "        profile_x_sublist = []\n",
    "        seq_x_sublist = []\n",
    "        ss_y_sublist = []\n",
    "\n",
    "        train_path = 'Data/training/cv/' + train_name + '.txt'\n",
    "        test_path = 'Data/training/cv/' + test_name + '.txt'\n",
    "        print(f'{train_name} is preprocessed.')\n",
    "\n",
    "        profile_x, seq_x, ss_y  = preprocess_data(train_path,\n",
    "                                                  'Data/training/',\n",
    "                                                  window_size = window_size,\n",
    "                                                  str_transformation = True\n",
    "                                                 )\n",
    "        profile_x_sublist.append(profile_x)\n",
    "        seq_x_sublist.append(seq_x)\n",
    "        ss_y_sublist.append(ss_y)\n",
    "\n",
    "        print(f'{test_name} is preprocessed.')\n",
    "\n",
    "        profile_x_test, seq_x_test, ss_y_test  = preprocess_data(test_path,\n",
    "                                                                 'Data/training/',\n",
    "                                                                 window_size = window_size,\n",
    "                                                                 str_transformation = True\n",
    "                                                                )\n",
    "        profile_x_sublist.append(profile_x_test)\n",
    "        seq_x_sublist.append(seq_x_test)\n",
    "        ss_y_sublist.append(ss_y_test)\n",
    "\n",
    "        profile_x_list.append(profile_x_sublist)\n",
    "        seq_x_list.append(seq_x_sublist)\n",
    "        ss_y_list.append(ss_y_sublist)\n",
    "\n",
    "    return profile_x_list, seq_x_list, ss_y_list\n",
    "\n",
    "\n",
    "def develop_user_ensemble_model(\n",
    "    profile_x,\n",
    "    ss_y,\n",
    "    seq_x,\n",
    "    profile_x_test,\n",
    "    seq_x_test,\n",
    "    accuracy_cv,\n",
    "    loss_cv,\n",
    "    history_cv,\n",
    "    ss_pred_cv\n",
    "):\n",
    "    \"\"\"\n",
    "    Develops an ensemble model for user data.\n",
    "\n",
    "    Args:\n",
    "        profile_x (numpy.ndarray): Profile features for training.\n",
    "        ss_y (numpy.ndarray): Secondary structure labels for training.\n",
    "        seq_x (numpy.ndarray): Sequence features for training.\n",
    "        profile_x_test (numpy.ndarray): Profile features for testing.\n",
    "        seq_x_test (numpy.ndarray): Sequence features for testing.\n",
    "        accuracy_cv (list): List to store accuracy scores.\n",
    "        loss_cv (list): List to store loss values.\n",
    "        history_cv (list): List to store training history.\n",
    "        ss_pred_cv (list): List to store predicted secondary structures.\n",
    "\n",
    "    Returns:\n",
    "        tuple: A tuple containing accuracy scores, loss values, training history, and predicted secondary structures.\n",
    "\n",
    "    \"\"\"\n",
    "    inputA = layers.Input(shape=(340))\n",
    "    inputB = layers.Input(shape=(340))\n",
    "\n",
    "    # First branch operates on the first input\n",
    "    x = layers.Dense(1024, activation='leaky_relu')(inputA)\n",
    "    x = layers.Dropout(0.5)(x)\n",
    "    x = layers.Dense(512, activation='relu')(x)\n",
    "    x = layers.Dropout(0.5)(x)\n",
    "    x = layers.Dense(256, activation='relu')(x)\n",
    "    x = layers.Dropout(0.5)(x)\n",
    "    x = layers.Dense(128, activation=\"relu\")(x)\n",
    "    x = layers.Dropout(0.5)(x)\n",
    "    x = Model(inputs=inputA, outputs=x)\n",
    "\n",
    "    # Second branch operates on the second input\n",
    "    y = Model(inputs=inputB, outputs=inputB)\n",
    "\n",
    "    combined = layers.concatenate([x.output, y.output])  # combine the output of the two branches\n",
    "    z = layers.Dense(64, activation=\"relu\")(combined)\n",
    "    z = layers.Dropout(0.5)(z)\n",
