fig3.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4c596922",
   "metadata": {},
   "outputs": [],
   "source": [
    "import gzip\n",
    "import io\n",
    "import pandas as pd\n",
    "import os\n",
    "from collections import Counter\n",
    "from umi_tools import UMIClusterer\n",
    "import matplotlib.pyplot as plt\n",
    "from Bio import SeqIO\n",
    "from collections import Counter\n",
    "import csv\n",
    "import re\n",
    "import numpy as np\n",
    "from scipy.stats import pearsonr\n",
    "import seaborn as sns\n",
    "from adjustText import adjust_text  # Import the adjust_text function\n",
    "from matplotlib.lines import Line2D  # Import Line2D\n",
    "from brokenaxes import brokenaxes\n",
    "from matplotlib.ticker import MaxNLocator\n",
    "from matplotlib.colors import LinearSegmentedColormap\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "37951933",
   "metadata": {},
   "outputs": [],
   "source": [
    "# loading files\n",
    "path_to_files = \"/corgi/martin/CRISPR_padlock/BK_FINAL_DATA/5_Johan_files/\"\n",
    "\n",
    "p_umi_mle = pd.read_csv(path_to_files+\"14d_mip_dedup#allreads#0.mageckmle.gene_summary.txt\", sep='\\t')\n",
    "p_mle = pd.read_csv(path_to_files+\"14d_mip_nodedup#allreads#0.mageckmle.gene_summary.txt\", sep='\\t')\n",
    "PCR_mle = pd.read_csv(path_to_files+\"14d_pcr#allreads#0.mageckmle.gene_summary.txt\", sep='\\t')\n",
    "\n",
    "p_umi_test = pd.read_csv(path_to_files+\"14d_mip_dedup#allreads#0.gene_summary.txt\", sep='\\t')\n",
    "p_test = pd.read_csv(path_to_files+\"14d_mip_nodedup#allreads#0.gene_summary.txt\", sep='\\t')\n",
    "PCR_test = pd.read_csv(path_to_files+\"14d_pcr#allreads#0.gene_summary.txt\", sep='\\t')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "52045e9a",
   "metadata": {},
   "source": [
    "### Figure 3A"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c5b5dec5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# MAGeCK test\n",
    "def plot_RRA_score2(df, gene_list, name):\n",
    "    # Filter genes with 'eff|fdr' value < 0.05\n",
    "    highlighted_genes = df[df['neg|fdr'] < 0.05]\n",
    "\n",
    "    # Identify genes in the gene_list\n",
    "    green_genes_mask = df['id'].isin(gene_list)\n",
    "\n",
    "    # Create the scatter plot for 'neg|score' with a logarithmic scale on the y-axis\n",
    "    plt.figure(figsize=(6, 6))\n",
    "    # All genes in dark grey with black edge\n",
    "    plt.scatter(range(len(df)), df['neg|score'], alpha=0.7, color='dimgrey', edgecolor='dimgrey', label='All Genes', s=20)\n",
    "    \n",
    "    # Set logarithmic scale on y-axis\n",
    "    plt.yscale('log')\n",
    "    \n",
    "    # Set the y-axis range if specified\n",
    "    plt.ylim(0.0000001, 2)\n",
    "\n",
    "    # Highlight filtered genes in red with black edge, excluding those in gene_list to prioritize green\n",
    "    plt.scatter(highlighted_genes[~highlighted_genes['id'].isin(gene_list)].index, highlighted_genes[~highlighted_genes['id'].isin(gene_list)]['neg|score'], color='red', edgecolor='black', label='Highlighted Genes (< 0.05 eff|fdr)',s=50)\n",
    "\n",
    "    # Highlight genes from gene_list in green\n",
    "    plt.scatter(df[green_genes_mask].index, df[green_genes_mask]['neg|score'], color='green', edgecolor='black', label='Gene List Highlight', s=50)\n",
    "\n",
    "    # Initialize a list to store text objects for adjust_text\n",
    "    texts = []\n",
