Pfam_protein_sequence_classification.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iAHSXYnksSBj"
      },
      "source": [
        "# Protein Sequence Classification on pfam dataset"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "y_dDnHNus7SL"
      },
      "source": [
        "## 1. Business Problem"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KmJ2E-Krs_xt"
      },
      "source": [
        "### 1.1 Description"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MDx-UEBztDpK"
      },
      "source": [
        "[Proteins](https://en.wikipedia.org/wiki/Protein) are large, complex biomolecules that play many critical roles in biological bodies. Proteins are made up of one or more long chains of [amino acids](https://en.wikipedia.org/wiki/Amino_acid) sequences. These Sequence are the arrangement of amino acids in a protein held together by [peptide bonds](https://en.wikipedia.org/wiki/Peptide_bond). Proteins can be made from [20](https://www.hornetjuice.com/amino-acids-types/) different kinds of amino acids, and the structure and function of each protein are determined by the kinds of amino acids used to make it and how they are arranged. \n",
        "\n",
        "Understanding this relationship between amino acid sequence and protein function is a long-standing problem in moleculer biology with far-reaching scientific implications. Can we use deep learning that learns the relationship between unaligned amino acid sequences and their functional annotations across all 17929 families of the Pfam database.\n",
        "\n",
        "[Pfam](https://en.m.wikipedia.org/wiki/Pfam) is a database of [protein families](https://en.m.wikipedia.org/wiki/Protein_family) that includes their annotations and multiple sequence alignments.\n",
        "\n",
        "__Problem Statement__\n",
        "- Classification of protein's amino acid sequence to one of the protein family accession, based on PFam dataset.\n",
        "- In other words, the task is: given the amino acid sequence of the protein domain, predict which class it belongs to.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jtyg15A6ATsS"
      },
      "source": [
        "### 1.2 Sources/Useful Links"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uPZiBF6ca_Sh"
      },
      "source": [
        "- Source: [Pfam seed random split](https://www.kaggle.com/googleai/pfam-seed-random-split)\n",
        "- Paper: [Using deep learning to annotate the protein universe](https://www.biorxiv.org/content/10.1101/626507v4.full)."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eF6y4ZtNBAaY"
      },
      "source": [
        "### 1.3 Real world/Business Objectives and Constraints"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0U6WEFA3BI0S"
      },
      "source": [
        "Objectives\n",
        "- Predict protein family accession from its amino acids sequence with high accuracy.\n",
        "\n",
        "Constraints\n",
        "- No strict latency concerns."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ylHPrtMlBvyB"
      },
      "source": [
        "## 2. Machine Learning Problem"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xdkPtpHiB45Y"
      },
      "source": [
        "### 2.1 Data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Vn03S7RtB8XS"
      },
      "source": [
        "#### 2.1.1 Data Overview"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0R9yCaZiB_76"
      },
      "source": [
        "- `sequence`: These are usually the input features to the model. Amino acid sequence for this domain. There are 20 very common amino acids (frequency > 1,000,000), and 4 amino acids that are quite uncommon: X, U, B, O, Z.\n",
        "\n",
        "- `family_accession`: These are usually the labels for the model. Accession number in form PFxxxxx.y (Pfam), where xxxxx is the family accession, and y is the version number. Some values of y are greater than ten, and so 'y' has two digits.\n",
        "\n",
        "- `sequence_name`: Sequence name, in the form \"uniprot_accession_id/start_index-end_index\". \n",
        "\n",
        "- `aligned_sequence`: Contains a single sequence from the multiple sequence alignment (with the rest of the members of the family in seed, with gaps retained.\n",
        "\n",
        "- `family_id`: One word name for family. "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "u12YlDWcCr66"
      },
      "source": [
        "#### 2.1.2 Example Data point"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OtOApClbCDfM"
      },
      "source": [
        "\n",
        "\n",
        "```\n",
        "sequence: HWLQMRDSMNTYNNMVNRCFATCIRSFQEKKVNAEEMDCTKRCVTKFVGYSQRVALRFAE \n",
        "family_accession: PF02953.15\n",
        "sequence_name: C5K6N5_PERM5/28-87\n",
        "aligned_sequence: ....HWLQMRDSMNTYNNMVNRCFATCI...........RS.F....QEKKVNAEE.....MDCT....KRCVTKFVGYSQRVALRFAE \n",
        "family_id: zf-Tim10_DDP\n",
        "```\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "z4tAp7Xrb5gX"
      },
      "source": [
        "#### 2.1.3 Data split"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Cq-WD5arGu7x"
      },
      "source": [
        "- We have been provided with already done random split(train, val, test) of pfam dataset.\n",
        "  - Train - 80% (For training the models).\n",
        "  - Val - 10% (For hyperparameter tuning/model validation).\n",
        "  - Test - 10% (For acessing the model performance)."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CUfCZp0ICwaF"
      },
      "source": [
        "### 2.2 Mapping the real world problem to an ML problem"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "24iT472YC2J-"
      },
      "source": [
        "#### 2.2.1 Type of Machine learning Problem"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "S7QCMBQJDIDw"
      },
      "source": [
        "It is a multi class classification problem, for a given sequence of amino acids we need to predict its family accession."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qnlBgZX8DVP2"
      },
      "source": [
        "#### 2.2.2 Performance Metric"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SZ5tn5EmDdYe"
      },
      "source": [
        "- Multi class log loss\n",
        "- Accuracy"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zQJnjfq1D6Js"
      },
      "source": [
        "## 3. Exploratory Data Analysis"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2YlZeBjv5ejk"
      },
      "source": [
        "### Importing Libraries"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 385
        },
        "id": "EqLtXwPchzZf",
        "outputId": "3fc64e79-3b9b-450d-8ce7-2df6b21e3d2c"
      },
      "outputs": [
        {
          "output_type": "error",
          "ename": "ImportError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mImportError\u001b[0m                               Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-1-6dd901083588>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     20\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0mkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mutils\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mto_categorical\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     21\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0mkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpreprocessing\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtext\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mTokenizer\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 22\u001b[0;31m \u001b[0;32mfrom\u001b[0m \u001b[0mkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpreprocessing\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msequence\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mpad_sequences\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     23\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0mkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcallbacks\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mEarlyStopping\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     24\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0mkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlayers\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mInput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mDense\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mDropout\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mFlatten\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mActivation\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mImportError\u001b[0m: cannot import name 'pad_sequences' from 'keras.preprocessing.sequence' (/usr/local/lib/python3.7/dist-packages/keras/preprocessing/sequence.py)",
            "",