    "    z = layers.Dense(3, activation='softmax')(z)\n",
    "    model = Model(inputs=[x.input, y.input], outputs=z)\n",
    "\n",
    "    model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "\n",
    "    history = model.fit([profile_x, seq_x], ss_y, epochs=10, batch_size=128)\n",
    "    history_cv.append(history)\n",
    "\n",
    "    ss_y_pred = model.predict([profile_x_test, seq_x_test])\n",
    "    ss_pred_cv.append(ss_y_pred)\n",
    "\n",
    "    return accuracy_cv, loss_cv, history_cv, ss_pred_cv\n",
    "\n",
    "\n",
    "def ensemble_model_user(profile_x_list, seq_x_list, ss_y_list, profile_x_blindtest, seq_x_blindtest):\n",
    "    \"\"\"\n",
    "    Trains an ensemble model for user data.\n",
    "\n",
    "    Args:\n",
    "        profile_x_list (list): List of profile features for each cross-validation set.\n",
    "        seq_x_list (list): List of sequence features for each cross-validation set.\n",
    "        ss_y_list (list): List of secondary structure labels for each cross-validation set.\n",
    "        profile_x_blindtest (numpy.ndarray): Profile features for the blind test set.\n",
    "        seq_x_blindtest (numpy.ndarray): Sequence features for the blind test set.\n",
    "\n",
    "    Returns:\n",
    "        tuple: A tuple containing accuracy scores, loss values, training history, and predicted secondary structures.\n",
    "\n",
    "    \"\"\"\n",
    "    accuracy_cv = []\n",
    "    loss_cv = []\n",
    "    history_cv = []\n",
    "    ss_pred_cv = []\n",
    "\n",
    "    for s in range(len(profile_x_list)):\n",
    "        print(f'Train/test cross-validation set.\\n-------------------------------')\n",
    "        profile_x = profile_x_list[s][0]\n",
    "        seq_x = seq_x_list[s][0]\n",
    "        ss_y = ss_y_list[s][0]\n",
    "        profile_x_test = profile_x_blindtest\n",
    "        seq_x_test = seq_x_blindtest\n",
    "\n",
    "        accuracy_cv, loss_cv, history_cv, ss_pred_cv = develop_user_ensemble_model(\n",
    "            profile_x,\n",
    "            ss_y,\n",
    "            seq_x,\n",
    "            profile_x_test,\n",
    "            seq_x_test,\n",
    "            accuracy_cv,\n",
    "            loss_cv,\n",
    "            history_cv,\n",
    "            ss_pred_cv\n",
    "        )\n",
    "\n",
    "    return accuracy_cv, loss_cv, history_cv, ss_pred_cv\n",
    "\n",
    "\n",
    "def convert(ss_pred_test):\n",
    "    \"\"\"\n",
    "    Convert predicted secondary structure probabilities into a sequence and probabilities.\n",
    "\n",
    "    Args:\n",
    "        ss_pred_test (numpy.ndarray): Predicted secondary structure probabilities.\n",
    "\n",
    "    Returns:\n",
    "        tuple: A tuple containing the predicted secondary structure sequence and corresponding probabilities.\n",
    "\n",
    "    \"\"\"\n",
    "    probs_array = ss_pred_test[0]\n",
    "    sequence = ''\n",
    "    probabilities = []\n",
    "\n",
    "    for prob in probs_array:\n",
    "        max_prob_index = np.argmax(prob)\n",
    "        probabilities.append(round(prob[max_prob_index], 2))\n",
    "\n",
    "        if max_prob_index == 0:\n",
    "            sequence += 'H'\n",
    "        elif max_prob_index == 1:\n",
    "            sequence += 'E'\n",
    "        else:\n",
    "            sequence += 'C'\n",
    "\n",
    "    return sequence, probabilities\n",
    "\n",
    "\n",
    "def plot_probabilities(sequence, probabilities):\n",
    "    \"\"\"\n",
    "    Plot the prediction probabilities of each secondary structure element along the sequence.\n",
    "\n",
    "    Args:\n",
    "        sequence (str): A string representing the sequence of the secondary structure elements.\n",
    "        probabilities (list): A list of prediction probabilities.\n",
    "\n",
    "    Returns:\n",
    "        None\n",
    "    \"\"\"\n",
    "    sequence_list = []\n",
    "    for char in sequence:\n",
