05_correlation_and_testing.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "cc3cd54d",
   "metadata": {},
   "source": [
    "### Correlation testing and evaluation"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "de240ddb",
   "metadata": {},
   "source": [
    "0. Get the expression matrices for each window\n",
    "1. Get two genes (vectors) from expression matrix of branch A\n",
    "2. Compute correlation between those two vectors $\\rightarrow w_1$\n",
    "3. Get the same genes (vectors) from expression matrix of branch B and compute correlation $\\rightarrow w_2$\n",
    "4. Get the absolut value $|w_1-w_2|$\n",
    "5. Do it for all windows\n",
    "6. Determine which correlation method to use\n",
    "7. Try to evaluate significance of correlation results with permutation test for p-value (not practicable yet)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "d68682ca",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/nfs/home/students/l.willruth/.conda/envs/myenv/lib/python3.6/site-packages/dask/config.py:161: YAMLLoadWarning: calling yaml.load() without Loader=... is deprecated, as the default Loader is unsafe. Please read https://msg.pyyaml.org/load for full details.\n",
      "  data = yaml.load(f.read()) or {}\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import scanpy as sc\n",
    "import anndata\n",
    "import glob\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "47c5e2d1",
   "metadata": {},
   "outputs": [],
   "source": [
    "PATH_TO_WINDOW_ANNDATA = 'data_files/windows_anndata/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "5058e445",
   "metadata": {},
   "outputs": [],
   "source": [
    "branchAwindows = []\n",
    "branchBwindows = []\n",
    "\n",
    "A_windows_files=(glob.glob(f'{PATH_TO_WINDOW_ANNDATA}*A.h5ad'))\n",
    "B_windows_files=(glob.glob(f'{PATH_TO_WINDOW_ANNDATA}*B.h5ad'))\n",
    "A_windows_files = sorted(A_windows_files)\n",
    "B_windows_files = sorted(B_windows_files)\n",
    "\n",
    "for i in range(len(A_windows_files)):\n",
    "    branchAwindows.append(sc.read_h5ad(A_windows_files[i]))\n",
    "    branchBwindows.append(sc.read_h5ad(B_windows_files[i]))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b7fb7f46",
   "metadata": {},
   "source": [
    " "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d8486166",
   "metadata": {},
   "source": [
    "#### 1. Get two genes (vectors) from expression matrix of branch A"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ef3e6017",
   "metadata": {},
   "source": [
    "Vector Xa is expression vector for gene X in branch A. <br>\n",
    "Vector Ya is expression vector for gene Y in branch A. <br>\n",
    "(Just random vectors for testing)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "14ab0e1a",
   "metadata": {},
   "outputs": [],
   "source": [
    "test = branchAwindows[4].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "3d3486ab",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['0610030E20Rik', '1110004F10Rik', '1700020I14Rik', '1810037I17Rik',\n",
       "       '1810058I24Rik', '2010111I01Rik', '2210016F16Rik', '2310011J03Rik',\n",
       "       '2410015M20Rik', '2510002D24Rik',\n",
       "       ...\n",
       "       'Zkscan3', 'Zmat2', 'Zmynd11', 'Zmynd8', 'Znrd1', 'Zpr1', 'Zranb2',\n",
       "       'Zrsr2', 'Zufsp', 'Zwint'],\n",
       "      dtype='object', name='index', length=2000)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.var_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "5ec7aeac",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "        1.3534448 , -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307,  5.056215  ,\n",