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0;32m\nNOTE: If your import is failing due to a missing package, you can\nmanually install dependencies using either !pip or !apt.\n\nTo view examples of installing some common dependencies, click the\n\"Open Examples\" button below.\n\u001b[0;31m---------------------------------------------------------------------------\u001b[0m\n"
          ],
          "errorDetails": {
            "actions": [
              {
                "action": "open_url",
                "actionText": "Open Examples",
                "url": "/notebooks/snippets/importing_libraries.ipynb"
              }
            ]
          }
        }
      ],
      "source": [
        "%matplotlib inline\n",
        "\n",
        "import os\n",
        "import gc\n",
        "\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "\n",
        "from collections import Counter\n",
        "from prettytable import PrettyTable\n",
        "from IPython.display import Image\n",
        "\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.regularizers import l2\n",
        "from keras.constraints import max_norm\n",
        "from keras.utils import to_categorical\n",
        "from keras.preprocessing.text import Tokenizer\n",
        "from keras.preprocessing.sequence import pad_sequences\n",
        "from keras.callbacks import EarlyStopping\n",
        "from keras.layers import Input, Dense, Dropout, Flatten, Activation\n",
        "from keras.layers import Conv1D, Add, MaxPooling1D, BatchNormalization\n",
        "from keras.layers import Embedding, Bidirectional, CuDNNLSTM, GlobalMaxPooling1D"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 181
        },
        "id": "a_k-q8gv_J2_",
        "outputId": "08e4f479-a2cb-493d-faa8-57c3556cc2c8"
      },
      "outputs": [
        {
          "output_type": "error",
          "ename": "AttributeError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-2-8a9785c3ca7b>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mtensorflow\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mtf\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mtf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlogging\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset_verbosity\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlogging\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mERROR\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;31mAttributeError\u001b[0m: module 'tensorflow' has no attribute 'logging'"
          ]
        }
      ],
      "source": [
        "import tensorflow as tf\n",
        "tf.logging.set_verbosity(tf.logging.ERROR)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "L4wU9AP3jYWU"
      },
      "outputs": [],
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GCPrxSoI5jQZ"
      },
      "source": [
        "### Loading Data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "id": "Is-kuTuGzYGN",
        "outputId": "270d71b8-6772-4a6a-b7b1-0cdbdee0353e"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Available data ['dev', 'test', 'train']\n"
          ]
        }
      ],
      "source": [
        "data_path = 'drive/My Drive/Case_Study/pfam/random_split/'\n",
        "print('Available data', os.listdir(data_path))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "4NrlxI8tiIww"
      },
      "outputs": [],
      "source": [
        "# https://www.kaggle.com/drewbryant/starter-pfam-seed-random-split\n",
        "\n",
        "# data is randomly splitted in three folders [train(80%), test(10%), dev(10%)]\n",
        "# reading and concatinating data for each folder.\n",
        "\n",
        "def read_data(partition):\n",
        "  data = []\n",
        "  for fn in os.listdir(os.path.join(data_path, partition)):\n",
        "    with open(os.path.join(data_path, partition, fn)) as f:\n",
        "      data.append(pd.read_csv(f, index_col=None))\n",
        "  return pd.concat(data)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Xy25Vu-K3MvW"
      },
      "outputs": [],
      "source": [
        "# reading all data_partitions\n",
        "\n",
        "df_train = read_data('train')\n",
        "df_val = read_data('dev')\n",
        "df_test = read_data('test')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lqRU2FBxezKq"
      },
      "source": [
        "### Basic Statistics"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 185
        },
        "id": "pj2uvMN6pw7Y",
        "outputId": "e9850a9f-44da-4804-940c-0dce9d7998d0"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "Int64Index: 1086741 entries, 0 to 13514\n",
            "Data columns (total 5 columns):\n",
            "family_id           1086741 non-null object\n",
            "sequence_name       1086741 non-null object\n",
            "family_accession    1086741 non-null object\n",
            "aligned_sequence    1086741 non-null object\n",
            "sequence            1086741 non-null object\n",
            "dtypes: object(5)\n",
            "memory usage: 49.7+ MB\n"
          ]
        }
      ],
      "source": [
        "df_train.info()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 195
        },
        "id": "fvvyiT4PTlew",
        "outputId": "e5d8d420-9d92-43bf-99a2-6f5408e5d398"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>family_id</th>\n",
              "      <th>sequence_name</th>\n",
              "      <th>family_accession</th>\n",
              "      <th>aligned_sequence</th>\n",
              "      <th>sequence</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>GMC_oxred_C</td>\n",
              "      <td>A4WZS5_RHOS5/416-539</td>\n",
              "      <td>PF05199.13</td>\n",
              "      <td>PHPE.SRIRLST.RRDAHGMP.....IP.RIESRLGP............</td>\n",
              "      <td>PHPESRIRLSTRRDAHGMPIPRIESRLGPDAFARLRFMARTCRAIL...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>DUF2887</td>\n",
              "      <td>K9QI92_9NOSO/3-203</td>\n",
              "      <td>PF11103.8</td>\n",
              "      <td>RDSIYYQIFKRFPALIFEL..VD.NRPPQAQNYRFESVEVKETAFR...</td>\n",
              "      <td>RDSIYYQIFKRFPALIFELVDNRPPQAQNYRFESVEVKETAFRIDG...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>zf-IS66</td>\n",
              "      <td>Q92LC9_RHIME/32-75</td>\n",
              "      <td>PF13005.7</td>\n",
              "      <td>.TCCPDCGG.E..LRLVGED.AS....EILDMI.AAQMKVIEVARL...</td>\n",
              "      <td>TCCPDCGGELRLVGEDASEILDMIAAQMKVIEVARLKKSCRCCE</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Asp_decarbox</td>\n",
              "      <td>X2GQZ4_9BACI/1-115</td>\n",
              "      <td>PF02261.16</td>\n",
              "      <td>MLRMMMNSKIHRATVTEADLNYVGSITIDEDILDAVGMLPNEKVHI...</td>\n",
              "      <td>MLRMMMNSKIHRATVTEADLNYVGSITIDEDILDAVGMLPNEKVHI...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Filamin</td>\n",
              "      <td>A7SQM3_NEMVE/342-439</td>\n",
              "      <td>PF00630.19</td>\n",
              "      <td>TACPKQ.CTA....RGLG.............LK.AAPVT.QPT..R...</td>\n",
              "      <td>TACPKQCTARGLGLKAAPVTQPTRFVVILNDCHGQPLGRSEGELEV...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "      family_id  ...                                           sequence\n",
              "0   GMC_oxred_C  ...  PHPESRIRLSTRRDAHGMPIPRIESRLGPDAFARLRFMARTCRAIL...\n",
              "1       DUF2887  ...  RDSIYYQIFKRFPALIFELVDNRPPQAQNYRFESVEVKETAFRIDG...\n",
              "2       zf-IS66  ...       TCCPDCGGELRLVGEDASEILDMIAAQMKVIEVARLKKSCRCCE\n",
              "3  Asp_decarbox  ...  MLRMMMNSKIHRATVTEADLNYVGSITIDEDILDAVGMLPNEKVHI...\n",
              "4       Filamin  ...  TACPKQCTARGLGLKAAPVTQPTRFVVILNDCHGQPLGRSEGELEV...\n",
              "\n",
              "[5 rows x 5 columns]"
            ]
          },
          "execution_count": 8,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        }
      ],
      "source": [
        "df_train.head()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "id": "FZD6gGJLd7uB",
        "outputId": "bf474ea4-43a3-4e4e-d154-350b18cef59f"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "'PHPESRIRLSTRRDAHGMPIPRIESRLGPDAFARLRFMARTCRAILAAAGCAAPFEEFSSADAFSSTHVFGTCRMGHDPMRNVVDGWGRSHRWPNLFVADASLFPSSGGGESPGLTIQALALRT'"
            ]
          },
          "execution_count": 9,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# ex: unaligned sequence\n",
        "# each character reperesents one of the 24(20 common + 4 uncommon) amino acids in the sequence.\n",
        "\n",
        "df_train.head(1)['sequence'].values[0]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 67
        },
        "id": "lYmykaXe3aKy",
        "outputId": "6df755e5-a1ca-4c71-e3eb-bbd939264790"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Train size:  1086741\n",
            "Val size:  126171\n",
            "Test size:  126171\n"
          ]
        }
      ],
      "source": [
        "# Given data size\n",
        "print('Train size: ', len(df_train))\n",
        "print('Val size: ', len(df_val))\n",
        "print('Test size: ', len(df_test))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "hmjg-CbQyioF"
