covid_primer_design.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "covid_primer_design.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "machine_shape": "hm",
      "mount_file_id": "1qXWpeRZmfFWGwMnAKtIQUkdSEPl7ASC5",
      "authorship_tag": "ABX9TyMVvDSmJjvc8Uk60WRbasmN",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/darchon30704/primers-sars-cov-2/blob/main/covid_primer_design.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hAYRWPBzlq2e"
      },
      "source": [
        "https://github.com/steppenwolf0/primers-sars-cov-2\n",
        "\n",
        "https://www.nature.com/articles/s41598-020-80363-5#Sec2\n",
        "\n",
        "![image.png](data:image/png;base64,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)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LAQAdsLyk85e"
      },
      "source": [
        "path = '/content/drive/MyDrive/Colab Notebooks/Genomic/primers-sars-cov-2'"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nNAyQAvjmF3s"
      },
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "import os\n",
        "import sys\n",
        "import tarfile\n",
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "geLqHGF4mF6p"
      },
      "source": [
        "# data_1503_featureSpace\n",
        "data_0 = pd.read_csv(path+'/data_1503_featureSpace/data/data_0.csv')\n",
        "features_0 = pd.read_csv(path+'/data_1503_featureSpace/data/features_0.csv')\n",
        "labels_0 = pd.read_csv(path+'/data_1503_featureSpace/data/labels.csv')\n",
        "\n",
        "sequences= pd.read_csv(path+'/CNN/data/sequences.csv')\n",
        "labels= pd.read_csv(path+'/CNN/data/labels.csv')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RpnTIT6WOjUY"
      },
      "source": [
        "ss= pd.read_csv(path+'/CNN/data/sequences.csv', names=[\"seq\"])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 419
        },
        "id": "YT5cyp0pRDWG",
        "outputId": "628c9762-57a8-4727-c6c8-81fe38b8e236"
      },
      "source": [
        "ss\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>seq</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>ATTAAAGGTTTATACCTTCCCAGGTAACAAACCAACCAACTTTCGA...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>GATTTAAGTGAATAGCTTGGCTATCTCACTTTCCCTCGTTCTCTTG...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>GATTTAAGTGAATAGCTTGGCTATCTCACTTTCCCTCGTTCTCTTG...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>GATTTAAGTGAATAGCTTGGCTATCTCACTTCCCCTCGTTCTCTTG...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>GATTTAAGTGAATAGCTTGGCTATCTCACTTCCCCTCGTTCTCTTG...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>548</th>\n",
              "      <td>ATTAAAGGTTTATACCTTCCCAGGTAACAAACCAACCAACTTTCGA...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>549</th>\n",
              "      <td>ATTAAAGGTTTATACCTTCCCAGGTAACAAACCAACCAACTTTCGA...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>550</th>\n",
              "      <td>TAAAGGTTTATACCTTCCCAGGTAACAAACCAACCAACTTTCGATC...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>551</th>\n",
              "      <td>CAGGTAACAAACCAACCAACTTTCGATCTCTTGTAGATCTGTTCTC...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>552</th>\n",
              "      <td>ACCTTCCCAGGTAACAAACCAACCAACTTTCGATCTCTTGTAGATC...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>553 rows × 1 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                   seq\n",
              "0    ATTAAAGGTTTATACCTTCCCAGGTAACAAACCAACCAACTTTCGA...\n",
              "1    GATTTAAGTGAATAGCTTGGCTATCTCACTTTCCCTCGTTCTCTTG...\n",
              "2    GATTTAAGTGAATAGCTTGGCTATCTCACTTTCCCTCGTTCTCTTG...\n",
              "3    GATTTAAGTGAATAGCTTGGCTATCTCACTTCCCCTCGTTCTCTTG...\n",
              "4    GATTTAAGTGAATAGCTTGGCTATCTCACTTCCCCTCGTTCTCTTG...\n",
              "..                                                 ...\n",
              "548  ATTAAAGGTTTATACCTTCCCAGGTAACAAACCAACCAACTTTCGA...\n",
              "549  ATTAAAGGTTTATACCTTCCCAGGTAACAAACCAACCAACTTTCGA...\n",
              "550  TAAAGGTTTATACCTTCCCAGGTAACAAACCAACCAACTTTCGATC...\n",
              "551  CAGGTAACAAACCAACCAACTTTCGATCTCTTGTAGATCTGTTCTC...\n",
              "552  ACCTTCCCAGGTAACAAACCAACCAACTTTCGATCTCTTGTAGATC...\n",
              "\n",
              "[553 rows x 1 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jiBuYGjyR56D",
        "outputId": "afbb375f-eb4b-42a3-da0a-4f40461811dd"
      },
      "source": [
        "max(ss.seq.str.len())"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "31029"
            ]
          },
          "metadata": {},
          "execution_count": 52
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VffLtFdMAWrt"
      },
      "source": [
        "## crossValB.py"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D6pC9BLRy0f2"
      },
      "source": [
        "def oneHot(array, size):\n",
        "\toutput=[]\n",
        "\t# print('len(array): ',len(array))\n",
        "\t# print('size: ',size)\n",
        "\t# print('array: ',array)\n",
        "\tfor i in range (len(array)):\n",
        "\t\ttemp=np.zeros(size)\n",
        "\t\ttemp[int(array[i])]=1\n",
        "\t\toutput.append(temp)\n",
        "\t# print('output len:',len(output))\n",
        "\t# print('output len 0:',len(output[0]))\n",
        "\treturn np.array(output)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PAgvL1bkmGAx"
      },
      "source": [
        "def get_Info(indexVar,dataFolder):\n",
        "\t#examples=8129\n",
        "\n",
        "\tdata = pd.read_csv(dataFolder+'sequences.csv', header=None).values.ravel()\n",
        "\tprint('data set', data.shape)\n",
        "\n",
        "\t\n",
        "\tlabels=pd.read_csv(dataFolder+'labels.csv', header=None).values.ravel()\n",
        "\t\n",
        "\tlabelSize=int(np.max(labels)+1)\n",
        "\t\n",
        "\tmax=0\n",
        "\tfor i in range (0,len(data)):\n",
        "\t\tif (len(data[i])>max):\n",
        "\t\t\tmax=len(data[i])\n",
        "\t\n",
        "\tvectorSize=max\n",
        "\tprint('vectorSize', vectorSize)\n",
        "\t\n",
        "\t#print(labels)\n",
        "\ttestIndex=pd.read_csv(dataFolder+'index/'+str(indexVar)+'test_index.txt', header=None).values.ravel()\n",
        "\tvalIndex=pd.read_csv(dataFolder+'index/'+str(indexVar)+'val_index.txt', header=None).values.ravel()\n",
        "\ttrainIndex=pd.read_csv(dataFolder+'index/'+str(indexVar)+'train_index.txt', header=None).values.ravel()\n",
        "\t\n",
        "\ttestIndex=testIndex.astype(int)\n",
        "\tvalIndex=valIndex.astype(int)\n",
        "\ttrainIndex=trainIndex.astype(int)\n",
        "\t\t\n",
        "\ttrain=[]\n",
        "\ttest=[]\n",
        "\tvalid=[]\n",
        "\t\n",
        "\t\n",
        "\ttrainLabels=[]\n",
        "\ttestLabels=[]\n",
        "\tvalidLabels=[]\n",
        "\t#test***************************************************************************\n",
        "\tfor i in range (0,len(testIndex)):\n",
        "\t\ttestLabels.append(labels[testIndex[i]])\n",
        "\t\ttest.append(data[testIndex[i]])\n",
        "\t#valid***************************************************************************\n",
        "\tfor i in range (0,len(valIndex)):\n",
        "\t\tvalidLabels.append(labels[valIndex[i]])\n",
        "\t\tvalid.append(data[valIndex[i]])\n",
        "\t#train***************************************************************************\n",
        "\tfor i in range (0,len(trainIndex)):\n",
        "\t\ttrainLabels.append(labels[trainIndex[i]])\n",
        "\t\ttrain.append(data[trainIndex[i]])\n",
        "\t\n",
        "\n",
        "\ttestLabels=np.array(testLabels)\n",
        "\tvalidLabels=np.array(validLabels)\n",
        "\ttrainLabels=np.array(trainLabels)\n",
        "\n",
        "\tprint('train length:',len(train))\n",
        "\tprint('train length 0:',len(train[0]))\n",
        "\tprint('train label shape:', trainLabels.shape)\n",
        "\tprint('valid length:', len(valid))\n",
        "\tprint('valid label shape:', validLabels.shape)\n",
        "\tprint('test length:',len(test))\n",
        "\tprint('test label shape:',testLabels.shape)\n",
        "\t\n",
        "\tprint(labelSize)\n",
        "\toneHot_train_labels=oneHot(trainLabels,labelSize)\n",
        "\tprint('oneHot_train_labels.shape: ',oneHot_train_labels.shape)\n",
        "\n",
        "\toneHot_valid_labels=oneHot(validLabels,labelSize)\n",
        "\tprint('oneHot_valid_labels.shape:',oneHot_valid_labels.shape)\n",
        "\n",
        "\toneHot_test_labels=oneHot(testLabels,labelSize)\n",
        "\tprint('oneHot_test_labels.shape',oneHot_test_labels.shape)\n",
        "\n",
        "\n",
        "\treturn(test,oneHot_test_labels,valid,oneHot_valid_labels,train,oneHot_train_labels,labelSize,vectorSize)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YBQUhjSEmF9z",
        "outputId": "46c968e3-1d40-4ed9-d4bb-1f9d4687c6fc"
      },
      "source": [
        "kfoldIndex=0\n",
        "dataFolder=path+'/CNN/data/'\n",
        "\n",
        "(test_dataset_Flat,oneHot_test_labels,valid_dataset_Flat,oneHot_valid_labels,\n",
        "\ttrain_dataset_Flat,oneHot_train_labels,labelSize,vectorSize)=get_Info(\n",
        "\tkfoldIndex,dataFolder)\n",
        " "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "data set (553,)\n",
            "vectorSize 31029\n",
            "train length: 441\n",
            "train length 0: 29727\n",
            "train label shape: (441,)\n",
            "valid length: 55\n",
            "valid label shape: (55,)\n",
            "test length: 57\n",
            "test label shape: (57,)\n",
            "5\n",
            "oneHot_train_labels.shape:  (441, 5)\n",
            "oneHot_valid_labels.shape: (55, 5)\n",
            "oneHot_test_labels.shape (57, 5)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NED2roo9vx85",
        "outputId": "fb8d7654-bfb4-4f29-d2fb-4f6ba1fc5a02"
      },
      "source": [
        "vectorSize"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "31029"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gt3_Ac56vyAR"
      },
      "source": [
        "def get_InfoTotal(dataFolder):\n",
        "\t#examples=8129\n",
        "\n",
        "\tdataSeq = pd.read_csv(dataFolder+'sequences.csv', header=None).values.ravel()\n",
        "\tdata=dataSeq\n",
        "\tprint('data set', data.shape)\n",
        "\n",
        "\t\n",
        "\tlabels=pd.read_csv(dataFolder+'labels.csv', header=None).values.ravel()\n",
        "\t\n",
        "\tlabelSize=int(np.max(labels)+1)\n",
        "\t\n",
        "\tmax=0\n",
        "\tfor i in range (0,len(data)):\n",
        "\t\tif (len(data[i])>max):\n",
        "\t\t\tmax=len(data[i])\n",
        "\t\n",
        "\tvectorSize=max\n",
        "\tprint('vectorSize', vectorSize)\n",
        "\t\n",
        "\toutData=[]\n",
        "\toutLabels=[]\n",
        "\tfor i in range (len(data)):\n",
        "\t\tsample=np.zeros(vectorSize)\n",
        "\t\tfor j in range (0, len(data[i])):\n",