       "       -0.21651307,  6.6066732 , -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307,  1.5155494 , -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "        3.2920623 , -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307,  6.0293612 , -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307,  2.3703604 ,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307,  5.997699  ,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307,  2.0703127 , -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307, -0.21651307, -0.21651307, -0.21651307, -0.21651307,\n",
       "       -0.21651307], dtype=float32)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.obs_vector('Znrd1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "e3534e1f",
   "metadata": {},
   "outputs": [],
   "source": [
    "Xa = test.obs_vector('Nrf1')\n",
    "Ya = test.obs_vector('Psmb1')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "596246d9",
   "metadata": {},
   "source": [
    "#### 2. Compute the correlation <br>\n",
    "Correlation is computed with numpy.corrcoef(Xa,Ya)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "262c00c2",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.20010931316835748"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "w1 = np.corrcoef(Xa, Ya)[0, 1]\n",
    "w1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "b76def0a",
   "metadata": {},
   "outputs": [],
   "source": [
    "import astropy.stats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "e9fcae87",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.2885177550846927"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "w3 = astropy.stats.biweight_midcorrelation(Xa, Ya)\n",
    "w3"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d877e8a6",
   "metadata": {},
   "source": [
    "#### 3. Get the same genes (vectors) from expression matrix of branch B and compute correlation"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6a1f9c0c",
   "metadata": {},
   "source": [
    "Vector Xb is expression vector for gene X in branch B. <br>\n",
    "Vector Yb is expression vector for gene Y in branch B."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "f709f34a",
   "metadata": {},
   "outputs": [],
   "source": [
    "test = branchBwindows[4].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "dee96fdb",
   "metadata": {},
   "outputs": [],
   "source": [
    "Xb = test.obs_vector('Nrf1')\n",
    "Yb = test.obs_vector('Psmb1')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "30aac84f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.13770408886638671"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "w2 = np.corrcoef(Xb, Yb)[0, 1]\n",
    "w2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "88a9eb55",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.29751276908013824"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "w4 = astropy.stats.biweight_midcorrelation(Xb, Yb)\n",
    "w4"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "acf6e13f",
   "metadata": {},
   "source": [
    "#### 4. Get the absolute value $|w_1-w_2|$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "da232f94",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.06240522430197076"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "abs(w1-w2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "0c5ab717",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.008995013995445555"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "abs(w3-w4)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "84a629a4",
   "metadata": {},
   "source": [
    "#### Do this for each pair in one window: (no filtering before)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "e2fe3aa8",
   "metadata": {},
   "outputs": [],
   "source": [
    "import math"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "17d31b05",
   "metadata": {},
   "outputs": [],
   "source": [
    "A = branchAwindows[4].copy()\n",
    "B = branchBwindows[4].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "bdb9bef7",
   "metadata": {},
   "outputs": [],