    "\n",
    "    # Annotate genes in the gene_list with green text\n",
    "    for _, row in df[green_genes_mask].iterrows():\n",
    "        texts.append(plt.text(row.name, row['neg|score'], row['id'], ha='center', va='bottom', color='green'))\n",
    "\n",
    "    #  Use adjust_text to repel labels with stronger repulsion parameters\n",
    "    adjust_text(texts, arrowprops=dict(arrowstyle='-', color='black'),\n",
    "                expand_points=(2.5, 2.5), expand_text=(1.5, 1.5),\n",
    "                force_points=0.5, force_text=0.5)\n",
    "\n",
    "    # Set labels and title\n",
    "    plt.xlabel('Gene Rank', fontsize=14)\n",
    "    plt.ylabel('RRA Score', fontsize=14)\n",
    "    plt.title('RRA Score by Gene Rank')\n",
    "\n",
    "    # Reverse values on y-axis\n",
    "    plt.gca().invert_yaxis()\n",
    "\n",
    "    # Increase the thickness of the axis lines and tick marks\n",
    "    ax = plt.gca()\n",
    "    ax.spines['top'].set_linewidth(2)\n",
    "    ax.spines['bottom'].set_linewidth(2)\n",
    "    ax.spines['left'].set_linewidth(2)\n",
    "    ax.spines['right'].set_linewidth(2)\n",
    "    ax.tick_params(axis='both', which='major', labelsize=12, width=2, length=6)\n",
    "\n",
    "    # save plot as svg\n",
    "    plt.savefig(f'/corgi/martin/CRISPR_padlock/BK_FINAL_DATA/5_Johan_files/figures/{name}.svg', format='svg')\n",
    "    \n",
    "    # save plot as png\n",
    "    plt.savefig(f'/corgi/martin/CRISPR_padlock/BK_FINAL_DATA/5_Johan_files/figures/{name}.png', format='png', dpi=600) \n",
    "    \n",
    "    # Show plot\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "    \n",
    "gene_list = [\"GUK1\", \"AURKB\", \"WEE1\", \"DOLK\"]\n",
    "plot_RRA_score2(PCR_test,gene_list,\"RRA_PCR\")\n",
    "plot_RRA_score2(p_umi_test,gene_list,\"RRA_padlock_umi\")\n",
    "\n",
    "# MAGeCK mle\n",
    "def plot_beta_score_genes2(df, gene_list, name):\n",
    "    # Filter genes with 'eff|fdr' value < 0.05\n",
    "    highlighted_genes = df[df['eff|fdr'] < 0.05]\n",
    "\n",
    "    # Create the scatter plot for 'eff|beta' with a logarithmic scale on the y-axis\n",
    "    plt.figure(figsize=(6, 6))\n",
    "    # All genes in dark grey with black edge\n",
    "    plt.scatter(range(len(df)), df['eff|beta'], alpha=0.7, color='dimgrey', edgecolor='dimgrey', label='All Genes', s=20)\n",
    "\n",
    "    # Highlight filtered genes in red with black edge, excluding those in gene_list to prioritize green\n",
    "    plt.scatter(highlighted_genes[~highlighted_genes['id'].isin(gene_list)].index, highlighted_genes[~highlighted_genes['id'].isin(gene_list)]['eff|beta'], color='red', edgecolor='black', label='Highlighted Genes (< 0.05 eff|fdr)',s=50)\n",
    "\n",
    "    # Highlight genes from gene_list in green\n",
    "    plt.scatter(df[df['id'].isin(gene_list)].index, df[df['id'].isin(gene_list)]['eff|beta'], color='green', edgecolor='black', label='Gene List Highlight', s=50)\n",
    "\n",
    "    # Initialize a list to store text objects for adjust_text\n",
    "    texts = []\n",
    "\n",
    "    # Annotate genes in the gene_list and add them to the texts list\n",
    "    for _, row in df[df['id'].isin(gene_list)].iterrows():\n",
    "        texts.append(plt.text(row.name, row['eff|beta'], row['id'], ha='center', va='bottom',color='green'))\n",
    "\n",
    "    #  Use adjust_text to repel labels with stronger repulsion parameters\n",