        "\t\t\tif(data[i][j]=='C'):\n",
        "\t\t\t\tsample[j]=0.25\n",
        "\t\t\telif(data[i][j]=='T'):\n",
        "\t\t\t\tsample[j]=0.50\n",
        "\t\t\telif(data[i][j]=='G'):\n",
        "\t\t\t\tsample[j]=0.75\n",
        "\t\t\telif(data[i][j]=='A'):\n",
        "\t\t\t\tsample[j]=1.0\n",
        "\t\t\telse:\n",
        "\t\t\t\tsample[j]=0.0\n",
        "\t\toutData.append(sample)\n",
        "\t\toutLabels.append(labels[i])\n",
        "\t\n",
        "\tdata=np.array(outData)\n",
        "\tlabels=np.array(outLabels)\n",
        "\t\n",
        "\tsort_index = np.argsort(labels)\n",
        "\n",
        "\tsize=data.shape\n",
        "\t\n",
        "\tdataSorted=np.zeros((size[0],size[1]))\n",
        "\tlabelsSorted=np.zeros(size[0])\n",
        "\tsequencesSorted=[]\n",
        "\n",
        "\tfor i in range(0,len(labels)):\n",
        "\t\tlabelsSorted[i]=labels[sort_index[i]]\n",
        "\t\tsequencesSorted.append(dataSeq[sort_index[i]])\n",
        "\t\tdataSorted[i]=data[sort_index[i]]\n",
        "\t\n",
        "\tpd.DataFrame(sequencesSorted).to_csv(path+'/CNN/data/filters/sequencesSorted.csv', header=None, index =None)\n",
        "\t#pd.DataFrame(dataSorted).to_csv(\"../data/filters/sortedData.csv\", header=None, index =None)\n",
        "\tpd.DataFrame(labelsSorted).to_csv(path+'/CNN/data/filters/labelsSorted.csv', header=None, index =None)\n",
        "\t\n",
        "\n",
        "\tprint(labelSize)\n",
        "\toneHot_labels=oneHot(labelsSorted,labelSize)\n",
        "\tprint(oneHot_labels.shape)\n",
        "\n",
        "\treturn(dataSorted,oneHot_labels)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8pKwoIFcvyDN",
        "outputId": "70142991-2f93-434c-bc76-f4ab39079e1e"
      },
      "source": [
        "(data,oneHotLabels)=get_InfoTotal(path+'/CNN/data/')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "data set (553,)\n",
            "vectorSize 31029\n",
            "5\n",
            "(553, 5)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mhM0vx8lpt_q",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 419
        },
        "outputId": "07c2528c-5592-42de-9952-b8c083c22d68"
      },
      "source": [
        "pd.DataFrame(oneHotLabels)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>548</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>549</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>550</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>551</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>552</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>553 rows × 5 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "       0    1    2    3    4\n",
              "0    1.0  0.0  0.0  0.0  0.0\n",
              "1    1.0  0.0  0.0  0.0  0.0\n",
              "2    1.0  0.0  0.0  0.0  0.0\n",
              "3    1.0  0.0  0.0  0.0  0.0\n",
              "4    1.0  0.0  0.0  0.0  0.0\n",
              "..   ...  ...  ...  ...  ...\n",
              "548  0.0  0.0  0.0  0.0  1.0\n",
              "549  0.0  0.0  0.0  0.0  1.0\n",
              "550  0.0  0.0  0.0  0.0  1.0\n",
              "551  0.0  0.0  0.0  0.0  1.0\n",
              "552  0.0  0.0  0.0  0.0  1.0\n",
              "\n",
              "[553 rows x 5 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WeFHhexty0Rj"
      },
      "source": [
        "filter_0 = pd.read_csv(path+'/CNN/data/filters/filter_0.csv', header=None)\n",
        "DNA_0 = pd.read_csv(path+'/CNN/data/filters/dataDNAFeatures_0.csv', header=None)\n",
        "featsV_0 = pd.read_csv(path+'/CNN/data/filters/featsVector_0.csv', header=None)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 439
        },
        "id": "aomKT0ujT8Ez",
        "outputId": "950e27e8-1a9a-4299-a498-0e5206a1ba12"
      },
      "source": [
        "filter_0"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "      <th>5</th>\n",
              "      <th>6</th>\n",
              "      <th>7</th>\n",
              "      <th>8</th>\n",
              "      <th>9</th>\n",
              "      <th>10</th>\n",
              "      <th>11</th>\n",
              "      <th>12</th>\n",
              "      <th>13</th>\n",
              "      <th>14</th>\n",
              "      <th>15</th>\n",
              "      <th>16</th>\n",
              "      <th>17</th>\n",
              "      <th>18</th>\n",
              "      <th>19</th>\n",
              "      <th>20</th>\n",
              "      <th>21</th>\n",
              "      <th>22</th>\n",
              "      <th>23</th>\n",
              "      <th>24</th>\n",
              "      <th>25</th>\n",
              "      <th>26</th>\n",
              "      <th>27</th>\n",
              "      <th>28</th>\n",
              "      <th>29</th>\n",
              "      <th>30</th>\n",
              "      <th>31</th>\n",
              "      <th>32</th>\n",
              "      <th>33</th>\n",
              "      <th>34</th>\n",
              "      <th>35</th>\n",
              "      <th>36</th>\n",
              "      <th>37</th>\n",
              "      <th>38</th>\n",
              "      <th>39</th>\n",
              "      <th>...</th>\n",
              "      <th>30989</th>\n",
              "      <th>30990</th>\n",
              "      <th>30991</th>\n",
              "      <th>30992</th>\n",
              "      <th>30993</th>\n",
              "      <th>30994</th>\n",
              "      <th>30995</th>\n",
              "      <th>30996</th>\n",
              "      <th>30997</th>\n",
              "      <th>30998</th>\n",
              "      <th>30999</th>\n",
              "      <th>31000</th>\n",
              "      <th>31001</th>\n",
              "      <th>31002</th>\n",
              "      <th>31003</th>\n",
              "      <th>31004</th>\n",
              "      <th>31005</th>\n",
              "      <th>31006</th>\n",
              "      <th>31007</th>\n",
              "      <th>31008</th>\n",
              "      <th>31009</th>\n",
              "      <th>31010</th>\n",
              "      <th>31011</th>\n",
              "      <th>31012</th>\n",
              "      <th>31013</th>\n",
              "      <th>31014</th>\n",
              "      <th>31015</th>\n",
              "      <th>31016</th>\n",
              "      <th>31017</th>\n",
              "      <th>31018</th>\n",
              "      <th>31019</th>\n",
              "      <th>31020</th>\n",
              "      <th>31021</th>\n",
              "      <th>31022</th>\n",
              "      <th>31023</th>\n",
              "      <th>31024</th>\n",
              "      <th>31025</th>\n",
              "      <th>31026</th>\n",
              "      <th>31027</th>\n",
              "      <th>31028</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.051684</td>\n",
              "      <td>0.091732</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.118354</td>\n",
              "      <td>0.011225</td>\n",
              "      <td>0.179532</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.041589</td>\n",
              "      <td>0.000341</td>\n",
              "      <td>0.04197</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.213936</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.194765</td>\n",
              "      <td>...</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.051684</td>\n",
              "      <td>0.091732</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.118354</td>\n",
              "      <td>0.011225</td>\n",
              "      <td>0.179532</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.041589</td>\n",
              "      <td>0.000341</td>\n",
              "      <td>0.04197</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.213936</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.194765</td>\n",
              "      <td>...</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.173915</td>\n",
              "      <td>0.052191</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.118354</td>\n",
              "      <td>0.011225</td>\n",
              "      <td>0.179532</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.041589</td>\n",
              "      <td>0.000341</td>\n",
              "      <td>0.041970</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.213936</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.194765</td>\n",
              "      <td>0.296246</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.031243</td>\n",
              "      <td>0.070362</td>\n",
              "      <td>0.098038</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.041589</td>\n",
              "      <td>0.000341</td>\n",
              "      <td>0.041970</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.213936</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.194765</td>\n",
              "      <td>0.296246</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.064015</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.09311</td>\n",
              "      <td>0.097173</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.054954</td>\n",
              "      <td>0.064544</td>\n",
              "      <td>...</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0.027175</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.006380</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.102870</td>\n",
              "      <td>0.146194</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.064015</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.093110</td>\n",
              "      <td>0.097173</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.054954</td>\n",
              "      <td>0.064544</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.027886</td>\n",
              "      <td>0.107019</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.088125</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.015039</td>\n",
              "      <td>0.128591</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>548</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.053295</td>\n",
              "      <td>0.030184</td>\n",
              "      <td>0.011094</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.021279</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.155960</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.018398</td>\n",
              "      <td>0.090009</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.179951</td>\n",
              "      <td>0.171052</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.098708</td>\n",
              "      <td>0.062487</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.116604</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.018398</td>\n",
              "      <td>0.105128</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>549</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.003754</td>\n",
              "      <td>0.040798</td>\n",
              "      <td>0.015384</td>\n",