   "source": [
    "PATH_TO_ABS_MATRICES='data_files/abs_networks/'\n",
    "dataframe = pd.read_csv(f'{PATH_TO_ABS_MATRICES}win3_abs.csv')\n",
    "tfs = set(dataframe['TF'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "7af498ad",
   "metadata": {},
   "outputs": [],
   "source": [
    "def filter_df(df, threshold):\n",
    "    df_new = df[df['abs'] > threshold]\n",
    "    df_new.reset_index(drop=True, inplace=True)\n",
    "    return df_new"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "0497cfcd",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "def get_corr_abs_corrcoef(tfs, A, B):\n",
    "    correlation_abs = []\n",
    "    wA=0.\n",
    "    wB=0.\n",
    "    for tf in tfs:\n",
    "        for gene in A.var_names:\n",
    "            if tf==gene: \n",
    "                w=0.\n",
    "            else:\n",
    "                wA = np.corrcoef(A.obs_vector(tf), A.obs_vector(gene))[0, 1]\n",
    "                wB = np.corrcoef(B.obs_vector(tf), B.obs_vector(gene))[0, 1]\n",
    "                w = abs(wA-wB)\n",
    "            correlation_abs.append(w)\n",
    "    return correlation_abs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "ac3fc074",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_corr_abs_bicor(tfs, A, B):    \n",
    "    correlation_abs = []\n",
    "    wA=0.\n",
    "    wB=0.\n",
    "    for tf in tfs:\n",
    "        for gene in A.var_names:\n",
    "            if tf==gene: \n",
    "                w=0.\n",
    "            else:\n",
    "                wA = astropy.stats.biweight_midcorrelation(A.obs_vector(tf), A.obs_vector(gene))\n",
    "                wB = astropy.stats.biweight_midcorrelation(B.obs_vector(tf), B.obs_vector(gene))\n",
    "                w = abs(wA-wB)\n",
    "            correlation_abs.append(w)\n",
    "    return correlation_abs\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "57a6ffa0",
   "metadata": {},
   "outputs": [],
   "source": [
    "def corr_abs_to_df(tfs, A, B):\n",
    "    rows = []\n",
    "    for tf in tfs:\n",
    "        c_tf = tf.split(\"_\")[0]\n",
    "        for gene in A.var_names:\n",
    "            if c_tf==gene: \n",
    "                w=0.\n",
    "            else:\n",
    "                wA = np.corrcoef(A.obs_vector(c_tf), A.obs_vector(gene))[0, 1]\n",
    "                wB = np.corrcoef(B.obs_vector(c_tf), B.obs_vector(gene))[0, 1]\n",
    "                w = abs(wA-wB)\n",
    "            rows.append([tf, gene, w])\n",
    "    return pd.DataFrame(rows, columns=['TF', 'GENE', 'cc_abs'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "9f375ad4",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/nfs/home/students/l.willruth/.conda/envs/myenv/lib/python3.6/site-packages/numpy/lib/function_base.py:2559: RuntimeWarning: invalid value encountered in true_divide\n",
      "  c /= stddev[:, None]\n",
      "/nfs/home/students/l.willruth/.conda/envs/myenv/lib/python3.6/site-packages/numpy/lib/function_base.py:2560: RuntimeWarning: invalid value encountered in true_divide\n",
      "  c /= stddev[None, :]\n",
      "/nfs/home/students/l.willruth/.conda/envs/myenv/lib/python3.6/site-packages/astropy/stats/biweight.py:710: RuntimeWarning: invalid value encountered in double_scalars\n",
      "  return bicorr[0, 1] / (np.sqrt(bicorr[0, 0] * bicorr[1, 1]))\n"
     ]
    }
   ],
   "source": [
    "correlation_abs_corrcoef = get_corr_abs_corrcoef(tfs, A, B)\n",
    "correlation_abs_bicor = get_corr_abs_bicor(tfs, A, B)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2bc454f3",
   "metadata": {},
   "source": [
    "Plot distribution of absolute correlation values:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "possible-graduate",
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "9088d04c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([1.7195e+04, 1.4801e+04, 1.1606e+04, 9.2600e+03, 8.1260e+03,\n",