    "    adjust_text(texts, arrowprops=dict(arrowstyle='-', color='black'),\n",
    "                expand_points=(2.5, 2.5), expand_text=(1.5, 1.5),\n",
    "                force_points=0.5, force_text=0.5)\n",
    "\n",
    "    # Set labels and title\n",
    "    plt.xlabel('Gene Rank', fontsize=14)\n",
    "    plt.ylabel('Beta Score', fontsize=14)\n",
    "    plt.title('Beta Score by Gene Rank')\n",
    "    \n",
    "    # Set the y-axis range if specified\n",
    "    plt.ylim(-2.1, 0.8)\n",
    "        \n",
    "    # Increase the thickness of the axis lines\n",
    "    ax = plt.gca()\n",
    "    ax.spines['top'].set_linewidth(2)\n",
    "    ax.spines['bottom'].set_linewidth(2)\n",
    "    ax.spines['left'].set_linewidth(2)\n",
    "    ax.spines['right'].set_linewidth(2)\n",
    "    \n",
    "    # Increase the thickness and size of scale labels and tick marks\n",
    "    ax.tick_params(axis='both', which='major', labelsize=12, width=2, length=6) \n",
    "    \n",
    "    # save plot as svg\n",
    "    plt.savefig(f'/corgi/martin/CRISPR_padlock/BK_FINAL_DATA/5_Johan_files/figures/{name}.svg', format='svg')\n",
    "    \n",
    "    # save plot as png\n",
    "    plt.savefig(f'/corgi/martin/CRISPR_padlock/BK_FINAL_DATA/5_Johan_files/figures/{name}.png', format='png', dpi=600) \n",
    "\n",
    "    # Show plot\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "    \n",
    "plot_beta_score_genes2(p_umi_mle, gene_list, \"beta_padlock_umi\")\n",
    "plot_beta_score_genes2(PCR_mle, gene_list, \"beta_PCR\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "64001d30",
   "metadata": {},
   "source": [
    "### Figure 3B"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "d9a00609",
   "metadata": {},
   "outputs": [],
   "source": [
    "# MAGecK test\n",
    "def correlation_rank_grad(df1, df2, gene_list, name):\n",
    "    # Merge the DataFrames on the 'id' column, including 'neg|fdr' columns\n",
    "    common_genes = pd.merge(df1[['id', 'neg|rank', 'neg|fdr']], df2[['id', 'neg|rank', 'neg|fdr']], on='id', suffixes=('_df1', '_df2'))\n",
    "\n",
    "    # Calculate the product of 'neg|fdr' values from df1 and df2 for the color gradient\n",
    "    score_product = common_genes['neg|fdr_df1'] * common_genes['neg|fdr_df2']\n",
    "\n",
    "    # Create scatter plot with gradient colors based on the score product using the red-white-blue color map\n",
    "    plt.figure(figsize=(8, 6))\n",
    "    scatter = plt.scatter(common_genes['neg|rank_df1'], common_genes['neg|rank_df2'], c=score_product, cmap=sns.dark_palette(\"#69d\", reverse=True, as_cmap=True), s=15)\n",
    "\n",
    "    # Add a color bar to the plot to represent the scale of the score product\n",
    "    plt.colorbar(scatter, label='Product of neg|fdr')\n",
    "\n",
    "    # Highlight genes from gene_list in green\n",
    "    green_genes_mask = common_genes['id'].isin(gene_list)\n",
    "    plt.scatter(common_genes[green_genes_mask]['neg|rank_df1'], common_genes[green_genes_mask]['neg|rank_df2'], color='green', edgecolor='black', s=40)\n",
    "\n",
    "    # Calculate Pearson correlation coefficient for 'neg|rank' columns\n",
    "    corr_coefficient, _ = pearsonr(common_genes['neg|rank_df1'], common_genes['neg|rank_df2'])\n",
    "    corr_coefficient = round(corr_coefficient, 2)\n",
    "    print(f\"Pearson Correlation Coefficient: {corr_coefficient}\")\n",
    "\n",
    "    # Fit a linear regression line to the 'neg|rank' data\n",