              "      <td>0.036627</td>\n",
              "      <td>0.063079</td>\n",
              "      <td>0.000572</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.027886</td>\n",
              "      <td>0.107019</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.088125</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.015039</td>\n",
              "      <td>0.128591</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.031467</td>\n",
              "      <td>0.089429</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.094154</td>\n",
              "      <td>...</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>550</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.005192</td>\n",
              "      <td>0.091989</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.018398</td>\n",
              "      <td>0.105128</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.301053</td>\n",
              "      <td>0.186335</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.021128</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.132411</td>\n",
              "      <td>0.108031</td>\n",
              "      <td>0.022969</td>\n",
              "      <td>0.025188</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.027886</td>\n",
              "      <td>0.107019</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.088125</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>551</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.024125</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.026186</td>\n",
              "      <td>0.131038</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.028760</td>\n",
              "      <td>0.006051</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.031671</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.074487</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.018375</td>\n",
              "      <td>0.242041</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.018398</td>\n",
              "      <td>0.105128</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.301053</td>\n",
              "      <td>0.186335</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>552</th>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.024125</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.026186</td>\n",
              "      <td>0.131038</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.028760</td>\n",
              "      <td>0.006051</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.031671</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.074487</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.018375</td>\n",
              "      <td>0.242041</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.018398</td>\n",
              "      <td>0.105128</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.00000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.301053</td>\n",
              "      <td>0.186335</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>...</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "      <td>0.1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>553 rows × 31029 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "        0      1         2         3      ...  31025  31026  31027  31028\n",
              "0    0.000000    0.0  0.000000  0.000000  ...    0.1    0.1    0.1    0.1\n",
              "1    0.000000    0.0  0.000000  0.000000  ...    0.1    0.1    0.1    0.1\n",
              "2    0.000000    0.0  0.000000  0.000000  ...    0.1    0.1    0.1    0.1\n",
              "3    0.000000    0.0  0.000000  0.031243  ...    0.1    0.1    0.1    0.1\n",
              "4    0.027175    0.0  0.000000  0.000000  ...    0.1    0.1    0.1    0.1\n",
              "..        ...    ...       ...       ...  ...    ...    ...    ...    ...\n",
              "548  0.000000    0.0  0.053295  0.030184  ...    0.1    0.1    0.1    0.1\n",
              "549  0.000000    0.0  0.000000  0.003754  ...    0.1    0.1    0.1    0.1\n",
              "550  0.000000    0.0  0.000000  0.000000  ...    0.1    0.1    0.1    0.1\n",
              "551  0.000000    0.0  0.000000  0.000000  ...    0.1    0.1    0.1    0.1\n",
              "552  0.000000    0.0  0.000000  0.000000  ...    0.1    0.1    0.1    0.1\n",
              "\n",
              "[553 rows x 31029 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 419
        },
        "id": "ZGzdSf4UUKFh",
        "outputId": "3b9d3251-5b38-42b8-fc16-fba35c262cbd"
      },
      "source": [
        "features_0"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>0</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>ACCAACCAACTTTCGATCTCT</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>TTCTGCAGGCTGCTTACGGTT</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>AACGAGAAAACACACGTCCAA</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>ACTTGAACAGCCCTATGTGTT</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>AGTGGCTTACCGCAAGGTTCT</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3883</th>\n",
              "      <td>ATCCGAAGGCTATCAAGTGTG</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3884</th>\n",
              "      <td>AACTTGAACTTACCAGGCTGT</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3885</th>\n",
              "      <td>AGGAACACAAATCCTATCCAG</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3886</th>\n",
              "      <td>TACTACAATTTGCCTATTCTA</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3887</th>\n",
              "      <td>ACTTGTATTTCTTTATGCAGT</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>3888 rows × 1 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                          0\n",
              "0     ACCAACCAACTTTCGATCTCT\n",
              "1     TTCTGCAGGCTGCTTACGGTT\n",
              "2     AACGAGAAAACACACGTCCAA\n",
              "3     ACTTGAACAGCCCTATGTGTT\n",
              "4     AGTGGCTTACCGCAAGGTTCT\n",
              "...                     ...\n",
              "3883  ATCCGAAGGCTATCAAGTGTG\n",
              "3884  AACTTGAACTTACCAGGCTGT\n",
              "3885  AGGAACACAAATCCTATCCAG\n",
              "3886  TACTACAATTTGCCTATTCTA\n",
              "3887  ACTTGTATTTCTTTATGCAGT\n",
              "\n",
              "[3888 rows x 1 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 419
        },
        "id": "ryegeKc0zaD-",
        "outputId": "582c1d29-6e7b-4d6d-9237-fc7bdbc8445c"
      },
      "source": [
        "featsV_0"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "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>0</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>ACCAACCAACTTTCGATCTCT</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>TTCTGCAGGCTGCTTACGGTT</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>AACGAGAAAACACACGTCCAA</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>ACTTGAACAGCCCTATGTGTT</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>AGTGGCTTACCGCAAGGTTCT</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3893</th>\n",
              "      <td>ATCCGAAGGCTATCAAGTGTG</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3894</th>\n",
              "      <td>AACTTGAACTTACCAGGCTGT</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3895</th>\n",
              "      <td>AGGAACACAAATCCTATCCAG</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3896</th>\n",
              "      <td>TACTACAATTTGCCTATTCTA</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3897</th>\n",
              "      <td>ACTTGTATTTCTTTATGCAGT</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>3898 rows × 1 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                          0\n",
              "0     ACCAACCAACTTTCGATCTCT\n",
              "1     TTCTGCAGGCTGCTTACGGTT\n",
              "2     AACGAGAAAACACACGTCCAA\n",
              "3     ACTTGAACAGCCCTATGTGTT\n",
              "4     AGTGGCTTACCGCAAGGTTCT\n",
              "...                     ...\n",
              "3893  ATCCGAAGGCTATCAAGTGTG\n",
              "3894  AACTTGAACTTACCAGGCTGT\n",
              "3895  AGGAACACAAATCCTATCCAG\n",
              "3896  TACTACAATTTGCCTATTCTA\n",
              "3897  ACTTGTATTTCTTTATGCAGT\n",
              "\n",
              "[3898 rows x 1 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GTD3D-S80c_a"
      },
      "source": [
        "# CNN \n",
        "## genCreateFilters.py"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_JxR2FBFfJ2-"
      },
      "source": [
        "#maximum number of iterations\n",
        "iterMax=1000\n",
        "limit=1.00\n",
        "#regularization on the weights\n",
        "beta=0.001\n",
        "#version of the code\n",
        "version='gen1'\n",
        "#size of batch\n",
        "batchSize=50\n",
        "#Parameters*******************************************************************************\n",
        "w1=12 # 12 filters\n",
        "w4=196 # 196 rectified linear units\n",
        "h1=148 # 148-position interval of the original genome\n",
        "wd1=21 # 21-bps-long sequences. \n",
        "# There are 210 (31,029 divided by 148) max pooling features, \n",
        "# each one identifying the 21-bps sequence that obtained the highest value from the convolutional filter\n",
        "index=0\n",
        "kfoldIndex=0\n",
        "generation=0"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Bo6_TS8ORz6q",
        "outputId": "dc340d39-21cf-4f64-dfc8-3082abf02c5e"
      },
      "source": [
        "import tensorflow as tf\n",
        "#declare interactive session\n",
        "import tensorflow.compat.v1 as tf\n",
        "tf.disable_v2_behavior()\n",
        "sess=tf.InteractiveSession()\n",
        "#INPUT->CONV LAYER->CONV LAYER->CONV LAYER->RECT FLAT->RECT DROPOUT\n",
        "\n",
        "#function to declare easily the weights only by shape\n",
        "def weight_variable(shape):\n",
        "\tinitial = tf.truncated_normal(shape, stddev=0.1)\n",
        "\treturn tf.Variable(initial)\n",
        "#function to declare easily the bias only by shape\n",
        "def bias_variable(shape):\n",
        "\tinitial = tf.constant(0.1, shape=shape)\n",
        "\treturn tf.Variable(initial)\n",
        "\n",
        "#input variable\n",
        "x = tf.placeholder(tf.float32, [None, vectorSize])\n",
        "#keep probability to change from dropout 0.50 to 1.0 in validation and test\n",
        "keep_prob = tf.placeholder(tf.float32)\n",
        "#expected outputs variable\n",
        "y_ = tf.placeholder(tf.float32, [None, labelSize])\n",
        "\n",
        "#arrange the tensor as an image (1*31029) 1 channel\n",
        "x_image0 = tf.reshape(x, [-1,1,vectorSize,1])\n",
        "x_image = tf.transpose(x_image0, perm=[0,3,2,1])\n",
        "#arrange the tensor into 1 channels (1*31029)\n",
        "\n",
        "#1 LAYER*************************************************************************************\n",
        "#1 Convolutional Layer Explicit for regularization of the weights\n",