      },
      "outputs": [],
      "source": [
        "def calc_unique_cls(train, test, val):\n",
        "  \"\"\"\n",
        "  Prints # unique classes in data sets.\n",
        "  \"\"\"\n",
        "  train_unq = np.unique(train['family_accession'].values)\n",
        "  val_unq = np.unique(val['family_accession'].values)\n",
        "  test_unq = np.unique(test['family_accession'].values)\n",
        "\n",
        "  print('Number of unique classes in Train: ', len(train_unq))\n",
        "  print('Number of unique classes in Val: ', len(val_unq))\n",
        "  print('Number of unique classes in Test: ', len(test_unq))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 67
        },
        "id": "-2r0wIsBiy-c",
        "outputId": "a13676ba-448c-467e-a2ec-42b81b1a83d3"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Number of unique classes in Train:  17929\n",
            "Number of unique classes in Val:  13071\n",
            "Number of unique classes in Test:  13071\n"
          ]
        }
      ],
      "source": [
        "# Unique classes in the given dataset : [df_train, df_val and df_test]\n",
        "\n",
        "calc_unique_cls(df_train, df_test, df_val)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7XvpRxecd_B1"
      },
      "source": [
        "### Sequence Counts"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ViaJx_Peu5y_"
      },
      "outputs": [],
      "source": [
        "# Length of sequence in train data.\n",
        "df_train['seq_char_count']= df_train['sequence'].apply(lambda x: len(x))\n",
        "df_val['seq_char_count']= df_val['sequence'].apply(lambda x: len(x))\n",
        "df_test['seq_char_count']= df_test['sequence'].apply(lambda x: len(x))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ATEmEmeAYoeB"
      },
      "outputs": [],
      "source": [
        "def plot_seq_count(df, data_name):\n",
        "  sns.distplot(df['seq_char_count'].values)\n",
        "  plt.title(f'Sequence char count: {data_name}')\n",
        "  plt.grid(True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "G1Mo0m3yFAFu",
        "outputId": "43fff7b5-7292-40c1-e8a5-cec23b7065ad"
      },
      "outputs": [
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABQ8AAAEJCAYAAAA6r9GhAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhyIiIiIiIiIiItKj/w9Z5KWI/B5hAwAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 3 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "plt.subplot(1, 3, 1)\n",
        "plot_seq_count(df_train, 'Train')\n",
        "\n",
        "plt.subplot(1, 3, 2)\n",
        "plot_seq_count(df_val, 'Val')\n",
        "\n",
        "plt.subplot(1, 3, 3)\n",
        "plot_seq_count(df_test, 'Test')\n",
        "\n",
        "plt.subplots_adjust(right=3.0)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-ZIi2DB6bLpc"
      },
      "source": [
        "__Observation__\n",
        "- Most of the unaligned amino acid sequences have character counts in the range of 50-250."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LFxzvA1oeJGm"
      },
      "source": [
        "### Sequence Code Frequency"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6qg7NNbNhdMH"
      },
      "source": [
        "Amino acid sequences are represented with their corresponding 1 letter code, for example, code for alanine is (A), arginine is (R) and so on. The complete list of amino acids with there code can be found <a href='http://www.cryst.bbk.ac.uk/education/AminoAcid/the_twenty.html'>here</a>."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Mm54TAoH_C8t"
      },
      "outputs": [],
      "source": [
        "def get_code_freq(df, data_name):\n",
        "  \n",
        "  df = df.apply(lambda x: \" \".join(x))\n",
        "  \n",
        "  codes = []\n",
        "  for i in df: # concatination of all codes\n",
        "    codes.extend(i)\n",
        "\n",
        "  codes_dict= Counter(codes)\n",
        "  codes_dict.pop(' ') # removing white space\n",
        "  \n",
        "  print(f'Codes: {data_name}')\n",
        "  print(f'Total unique codes: {len(codes_dict.keys())}')\n",
        "\n",
        "  df = pd.DataFrame({'Code': list(codes_dict.keys()), 'Freq': list(codes_dict.values())})\n",
        "  return df.sort_values('Freq', ascending=False).reset_index()[['Code', 'Freq']]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 821
        },
        "id": "lekmtCh9KXrY",
        "outputId": "c2bcd5c7-0cb4-448e-cb28-ec8786f81fbc"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Codes: Train\n",
            "Total unique codes: 25\n"
          ]
        },
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Code</th>\n",
              "      <th>Freq</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>L</td>\n",
              "      <td>17062816</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>A</td>\n",
              "      <td>14384873</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>V</td>\n",
              "      <td>11913147</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>G</td>\n",
              "      <td>11845579</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>E</td>\n",
              "      <td>10859966</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>S</td>\n",
              "      <td>10597822</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>I</td>\n",
              "      <td>10234455</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>R</td>\n",
              "      <td>9406165</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>D</td>\n",
              "      <td>9371097</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>K</td>\n",
              "      <td>9127832</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>T</td>\n",
              "      <td>9034110</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>P</td>\n",
              "      <td>7441084</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>F</td>\n",
              "      <td>7130287</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>N</td>\n",
              "      <td>6616976</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>Q</td>\n",
              "      <td>6250389</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>Y</td>\n",
              "      <td>5556597</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>M</td>\n",
              "      <td>3708948</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>H</td>\n",
              "      <td>3704587</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>C</td>\n",
              "      <td>2316115</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>W</td>\n",
              "      <td>2293257</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>X</td>\n",
              "      <td>1505</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>21</th>\n",
              "      <td>U</td>\n",
              "      <td>119</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>22</th>\n",
              "      <td>B</td>\n",
              "      <td>33</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>23</th>\n",
              "      <td>O</td>\n",
              "      <td>18</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>24</th>\n",
              "      <td>Z</td>\n",
              "      <td>8</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   Code      Freq\n",
              "0     L  17062816\n",
              "1     A  14384873\n",
              "2     V  11913147\n",
              "3     G  11845579\n",
              "4     E  10859966\n",
              "5     S  10597822\n",
              "6     I  10234455\n",
              "7     R   9406165\n",
              "8     D   9371097\n",
              "9     K   9127832\n",
              "10    T   9034110\n",
              "11    P   7441084\n",
              "12    F   7130287\n",
              "13    N   6616976\n",
              "14    Q   6250389\n",
              "15    Y   5556597\n",
              "16    M   3708948\n",
              "17    H   3704587\n",
              "18    C   2316115\n",
              "19    W   2293257\n",
              "20    X      1505\n",
              "21    U       119\n",
              "22    B        33\n",
              "23    O        18\n",
              "24    Z         8"
            ]
          },
          "execution_count": 60,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# train code sequence\n",
        "train_code_freq = get_code_freq(df_train['sequence'], 'Train')\n",
        "train_code_freq"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 732
        },
        "id": "26OQZxJbDABI",
        "outputId": "cf8fbd67-a4c1-44f4-ada0-12b5529c9f20"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Codes: Val\n",
            "Total unique codes: 22\n"
          ]
        },
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Code</th>\n",