    "        sequence_list.append(char)\n",
    "    position = list(range(1, len(sequence) + 1))\n",
    "\n",
    "    plt.figure(figsize=(12, 4))\n",
    "    sns.set(style='ticks')\n",
    "    sns.color_palette('deep')\n",
    "    sns.lineplot(x=position, y=probabilities, drawstyle='steps-pre')\n",
    "    sns.scatterplot(x=position, y=[0]*len(position), hue=sequence_list, s=5)\n",
    "    plt.legend(bbox_to_anchor=(1.05, 0.7), loc='upper left')\n",
    "    plt.xlabel('Amino acid position', fontname='Arial', fontsize=11)\n",
    "    plt.ylabel('Prediction probability',  fontname='Arial', fontsize=11)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "572afce2-0f75-43fb-a2bb-b1bf0e626960",
   "metadata": {},
   "source": [
    "Example file path: Data/blindTest/pssm/4S1H_A.pssm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "1574b7d2-7291-48d1-87ce-d06d13ad14a8",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdin",
     "output_type": "stream",
     "text": [
      "Provide a file path for a PSSM file:  Data/blindTest/pssm/4S1H_A.pssm\n",
      "Provide a window size (7-17):  17\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train3 is preprocessed.\n",
      "test3 is preprocessed.\n",
      "Train/test cross-validation set.\n",
      "-------------------------------\n",
      "Epoch 1/10\n",
      "1238/1238 [==============================] - 18s 13ms/step - loss: 0.7582 - accuracy: 0.6803\n",
      "Epoch 2/10\n",
      "1238/1238 [==============================] - 18s 14ms/step - loss: 0.6488 - accuracy: 0.7373\n",
      "Epoch 3/10\n",
      "1238/1238 [==============================] - 19s 15ms/step - loss: 0.6178 - accuracy: 0.7507\n",
      "Epoch 4/10\n",
      "1238/1238 [==============================] - 17s 14ms/step - loss: 0.5999 - accuracy: 0.7588\n",
      "Epoch 5/10\n",
      "1238/1238 [==============================] - 18s 14ms/step - loss: 0.5869 - accuracy: 0.7639\n",
      "Epoch 6/10\n",
      "1238/1238 [==============================] - 20s 16ms/step - loss: 0.5775 - accuracy: 0.7664\n",
      "Epoch 7/10\n",
      "1238/1238 [==============================] - 24s 20ms/step - loss: 0.5699 - accuracy: 0.7696\n",
      "Epoch 8/10\n",
      "1238/1238 [==============================] - 19s 15ms/step - loss: 0.5636 - accuracy: 0.7727\n",
      "Epoch 9/10\n",
      "1238/1238 [==============================] - 21s 17ms/step - loss: 0.5566 - accuracy: 0.7757\n",
      "Epoch 10/10\n",
      "1238/1238 [==============================] - 21s 17ms/step - loss: 0.5529 - accuracy: 0.7782\n",
      "9/9 [==============================] - 0s 2ms/step\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predicted secondary structure: CCCCEEEEEECCCCCCCCCCCCCCEHHHCCCEEEEEEEEEEECCCCCEEECCCCCCHHHHHHHHHHHHHCCCCCCCHEEEECCCCCHHHHHHHHHHHHHHHCCCCCEEEECCCCCCCCCEECCHHHHHHHHHHHCCCEEECCCHHHHHHHHCCCCCCHHHHHHHHHHHHHCCCCEEEEEECCCCCCCEEECCCCCCCCCCCCEECCCCEECCCCCCCHHHHHHHHHHHHCCCCHHHHHHHHHHHHHHHHEHCCCCCEHHHHHHHHCCCCCCCCHHHHHHHHHHCCCC\n"
     ]
    }
   ],
   "source": [
    "user_input = input('Provide a file path for a PSSM file: ')\n",
    "window_size = int(input('Provide a window size (7-17): '))\n",
    "\n",
    "profile_x_list, seq_x_list, ss_y_list = create_user_validation_sets(window_size = window_size)\n",
    "profile_x_blindtest, seq_x_blindtest = preprocess_user_data(\n",
    "    user_input,\n",
    "    window_size = 17,\n",
    "    str_transformation = True\n",
    ")\n",
    "\n",
    "accuracy_test, loss_test, history_test, ss_pred_test = ensemble_model_user(\n",
    "    profile_x_list, \n",
    "    seq_x_list, \n",
    "    ss_y_list, \n",
    "    profile_x_blindtest, \n",
    "    seq_x_blindtest, \n",
    ")\n",
    "\n",
    "sequence, probabilities = convert(ss_pred_test)\n",
    "plot_probabilities(sequence, probabilities)\n",
    "print(f'Predicted secondary structure: {sequence}')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}