        "#weigth first layer 1 input channels, 12 output channels, 1x21 filter window size\n",
        "W_conv1 = weight_variable([1, wd1, 1, w1])\n",
        "#bias declaration the size has to be the same as the output channels 12\n",
        "b_conv1 = bias_variable([w1])\n",
        "#convolution (input weights) moving 1 step each time with a relu\n",
        "h_conv1 = tf.nn.relu(tf.nn.conv2d(x_image, W_conv1, \n",
        "\tstrides=[1, 1, 1, 1], padding='SAME') + b_conv1)\n",
        "#max pooling with a 148 width window size, moving 148 in width by step\n",
        "h_pool1 = tf.nn.max_pool(h_conv1, ksize=[1, 1, h1, 1],\n",
        "\tstrides=[1, 1, h1, 1], padding='SAME')\n",
        "#output=545/4\n",
        "#1 LAYER*************************************************************************************\n",
        "\n",
        "#Rectifier LAYER*****************************************************************************\n",
        "#calculated coefficient for the flattening from the size of the 3 convolutional layer\n",
        "coef=int (h_pool1.get_shape()[1]*h_pool1.get_shape()[2]*h_pool1.get_shape()[3])\n",
        "h_pool2_flat = tf.reshape(h_pool1, [-1, coef])\n",
        "#declare the weights considering the constants and 256 output \n",
        "W_fc1 = weight_variable([coef, w4])\n",
        "b_fc1 = bias_variable([w4])\n",
        "\n",
        "#rectifier (matmul)\n",
        "h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)\n",
        "#Rectifier LAYER*****************************************************************************\n",
        "\n",
        "#Rectifier-Dropout LAYER**********************************************************************\n",
        "#dropout\n",
        "h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)\n",
        "#declare weights with the ouput layer in this case 2 (labelSize)\n",
        "W_fc2 = weight_variable([w4, labelSize])\n",
        "b_fc2 = bias_variable([labelSize])\n",
        "#output\n",
        "y_conv = tf.matmul(h_fc1_drop, W_fc2) + b_fc2\n",
        "#Rectifier-Dropout LAYER**********************************************************************\n",
        "\n",
        "#Loss Function********************************************************************************\n",
        "#cross_entropy = tf.reduce_mean(-tf.reduce_sum(y_ * tf.log(y_conv), reduction_indices=[0]))\n",
        "cross_entropy = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits_v2(logits=y_conv, labels=y_)+\n",
        "\tbeta*tf.nn.l2_loss(W_conv1))\n",
        "#Optimizer Adam at 1e-5 (literature)**********************************************************\n",
        "train_step = tf.train.AdamOptimizer(1e-5).minimize(cross_entropy)\n",
        "#softmax prediction remember we are using one hot labels\n",
        "correct_prediction = tf.equal(tf.argmax(y_conv,1), tf.argmax(y_,1))\n",
        "\n",
        "trueResult=tf.argmax(y_conv,1)\n",
        "trueTest=tf.argmax(y_,1)\n",
        "#accuracy\n",
        "accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))\n",
        "#Loss Function********************************************************************************\n",
        "valid_accuracy_global=0.0\n",
        "test_accuracy_global=0.0\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "WARNING:tensorflow:From /usr/local/lib/python3.7/dist-packages/tensorflow/python/compat/v2_compat.py:101: disable_resource_variables (from tensorflow.python.ops.variable_scope) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "non-resource variables are not supported in the long term\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.7/dist-packages/tensorflow/python/util/dispatch.py:206: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8o8qHWCrRz9R",
        "outputId": "ddacfb92-d9c4-408a-cf02-e9ec6044980f"
      },
      "source": [
        "#start\n",
        "sess.run(tf.initialize_all_variables())\n",
        "# Add ops to save and restore all the variables.\n",
        "saver = tf.train.Saver()\n",
        "saver.restore(sess, path+'/CNN/data/model/model.ckpt')\n",
        "#Extra to verify sizes************************************************************************\n",
        "print(x_image.get_shape())\n",
        "#print(h_conv1.get_shape())\n",
        "print(h_pool1.get_shape())\n",
        "\n",
        "print(y_conv.get_shape())\n",
        "print(y_.get_shape())\n",
        "#Extra to verify sizes************************************************************************\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "WARNING:tensorflow:From /usr/local/lib/python3.7/dist-packages/tensorflow/python/util/tf_should_use.py:247: initialize_all_variables (from tensorflow.python.ops.variables) is deprecated and will be removed after 2017-03-02.\n",
            "Instructions for updating:\n",
            "Use `tf.global_variables_initializer` instead.\n",
            "INFO:tensorflow:Restoring parameters from /content/drive/MyDrive/Colab Notebooks/Genomic/primers-sars-cov-2/CNN/data/model/model.ckpt\n",
            "(?, 1, 31029, 1)\n",
            "(?, 1, 210, 12)\n",
            "(?, 5)\n",
            "(?, 5)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "B8vcT0lMhweR"
      },
      "source": [
        "#Output:ay amo\n",
        "units = sess.run(W_conv1,feed_dict={x:data, \n",
        "\t\t\ty_: oneHotLabels, keep_prob: 1.0})\n",
        "print(units.shape)\n",
        "units = sess.run(h_conv1,feed_dict={x:data, \n",
        "\t\t\ty_: oneHotLabels, keep_prob: 1.0})\n",
        "print(units.shape)\n",
        "\n",
        "dataSize=data.shape\n",
        "sampleSize=int(dataSize[0])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "aK4rEoPORz_8",
        "outputId": "73add8b7-893b-4a54-fa2c-f3faaa9dcda1"
      },
      "source": [
        "# Practice\n",
        "#Output:ay amo\n",
        "units_w = sess.run(W_conv1,feed_dict={x:data, \n",
        "\t\t\ty_: oneHotLabels, keep_prob: 1.0})\n",
        "print(units_w.shape)\n",
        "units_h = sess.run(h_conv1,feed_dict={x:data, \n",
        "\t\t\ty_: oneHotLabels, keep_prob: 1.0})\n",
        "print(units_h.shape)\n",
        "\n",
        "dataSize=data.shape\n",
        "sampleSize=int(dataSize[0])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(1, 21, 1, 12)\n",
            "(553, 1, 31029, 12)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "URnORztwheoi",
        "outputId": "fb45a07c-9d59-441d-c1e9-aeb0d7f364ee"
      },
      "source": [
        "units_w[0][0][0]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([ 0.00614528, -0.11750853,  0.1115165 ,  0.04802587, -0.12801747,\n",
              "       -0.01048534, -0.11457278,  0.06753053,  0.09112253,  0.02589372,\n",
              "       -0.01969883,  0.05848107], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 40
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lNdfIk9sh9Vx",
        "outputId": "7f6923bd-f7d7-4490-f75e-6b2269119a86"
      },
      "source": [
        "units_h[0][0][0]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.        , 0.25112575, 0.17467204, 0.        , 0.        ,\n",
              "       0.11897328, 0.22199348, 0.22719112, 0.02040457, 0.2067591 ,\n",
              "       0.        , 0.        ], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 39
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5EGFhjL9R0Cu",
        "outputId": "4cc139c0-42c3-4ec3-e5de-bd7dbf745062"
      },
      "source": [
        "Mat=np.zeros((sampleSize, vectorSize))\n",
        "\n",
        "for filterIndex in range(0,units.shape[3]):\n",
        "\tfor testSize in range(0,sampleSize):\n",
        "\t\tfor inputSize in range (0,vectorSize):\n",
        "\t\t\tMat[testSize][inputSize]=units[testSize][0][inputSize][filterIndex]\n",
        "\tpd.DataFrame(Mat).to_csv(path+'/CNN/data/filters/filter_'+str(filterIndex)+\".csv\", header=None, index =None)\n",
        "\n",
        "units = sess.run(h_pool1,feed_dict={x:data, \n",
        "\t\t\ty_: oneHotLabels, keep_prob: 1.0})\n",
        "print(units.shape)\n",
        "\n",
        "\n",
        "\n",
        "Mat=np.zeros((sampleSize, units.shape[2]))\n",
        "\n",
        "for filterIndex in range(0,units.shape[3]):\n",
        "\tfor testSize in range(0,sampleSize):\n",
        "\t\tfor inputSize in range (0,units.shape[2]):\n",
        "\t\t\tMat[testSize][inputSize]=units[testSize][0][inputSize][filterIndex]\n",
        "\tpd.DataFrame(Mat).to_csv(path+'/CNN/data/filters/maxPool_'+str(filterIndex)+\".csv\", header=None, index =None)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(553, 1, 210, 12)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "btPuSbalz61X"
      },
      "source": [
        "# CNN \n",
        "## funcCNN.py"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1g1w_AW9kHHM"
      },
      "source": [
        "def openMatrix (name) :\n",
        "\tdataArray = np.genfromtxt(name, delimiter=' ')\n",
        "\tdata=dataArray\n",
        "\tprint(dataArray.shape)\n",
        "\tmaxData=np.nanmax(dataArray)\n",
        "\tminData=np.nanmin(dataArray)\n",
        "\tmeanData=np.nanmean(dataArray)\n",
        "\tdistance=maxData-minData\n",
        "\tprint('Max ',maxData)\n",
        "\tprint('Min ',minData)\n",
        "\tprint('meanData ',meanData)\n",
        "\tfor i in range (len(dataArray)):\n",
        "\t\tfor j in range (len(dataArray[0])):\n",
        "\t\t\tdata[i][j]=(data[i][j]-minData)/distance\n",
        "\t\t\t#data[i][j]=data[i][j]\n",
        "\t\t\tif math.isnan(data[i][j]):\n",
        "\t\t\t\tdata[i][j]=-1\n",
        "\treturn data\n",
        "\n",
        "def openVector (name) :\n",
        "\tdat= np.genfromtxt(name, delimiter=' ')\n",
        "\t#data = pd.read_csv(name, header=None)\n",
        "\tprint(dat.shape)\n",
        "\treturn dat\n",
        "\n",
        "def saveMatrix(name,var):\n",
        "\tnp.savetxt(name, var, fmt='%1.3f', delimiter=' ')\n",
        "\n",
        "def saveMatrixInt(name,var):\n",
        "\tnp.savetxt(name,var, fmt='%i', delimiter=' ')\n",
        "\n",
        "def saveVectorInt(name,var):\n",
        "\tnp.savetxt(name, var, fmt='%i', delimiter=' ')\n",
        "\n",
        "#Pre Processing*********************************************************************\n",
        "def oneHot (array, size):\n",
        "\toutput=[]\n",
        "\t#print(len(array))\n",
        "\t#print(size)\n",
        "\t#print(array)\n",
        "\tfor i in range (len(array)):\n",
        "\t\ttemp=np.zeros(size)\n",
        "\t\ttemp[int(array[i])]=1\n",
        "\t\toutput.append(temp)\n",
        "\tprint(len(output))\n",
        "\tprint(len(output[0]))\n",
        "\treturn np.array(output)\n",
        "\n",
        "def compressArray(array,size):\n",
        "\toutputArray=[]\n",
        "\tfor i in range(len(array)):\n",
        "\t\ttemp=[]\n",
        "\t\tfor j in range(size):\n",
        "\t\t\tif (j<len(array[0])):\n",
        "\t\t\t\ttemp.append(array[i][j])\n",