              "      <th>Freq</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>L</td>\n",
              "      <td>1967025</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>A</td>\n",
              "      <td>1667703</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>V</td>\n",
              "      <td>1382128</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>G</td>\n",
              "      <td>1376124</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>E</td>\n",
              "      <td>1249356</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>S</td>\n",
              "      <td>1210750</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>I</td>\n",
              "      <td>1185722</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>R</td>\n",
              "      <td>1085950</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>D</td>\n",
              "      <td>1080572</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>K</td>\n",
              "      <td>1047638</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>T</td>\n",
              "      <td>1039590</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>P</td>\n",
              "      <td>850937</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>F</td>\n",
              "      <td>820778</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>N</td>\n",
              "      <td>757315</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>Q</td>\n",
              "      <td>714424</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>Y</td>\n",
              "      <td>639252</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>M</td>\n",
              "      <td>428275</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>H</td>\n",
              "      <td>426922</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>C</td>\n",
              "      <td>264434</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>W</td>\n",
              "      <td>263317</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>X</td>\n",
              "      <td>146</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>21</th>\n",
              "      <td>U</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   Code     Freq\n",
              "0     L  1967025\n",
              "1     A  1667703\n",
              "2     V  1382128\n",
              "3     G  1376124\n",
              "4     E  1249356\n",
              "5     S  1210750\n",
              "6     I  1185722\n",
              "7     R  1085950\n",
              "8     D  1080572\n",
              "9     K  1047638\n",
              "10    T  1039590\n",
              "11    P   850937\n",
              "12    F   820778\n",
              "13    N   757315\n",
              "14    Q   714424\n",
              "15    Y   639252\n",
              "16    M   428275\n",
              "17    H   426922\n",
              "18    C   264434\n",
              "19    W   263317\n",
              "20    X      146\n",
              "21    U       12"
            ]
          },
          "execution_count": 61,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# val code sequence\n",
        "val_code_freq = get_code_freq(df_val['sequence'], 'Val')\n",
        "val_code_freq"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 791
        },
        "id": "CEcQG5Y2CVvA",
        "outputId": "8e7f8cd4-adc1-415c-9ed4-e0a0dfd50cd0"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Codes: Test\n",
            "Total unique codes: 24\n"
          ]
        },
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Code</th>\n",
              "      <th>Freq</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>L</td>\n",
              "      <td>1967046</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>A</td>\n",
              "      <td>1668137</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>V</td>\n",
              "      <td>1380962</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>G</td>\n",
              "      <td>1375349</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>E</td>\n",
              "      <td>1251000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>S</td>\n",
              "      <td>1210559</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>I</td>\n",
              "      <td>1184239</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>R</td>\n",
              "      <td>1085786</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>D</td>\n",
              "      <td>1078379</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>K</td>\n",
              "      <td>1045957</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>T</td>\n",
              "      <td>1038682</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>P</td>\n",
              "      <td>851574</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>F</td>\n",
              "      <td>822738</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>N</td>\n",
              "      <td>756549</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>Q</td>\n",
              "      <td>712317</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>Y</td>\n",
              "      <td>639218</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>M</td>\n",
              "      <td>428892</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>H</td>\n",
              "      <td>425862</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>C</td>\n",
              "      <td>264168</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>W</td>\n",
              "      <td>263755</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>20</th>\n",
              "      <td>X</td>\n",
              "      <td>198</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>21</th>\n",
              "      <td>U</td>\n",
              "      <td>12</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>22</th>\n",
              "      <td>Z</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>23</th>\n",
              "      <td>B</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   Code     Freq\n",
              "0     L  1967046\n",
              "1     A  1668137\n",
              "2     V  1380962\n",
              "3     G  1375349\n",
              "4     E  1251000\n",
              "5     S  1210559\n",
              "6     I  1184239\n",
              "7     R  1085786\n",
              "8     D  1078379\n",
              "9     K  1045957\n",
              "10    T  1038682\n",
              "11    P   851574\n",
              "12    F   822738\n",
              "13    N   756549\n",
              "14    Q   712317\n",
              "15    Y   639218\n",
              "16    M   428892\n",
              "17    H   425862\n",
              "18    C   264168\n",
              "19    W   263755\n",
              "20    X      198\n",
              "21    U       12\n",
              "22    Z        4\n",
              "23    B        2"
            ]
          },
          "execution_count": 62,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# test code sequence\n",
        "test_code_freq = get_code_freq(df_test['sequence'], 'Test')\n",
        "test_code_freq"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Vsp8Bd37DQi4"
      },
      "outputs": [],
      "source": [
        "def plot_code_freq(df, data_name):\n",
        "  \n",
        "  plt.title(f'Code frequency: {data_name}')\n",
        "  sns.barplot(x='Code', y='Freq', data=df)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "id": "hqpczOVOF8sA",
        "outputId": "f1917b77-bdc4-405f-bd62-4d1ffebc1069"
      },
      "outputs": [
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABRMAAAEaCAYAAACCf8DyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 3 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "plt.subplot(1, 3, 1)\n",
        "plot_code_freq(train_code_freq, 'Train')\n",
        "\n",
        "plt.subplot(1, 3, 2)\n",
        "plot_code_freq(val_code_freq, 'Val')\n",
        "\n",
        "plt.subplot(1, 3, 3)\n",
        "plot_code_freq(test_code_freq, 'Test')\n",
        "\n",
        "plt.subplots_adjust(right=3.0)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "J_SB86pZzCIq"
      },
      "source": [
        "__Observations__\n",
        "- Most frequent amino acid code is L followed by A, V, G.\n",
        "- As we can see, that the uncommon amino acids (i.e., X, U, B, O, Z) are present in very less quantity. Therefore we can consider only 20 common natural amino acids for sequence encoding."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "08mVHtLVfCmp"
      },
      "source": [
        "### Protein families with most sequences(No. of observations)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 386
        },
        "id": "VOx7dLl4uVwd",
        "outputId": "f81e1563-a81e-4899-8256-074d6d8d5ffc"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "family_id\n",
              "Methyltransf_25    3637\n",