    "    m, b = np.polyfit(common_genes['neg|rank_df1'], common_genes['neg|rank_df2'], 1)\n",
    "    x_values = np.linspace(common_genes['neg|rank_df1'].min(), common_genes['neg|rank_df1'].max(), 100)\n",
    "    y_values = m * x_values + b\n",
    "    plt.plot(x_values, y_values, '-', color='red', label=f'Linear Regression (r={corr_coefficient})')\n",
    "\n",
    "    plt.xlabel('neg|rank_df1')\n",
    "    plt.ylabel('neg|rank_df2')\n",
    "    plt.title('Visualization of neg|rank values for Common Genes')\n",
    "\n",
    "    # Increase the thickness of the axis lines and tick marks\n",
    "    ax = plt.gca()\n",
    "    ax.spines['top'].set_linewidth(2)\n",
    "    ax.spines['bottom'].set_linewidth(2)\n",
    "    ax.spines['left'].set_linewidth(2)\n",
    "    ax.spines['right'].set_linewidth(2)\n",
    "    ax.tick_params(axis='both', which='major', labelsize=12, width=2, length=6)\n",
    "    \n",
    "    # save plot as svg\n",
    "    plt.savefig(f'/corgi/martin/CRISPR_padlock/BK_FINAL_DATA/5_Johan_files/figures/{name}.svg', format='svg')\n",
    "    \n",
    "    # save plot as png\n",
    "    plt.savefig(f'/corgi/martin/CRISPR_padlock/BK_FINAL_DATA/5_Johan_files/figures/{name}.png', format='png', dpi=600) \n",
    "\n",
    "    plt.show()\n",
    "    \n",
    "gene_list = [\"GUK1\", \"AURKB\", \"WEE1\", \"DOLK\"]\n",
    "correlation_rank_grad(p_umi_test, PCR_test,gene_list, \"test_corr\")\n",
    "\n",
    "# MAGecK mle\n",
    "def correlation_rank_grad(df1, df2, gene_list, name):\n",
    "    # Merge the DataFrames on the 'id' column, including 'neg|fdr' columns\n",
    "    common_genes = pd.merge(df1[['id', 'neg|rank', 'eff|fdr']], df2[['id', 'neg|rank', 'eff|fdr']], on='id', suffixes=('_df1', '_df2'))\n",
    "\n",
    "    # Calculate the product of 'neg|fdr' values from df1 and df2 for the color gradient\n",
    "    score_product = common_genes['eff|fdr_df1'] * common_genes['eff|fdr_df2']\n",
    "\n",
    "    # Create scatter plot with gradient colors based on the score product using the red-white-blue color map\n",
    "    plt.figure(figsize=(8, 6))\n",
    "    scatter = plt.scatter(common_genes['neg|rank_df1'], common_genes['neg|rank_df2'], c=score_product, cmap=sns.dark_palette(\"#69d\", reverse=True, as_cmap=True), s=15)\n",
    "\n",
    "    # Add a color bar to the plot to represent the scale of the score product\n",
    "    plt.colorbar(scatter, label='Product of neg|fdr')\n",
    "\n",
    "    # Highlight genes from gene_list in green\n",
    "    green_genes_mask = common_genes['id'].isin(gene_list)\n",
    "    plt.scatter(common_genes[green_genes_mask]['neg|rank_df1'], common_genes[green_genes_mask]['neg|rank_df2'], color='green', edgecolor='black', s=40)\n",
    "\n",
    "    # Calculate Pearson correlation coefficient for 'neg|rank' columns\n",
    "    corr_coefficient, _ = pearsonr(common_genes['neg|rank_df1'], common_genes['neg|rank_df2'])\n",
    "    corr_coefficient = round(corr_coefficient, 2)\n",
    "    print(f\"Pearson Correlation Coefficient: {corr_coefficient}\")\n",
    "\n",
    "    # Fit a linear regression line to the 'neg|rank' data\n",
    "    m, b = np.polyfit(common_genes['neg|rank_df1'], common_genes['neg|rank_df2'], 1)\n",
    "    x_values = np.linspace(common_genes['neg|rank_df1'].min(), common_genes['neg|rank_df1'].max(), 100)\n",
    "    y_values = m * x_values + b\n",