       "        7.3530e+03, 6.7920e+03, 6.1720e+03, 5.7690e+03, 5.4280e+03,\n",
       "        5.0330e+03, 4.5880e+03, 4.3340e+03, 3.9700e+03, 3.6210e+03,\n",
       "        3.1770e+03, 2.9570e+03, 2.6240e+03, 2.3640e+03, 2.1130e+03,\n",
       "        1.8810e+03, 1.5840e+03, 1.4760e+03, 1.2080e+03, 1.1200e+03,\n",
       "        9.7700e+02, 8.5600e+02, 7.3300e+02, 6.3000e+02, 5.6600e+02,\n",
       "        4.9000e+02, 4.5200e+02, 3.9800e+02, 3.3600e+02, 2.8400e+02,\n",
       "        2.5900e+02, 1.9800e+02, 1.8600e+02, 2.1700e+02, 1.7300e+02,\n",
       "        1.6700e+02, 1.6000e+02, 1.3900e+02, 1.0300e+02, 8.9000e+01,\n",
       "        9.7000e+01, 8.1000e+01, 7.1000e+01, 7.2000e+01, 6.5000e+01,\n",
       "        6.3000e+01, 4.9000e+01, 4.5000e+01, 4.4000e+01, 2.7000e+01,\n",
       "        3.5000e+01, 2.3000e+01, 1.7000e+01, 2.0000e+01, 2.1000e+01,\n",
       "        1.3000e+01, 1.4000e+01, 1.5000e+01, 7.0000e+00, 5.0000e+00,\n",
       "        1.0000e+01, 6.0000e+00, 8.0000e+00, 6.0000e+00, 6.0000e+00,\n",
       "        9.0000e+00, 5.0000e+00, 4.0000e+00, 7.0000e+00, 3.0000e+00,\n",
       "        2.0000e+00, 3.0000e+00, 7.0000e+00, 2.0000e+00, 0.0000e+00,\n",
       "        3.0000e+00, 1.0000e+00, 5.0000e+00, 0.0000e+00, 1.0000e+00,\n",
       "        0.0000e+00, 0.0000e+00, 1.0000e+00, 1.0000e+00, 1.0000e+00,\n",
       "        2.0000e+00, 0.0000e+00, 0.0000e+00, 1.0000e+00, 1.0000e+00,\n",
       "        0.0000e+00, 1.0000e+00, 0.0000e+00, 1.0000e+00, 2.0000e+00]),\n",
       " array([0.        , 0.01012076, 0.02024152, 0.03036228, 0.04048304,\n",
       "        0.0506038 , 0.06072456, 0.07084533, 0.08096609, 0.09108685,\n",
       "        0.10120761, 0.11132837, 0.12144913, 0.13156989, 0.14169065,\n",
       "        0.15181141, 0.16193217, 0.17205293, 0.18217369, 0.19229445,\n",
       "        0.20241522, 0.21253598, 0.22265674, 0.2327775 , 0.24289826,\n",
       "        0.25301902, 0.26313978, 0.27326054, 0.2833813 , 0.29350206,\n",
       "        0.30362282, 0.31374358, 0.32386434, 0.3339851 , 0.34410587,\n",
       "        0.35422663, 0.36434739, 0.37446815, 0.38458891, 0.39470967,\n",
       "        0.40483043, 0.41495119, 0.42507195, 0.43519271, 0.44531347,\n",
       "        0.45543423, 0.46555499, 0.47567576, 0.48579652, 0.49591728,\n",
       "        0.50603804, 0.5161588 , 0.52627956, 0.53640032, 0.54652108,\n",
       "        0.55664184, 0.5667626 , 0.57688336, 0.58700412, 0.59712488,\n",
       "        0.60724565, 0.61736641, 0.62748717, 0.63760793, 0.64772869,\n",
       "        0.65784945, 0.66797021, 0.67809097, 0.68821173, 0.69833249,\n",
       "        0.70845325, 0.71857401, 0.72869477, 0.73881553, 0.7489363 ,\n",
       "        0.75905706, 0.76917782, 0.77929858, 0.78941934, 0.7995401 ,\n",
       "        0.80966086, 0.81978162, 0.82990238, 0.84002314, 0.8501439 ,\n",
       "        0.86026466, 0.87038542, 0.88050619, 0.89062695, 0.90074771,\n",
       "        0.91086847, 0.92098923, 0.93110999, 0.94123075, 0.95135151,\n",
       "        0.96147227, 0.97159303, 0.98171379, 0.99183455, 1.00195531,\n",
       "        1.01207608]),\n",
       " <BarContainer object of 100 artists>)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "ax.hist(correlation_abs_corrcoef, bins=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "ec4901fc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([2.0171e+04, 7.9200e+03, 7.6570e+03, 7.2950e+03, 6.9120e+03,\n",
       "        6.4830e+03, 6.3230e+03, 6.2450e+03, 6.1840e+03, 5.8980e+03,\n",
       "        5.8340e+03, 5.3010e+03, 5.0810e+03, 4.5730e+03, 4.2530e+03,\n",
       "        3.8700e+03, 3.4090e+03, 3.2680e+03, 2.8470e+03, 2.6450e+03,\n",
       "        2.3570e+03, 2.0550e+03, 1.8340e+03, 1.6860e+03, 1.4940e+03,\n",
       "        1.2710e+03, 1.1340e+03, 1.0300e+03, 8.5000e+02, 7.8600e+02,\n",