              "LRR_1              1927\n",
              "Acetyltransf_7     1761\n",
              "His_kinase         1537\n",
              "Bac_transf         1528\n",
              "Lum_binding        1504\n",
              "DNA_binding_1      1345\n",
              "Chromate_transp    1265\n",
              "Lipase_GDSL_2      1252\n",
              "DnaJ_CXXCXGXG      1210\n",
              "SRP54_N            1185\n",
              "WD40               1173\n",
              "OTCace_N           1171\n",
              "PEP-utilizers      1147\n",
              "Glycos_trans_3N    1138\n",
              "THF_DHG_CYH        1113\n",
              "Prenyltransf       1104\n",
              "HTH_1              1064\n",
              "Maf                1061\n",
              "DHH                1057\n",
              "dtype: int64"
            ]
          },
          "execution_count": 65,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        }
      ],
      "source": [
        "df_train.groupby('family_id').size().sort_values(ascending=False).head(20)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 386
        },
        "id": "PFc1swdQ7fA3",
        "outputId": "5a2628fd-adff-4951-a8e2-2aa1d1317dfc"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "family_id\n",
              "Methyltransf_25    454\n",
              "LRR_1              240\n",
              "Acetyltransf_7     219\n",
              "His_kinase         192\n",
              "Bac_transf         190\n",
              "Lum_binding        187\n",
              "DNA_binding_1      168\n",
              "Chromate_transp    157\n",
              "Lipase_GDSL_2      156\n",
              "DnaJ_CXXCXGXG      151\n",
              "SRP54_N            148\n",
              "OTCace_N           146\n",
              "WD40               146\n",
              "PEP-utilizers      143\n",
              "Glycos_trans_3N    142\n",
              "Prenyltransf       138\n",
              "THF_DHG_CYH        138\n",
              "HTH_1              133\n",
              "Maf                132\n",
              "DHH                131\n",
              "dtype: int64"
            ]
          },
          "execution_count": 67,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        }
      ],
      "source": [
        "df_val.groupby('family_id').size().sort_values(ascending=False).head(20)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 386
        },
        "id": "XYKZ-oedp5JQ",
        "outputId": "8948b589-b5f0-4cb4-ab1e-320cfd73591b"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "family_id\n",
              "Methyltransf_25    454\n",
              "LRR_1              240\n",
              "Acetyltransf_7     219\n",
              "His_kinase         192\n",
              "Bac_transf         190\n",
              "Lum_binding        187\n",
              "DNA_binding_1      168\n",
              "Chromate_transp    157\n",
              "Lipase_GDSL_2      156\n",
              "DnaJ_CXXCXGXG      151\n",
              "SRP54_N            148\n",
              "OTCace_N           146\n",
              "WD40               146\n",
              "PEP-utilizers      143\n",
              "Glycos_trans_3N    142\n",
              "Prenyltransf       138\n",
              "THF_DHG_CYH        138\n",
              "HTH_1              133\n",
              "Maf                132\n",
              "DHH                131\n",
              "dtype: int64"
            ]
          },
          "execution_count": 66,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        }
      ],
      "source": [
        "df_test.groupby('family_id').size().sort_values(ascending=False).head(20)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XRyUdzzxhdMU"
      },
      "source": [
        "__Observation__\n",
        "\n",
        "- Top 20 classes are same across all the sets [train, test, val].\n",
        "- Test and Val sets have almost same frequency for the top 20 classes."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TGQw5qydpFw5"
      },
      "source": [
        "### Considering 1000 classes based on no. of observations."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "id": "bOGBUMFw-CdY",
        "outputId": "07b4ac68-da22-447c-f504-ec3d28c79552"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "1000"
            ]
          },
          "execution_count": 16,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# Considering top 1000 classes based on most observations because of limited computational power.\n",
        "\n",
        "classes = df_train['family_accession'].value_counts()[:1000].index.tolist()\n",
        "len(classes)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 84
        },
        "id": "Ctv5whnYk8FL",
        "outputId": "fc992b26-c07c-453c-cb1a-d1da2f9e54c8"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Data size after considering 1000 classes for each data split:\n",
            "Train size : 439493\n",
            "Val size : 54378\n",
            "Test size : 54378\n"
          ]
        }
      ],
      "source": [
        "# Filtering data based on considered 1000 classes.\n",
        "train_sm = df_train.loc[df_train['family_accession'].isin(classes)].reset_index()\n",
        "val_sm = df_val.loc[df_val['family_accession'].isin(classes)].reset_index()\n",
        "test_sm = df_test.loc[df_test['family_accession'].isin(classes)].reset_index()\n",
        "\n",
        "print('Data size after considering 1000 classes for each data split:')\n",
        "print('Train size :', len(train_sm))\n",
        "print('Val size :', len(val_sm))\n",
        "print('Test size :', len(test_sm))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 67
        },
        "id": "0sRhsRRqUGGV",
        "outputId": "bee31926-b35b-4d12-eb75-e35109705744"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Number of unique classes in Train:  1000\n",
            "Number of unique classes in Val:  1000\n",
            "Number of unique classes in Test:  1000\n"
          ]
        }
      ],
      "source": [
        "# No. of unique classes after reducing the data size.\n",
        "\n",
        "calc_unique_cls(train_sm, test_sm, val_sm)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_aPR87m9hYV2"
      },
      "source": [
        "## 4. Deep Learning Models"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LHh3AXwNp7Yv"
      },
      "source": [
        "### Text Preprocessing"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 50
        },
        "id": "D16pJrHvqDdU",
        "outputId": "0ff0f88d-18b2-4fff-c65f-1108c58707aa"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "{'A': 1, 'C': 2, 'D': 3, 'E': 4, 'F': 5, 'G': 6, 'H': 7, 'I': 8, 'K': 9, 'L': 10, 'M': 11, 'N': 12, 'P': 13, 'Q': 14, 'R': 15, 'S': 16, 'T': 17, 'V': 18, 'W': 19, 'Y': 20}\n",
            "Dict Length: 20\n"
          ]
        }
      ],
      "source": [
        "# https://dmnfarrell.github.io/bioinformatics/mhclearning\n",
        "# http://www.cryst.bbk.ac.uk/education/AminoAcid/the_twenty.html\n",
        "# 1 letter code for 20 natural amino acids\n",
        "\n",
        "codes = ['A', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'K', 'L',\n",
        "         'M', 'N', 'P', 'Q', 'R', 'S', 'T', 'V', 'W', 'Y']\n",
        "\n",
        "def create_dict(codes):\n",
        "  char_dict = {}\n",
        "  for index, val in enumerate(codes):\n",
        "    char_dict[val] = index+1\n",
        "\n",
        "  return char_dict\n",
        "\n",
        "char_dict = create_dict(codes)\n",
        "\n",
        "print(char_dict)\n",
        "print(\"Dict Length:\", len(char_dict))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "dN92LSEG62zX"
      },
      "outputs": [],
      "source": [
        "def integer_encoding(data):\n",
        "  \"\"\"\n",
        "  - Encodes code sequence to integer values.\n",
        "  - 20 common amino acids are taken into consideration\n",
        "    and rest 4 are categorized as 0.\n",
        "  \"\"\"\n",
        "  \n",
        "  encode_list = []\n",
        "  for row in data['sequence'].values:\n",
        "    row_encode = []\n",
        "    for code in row:\n",
        "      row_encode.append(char_dict.get(code, 0))\n",
        "    encode_list.append(np.array(row_encode))\n",
        "  \n",
        "  return encode_list"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "2SW62zkr7Xzf"
      },
      "outputs": [],
      "source": [
        "train_encode = integer_encoding(train_sm) \n",
        "val_encode = integer_encoding(val_sm) \n",
        "test_encode = integer_encoding(test_sm) "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "id": "0_vX2hJJ7noX",
        "outputId": "6175a913-7b43-4283-9353-477f809f23d7"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "((439493, 100), (54378, 100), (54378, 100))"
            ]
          },