        "\t\t\telse:\n",
        "\t\t\t\ttemp.append(-1)\n",
        "\t\toutputArray.append(temp)\n",
        "\treturn np.array(outputArray)\n",
        "\n",
        "#def getBatch(data, labels, size):\n",
        "#\tindex=[]\n",
        "#\tfor i in range (len(data)):\n",
        "#\t\tindex.append(i)\n",
        "#\tbatch=random.sample(index,  size)\n",
        "#\toutData=[]\n",
        "#\toutLabels=[]\n",
        "#\tfor i in range (len(batch)):\n",
        "#\t\toutData.append(data[batch[i]])\n",
        "#\t\toutLabels.append(labels[batch[i]])\n",
        "#\treturn np.array(outData), np.array(outLabels)\n",
        "\n",
        "def getBatch(data, labels, size, sampleSize):\n",
        "\tindex=[]\n",
        "\tfor i in range (len(data)):\n",
        "\t\tindex.append(i)\n",
        "\tbatch=random.sample(index,  size)\n",
        "\toutData=[]\n",
        "\toutLabels=[]\n",
        "\tfor i in range (len(batch)):\n",
        "\t\tsample=np.zeros(sampleSize)\n",
        "\t\tfor j in range (0, len(data[batch[i]])):\n",
        "\t\t\tif(data[batch[i]][j]=='C'):\n",
        "\t\t\t\tsample[j]=0.25\n",
        "\t\t\telif(data[batch[i]][j]=='T'):\n",
        "\t\t\t\tsample[j]=0.50\n",
        "\t\t\telif(data[batch[i]][j]=='G'):\n",
        "\t\t\t\tsample[j]=0.75\n",
        "\t\t\telif(data[batch[i]][j]=='A'):\n",
        "\t\t\t\tsample[j]=1.0\n",
        "\t\t\telse:\n",
        "\t\t\t\tsample[j]=0.0\n",
        "\t\toutData.append(sample)\n",
        "\t\toutLabels.append(labels[batch[i]])\n",
        "\treturn np.array(outData), np.array(outLabels)\n",
        "\n",
        "def getBatch_c(data, labels, size, constants):\n",
        "\tindex=[]\n",
        "\tfor i in range (len(data)):\n",
        "\t\tindex.append(i)\n",
        "\tbatch=random.sample(index,  size)\n",
        "\t#print(batch)\n",
        "\toutData=[]\n",
        "\toutLabels=[]\n",
        "\toutConstants=[]\n",
        "\tfor i in range (len(batch)):\n",
        "\t\toutData.append(data[batch[i]])\n",
        "\t\toutLabels.append(labels[batch[i]])\n",
        "\t\toutConstants.append(constants[batch[i]])\n",
        "\treturn np.array(outData), np.array(outLabels), np.array(outConstants)\n",
        "\n",
        "def getBatch_run_c(data, labels, size, constants,run,vector):\n",
        "\tinfLimit=run*size\n",
        "\tsupLimit=infLimit+size\n",
        "\tif supLimit > len(data):\n",
        "\t\tsupLimit=len(data)\n",
        "\tbatch=[]\n",
        "\tfor i in range (infLimit,supLimit):\n",
        "\t\tbatch.append(vector[i])\n",
        "\toutData=[]\n",
        "\toutLabels=[]\n",
        "\toutConstants=[]\n",
        "\tfor i in range (len(batch)):\n",
        "\t\toutData.append(data[batch[i]])\n",
        "\t\toutLabels.append(labels[batch[i]])\n",
        "\t\toutConstants.append(constants[batch[i]])\n",
        "\treturn np.array(outData), np.array(outLabels), np.array(outConstants)\n",
        "\n",
        "\treturn np.array(outData), np.array(outLabels)\n",
        "\n",
        "def print_no_newline(string):\n",
        "\timport sys\n",
        "\tsys.stdout.write(string)\n",
        "\tsys.stdout.flush()\n",
        "\n",
        "#def getBatch_run(data, labels, size,run,vector):\n",
        "#\tinfLimit=run*size\n",
        "#\tsupLimit=infLimit+size\n",
        "#\tif supLimit > len(data):\n",
        "#\t\tsupLimit=len(data)\n",
        "#\tbatch=[]\n",
        "#\tfor i in range (infLimit,supLimit):\n",
        "#\t\tbatch.append(vector[i])\n",
        "#\toutData=[]\n",
        "#\toutLabels=[]\n",
        "#\tfor i in range (len(batch)):\n",
        "#\t\toutData.append(data[batch[i]])\n",
        "#\t\toutLabels.append(labels[batch[i]])\n",
        "#\treturn np.array(outData), np.array(outLabels)\n",
        " \n",
        "def getBatch_run(data, labels, size,run,vector, sampleSize):\n",
        "\tinfLimit=run*size\n",
        "\tsupLimit=infLimit+size\n",
        "\tif supLimit > len(data):\n",
        "\t\tsupLimit=len(data)\n",
        "\tbatch=[]\n",
        "\tfor i in range (infLimit,supLimit):\n",
        "\t\tbatch.append(vector[i])\n",
        "\toutData=[]\n",
        "\toutLabels=[]\n",
        "\tfor i in range (len(batch)):\n",
        "\t\tsample=np.zeros(sampleSize)\n",
        "\t\tfor j in range (0, len(data[batch[i]])):\n",
        "\t\t\tif(data[batch[i]][j]=='C'):\n",
        "\t\t\t\tsample[j]=0.25\n",
        "\t\t\telif(data[batch[i]][j]=='T'):\n",
        "\t\t\t\tsample[j]=0.50\n",
        "\t\t\telif(data[batch[i]][j]=='G'):\n",
        "\t\t\t\tsample[j]=0.75\n",
        "\t\t\telif(data[batch[i]][j]=='A'):\n",
        "\t\t\t\tsample[j]=1.0\n",
        "\t\t\telse:\n",
        "\t\t\t\tsample[j]=0.0\n",
        "\t\toutData.append(sample)\n",
        "\t\toutLabels.append(labels[batch[i]])\n",
        "\treturn np.array(outData), np.array(outLabels)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Vx_GkqWeBDu8"
      },
      "source": [
        "## posPool.py"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gP9ItkR3z4u_",
        "outputId": "487cd6b0-b165-4555-94e4-556c86bf3af0"
      },
      "source": [
        "import numpy as np\n",
        "import math\n",
        "import random\n",
        "import pandas as pd \n",
        "\n",
        "for filterIndex in range(0,12):\n",
        "\t\n",
        "\tdata = pd.read_csv(path+'/CNN/data/filters/'+\"filter_\"+str(filterIndex)+\".csv\", header=None).values\n",
        "\tnumberWindows = 210\n",
        "\n",
        "\tsizeData=np.shape(data)\n",
        "\n",
        "\tprint(sizeData)\n",
        "\n",
        "\tmaxPool = np.zeros(shape=(sizeData[0],numberWindows))\n",
        "\tposPool = np.zeros(shape=(sizeData[0],numberWindows))\n",
        "\n",
        "\tfor  i in range(0, sizeData[0]):\n",
        "\t\tmaxPool_windowSize = 148\n",
        "\t\tpad_left_HPool = 25\n",
        "\t\tmax = -1e6\n",
        "\t\tindex = pad_left_HPool\n",
        "\t\tposition = -1\n",
        "\t\tindexMax = 0\n",
        "\t\tfor j in range (0, sizeData[1]):\n",
        "\t\t\tif data[i][j]>max:\n",
        "\t\t\t\tmax=data[i][j]\n",
        "\t\t\t\tposition=j\n",
        "\t\t\tindex=index+1\n",
        "\t\t\tif (index == maxPool_windowSize) or (j == sizeData[1] - 1):\n",
        "\t\t\t\tmaxPool[i][indexMax] = max\n",
        "\t\t\t\tposPool[i][indexMax] = position\n",
        "\t\t\t\tmax = -1e6\n",
        "\t\t\t\tposition = -1\n",
        "\t\t\t\tindex = 0\n",
        "\t\t\t\tindexMax=indexMax+1\n",
        "\t\t\t\t\n",
        "\tpd.DataFrame(maxPool).to_csv(path+'/CNN/data/filters/'+\"maxPool_\"+str(filterIndex)+\".csv\", header=None, index =None)\n",
        "\tpd.DataFrame(posPool).to_csv(path+'/CNN/data/filters/'+\"posPool_\"+str(filterIndex)+\".csv\", header=None, index =None)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(553, 31029)\n",
            "(553, 31029)\n",
            "(553, 31029)\n",
            "(553, 31029)\n",
            "(553, 31029)\n",
            "(553, 31029)\n",
            "(553, 31029)\n",
            "(553, 31029)\n",
            "(553, 31029)\n",
            "(553, 31029)\n",
            "(553, 31029)\n",
            "(553, 31029)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jQ96mlVZCp5e"
      },
      "source": [
        "## createFeatVector.py"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OT3IsfFSz4x5",
        "outputId": "84200c60-6f2b-4d57-e1f5-4c8b3a427360"
      },
      "source": [
        "import numpy as np\n",
        "import math\n",
        "import random\n",
        "import pandas as pd \n",
        "\n",
        "filterIndex=0\n",
        "posMatrix = pd.read_csv(path+'/CNN/data/filters/'+\"posPool_\"+str(filterIndex)+\".csv\", header=None).values\n",
        "matrix = pd.read_csv(path+'/CNN/data/filters/'+\"sequencesSorted.csv\", header=None).values.ravel()\n",
        "\n",
        "max=0\n",
        "for i in range (0,len(matrix)):\n",
        "\tif (len(matrix[i])>max):\n",
        "\t\tmax=len(matrix[i])\n",
        "\n",
        "vectorSize=max\n",
        "print('vectorSize', vectorSize)\n",
        "\n",
        "outData=[]\n",
        "for i in range (len(matrix)):\n",
        "\tsample=np.zeros(vectorSize)\n",
        "\tfor j in range (0, len(matrix[i])):\n",
        "\t\tif(matrix[i][j]=='C'):\n",
        "\t\t\tsample[j]=0.25\n",
        "\t\telif(matrix[i][j]=='T'):\n",
        "\t\t\tsample[j]=0.50\n",
        "\t\telif(matrix[i][j]=='G'):\n",
        "\t\t\tsample[j]=0.75\n",
        "\t\telif(matrix[i][j]=='A'):\n",
        "\t\t\tsample[j]=1.0\n",
        "\t\telse:\n",
        "\t\t\tsample[j]=0.0\n",
        "\toutData.append(sample)\n",
        "\t\n",
        "matrix=np.array(outData)\n",
        "\n",
        "sizePosMatrix=np.shape(posMatrix)\n",
        "numberFilters = 21\n",
        "padding = 10\n",
        "dataDNA = [[0 for i in range(210 * numberFilters)] for j in range(sizePosMatrix[0])] \n",
        "#dataDNA = np.array(shape=(sizePosMatrix[0],210 * numberFilters))\n",
        "\n",
        "sizeDNAMatrix=np.shape(matrix)\n",
        "temp=((matrix[0]))\n",
        "#temp=(str(matrix[0]))\n",
        "print(temp)\n",
        "for i in range (0,sizePosMatrix[0]):\n",
        "\ttemp=((matrix[i]))\n",
        "\tfor j in range ( 0, sizePosMatrix[1]):\n",
        "\t\tcoef=int(posMatrix[i][j])\n",
        "\t\tfor k in range(0, padding+1):\n",
        "\t\t\tif ((coef + k) < len(temp)):\n",
        "\t\t\t\tdataDNA[i][ j * numberFilters + padding + k] = temp[ coef + k]\n",
        "\t\t\tif ((coef - k) >= 0 and (coef - k)< len(temp)):\n",
        "\t\t\t\tdataDNA[i][ j * numberFilters + padding - k] = temp[ coef - k]\n",
        "\n",
        "\n",
        "dataDNAString = [[0 for i in range(210 * numberFilters)] for j in range(sizePosMatrix[0])] \n",
        "\n",
        "for i in range (0,sizePosMatrix[0]):\t\n",
        "\tfor j in range ( 0, 210 * numberFilters):\n",
        "\t\tif (dataDNA[i][ j] == 0.25):\n",
        "\t\t\n",
        "\t\t\tdataDNAString[i][ j] = \"C\"\n",
        "\t\t\n",
        "\t\telif (dataDNA[i][ j] == 0.50):\n",
        "\t\t\n",
        "\t\t\tdataDNAString[i][ j] = \"T\"\n",
        "\t\t\n",
        "\t\telif (dataDNA[i][ j] == 0.75):\n",
        "\t\t\n",
        "\t\t\tdataDNAString[i][ j] = \"G\"\n",
        "\t\t\n",
        "\t\telif (dataDNA[i][ j] == 1.00):\n",
        "\t\t\n",
        "\t\t\tdataDNAString[i][ j] = \"A\"\n",
        "\t\t\n",
        "\t\telse:\n",
        "\t\t\n",
        "\t\t\tdataDNAString[i][ j] = \"N\"\n",
        "\t\t\n",
        "dataDNAFeatures = [[0 for i in range(210)] for j in range(sizePosMatrix[0])] \n",
        "for i in range (0,sizePosMatrix[0]):\n",
        "\tfor j in range(0,210):\n",
        "\t\tdataDNAFeatures[i][ j] =str(\"\")\n",
        "\n",
        "for i in range (0,sizePosMatrix[0]):\n",
        "\tindexFeature = 0\n",
        "\tfeature = 0\n",
        "\tfor j in range(0,210 * numberFilters):\n",
        "\t\tdataDNAFeatures[i][ feature] =str(dataDNAFeatures[i][ feature])+str(dataDNAString[i][ j])\n",