       "        6.9600e+02, 6.2800e+02, 5.0100e+02, 4.8000e+02, 4.7600e+02,\n",
       "        3.5200e+02, 3.0600e+02, 2.9800e+02, 2.4700e+02, 2.7300e+02,\n",
       "        2.3000e+02, 1.9400e+02, 1.6100e+02, 1.5000e+02, 1.5200e+02,\n",
       "        1.2000e+02, 9.7000e+01, 1.3700e+02, 8.2000e+01, 7.4000e+01,\n",
       "        7.1000e+01, 4.9000e+01, 4.5000e+01, 3.8000e+01, 4.4000e+01,\n",
       "        3.4000e+01, 4.0000e+01, 2.3000e+01, 1.4000e+01, 1.8000e+01,\n",
       "        2.1000e+01, 1.6000e+01, 1.2000e+01, 1.2000e+01, 1.3000e+01,\n",
       "        7.0000e+00, 5.0000e+00, 7.0000e+00, 7.0000e+00, 5.0000e+00,\n",
       "        6.0000e+00, 2.0000e+00, 3.0000e+00, 0.0000e+00, 4.0000e+00,\n",
       "        2.0000e+00, 7.0000e+00, 1.0000e+00, 1.0000e+00, 1.0000e+00,\n",
       "        1.0000e+00, 2.0000e+00, 4.0000e+00, 1.0000e+00, 2.0000e+00,\n",
       "        1.0000e+00, 1.0000e+00, 1.0000e+00, 1.0000e+00, 0.0000e+00,\n",
       "        0.0000e+00, 1.0000e+00, 1.0000e+00, 0.0000e+00, 1.0000e+00,\n",
       "        0.0000e+00, 0.0000e+00, 0.0000e+00, 1.0000e+00, 2.0000e+00]),\n",
       " array([0.        , 0.00893307, 0.01786614, 0.0267992 , 0.03573227,\n",
       "        0.04466534, 0.05359841, 0.06253148, 0.07146455, 0.08039761,\n",
       "        0.08933068, 0.09826375, 0.10719682, 0.11612989, 0.12506296,\n",
       "        0.13399602, 0.14292909, 0.15186216, 0.16079523, 0.1697283 ,\n",
       "        0.17866136, 0.18759443, 0.1965275 , 0.20546057, 0.21439364,\n",
       "        0.22332671, 0.23225977, 0.24119284, 0.25012591, 0.25905898,\n",
       "        0.26799205, 0.27692512, 0.28585818, 0.29479125, 0.30372432,\n",
       "        0.31265739, 0.32159046, 0.33052352, 0.33945659, 0.34838966,\n",
       "        0.35732273, 0.3662558 , 0.37518887, 0.38412193, 0.393055  ,\n",
       "        0.40198807, 0.41092114, 0.41985421, 0.42878728, 0.43772034,\n",
       "        0.44665341, 0.45558648, 0.46451955, 0.47345262, 0.48238568,\n",
       "        0.49131875, 0.50025182, 0.50918489, 0.51811796, 0.52705103,\n",
       "        0.53598409, 0.54491716, 0.55385023, 0.5627833 , 0.57171637,\n",
       "        0.58064944, 0.5895825 , 0.59851557, 0.60744864, 0.61638171,\n",
       "        0.62531478, 0.63424785, 0.64318091, 0.65211398, 0.66104705,\n",
       "        0.66998012, 0.67891319, 0.68784625, 0.69677932, 0.70571239,\n",
       "        0.71464546, 0.72357853, 0.7325116 , 0.74144466, 0.75037773,\n",
       "        0.7593108 , 0.76824387, 0.77717694, 0.78611001, 0.79504307,\n",
       "        0.80397614, 0.81290921, 0.82184228, 0.83077535, 0.83970841,\n",
       "        0.84864148, 0.85757455, 0.86650762, 0.87544069, 0.88437376,\n",
       "        0.89330682]),\n",
       " <BarContainer object of 100 artists>)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "ax.hist(correlation_abs_bicor, bins=100)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "67949fc8",
   "metadata": {},
   "source": [
    "### 5. Do it for all windows\n",
    "#### Filtered dataframe before and just compute correlation for remaining tf-gene entries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "9ad27ac8",
   "metadata": {},
   "outputs": [],
   "source": [
    "PATH_TO_ABS_MATRICES = 'data_files/abs_wo0/'\n",
    "\n",
    "dataframe = pd.read_csv(f'{PATH_TO_ABS_MATRICES}win3_abs.csv')\n",
    "tfs = set(dataframe['TF'])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "991b8c7e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TF</th>\n",
       "      <th>GENE</th>\n",
       "      <th>abs</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>4833420G17Rik</td>\n",
       "      <td>4.151750</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>AI467606</td>\n",
       "      <td>0.082890</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>Ablim1</td>\n",