          "execution_count": 23,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# padding sequences\n",
        "\n",
        "max_length = 100\n",
        "train_pad = pad_sequences(train_encode, maxlen=max_length, padding='post', truncating='post')\n",
        "val_pad = pad_sequences(val_encode, maxlen=max_length, padding='post', truncating='post')\n",
        "test_pad = pad_sequences(test_encode, maxlen=max_length, padding='post', truncating='post')\n",
        "\n",
        "train_pad.shape, val_pad.shape, test_pad.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "id": "3MNzmuhn9zTJ",
        "outputId": "55c143ec-aed3-4bb4-d88b-7720a7651ae5"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "((439493, 100, 21), (54378, 100, 21), (54378, 100, 21))"
            ]
          },
          "execution_count": 24,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# One hot encoding of sequences\n",
        "\n",
        "train_ohe = to_categorical(train_pad)\n",
        "val_ohe = to_categorical(val_pad)\n",
        "test_ohe = to_categorical(test_pad)\n",
        "\n",
        "train_ohe.shape, test_ohe.shape, test_ohe.shape "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "REZ9V0mt_v8Y"
      },
      "outputs": [],
      "source": [
        "# del train_pad, val_pad, test_pad\n",
        "# del train_encode, val_encode, test_encode\n",
        "\n",
        "# gc.collect()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "id": "cOdjjL1mLMko",
        "outputId": "207f7985-fe44-49b4-d9a6-161d4eaf582b"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "((439493,), (54378,), (54378,))"
            ]
          },
          "execution_count": 26,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# label/integer encoding output variable: (y)\n",
        "le = LabelEncoder()\n",
        "\n",
        "y_train_le = le.fit_transform(train_sm['family_accession'])\n",
        "y_val_le = le.transform(val_sm['family_accession'])\n",
        "y_test_le = le.transform(test_sm['family_accession'])\n",
        "\n",
        "y_train_le.shape, y_val_le.shape, y_test_le.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "id": "1r-g3zHcNe5o",
        "outputId": "03227960-c615-4a41-d315-c522151c350a"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Total classes:  1000\n"
          ]
        }
      ],
      "source": [
        "print('Total classes: ', len(le.classes_))\n",
        "# le.classes_"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "id": "er4ArjShGnV3",
        "outputId": "79a9fc8c-b430-4971-e338-b8cef5868d42"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "((439493, 1000), (54378, 1000), (54378, 1000))"
            ]
          },
          "execution_count": 28,
          "metadata": {
            "tags": []
          },
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# One hot encoding of outputs\n",
        "y_train = to_categorical(y_train_le)\n",
        "y_val = to_categorical(y_val_le)\n",
        "y_test = to_categorical(y_test_le)\n",
        "\n",
        "y_train.shape, y_val.shape, y_test.shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ADeDl5u7whSp"
      },
      "outputs": [],
      "source": [
        "# Utility function: plot model's accuracy and loss\n",
        "\n",
        "# https://realpython.com/python-keras-text-classification/\n",
        "plt.style.use('ggplot')\n",
        "\n",
        "def plot_history(history):\n",
        "  acc = history.history['acc']\n",
        "  val_acc = history.history['val_acc']\n",
        "  loss = history.history['loss']\n",
        "  val_loss = history.history['val_loss']\n",
        "  x = range(1, len(acc) + 1)\n",
        "\n",
        "  plt.figure(figsize=(12, 5))\n",
        "  plt.subplot(1, 2, 1)\n",
        "  plt.plot(x, acc, 'b', label='Training acc')\n",
        "  plt.plot(x, val_acc, 'r', label='Validation acc')\n",
        "  plt.title('Training and validation accuracy')\n",
        "  plt.legend()\n",
        "\n",
        "  plt.subplot(1, 2, 2)\n",
        "  plt.plot(x, loss, 'b', label='Training loss')\n",
        "  plt.plot(x, val_loss, 'r', label='Validation loss')\n",
        "  plt.title('Training and validation loss')\n",
        "  plt.legend()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "GG5UJlgsK0z6"
      },
      "outputs": [],
      "source": [
        "# Utility function: Display model score(Loss & Accuracy) across all sets.\n",
        "\n",
        "def display_model_score(model, train, val, test, batch_size):\n",
        "\n",
        "  train_score = model.evaluate(train[0], train[1], batch_size=batch_size, verbose=1)\n",
        "  print('Train loss: ', train_score[0])\n",
        "  print('Train accuracy: ', train_score[1])\n",
        "  print('-'*70)\n",
        "\n",
        "  val_score = model.evaluate(val[0], val[1], batch_size=batch_size, verbose=1)\n",
        "  print('Val loss: ', val_score[0])\n",
        "  print('Val accuracy: ', val_score[1])\n",
        "  print('-'*70)\n",
        "  \n",
        "  test_score = model.evaluate(test[0], test[1], batch_size=batch_size, verbose=1)\n",
        "  print('Test loss: ', test_score[0])\n",
        "  print('Test accuracy: ', test_score[1])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0GQ3x4L6hKOY"
      },
      "source": [
        "### Model 1: Bidirectional LSTM"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 319
        },
        "id": "ERfM44Zk0dcz",
        "outputId": "e8fa2245-e20d-422f-ec08-9b5e7c54f558"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model: \"model_2\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_3 (InputLayer)         (None, 100)               0         \n",
            "_________________________________________________________________\n",
            "embedding_3 (Embedding)      (None, 100, 128)          2688      \n",
            "_________________________________________________________________\n",
            "bidirectional_2 (Bidirection (None, 128)               99328     \n",
            "_________________________________________________________________\n",
            "dropout_2 (Dropout)          (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 1000)              129000    \n",
            "=================================================================\n",
            "Total params: 231,016\n",
            "Trainable params: 231,016\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "x_input = Input(shape=(100,))\n",
        "emb = Embedding(21, 128, input_length=max_length)(x_input)\n",
        "bi_rnn = Bidirectional(CuDNNLSTM(64, kernel_regularizer=l2(0.01), recurrent_regularizer=l2(0.01), bias_regularizer=l2(0.01)))(emb)\n",
        "x = Dropout(0.3)(bi_rnn)\n",
        "\n",
        "# softmax classifier\n",
        "x_output = Dense(1000, activation='softmax')(x)\n",
        "\n",
        "model1 = Model(inputs=x_input, outputs=x_output)\n",
        "model1.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n",
        "\n",
        "model1.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "DbpNEd2nmySw"
      },
      "outputs": [],
      "source": [
        "# Early Stopping\n",
        "es = EarlyStopping(monitor='val_loss', patience=3, verbose=1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "rojzazHe0hlN",
        "outputId": "17222b58-54be-47ac-90eb-a9cdd75bffe5"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Train on 439493 samples, validate on 54378 samples\n",
            "Epoch 1/50\n",
            "439493/439493 [==============================] - 78s 178us/step - loss: 5.5381 - acc: 0.1120 - val_loss: 3.5944 - val_acc: 0.3442\n",
            "Epoch 2/50\n",
            "439493/439493 [==============================] - 77s 176us/step - loss: 2.8564 - acc: 0.4733 - val_loss: 2.0159 - val_acc: 0.6706\n",
            "Epoch 3/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 1.9031 - acc: 0.6607 - val_loss: 1.3857 - val_acc: 0.7869\n",
            "Epoch 4/50\n",
            "439493/439493 [==============================] - 77s 176us/step - loss: 1.4633 - acc: 0.7459 - val_loss: 1.0834 - val_acc: 0.8410\n",
            "Epoch 5/50\n",
            "439493/439493 [==============================] - 77s 176us/step - loss: 1.2147 - acc: 0.7920 - val_loss: 0.8749 - val_acc: 0.8773\n",
            "Epoch 6/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 1.0633 - acc: 0.8202 - val_loss: 0.8003 - val_acc: 0.8885\n",
            "Epoch 7/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.9566 - acc: 0.8389 - val_loss: 0.6798 - val_acc: 0.9117\n",
            "Epoch 8/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.8871 - acc: 0.8519 - val_loss: 0.6555 - val_acc: 0.9140\n",
            "Epoch 9/50\n",
            "439493/439493 [==============================] - 77s 176us/step - loss: 0.8342 - acc: 0.8612 - val_loss: 0.6059 - val_acc: 0.9209\n",
            "Epoch 10/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.7958 - acc: 0.8671 - val_loss: 0.5734 - val_acc: 0.9268\n",