        "\t\tindexFeature=indexFeature+1\n",
        "\t\tif (indexFeature == numberFilters):\n",
        "\t\t\tfeature=feature+1\n",
        "\t\t\tindexFeature = 0\n",
        "\n",
        "featsVector=[]            \n",
        "for i in range (0,sizePosMatrix[0]):\n",
        "\tfor j in range(0,210):\n",
        "\t\tcount=featsVector.count(dataDNAFeatures[i][j])\n",
        "\t\tif (count==0):\n",
        "\t\t\tif (\"N\" not in dataDNAFeatures[i][j]):\n",
        "\t\t\t\tfeatsVector.append(dataDNAFeatures[i][j])\n",
        "\n",
        "\n",
        "\t\t\n",
        "\t\t\n",
        "#pd.DataFrame(dataDNA).to_csv(\"dataDNA_\"+str(filterIndex)+\".csv\", header=None, index =None)\n",
        "#pd.DataFrame(dataDNAString).to_csv(\"dataDNAString_\"+str(filterIndex)+\".csv\", header=None, index =None)\n",
        "pd.DataFrame(featsVector).to_csv(path+'/CNN/data/filters/'+\"featsVector_\"+str(filterIndex)+\".csv\", header=None, index =None)\n",
        "pd.DataFrame(dataDNAFeatures).to_csv(path+'/CNN/data/filters/'+\"dataDNAFeatures_\"+str(filterIndex)+\".csv\", header=None, index =None)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "vectorSize 31029\n",
            "[1.  0.5 0.5 ... 0.  0.  0. ]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0p4oazPQE7xW"
      },
      "source": [
        "## getFeatures.py"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EDE3BtSTz40t",
        "outputId": "671c8744-e1db-45df-b011-13228269d534"
      },
      "source": [
        "filterIndex=0\n",
        "vector = pd.read_csv(path+'/CNN/data/filters/'+\"featsVector_\"+str(filterIndex)+\".csv\", header=None).values.ravel()\n",
        "sequences = pd.read_csv(path+'/CNN/data/'+\"sequences.csv\", header=None).values.ravel()\n",
        "\n",
        "print(len(sequences))\n",
        "print(len(vector))\n",
        "\n",
        "sizeSeq=int(len(sequences))\n",
        "sizeVec=int(len(vector))\n",
        "\n",
        "freqMatrix = np.zeros((sizeSeq,sizeVec))\n",
        "\n",
        "for i in range (0,sizeSeq):\t\n",
        "\tfor j in range ( 0, sizeVec):\n",
        "\t\tif (vector[j] in sequences[i]):\n",
        "\t\t\tfreqMatrix[i][j]=sequences[i].count(vector[j])\n",
        "\t\telse:\n",
        "\t\t\tfreqMatrix[i][j]=0\n",
        "\n",
        "#pd.DataFrame(freqMatrix).to_csv(\"data_\"+str(filterIndex)+\".csv\", header=None, index =None)\n",
        "\n",
        "repeatedList=[]\n",
        "for i in range (0,sizeSeq):\t\n",
        "\tfor j in range ( 0, sizeVec):\n",
        "\t\tif (freqMatrix[i][j]>1):\n",
        "\t\t\tif(vector[j] not in repeatedList):\n",
        "\t\t\t\trepeatedList.append(vector[j])\n",
        "\n",
        "print(repeatedList)\n",
        "\n",
        "nonRepeatedList=[]\n",
        "for i in range ( 0, sizeVec):\n",
        "\tif (vector[i] not in repeatedList):\n",
        "\t\tnonRepeatedList.append(vector[i])\n",
        "\n",
        "print(\"Repeated in the same sample: \"+str(len(repeatedList)))\n",
        "print(\"Non repeated in the same sample: \"+str(len(nonRepeatedList)))\n",
        "\n",
        "\n",
        "sizeSeq=int(len(sequences))\n",
        "sizeVec=int(len(nonRepeatedList))\n",
        "\n",
        "freqMatrix = np.zeros((sizeSeq,sizeVec))\n",
        "\n",
        "for i in range (0,sizeSeq):\t\n",
        "\tfor j in range ( 0, sizeVec):\n",
        "\t\tif (nonRepeatedList[j] in sequences[i]):\n",
        "\t\t\tfreqMatrix[i][j]=sequences[i].count(nonRepeatedList[j])\n",
        "\t\telse:\n",
        "\t\t\tfreqMatrix[i][j]=0\n",
        "\t\t\t\n",
        "pd.DataFrame(nonRepeatedList).to_csv(path+'/CNN/data/filters/'+\"features_\"+str(filterIndex)+\".csv\", header=None, index =None)\n",
        "pd.DataFrame(freqMatrix).to_csv(path+'/CNN/data/filters/'+\"data_\"+str(filterIndex)+\".csv\", header=None, index =None)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "553\n",
            "3898\n",
            "['GACGATGAAGATGTTGTTACT', 'TGTCAGATCTCATTAAATCTA', 'GTGTCTACTCTTCTCAACTAA', 'AATTATGTTCTTCAGGACTTT', 'AACAACATATCTAAAAATGTT', 'GGCCATGGTGCAGGAAGTGTG', 'ACTTAAAGATCGCACAGGCGC', 'ATTGAAATTTCGTCAAGTTTG', 'GTCTATATGCCTGCTAGTTGG', 'TTTTAATCAGCACGAAGTTCT']\n",
            "Repeated in the same sample: 10\n",
            "Non repeated in the same sample: 3888\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VcXT6sLCE93x"
      },
      "source": [
        "## getFreqMatrix.py"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "apzbgcSYEDoH",
        "outputId": "8d9ebb2a-8ac2-4eb2-91e2-61c65c38bed2"
      },
      "source": [
        "filterIndex=0\n",
        "vector = pd.read_csv(path+'/CNN/data/filters/'+\"features_\"+str(filterIndex)+\".csv\", header=None).values.ravel()\n",
        "sequences = pd.read_csv(path+'/CNN/data/'+\"sequences.csv\", header=None).values.ravel()\n",
        "\n",
        "print(len(sequences))\n",
        "print(len(vector))\n",
        "\n",
        "sizeSeq=int(len(sequences))\n",
        "sizeVec=int(len(vector))\n",
        "\n",
        "freqMatrix = np.zeros((sizeSeq,sizeVec))\n",
        "\n",
        "for i in range (0,sizeSeq):\t\n",
        "\tfor j in range ( 0, sizeVec):\n",
        "\t\tif (vector[j] in sequences[i]):\n",
        "\t\t\tfreqMatrix[i][j]=sequences[i].count(vector[j])\n",
        "\t\telse:\n",
        "\t\t\tfreqMatrix[i][j]=0\n",
        "\n",
        "pd.DataFrame(freqMatrix).to_csv(path+'/CNN/data/filters/'+\"data_\"+str(filterIndex)+\".csv\", header=None, index =None)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "553\n",
            "3888\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Gmvcb4qR6NUI",
        "outputId": "729cdfba-a537-4e09-a370-711edff29bae"
      },
      "source": [
        "import tensorflow.compat.v1 as tf\n",
        "print(tf.__version__, tf.config.list_physical_devices('GPU'))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "2.6.0 [PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jt5-hDvbyvjO"
      },
      "source": [
        "## genRunTrain.py"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oBLu0lAwEDrd",
        "outputId": "e18ca3dd-5f5d-4a20-8414-e813678d1269"
      },
      "source": [
        "dataFolder=path+'/CNN/data/'\n",
        "\n",
        "#Output:\n",
        "f=open(dataFolder+'results/outputVector.txt', 'w')\n",
        "f.write('1\\n')\n",
        "f.write('1\\n')\n",
        "temp=1.0\n",
        "trueAcc=str(temp)\n",
        "print(trueAcc)\n",
        "f.write(trueAcc+'\\n')\n",
        "f.close()\n",
        "#Parameters*******************************************************************************\n",
        "#Maximum number of iterations\n",
        "iterMax=100\n",
        "## iterMax=int(sys.argv[1])\n",
        "\n",
        "#maximum number of iterations\n",
        "limit=1.01\n",
        "#regularization on the weights\n",
        "beta=0.001\n",
        "#version of the code\n",
        "version='gen1'\n",
        "#size of batch\n",
        "batchSize=50\n",
        "#Parameters*******************************************************************************\n",
        "\n",
        "w1=12 #12\n",
        "## w1=int(sys.argv[2]) #12\n",
        "\n",
        "w4=196\n",
        "## w4=int(sys.argv[3]) #196\n",
        "\n",
        "h1=148\n",
        "## h1=int(sys.argv[4]) #148\n",
        "\n",
        "wd1=21\n",
        "##wd1=int(sys.argv[5]) #21\n",
        "\n",
        "index=0\n",
        "## index=int(sys.argv[6]) #0\n",
        "\n",
        "kfoldIndex=0\n",
        "## kfoldIndex=int(sys.argv[7]) #0\n",
        "\n",
        "generation=0\n",
        "## generation=int(sys.argv[8]) #0\n",
        "#Input Data***********************************************************************************\n",
        "(test_dataset_Flat,oneHot_test_labels,valid_dataset_Flat,oneHot_valid_labels,\n",
        "\ttrain_dataset_Flat,oneHot_train_labels,labelSize,vectorSize)=get_Info(\n",
        "\tkfoldIndex,dataFolder)\n",
        "runs=int(len(oneHot_train_labels)/batchSize)\n",
        "print(runs)\n",
        "#Model declaration************************************************************************\n",
        "#import tensorflow as tf\n",
        "#declare interactive session\n",
        "import tensorflow.compat.v1 as tf\n",
        "#tf.disable_v2_behavior()\n",
        "sess=tf.InteractiveSession()\n",
        "\n",
        "\n",
        "#INPUT->CONV LAYER->CONV LAYER->CONV LAYER->RECT FLAT->RECT DROPOUT\n",
        "\n",
        "#function to declare easily the weights only by shape\n",
        "def weight_variable(shape):\n",
        "\tinitial = tf.truncated_normal(shape, stddev=0.1)\n",
        "\treturn tf.Variable(initial)\n",
        "#function to declare easily the bias only by shape\n",
        "def bias_variable(shape):\n",
        "\tinitial = tf.constant(0.1, shape=shape)\n",
        "\treturn tf.Variable(initial)\n",
        "\n",
        "#input variable\n",
        "x = tf.placeholder(tf.float32, [None, vectorSize])\n",
        "#keep probability to change from dropout 0.50 to 1.0 in validation and test\n",
        "keep_prob = tf.placeholder(tf.float32)\n",
        "#expected outputs variable\n",
        "y_ = tf.placeholder(tf.float32, [None, labelSize])\n",
        "\n",
        "#arrange the tensor as an image (1*31029) 1 channel\n",
        "x_image0 = tf.reshape(x, [-1,1,vectorSize,1])\n",
        "x_image = tf.transpose(x_image0, perm=[0,3,2,1])\n",
        "#arrange the tensor into 1 channels (1*31029)\n",
        "\n",
        "#1 LAYER*************************************************************************************\n",
        "#1 Convolutional Layer Explicit for regularization of the weights\n",
        "#weigth first layer 1 input channels, 12 output channels, 1x21 filter window size\n",
        "W_conv1 = weight_variable([1, wd1, 1, w1])\n",
        "#bias declaration the size has to be the same as the output channels 12\n",
        "b_conv1 = bias_variable([w1])\n",
        "#convolution (input weights) moving 1 step each time with a relu\n",
        "h_conv1 = tf.nn.relu(tf.nn.conv2d(x_image, W_conv1, \n",
        "\tstrides=[1, 1, 1, 1], padding='SAME') + b_conv1)\n",
        "#max pooling with a 148 width window size, moving 148 in width by step\n",
        "h_pool1 = tf.nn.max_pool(h_conv1, ksize=[1, 1, h1, 1],\n",
        "\tstrides=[1, 1, h1, 1], padding='SAME')\n",
        "#output=545/4\n",
        "#1 LAYER*************************************************************************************\n",
        "\n",
        "#Rectifier LAYER*****************************************************************************\n",
        "#calculated coefficient for the flattening from the size of the 3 convolutional layer\n",
        "coef=int (h_pool1.get_shape()[1]*h_pool1.get_shape()[2]*h_pool1.get_shape()[3])\n",
        "h_pool2_flat = tf.reshape(h_pool1, [-1, coef])\n",
        "#declare the weights considering the constants and 256 output \n",
