.ipynb_checkpoints/PTCL-TMA restain analysis-checkpoint.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Analyse the images for DNA dropping\n",
    "```\n",
    "# Request for resources on O2\n",
    "srun --pty -p interactive --mem 170G -t 0-02:00 /bin/bash\n",
    "\n",
    "conda create -n dna_corr python=3\n",
    "\n",
    "# Copy python function inside the environment for future use\n",
    "~/.conda/envs/dna_corr/bin/\n",
    "\n",
    "# Activate the conda environment\n",
    "conda activate dna_corr\n",
    "\n",
    "# Run the function\n",
    "cd /n/scratch2/ajit/ptcl_tma/PTCL8/registration\n",
    "dna_corr(image_filepath = sys.argv[1], output_filepath = sys.argv[2]) # Format for running function\n",
    "python /home/ajn16/.conda/envs/dna_corr/bin/dna_corr.py PTCL7_484.ome.tif /n/scratch2/ajit/ptcl_tma/dna_corr\n",
    "\n",
    "```\n",
    "**Function used for generating DNA correlation**\n",
    "\n",
    "```\n",
    "# Import necessary packages\n",
    "import os\n",
    "import sys\n",
    "import numpy as np\n",
    "from skimage.external import tifffile\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "\n",
    "\n",
    "def dna_corr(image_filepath, output_filepath):\n",
    "    # read image\n",
    "    image = tifffile.imread(image_filepath)\n",
    "    # get number of cycles, assuming 4 channels per cycle\n",
    "    num_cycle = image.shape[0] // 4\n",
    "    # empty array to store correlation coefficient result\n",
    "    corrcoef = np.zeros(num_cycle)\n",
    "    # reference DNA intensity, hard-coded to be the DNA channel of the first cycle\n",
    "    reference_dna = image[0]\n",
    "    # loop over each cycle\n",
    "    for cycle_index in range(num_cycle):\n",
    "        # get the current DNA intensity\n",
    "        current_dna = image[cycle_index*4]\n",
    "        # calculate correlation coefficient. value range: [-1,+1]\n",
    "        corrcoef[cycle_index] = np.corrcoef(current_dna.flatten(), reference_dna.flatten())[0,1]\n",
    "\n",
    "    ## Generate Figure\n",
    "    plt.plot(corrcoef)\n",
    "    plt.ylim(0, 1)\n",
    "    plt.title(\"Correlation coefficient over cycles\")\n",
    "    plt.xlabel('Cycles')\n",
    "    plt.ylabel('Correlation Coefficient')\n",
    "    plt.savefig(output_filepath + \"/correlation coefficient.png\")\n",
    "\n",
    "    ## Generate the excel sheet and save it\n",
    "    corrcoef_df = pd.DataFrame(corrcoef)\n",
    "    corrcoef_df = corrcoef_df.rename(columns={0: \"correlation coefficient\"})\n",
    "    corrcoef_df.insert(0, 'Cycle', range(1, 1 + len(corrcoef_df)))\n",
    "    corrcoef_df.to_csv(output_filepath + \"/correlation coefficient.csv\", index=False)\n",
    "\n",
    "# Run the function\n",
    "dna_corr(image_filepath = sys.argv[1], output_filepath = sys.argv[2])\n",
    "\n",
    "```\n",
    "\n",
    "#### Clean up the CSV file to remove DNA channels and trim channel names\n",
    "\n",
    "```\n",
    "# Preprocess the data\n",
    "def preprocess (data, file_name=None,drop_markers=None):\n",
    "    '''\n",
    "    Parameters:\n",
    "        data: Data matrix generated by histocat\n",
    "        file_name: String. If a value is provided, the program will remove the sting from the marker names\n",
    "        drop_markers: Arrray. List of markers to drop from the analysis. e.g. [\"CD3D\", \"CD20\"]\n",
    "    Returns:\n",
    "        Data and meta data\n",
    "    Example:\n",
    "        preprocess (data, file_name=None,islog=True)\n",
    "    '''\n",
    "    # Unique name for the data\n",
    "    data.index = data.index.astype(str) + np.repeat('_',len(data.index)).astype(str) + data['CellId'].astype(str)\n",
    "    data = data.drop(['CellId'], axis=1)\n",
    "    # If there is INF replace with zero\n",
    "    data = data.replace([np.inf, -np.inf], 0)\n",
    "    # Remove DNA channels\n",
    "    data = data.loc[:,~data.columns.str.contains('dna', case=False)] \n",
    "    # Split the data into expression data and meta data\n",
    "    # Step-1 (Find the index of the column with name Area)\n",
    "    split_idx = data.columns.get_loc(\"Area\")\n",
    "    meta = data.iloc [:,split_idx:]\n",
    "    # Step-2 (select only the expression values)\n",
    "    d = data.iloc [:,:split_idx]        \n",
    "    # Rename the columns of the data\n",
    "    if file_name != None:\n",
    "        d.columns = list(map(lambda x: x.replace(file_name,''),list(d.columns)))\n",
    "    # Drop unnecessary markers\n",
    "    if drop_markers != None:\n",
    "        d = d.drop(columns = drop_markers)\n",
    "    # Return data\n",
    "    return d\n",
    "\n",
    "# Run the pre processing function.\n",
    "data = pd.read_csv('//research.files.med.harvard.edu/ImStor/sorger/data/RareCyte/Connor/Z155_PTCL/whole_section_csv_files/raw/Ton_192.csv')\n",
    "file_name= \"Cell_Ton_192\"\n",
    "data_clean = preprocess_d (data, file_name ,drop_markers= [\"PERK\", \"NOS2\"])\n",
    "meta_clean = preprocess_m (data, file_name ,drop_markers= [\"PERK\", \"NOS2\"]) # change return to meta\n",
    "\n",
    "# Write the data out\n",
    "output_filepath = \"//research.files.med.harvard.edu/ImStor/sorger/data/RareCyte/Connor/Z155_PTCL/whole_section_csv_files/\"\n",
    "file_name= \"TON\"\n",
    "data_clean.to_csv(output_filepath + file_name + \"_data.csv\", index=False)\n",
    "meta_clean.to_csv(output_filepath + file_name + \"_meta.csv\", index=False)\n",
    "```\n",
    "\n",
    "#### Cell type calling (using R)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}