            "Epoch 11/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.7650 - acc: 0.8735 - val_loss: 0.5454 - val_acc: 0.9326\n",
            "Epoch 12/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.7413 - acc: 0.8780 - val_loss: 0.5222 - val_acc: 0.9375\n",
            "Epoch 13/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.7192 - acc: 0.8821 - val_loss: 0.5129 - val_acc: 0.9375\n",
            "Epoch 14/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.6990 - acc: 0.8856 - val_loss: 0.5042 - val_acc: 0.9390\n",
            "Epoch 15/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.6847 - acc: 0.8882 - val_loss: 0.4822 - val_acc: 0.9442\n",
            "Epoch 16/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.6703 - acc: 0.8914 - val_loss: 0.5056 - val_acc: 0.9357\n",
            "Epoch 17/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.6580 - acc: 0.8935 - val_loss: 0.4658 - val_acc: 0.9478\n",
            "Epoch 18/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.6470 - acc: 0.8962 - val_loss: 0.4405 - val_acc: 0.9522\n",
            "Epoch 19/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.6367 - acc: 0.8976 - val_loss: 0.4493 - val_acc: 0.9489\n",
            "Epoch 20/50\n",
            "439493/439493 [==============================] - 77s 176us/step - loss: 0.6283 - acc: 0.8991 - val_loss: 0.4332 - val_acc: 0.9523\n",
            "Epoch 21/50\n",
            "439493/439493 [==============================] - 77s 176us/step - loss: 0.6163 - acc: 0.9015 - val_loss: 0.4241 - val_acc: 0.9548\n",
            "Epoch 22/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.6093 - acc: 0.9026 - val_loss: 0.4167 - val_acc: 0.9549\n",
            "Epoch 23/50\n",
            "439493/439493 [==============================] - 77s 176us/step - loss: 0.6004 - acc: 0.9043 - val_loss: 0.4188 - val_acc: 0.9547\n",
            "Epoch 24/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.5955 - acc: 0.9054 - val_loss: 0.4193 - val_acc: 0.9547\n",
            "Epoch 25/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.5892 - acc: 0.9063 - val_loss: 0.4024 - val_acc: 0.9585\n",
            "Epoch 26/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.5794 - acc: 0.9081 - val_loss: 0.4065 - val_acc: 0.9560\n",
            "Epoch 27/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.5737 - acc: 0.9090 - val_loss: 0.3896 - val_acc: 0.9599\n",
            "Epoch 28/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.5653 - acc: 0.9106 - val_loss: 0.3972 - val_acc: 0.9584\n",
            "Epoch 29/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.5636 - acc: 0.9113 - val_loss: 0.3953 - val_acc: 0.9582\n",
            "Epoch 30/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.5598 - acc: 0.9121 - val_loss: 0.3765 - val_acc: 0.9623\n",
            "Epoch 31/50\n",
            "439493/439493 [==============================] - 77s 174us/step - loss: 0.5529 - acc: 0.9135 - val_loss: 0.3907 - val_acc: 0.9575\n",
            "Epoch 32/50\n",
            "439493/439493 [==============================] - 77s 175us/step - loss: 0.5441 - acc: 0.9150 - val_loss: 0.3771 - val_acc: 0.9606\n",
            "Epoch 33/50\n",
            "439493/439493 [==============================] - 77s 176us/step - loss: 0.5418 - acc: 0.9150 - val_loss: 0.3870 - val_acc: 0.9577\n",
            "Epoch 00033: early stopping\n"
          ]
        }
      ],
      "source": [
        "history1 = model1.fit(\n",
        "    train_pad, y_train,\n",
        "    epochs=50, batch_size=256,\n",
        "    validation_data=(val_pad, y_val),\n",
        "    callbacks=[es]\n",
        "    )"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "AzBvM1RHqxgg"
      },
      "outputs": [],
      "source": [
        "# saving model weights.\n",
        "model1.save_weights('drive/My Drive/Case_Study/pfam/model1.h5')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 337
        },
        "id": "FvmHIj9gG9VP",
        "outputId": "d66c6376-d696-4bb6-f11f-caff6e8783ab"
      },
      "outputs": [
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsYAAAFACAYAAAC/abrtAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 864x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_history(history1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 202
        },
        "id": "VlW0i0DiLVI6",
        "outputId": "30afcf5f-1213-4f66-ff37-4faf92fdd404"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "439493/439493 [==============================] - 28s 65us/step\n",
            "Train loss:  0.36330516427409587\n",
            "Train accuracy:  0.9645910173696531\n",
            "----------------------------------------------------------------------\n",
            "54378/54378 [==============================] - 3s 63us/step\n",
            "Val loss:  0.3869630661736021\n",
            "Val accuracy:  0.9577034830108782\n",
            "----------------------------------------------------------------------\n",
            "54378/54378 [==============================] - 3s 64us/step\n",
            "Test loss:  0.3869193921893196\n",
            "Test accuracy:  0.9587149214887501\n"
          ]
        }
      ],
      "source": [
        "display_model_score(model1,\n",
        "    [train_pad, y_train],\n",
        "    [val_pad, y_val],\n",
        "    [test_pad, y_test],\n",
        "    256)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8Lip5vCthFn_"
      },
      "source": [
        "### Model 2: ProtCNN (https://www.biorxiv.org/content/10.1101/626507v4.full)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wxdIXVy-fDyz"
      },
      "source": [
        "<br/>\n",
        "<center><img src='https://www.biorxiv.org/content/biorxiv/early/2019/07/15/626507/F6.large.jpg' height='470' width='700'></center>\n",
        "\n",
        "<br/>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BgX7fQhjkkcD"
      },
      "source": [
        "- One hot encoded unaligned sequence of amino acids is passed as the input to the network with zero padding.\n",
        "\n",
        "- This network uses residual blocks inspired from [ResNet](https://arxiv.org/abs/1512.03385) architecture which also includes dilated convolutions offering larger receptive field without increasing number of model parameters."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "otNM0HBpZ_CL"
      },
      "outputs": [],
      "source": [
        "def residual_block(data, filters, d_rate):\n",
        "  \"\"\"\n",
        "  _data: input\n",
        "  _filters: convolution filters\n",
        "  _d_rate: dilation rate\n",
        "  \"\"\"\n",
        "\n",
        "  shortcut = data\n",
        "\n",
        "  bn1 = BatchNormalization()(data)\n",
        "  act1 = Activation('relu')(bn1)\n",
        "  conv1 = Conv1D(filters, 1, dilation_rate=d_rate, padding='same', kernel_regularizer=l2(0.001))(act1)\n",
        "\n",
        "  #bottleneck convolution\n",
        "  bn2 = BatchNormalization()(conv1)\n",
        "  act2 = Activation('relu')(bn2)\n",
        "  conv2 = Conv1D(filters, 3, padding='same', kernel_regularizer=l2(0.001))(act2)\n",
        "\n",
        "  #skip connection\n",
        "  x = Add()([conv2, shortcut])\n",
        "\n",
        "  return x"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 857
        },
        "id": "56zf2x_xDCY4",
        "outputId": "a2e8a399-2362-4c01-b395-ebb58b5ffbbe"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model: \"model_3\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_5 (InputLayer)            (None, 100, 21)      0                                            \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_6 (Conv1D)               (None, 100, 128)     2816        input_5[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_5 (BatchNor (None, 100, 128)     512         conv1d_6[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "activation_5 (Activation)       (None, 100, 128)     0           batch_normalization_5[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_7 (Conv1D)               (None, 100, 128)     16512       activation_5[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_6 (BatchNor (None, 100, 128)     512         conv1d_7[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "activation_6 (Activation)       (None, 100, 128)     0           batch_normalization_6[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_8 (Conv1D)               (None, 100, 128)     49280       activation_6[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "add_3 (Add)                     (None, 100, 128)     0           conv1d_8[0][0]                   \n",
            "                                                                 conv1d_6[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_7 (BatchNor (None, 100, 128)     512         add_3[0][0]                      \n",