        "W_fc1 = weight_variable([coef, w4])\n",
        "b_fc1 = bias_variable([w4])\n",
        "\n",
        "#rectifier (matmul)\n",
        "h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)\n",
        "#Rectifier LAYER*****************************************************************************\n",
        "\n",
        "#Rectifier-Dropout LAYER**********************************************************************\n",
        "#dropout\n",
        "h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)\n",
        "#declare weights with the ouput layer in this case 2 (labelSize)\n",
        "W_fc2 = weight_variable([w4, labelSize])\n",
        "b_fc2 = bias_variable([labelSize])\n",
        "#output\n",
        "y_conv = tf.matmul(h_fc1_drop, W_fc2) + b_fc2\n",
        "#Rectifier-Dropout LAYER**********************************************************************\n",
        "\n",
        "#Loss Function********************************************************************************\n",
        "#cross_entropy = tf.reduce_mean(-tf.reduce_sum(y_ * tf.log(y_conv), reduction_indices=[0]))\n",
        "cross_entropy = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits_v2(logits=y_conv, labels=y_)+\n",
        "\tbeta*tf.nn.l2_loss(W_conv1))\n",
        "#Optimizer Adam at 1e-5 (literature)**********************************************************\n",
        "train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)\n",
        "#softmax prediction remember we are using one hot labels\n",
        "correct_prediction = tf.equal(tf.argmax(y_conv,1), tf.argmax(y_,1))\n",
        "\n",
        "trueResult=tf.argmax(y_conv,1)\n",
        "trueTest=tf.argmax(y_,1)\n",
        "#accuracy\n",
        "accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))\n",
        "#Loss Function********************************************************************************\n",
        "valid_accuracy_global=0.0\n",
        "test_accuracy_global=0.0\n",
        "\n",
        "\n",
        "#start\n",
        "sess.run(tf.initialize_all_variables())\n",
        "# Add ops to save and restore all the variables.\n",
        "saver = tf.train.Saver()\n",
        "#Extra to verify sizes************************************************************************\n",
        "print(x_image.get_shape())\n",
        "#print(h_conv1.get_shape())\n",
        "print(h_pool1.get_shape())\n",
        "\n",
        "print(y_conv.get_shape())\n",
        "print(y_.get_shape())\n",
        "#Extra to verify sizes************************************************************************\n",
        "\n",
        "print('test ',str(oneHot_valid_labels.shape[0]))\n",
        "\n",
        "xaV,yaV=getBatch(valid_dataset_Flat,oneHot_valid_labels,oneHot_valid_labels.shape[0],vectorSize)\n",
        "print(xaV[1])\n",
        "print(yaV[1])\n",
        "\n",
        "yResult=[]\n",
        "yTest=[]\n",
        "#Ouput Data Variables*************************************************************************\n",
        "name=str(str(kfoldIndex)+'_'+str(iterMax)+'_'+str(w1)+'_'+str(w4)+'_'+str(h1)+'_'+str(wd1))\n",
        "f = open(dataFolder+'results/'+name+'.txt', 'a')\n",
        "#Ouput Data Variables*************************************************************************\n",
        "\n",
        "#Main Loop************************************************************************************\n",
        "#initialize variables\n",
        "iter=0\n",
        "train_accuracy=0.0\n",
        "valid_accuracy=0.0\n",
        "test_accuracy=0.0\n",
        "#best validation accuracy\n",
        "best=0\n",
        "validWindow=[0,0,0]\n",
        "repeatWindow=[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]\n",
        "validBest=1e6\n",
        "#limit (normally a dummy value 0.80) and iterations...\n",
        "while ((best<limit) & (iter<iterMax)):\n",
        "\tindexBatch=[]\n",
        "\tfor iB in range (0,len(oneHot_train_labels)):\n",
        "\t\tindexBatch.append(iB)\n",
        "\trandom.shuffle(indexBatch)\n",
        "\tfor run in range (0,runs):\n",
        "\t\t#Get data from train set and store in xa (inputs),ya(labels),ca(constants) 100 batch size\n",
        "\t\txa,ya=getBatch_run(train_dataset_Flat,oneHot_train_labels,batchSize,run,indexBatch,vectorSize)\n",
        "\t\t#pass the values and 0.50 dropout (literature)\n",
        "\t\ttrain_step.run(feed_dict={x: xa, y_: ya, keep_prob: 0.5})\n",
        "\t\t#print each 10 iterations\n",
        "\t#calculate train accuracy\n",
        "\txa,ya=getBatch(train_dataset_Flat,oneHot_train_labels,batchSize,vectorSize)\n",
        "\ttrain_accuracy = accuracy.eval(feed_dict={\n",
        "\t\tx:xa, y_: ya, keep_prob: 1.0})\n",
        "\t#calculate validation accuracy\n",
        "\txaV,yaV=getBatch(valid_dataset_Flat,oneHot_valid_labels,oneHot_valid_labels.shape[0],vectorSize)\n",
        "\t\n",
        "\tvalid_accuracy=accuracy.eval(feed_dict={\n",
        "\t\tx: xaV, y_: yaV, keep_prob: 1.0})\n",
        "\t#calculate validation loss\n",
        "\tcross_entropyVal=cross_entropy.eval(feed_dict={\n",
        "\t\tx: xaV, y_: yaV, keep_prob: 1.0})\n",
        "\t#calculate train loss\n",
        "\tcross_entropyTrain=cross_entropy.eval(feed_dict={\n",
        "\t\tx: xa, y_: ya, keep_prob: 1.0})\n",
        "\t#append values for graphs\n",
        "\t#if valid accuracy is better than the best accuracy then calculate test accuracy\n",
        "\t#if valid_accuracy>best:\n",
        "\tvalidWindowValue=0\n",
        "\ttempValid=validWindow\n",
        "\tfor i in range(0,len(validWindow)-1):\n",
        "\t\ttempValid[i]=validWindow[i+1]\n",
        "\tfor i in range(0,len(validWindow)):\n",
        "\t\tvalidWindow[i]=tempValid[i]\n",
        "\tvalidWindow[len(validWindow)-1]=valid_accuracy\n",
        "\tfor i in range(0,len(validWindow)):\n",
        "\t\tvalidWindowValue=validWindowValue+validWindow[i]\n",
        "\tvalidWindowValue=validWindowValue/len(validWindow)\n",
        "\ttempValid=repeatWindow\n",
        "\tfor i in range(0,len(repeatWindow)-1):\n",
        "\t\ttempValid[i]=repeatWindow[i+1]\n",
        "\tfor i in range(0,len(repeatWindow)):\n",
        "\t\trepeatWindow[i]=tempValid[i]\n",
        "\trepeatWindow[len(repeatWindow)-1]=valid_accuracy\n",
        "\tif np.var(repeatWindow)==0 and iter>10:\n",
        "\t\titer=iter\n",
        "\tif (validWindowValue)>best or cross_entropyVal<validBest:\n",
        "\t\tvalidBest=cross_entropyVal\n",
        "\t\tbest=(validWindowValue)\n",
        "\t\t\n",
        "\t\txaT,yaT=getBatch(test_dataset_Flat,oneHot_test_labels,oneHot_test_labels.shape[0],vectorSize)\n",
        "\t\t#calculate test accuracy\n",
        "\t\ttest_accuracy= accuracy.eval(feed_dict={x:xaT, \n",
        "\t\t\ty_: yaT, keep_prob: 1.0})\n",
        "\t\t# Save the variables to disk.\n",
        "\t\tif (kfoldIndex==0):\n",
        "\t\t\tsave_path = saver.save(sess, dataFolder+\"model/model.ckpt\")\n",
        "\t\t#plotNNFilter(units)\n",
        "\t\t#calculate the results of the whole model, probabilities in one hot format\n",
        "\t\tresults=correct_prediction.eval(feed_dict={x:xaT, y_: yaT, keep_prob: 1.0})\n",
        "\t\tyResult=trueResult.eval(feed_dict={x:xaT, y_: yaT, keep_prob: 1.0})\n",
        "\t\tyTest=trueTest.eval(feed_dict={x:xaT, y_: yaT, keep_prob: 1.0})\n",
        "\t\tfOut=open(dataFolder+'results/outputVector.txt', 'w')\n",
        "\t\tfOut.write('1\\n')\n",
        "\t\tfOut.write('1\\n')\n",
        "\t\ttemp=1.0-best\n",
        "\t\ttrueAcc=str(temp)\n",
        "\t\tprint(trueAcc)\n",
        "\t\tfOut.write(trueAcc+'\\n')\n",
        "\t\tfOut.close()\t\t\t\t\n",
        "\t#append everything to a log for retrieving results\n",
        "\tlog=\"%d\t%d\t%g\t%g\t%g\t%g\t%g\t%g\"%(iter,kfoldIndex,train_accuracy,valid_accuracy,best, test_accuracy,cross_entropyVal,cross_entropyTrain)\n",
        "\tprint(log)\n",
        "\tf.write(log+'\\n')\n",
        "\titer=iter+1\n",
        "#Main Loop************************************************************************************\n",
        "\n",
        "f.close()\n",
        "saveVectorInt(dataFolder+'results/results'+name+'.txt',yResult)\n",
        "saveVectorInt(dataFolder+'results/test'+name+'.txt',yTest)\n",
        "f = open(dataFolder+'log3.txt', 'a')\n",
        "name=str(str(index)+'_'+str(kfoldIndex)+'_'+str(iterMax)+'_'+str(test_accuracy)+'_'+str(valid_accuracy)+'_'+str(best)+'_'+str(w1)+'_'+\n",
        "\t\t str(w4)+'_'+str(h1)+'_'+str(wd1)+'_'+str(generation))\n",
        "f.write(name+'\\n')\n",
        "f.close()\n",
        "\n",
        "#f=open(str(index)+'.index','a')\n",
        "#name=str(str(index)+'_'+str(kfoldIndex)+'_'+str(iterMax)+'_'+str(test_accuracy)+'_'+str(valid_accuracy)+'_'+str(best)+'_'+str(w1)+'_'+#\n",
        "#\t\t str(w4)+'_'+str(h1)+'_'+str(wd1)+'_'+str(generation))\n",
        "#f.write(name+'\\n')\n",
        "#f.close()\n",
        "#Input Data***********************************************************************************\n",
        "#close session\n",
        "f=open(dataFolder+'results/outputVector.txt', 'w')\n",
        "f.write('1\\n')\n",
        "f.write('1\\n')\n",
        "temp=1.0-best\n",
        "trueAcc=str(temp)\n",
        "print(trueAcc)\n",
        "f.write(trueAcc+'\\n')\n",
        "f.close()\n",
        "\n",
        "xaT,yaT=getBatch(test_dataset_Flat,oneHot_test_labels,oneHot_test_labels.shape[0],vectorSize)\n",
        "\n",
        "units = sess.run(h_conv1,feed_dict={x:xaT, \n",
        "\t\t\ty_: yaT, keep_prob: 1.0})\n",
        "print(units.shape)\n",
        "units = sess.run(h_pool1,feed_dict={x:xaT, \n",
        "\t\t\ty_: yaT, keep_prob: 1.0})\n",
        "print(units.shape)\n",
        "\n",
        "sess.close()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1.0\n",
            "data set (553,)\n",
            "vectorSize 31029\n",
            "train length: 441\n",
            "train length 0: 29727\n",
            "train label shape: (441,)\n",
            "valid length: 55\n",
            "valid label shape: (55,)\n",
            "test length: 57\n",
            "test label shape: (57,)\n",
            "5\n",
            "441\n",
            "5\n",
            "oneHot_train_labels.shape:  (441, 5)\n",
            "55\n",
            "5\n",
            "oneHot_valid_labels.shape: (55, 5)\n",
            "57\n",
            "5\n",
            "oneHot_test_labels.shape (57, 5)\n",
            "8\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/tensorflow/python/client/session.py:1766: UserWarning: An interactive session is already active. This can cause out-of-memory errors in some cases. You must explicitly call `InteractiveSession.close()` to release resources held by the other session(s).\n",
            "  warnings.warn('An interactive session is already active. This can '\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(?, 1, 31029, 1)\n",
            "(?, 1, 210, 12)\n",