    "    plt.plot(x_values, y_values, '-', color='red', label=f'Linear Regression (r={corr_coefficient})')\n",
    "\n",
    "    plt.xlabel('neg|rank_df1')\n",
    "    plt.ylabel('neg|rank_df2')\n",
    "    plt.title('Visualization of neg|rank values for Common Genes')\n",
    "\n",
    "    # Increase the thickness of the axis lines and tick marks\n",
    "    ax = plt.gca()\n",
    "    ax.spines['top'].set_linewidth(2)\n",
    "    ax.spines['bottom'].set_linewidth(2)\n",
    "    ax.spines['left'].set_linewidth(2)\n",
    "    ax.spines['right'].set_linewidth(2)\n",
    "    ax.tick_params(axis='both', which='major', labelsize=12, width=2, length=6)\n",
    "    \n",
    "    # save plot as svg\n",
    "    plt.savefig(f'/corgi/martin/CRISPR_padlock/BK_FINAL_DATA/5_Johan_files/figures/{name}.svg', format='svg')\n",
    "    \n",
    "    # save plot as png\n",
    "    plt.savefig(f'/corgi/martin/CRISPR_padlock/BK_FINAL_DATA/5_Johan_files/figures/{name}.png', format='png', dpi=600) \n",
    "\n",
    "    plt.show()\n",
    "    \n",
    "gene_list = [\"GUK1\", \"AURKB\", \"WEE1\", \"DOLK\"]\n",
    "correlation_rank_grad(p_umi_mle, PCR_mle,gene_list, \"mle_corr\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8827f1af",
   "metadata": {},
   "source": [
    "### Figure 3C"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "efb3692c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# load files\n",
    "sgrna_p = pd.read_csv(path_to_files+\"14d_mip_dedup#allreads#0.sgrna_summary.txt\", sep='\\t')\n",
    "sgrna_PCR = pd.read_csv(path_to_files+\"14d_pcr#allreads#0.sgrna_summary.txt\", sep='\\t')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3043bb33",
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_summary_df(df, gene_list):\n",
    "    # Initialize the new DataFrame\n",
    "    summary_df = pd.DataFrame(columns=[\"Gene\", \"control_count\", \"treatment_count\"])\n",
    "    \n",
    "    # Dictionary to keep track of the count for each gene\n",
    "    gene_counts = {gene: 0 for gene in gene_list}\n",
    "\n",
    "    # Iterate through the DataFrame\n",
    "    for _, row in df.iterrows():\n",
    "        gene = row['Gene']\n",
    "        # Check if the gene is in the gene_list\n",
    "        if gene in gene_list:\n",
    "            # Update the count for this gene\n",
    "            gene_counts[gene] += 1\n",
    "            # Modify the gene name to include the count\n",
    "            gene_name = f\"{gene}_{gene_counts[gene]}\"\n",
    "\n",
    "            # Calculate the mean of values in 'control_count' and 'treatment_count', assuming they are separated by \"/\"\n",
    "            control_counts = [float(x) for x in row['control_count'].split('/')]\n",
    "            treatment_counts = [float(x) for x in row['treatment_count'].split('/')]\n",
    "            control_mean = sum(control_counts) / len(control_counts)\n",
    "            treatment_mean = sum(treatment_counts) / len(treatment_counts)\n",
    "\n",
    "            # Add the calculated data to the summary DataFrame\n",
    "            summary_df = summary_df.append({\n",
    "                \"Gene\": gene_name,\n",
    "                \"control_count\": control_mean,\n",
    "                \"treatment_count\": treatment_mean\n",
    "            }, ignore_index=True)\n",
    "    \n",
    "    # Sort the summary DataFrame by the 'Gene' column\n",
    "    summary_df = summary_df.sort_values(by=\"Gene\")\n",
    "    \n",