            "__________________________________________________________________________________________________\n",
            "activation_7 (Activation)       (None, 100, 128)     0           batch_normalization_7[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_9 (Conv1D)               (None, 100, 128)     16512       activation_7[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "batch_normalization_8 (BatchNor (None, 100, 128)     512         conv1d_9[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "activation_8 (Activation)       (None, 100, 128)     0           batch_normalization_8[0][0]      \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_10 (Conv1D)              (None, 100, 128)     49280       activation_8[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "add_4 (Add)                     (None, 100, 128)     0           conv1d_10[0][0]                  \n",
            "                                                                 add_3[0][0]                      \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_1 (MaxPooling1D)  (None, 33, 128)      0           add_4[0][0]                      \n",
            "__________________________________________________________________________________________________\n",
            "dropout_3 (Dropout)             (None, 33, 128)      0           max_pooling1d_1[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "flatten_1 (Flatten)             (None, 4224)         0           dropout_3[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_3 (Dense)                 (None, 1000)         4225000     flatten_1[0][0]                  \n",
            "==================================================================================================\n",
            "Total params: 4,361,448\n",
            "Trainable params: 4,360,424\n",
            "Non-trainable params: 1,024\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "# model\n",
        "\n",
        "x_input = Input(shape=(100, 21))\n",
        "\n",
        "#initial conv\n",
        "conv = Conv1D(128, 1, padding='same')(x_input) \n",
        "\n",
        "# per-residue representation\n",
        "res1 = residual_block(conv, 128, 2)\n",
        "res2 = residual_block(res1, 128, 3)\n",
        "\n",
        "x = MaxPooling1D(3)(res2)\n",
        "x = Dropout(0.5)(x)\n",
        "\n",
        "# softmax classifier\n",
        "x = Flatten()(x)\n",
        "x_output = Dense(1000, activation='softmax', kernel_regularizer=l2(0.0001))(x)\n",
        "\n",
        "model2 = Model(inputs=x_input, outputs=x_output)\n",
        "model2.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n",
        "\n",
        "model2.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "_gJi_tLlxvfC"
      },
      "outputs": [],
      "source": [
        "# Early Stopping\n",
        "es = EarlyStopping(monitor='val_loss', patience=3, verbose=1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 370
        },
        "id": "GeRJYs0wxzBM",
        "outputId": "8dbf7565-ef7e-49bb-a063-94a92ac61917"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Train on 439493 samples, validate on 54378 samples\n",
            "Epoch 1/10\n",
            "439493/439493 [==============================] - 120s 272us/step - loss: 0.9157 - acc: 0.9294 - val_loss: 0.4761 - val_acc: 0.9838\n",
            "Epoch 2/10\n",
            "439493/439493 [==============================] - 114s 260us/step - loss: 0.4438 - acc: 0.9788 - val_loss: 0.4545 - val_acc: 0.9831\n",
            "Epoch 3/10\n",
            "439493/439493 [==============================] - 114s 260us/step - loss: 0.4331 - acc: 0.9814 - val_loss: 0.4443 - val_acc: 0.9848\n",
            "Epoch 4/10\n",
            "439493/439493 [==============================] - 114s 260us/step - loss: 0.4198 - acc: 0.9825 - val_loss: 0.4279 - val_acc: 0.9863\n",
            "Epoch 5/10\n",
            "439493/439493 [==============================] - 114s 260us/step - loss: 0.4098 - acc: 0.9830 - val_loss: 0.4314 - val_acc: 0.9859\n",
            "Epoch 6/10\n",
            "439493/439493 [==============================] - 114s 260us/step - loss: 0.4033 - acc: 0.9834 - val_loss: 0.4181 - val_acc: 0.9867\n",
            "Epoch 7/10\n",
            "439493/439493 [==============================] - 114s 260us/step - loss: 0.3943 - acc: 0.9840 - val_loss: 0.4180 - val_acc: 0.9862\n",
            "Epoch 8/10\n",
            "439493/439493 [==============================] - 114s 259us/step - loss: 0.3906 - acc: 0.9842 - val_loss: 0.4086 - val_acc: 0.9858\n",
            "Epoch 9/10\n",
            "439493/439493 [==============================] - 114s 260us/step - loss: 0.3829 - acc: 0.9845 - val_loss: 0.4015 - val_acc: 0.9866\n",
            "Epoch 10/10\n",
            "439493/439493 [==============================] - 114s 260us/step - loss: 0.3841 - acc: 0.9844 - val_loss: 0.3962 - val_acc: 0.9887\n"
          ]
        }
      ],
      "source": [
        "history2 = model2.fit(\n",
        "    train_ohe, y_train,\n",
        "    epochs=10, batch_size=256,\n",
        "    validation_data=(val_ohe, y_val),\n",
        "    callbacks=[es]\n",
        "    )"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "kK7sr-pBIR2w"
      },
      "outputs": [],
      "source": [
        "# saving model weights.\n",
        "model2.save_weights('drive/My Drive/Case_Study/pfam/model2.h5')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 337
        },
        "id": "4uXtCBnR2RGv",
        "outputId": "965f54fb-113b-4a9a-f4d6-024698f0f82d"
      },
      "outputs": [
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAswAAAFACAYAAACoSyokAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nERERUSf+P47UVp0dd25vAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 864x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_history(history2)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 202
        },
        "id": "z-T9Ula42U11",
        "outputId": "1a39a91d-68a6-4245-fcd2-b5f1818324b8"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "439493/439493 [==============================] - 38s 85us/step\n",
            "Train loss:  0.3558084576734698\n",
            "Train accuracy:  0.9969123512774948\n",
            "----------------------------------------------------------------------\n",
            "54378/54378 [==============================] - 5s 85us/step\n",
            "Val loss:  0.39615299251274316\n",
            "Val accuracy:  0.9886718893955224\n",
            "----------------------------------------------------------------------\n",
            "54378/54378 [==============================] - 5s 85us/step\n",
            "Test loss:  0.3949931418234982\n",
            "Test accuracy:  0.9882489242257847\n"
          ]
        }
      ],
      "source": [
        "display_model_score(\n",
        "    model2,\n",
        "    [train_ohe, y_train],\n",
        "    [val_ohe, y_val],\n",
        "    [test_ohe, y_test],\n",
        "    256)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3u42Zf2bLxmM"
      },
      "source": [
        "## 5. Conclusion"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 118
        },
        "id": "MyPhPwMnL0af",
        "outputId": "68557968-a045-4171-afd1-d5bb435d793e"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "+-------+--------------------+-----------+---------+----------+\n",
            "| Sr.no |       Model        | Train Acc | Val Acc | Test Acc |\n",
            "+-------+--------------------+-----------+---------+----------+\n",
            "|   1.  | Bidirectional LSTM |   0.964   |  0.957  |  0.958   |\n",
            "|   2.  |      ProtCNN       |   0.996   |  0.988  |  0.988   |\n",
            "+-------+--------------------+-----------+---------+----------+\n"
          ]
        }
      ],
      "source": [
        "x = PrettyTable()\n",
        "x.field_names = ['Sr.no', 'Model', 'Train Acc', 'Val Acc','Test Acc']\n",
        "\n",
        "x.add_row(['1.', 'Bidirectional LSTM', '0.964', '0.957', '0.958'])\n",
        "x.add_row(['2.', 'ProtCNN', '0.996', '0.988', '0.988'])\n",
        "\n",
        "print(x)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OY0HAjgRQwii"
      },
      "source": [
        "### Reference:\n",
        "\n",
        "- https://www.biorxiv.org/content/10.1101/626507v4.full"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "collapsed_sections": [
        "y_dDnHNus7SL",
        "KmJ2E-Krs_xt",
        "jtyg15A6ATsS",
        "ylHPrtMlBvyB",
        "xdkPtpHiB45Y",
        "CUfCZp0ICwaF",
        "GCPrxSoI5jQZ",
        "lqRU2FBxezKq",
        "sRXFL8RwlZGZ",
        "7XvpRxecd_B1",
        "LFxzvA1oeJGm",
        "08mVHtLVfCmp",
        "LHh3AXwNp7Yv"
      ],
      "machine_shape": "hm",
      "name": "Copy of Pfam_protein_sequence_classification.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.5"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}