            "(?, 5)\n",
            "(?, 5)\n",
            "test  55\n",
            "[0.75 1.   0.75 ... 0.   0.   0.  ]\n",
            "[0. 0. 1. 0. 0.]\n",
            "0.8606060643990835\n",
            "0\t0\t0.52\t0.418182\t0.139394\t0.421053\t1.3869\t1.12948\n",
            "0.721212128798167\n",
            "1\t0\t0.4\t0.418182\t0.278788\t0.421053\t1.23546\t1.21946\n",
            "0.5575757622718811\n",
            "2\t0\t0.56\t0.490909\t0.442424\t0.561404\t1.15226\t1.09075\n",
            "0.4727272689342499\n",
            "3\t0\t0.76\t0.672727\t0.527273\t0.701754\t1.0855\t0.951955\n",
            "0.3757575750350952\n",
            "4\t0\t0.76\t0.709091\t0.624242\t0.719298\t1.03052\t0.954325\n",
            "0.2909090916315714\n",
            "5\t0\t0.76\t0.745455\t0.709091\t0.719298\t0.979467\t0.897524\n",
            "0.2666666706403097\n",
            "6\t0\t0.74\t0.745455\t0.733333\t0.719298\t0.918081\t0.881944\n",
            "0.23636362950007117\n",
            "7\t0\t0.8\t0.8\t0.763636\t0.807018\t0.87136\t0.806802\n",
            "0.21818180878957116\n",
            "8\t0\t0.74\t0.8\t0.781818\t0.807018\t0.820673\t0.917637\n",
            "0.19999998807907104\n",
            "9\t0\t0.88\t0.8\t0.8\t0.807018\t0.770379\t0.61302\n",
            "0.16363634665807092\n",
            "10\t0\t0.9\t0.909091\t0.836364\t0.807018\t0.722896\t0.699053\n",
            "0.12727270523707068\n",
            "11\t0\t0.96\t0.909091\t0.872727\t0.859649\t0.680634\t0.563767\n",
            "0.08484846353530884\n",
            "12\t0\t0.96\t0.927273\t0.915152\t0.894737\t0.653357\t0.54796\n",
            "0.08484846353530884\n",
            "13\t0\t0.92\t0.909091\t0.915152\t0.824561\t0.610811\t0.545928\n",
            "0.07878786325454712\n",
            "14\t0\t0.84\t0.927273\t0.921212\t0.824561\t0.582327\t0.631047\n",
            "0.07878786325454712\n",
            "15\t0\t0.84\t0.927273\t0.921212\t0.824561\t0.556202\t0.630948\n",
            "0.0727272629737854\n",
            "16\t0\t0.94\t0.927273\t0.927273\t0.824561\t0.529253\t0.435083\n",
            "0.0727272629737854\n",
            "17\t0\t0.92\t0.927273\t0.927273\t0.842105\t0.514103\t0.480775\n",
            "0.0727272629737854\n",
            "18\t0\t0.8\t0.927273\t0.927273\t0.824561\t0.483014\t0.56776\n",
            "0.0727272629737854\n",
            "19\t0\t0.88\t0.927273\t0.927273\t0.824561\t0.462723\t0.50668\n",
            "0.0727272629737854\n",
            "20\t0\t0.94\t0.927273\t0.927273\t0.824561\t0.446322\t0.381685\n",
            "0.0727272629737854\n",
            "21\t0\t0.92\t0.927273\t0.927273\t0.824561\t0.431831\t0.364707\n",
            "0.0727272629737854\n",
            "22\t0\t0.94\t0.927273\t0.927273\t0.824561\t0.414275\t0.359826\n",
            "0.0727272629737854\n",
            "23\t0\t0.94\t0.927273\t0.927273\t0.824561\t0.396541\t0.271362\n",
            "0.0727272629737854\n",
            "24\t0\t0.92\t0.927273\t0.927273\t0.824561\t0.381592\t0.384207\n",
            "0.0727272629737854\n",
            "25\t0\t0.88\t0.927273\t0.927273\t0.824561\t0.364017\t0.372642\n",
            "26\t0\t0.9\t0.927273\t0.927273\t0.824561\t0.369345\t0.398295\n",
            "0.0727272629737854\n",
            "27\t0\t0.84\t0.927273\t0.927273\t0.842105\t0.347052\t0.454724\n",
            "0.0727272629737854\n",
            "28\t0\t0.88\t0.927273\t0.927273\t0.824561\t0.334803\t0.37901\n",
            "0.0727272629737854\n",
            "29\t0\t0.92\t0.927273\t0.927273\t0.824561\t0.334146\t0.23485\n",
            "0.0727272629737854\n",
            "30\t0\t0.86\t0.927273\t0.927273\t0.842105\t0.320164\t0.392594\n",
            "0.0727272629737854\n",
            "31\t0\t0.96\t0.927273\t0.927273\t0.842105\t0.301431\t0.191695\n",
            "32\t0\t0.86\t0.927273\t0.927273\t0.842105\t0.30686\t0.371982\n",
            "0.0727272629737854\n",
            "33\t0\t0.92\t0.927273\t0.927273\t0.877193\t0.294367\t0.256872\n",
            "0.0727272629737854\n",
            "34\t0\t0.92\t0.927273\t0.927273\t0.842105\t0.282276\t0.252684\n",
            "0.0727272629737854\n",
            "35\t0\t0.92\t0.927273\t0.927273\t0.842105\t0.277855\t0.248452\n",
            "0.0727272629737854\n",
            "36\t0\t0.96\t0.927273\t0.927273\t0.894737\t0.274536\t0.218768\n",
            "37\t0\t0.94\t0.927273\t0.927273\t0.894737\t0.279636\t0.217197\n",
            "0.0727272629737854\n",
            "38\t0\t0.88\t0.927273\t0.927273\t0.912281\t0.268186\t0.370661\n",
            "0.06666666269302368\n",
            "39\t0\t0.96\t0.945455\t0.933333\t0.894737\t0.247872\t0.169668\n",
            "40\t0\t0.98\t0.927273\t0.933333\t0.894737\t0.250876\t0.217999\n",
            "0.06666666269302368\n",
            "41\t0\t0.94\t0.927273\t0.933333\t0.912281\t0.24392\t0.197619\n",
            "0.0727272629737854\n",
            "42\t0\t0.98\t0.927273\t0.927273\t0.947368\t0.234771\t0.200296\n",
            "0.0727272629737854\n",
            "43\t0\t0.92\t0.927273\t0.927273\t0.912281\t0.229772\t0.266169\n",
            "0.06666666269302368\n",
            "44\t0\t0.94\t0.945455\t0.933333\t0.964912\t0.225805\t0.172826\n",
            "0.06060606241226196\n",
            "45\t0\t0.94\t0.945455\t0.939394\t0.964912\t0.22752\t0.291759\n",
            "0.06060606241226196\n",
            "46\t0\t0.98\t0.927273\t0.939394\t0.947368\t0.22745\t0.16955\n",
            "0.06666666269302368\n",
            "47\t0\t0.88\t0.927273\t0.933333\t0.947368\t0.213871\t0.337452\n",
            "0.06666666269302368\n",
            "48\t0\t0.92\t0.945455\t0.933333\t0.947368\t0.209701\t0.244288\n",
            "0.06060606241226196\n",
            "49\t0\t0.96\t0.945455\t0.939394\t0.947368\t0.204641\t0.183965\n",
            "0.048484861850738525\n",
            "50\t0\t1\t0.963636\t0.951515\t0.964912\t0.203103\t0.173223\n",
            "0.048484861850738525\n",
            "51\t0\t0.96\t0.945455\t0.951515\t0.964912\t0.198376\t0.158038\n",
            "0.04242426156997681\n",
            "52\t0\t0.98\t0.963636\t0.957576\t0.964912\t0.193123\t0.157777\n",
            "53\t0\t0.96\t0.963636\t0.957576\t0.964912\t0.19341\t0.158512\n",
            "0.03636366128921509\n",
            "54\t0\t0.96\t0.963636\t0.963636\t0.964912\t0.185116\t0.134093\n",
            "55\t0\t0.96\t0.963636\t0.963636\t0.964912\t0.188401\t0.183919\n",
            "0.03636366128921509\n",
            "56\t0\t0.98\t0.963636\t0.963636\t0.964912\t0.179624\t0.14751\n",
            "57\t0\t1\t0.963636\t0.963636\t0.964912\t0.180232\t0.105188\n",
            "0.03636366128921509\n",
            "58\t0\t0.96\t0.963636\t0.963636\t0.964912\t0.170323\t0.19119\n",
            "0.03636366128921509\n",
            "59\t0\t0.98\t0.963636\t0.963636\t0.964912\t0.169333\t0.19854\n",
            "0.03636366128921509\n",
            "60\t0\t0.96\t0.963636\t0.963636\t0.964912\t0.168239\t0.146994\n",
            "0.03636366128921509\n",
            "61\t0\t1\t0.963636\t0.963636\t0.947368\t0.166096\t0.110341\n",
            "0.030303041140238407\n",
            "62\t0\t1\t0.981818\t0.969697\t0.964912\t0.14831\t0.0904775\n",
            "63\t0\t1\t0.963636\t0.969697\t0.964912\t0.15234\t0.132264\n",
            "64\t0\t1\t0.963636\t0.969697\t0.964912\t0.153973\t0.115359\n",
            "0.03636366128921509\n",
            "65\t0\t1\t0.963636\t0.963636\t0.964912\t0.14768\t0.0853264\n",
            "0.03636366128921509\n",
            "66\t0\t1\t0.963636\t0.963636\t0.964912\t0.144725\t0.0735419\n",
            "0.03636366128921509\n",
            "67\t0\t0.98\t0.963636\t0.963636\t0.964912\t0.144175\t0.133891\n",
            "0.03636366128921509\n",
            "68\t0\t1\t0.963636\t0.963636\t0.964912\t0.141272\t0.0984805\n",
            "0.03636366128921509\n",
            "69\t0\t0.98\t0.963636\t0.963636\t0.964912\t0.139606\t0.116327\n",
            "0.030303041140238407\n",
            "70\t0\t0.96\t0.981818\t0.969697\t0.964912\t0.131995\t0.173112\n",
            "71\t0\t1\t0.963636\t0.969697\t0.964912\t0.135057\t0.0895048\n",
            "72\t0\t0.98\t0.963636\t0.969697\t0.964912\t0.135121\t0.139915\n",
            "0.03636366128921509\n",
            "73\t0\t0.98\t0.963636\t0.963636\t0.964912\t0.129381\t0.167412\n",
            "0.030303041140238407\n",
            "74\t0\t1\t0.981818\t0.969697\t0.964912\t0.120996\t0.0914653\n",
            "0.024242420991261837\n",
            "75\t0\t1\t0.981818\t0.975758\t0.964912\t0.12053\t0.136146\n",
            "76\t0\t0.98\t0.963636\t0.975758\t0.964912\t0.126792\t0.0912557\n",
            "77\t0\t0.98\t0.963636\t0.975758\t0.964912\t0.125605\t0.111045\n",
            "0.030303041140238407\n",
            "78\t0\t1\t0.981818\t0.969697\t0.964912\t0.115285\t0.0853771\n",
            "0.024242420991261837\n",
            "79\t0\t1\t0.981818\t0.975758\t0.964912\t0.115138\t0.0662295\n",
            "0.018181800842285156\n",
            "80\t0\t1\t0.981818\t0.981818\t0.964912\t0.116366\t0.078017\n",
            "81\t0\t1\t0.981818\t0.981818\t0.964912\t0.117857\t0.0402265\n",
            "0.018181800842285156\n",
            "82\t0\t1\t0.981818\t0.981818\t0.964912\t0.1155\t0.140197\n",
            "0.018181800842285156\n",
            "83\t0\t1\t0.981818\t0.981818\t0.964912\t0.111353\t0.094042\n",
            "0.018181800842285156\n",
            "84\t0\t1\t0.981818\t0.981818\t0.964912\t0.105418\t0.0767073\n",
            "0.018181800842285156\n",
            "85\t0\t1\t0.981818\t0.981818\t0.964912\t0.10295\t0.0784752\n",
            "0.018181800842285156\n",
            "86\t0\t1\t0.981818\t0.981818\t0.964912\t0.102326\t0.106186\n",
            "0.012121200561523438\n",
            "87\t0\t1\t1\t0.987879\t0.982456\t0.0964707\t0.0815298\n",
            "0.006060600280761719\n",
            "88\t0\t1\t1\t0.993939\t0.982456\t0.0957777\t0.0645292\n",
            "89\t0\t1\t0.981818\t0.993939\t0.982456\t0.102071\t0.0690184\n",
            "90\t0\t1\t0.981818\t0.993939\t0.982456\t0.0992033\t0.0801687\n",
            "0.012121200561523438\n",
            "91\t0\t1\t1\t0.987879\t0.982456\t0.0913661\t0.0714617\n",
            "0.006060600280761719\n",
            "92\t0\t1\t1\t0.993939\t0.982456\t0.0894268\t0.0519147\n",
            "93\t0\t1\t0.981818\t0.993939\t0.982456\t0.0900129\t0.0317081\n",
            "94\t0\t1\t0.981818\t0.993939\t0.982456\t0.0957345\t0.0479729\n",
            "95\t0\t1\t0.981818\t0.993939\t0.982456\t0.0901403\t0.0682043\n",
            "0.012121200561523438\n",
            "96\t0\t1\t1\t0.987879\t0.982456\t0.0817127\t0.0416421\n",
            "0.006060600280761719\n",
            "97\t0\t1\t1\t0.993939\t0.982456\t0.0805691\t0.0421676\n",
            "98\t0\t1\t0.981818\t0.993939\t0.982456\t0.0898784\t0.065275\n",
            "99\t0\t1\t1\t0.993939\t0.982456\t0.0834449\t0.0852111\n",
            "0.006060600280761719\n",
            "(57, 1, 31029, 12)\n",
            "(57, 1, 210, 12)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0yoHCEyIH85N"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BIhB-uZXH89H"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}