    "    return summary_df\n",
    "\n",
    "gene_list = [\"CDK1\", \"DOLK\", \"GUK1\", \"AURKB\", \"WEE1\"]\n",
    "padlock = create_summary_df(sgrna_p, gene_list)\n",
    "PCR = create_summary_df(sgrna_PCR, gene_list)\n",
    "\n",
    "def plot_heatmap(df,name,max_value):\n",
    "    # Create a new DataFrame with genes as columns and 'Control' and 'Treatment' as index\n",
    "    reshaped_df = pd.DataFrame(index=['Control', 'Treatment'])\n",
    "\n",
    "    for gene in df['Gene']:\n",
    "        reshaped_df[gene] = [df[df['Gene'] == gene]['control_count'].values[0], \n",
    "                             df[df['Gene'] == gene]['treatment_count'].values[0]]\n",
    "\n",
    "    # Plotting the heatmap\n",
    "    plt.figure(figsize=(10, 2))  # Adjust the size as needed\n",
    "    ax = sns.heatmap(reshaped_df, annot=False, vmin=0, vmax=max_value, cmap=\"coolwarm\", cbar=True, linewidths=0.3, linecolor='black', robust=True, fmt='g')\n",
    "    \n",
    "    # Customizing the colorbar\n",
    "    cbar = ax.collections[0].colorbar\n",
    "    # Set the number of ticks in the colorbar\n",
    "    cbar.locator = MaxNLocator(nbins=5)  # Adjust 'nbins' to change the number of ticks\n",
    "    cbar.update_ticks()\n",
    "    \n",
    "    plt.title('Control vs. Treatment Counts per Gene')\n",
    "    plt.ylabel('Condition')\n",
    "    plt.xticks(rotation=45)  # Rotate gene names for better readability\n",
    "\n",
    "    plt.tight_layout()\n",
    "    plt.savefig(f'/corgi/martin/CRISPR_padlock/BK_FINAL_DATA/5_Johan_files/figures/{name}.svg', format='svg')\n",
    "    plt.show()\n",
    "    \n",
    "def plot_heatmap2(df,name):\n",
    "    # Create a new DataFrame with genes as columns and 'Control' and 'Treatment' as index\n",
    "    reshaped_df = pd.DataFrame(index=['CRISPR-MIP', 'PCR'])\n",
    "\n",
    "    for gene in df['Gene']:\n",
    "        reshaped_df[gene] = [df[df['Gene'] == gene]['CRISPR-MIP'].values[0], \n",
    "                             df[df['Gene'] == gene]['PCR'].values[0]]\n",
    "\n",
    "    # Plotting the heatmap\n",
    "    plt.figure(figsize=(10, 2))  # Adjust the size as needed\n",
    "    ax = sns.heatmap(reshaped_df, annot=False, cmap=\"coolwarm\", cbar=True, linewidths=0.05, linecolor='grey', vmin=-3.5, vmax=3.5, robust=True, fmt='g')\n",
    "    \n",
    "    # Customizing the colorbar\n",
    "    cbar = ax.collections[0].colorbar\n",
    "    # Set the number of ticks in the colorbar\n",
    "    cbar.locator = MaxNLocator(nbins=5)  # Adjust 'nbins' to change the number of ticks\n",
    "    cbar.update_ticks()\n",
    "    \n",
    "    plt.title('Control vs. Treatment Counts per Gene')\n",
    "    plt.ylabel('Condition')\n",
    "    plt.xticks(rotation=45)  # Rotate gene names for better readability\n",
    "\n",
    "    plt.tight_layout()\n",
    "    #plt.savefig(f'/corgi/martin/CRISPR_padlock/BK_FINAL_DATA/5_Johan_files/figures/{name}.svg', format='svg')\n",
    "    plt.savefig(f'/corgi/martin/CRISPR_padlock/BK_FINAL_DATA/5_Johan_files/figures/{name}.png', format='png', dpi=600) \n",
    "    plt.show()\n",
    "\n",
    "plot_heatmap(padlock,\"heatmap_sgrnas_padlock\",450)\n",
    "plot_heatmap(PCR,\"heatmap_sgrnas_PCR\", 6000)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "CRISPR",
   "language": "python",
   "name": "crispr"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.15"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}