       "      <td>4.325007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>Acap1</td>\n",
       "      <td>2.208852</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>Acap2</td>\n",
       "      <td>0.129101</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13852</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp518a</td>\n",
       "      <td>3.991905</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13853</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp707</td>\n",
       "      <td>0.763305</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13854</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfx</td>\n",
       "      <td>1.284463</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13855</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zkscan3</td>\n",
       "      <td>0.432063</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13856</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zmynd8</td>\n",
       "      <td>1.866047</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>13857 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         TF           GENE       abs\n",
       "0      Ezh2  4833420G17Rik  4.151750\n",
       "1      Ezh2       AI467606  0.082890\n",
       "2      Ezh2         Ablim1  4.325007\n",
       "3      Ezh2          Acap1  2.208852\n",
       "4      Ezh2          Acap2  0.129101\n",
       "...     ...            ...       ...\n",
       "13852  Rest        Zfp518a  3.991905\n",
       "13853  Rest         Zfp707  0.763305\n",
       "13854  Rest            Zfx  1.284463\n",
       "13855  Rest        Zkscan3  0.432063\n",
       "13856  Rest         Zmynd8  1.866047\n",
       "\n",
       "[13857 rows x 3 columns]"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "5032c74e",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/nfs/home/students/l.willruth/.conda/envs/myenv/lib/python3.6/site-packages/astropy/stats/biweight.py:710: RuntimeWarning: invalid value encountered in double_scalars\n",
      "  return bicorr[0, 1] / (np.sqrt(bicorr[0, 0] * bicorr[1, 1]))\n",
      "/nfs/home/students/l.willruth/.conda/envs/myenv/lib/python3.6/site-packages/numpy/lib/function_base.py:2559: RuntimeWarning: invalid value encountered in true_divide\n",
      "  c /= stddev[:, None]\n",
      "/nfs/home/students/l.willruth/.conda/envs/myenv/lib/python3.6/site-packages/numpy/lib/function_base.py:2560: RuntimeWarning: invalid value encountered in true_divide\n",
      "  c /= stddev[None, :]\n",
      "/nfs/home/students/l.willruth/.conda/envs/myenv/lib/python3.6/site-packages/astropy/stats/biweight.py:710: RuntimeWarning: invalid value encountered in double_scalars\n",
      "  return bicorr[0, 1] / (np.sqrt(bicorr[0, 0] * bicorr[1, 1]))\n",
      "/nfs/home/students/l.willruth/.conda/envs/myenv/lib/python3.6/site-packages/numpy/lib/function_base.py:2559: RuntimeWarning: invalid value encountered in true_divide\n",
      "  c /= stddev[:, None]\n",
      "/nfs/home/students/l.willruth/.conda/envs/myenv/lib/python3.6/site-packages/numpy/lib/function_base.py:2560: RuntimeWarning: invalid value encountered in true_divide\n",
      "  c /= stddev[None, :]\n",
      "/nfs/home/students/l.willruth/.conda/envs/myenv/lib/python3.6/site-packages/astropy/stats/biweight.py:710: RuntimeWarning: invalid value encountered in double_scalars\n",
      "  return bicorr[0, 1] / (np.sqrt(bicorr[0, 0] * bicorr[1, 1]))\n",
      "/nfs/home/students/l.willruth/.conda/envs/myenv/lib/python3.6/site-packages/numpy/lib/function_base.py:2559: RuntimeWarning: invalid value encountered in true_divide\n",
      "  c /= stddev[:, None]\n",
      "/nfs/home/students/l.willruth/.conda/envs/myenv/lib/python3.6/site-packages/numpy/lib/function_base.py:2560: RuntimeWarning: invalid value encountered in true_divide\n",
      "  c /= stddev[None, :]\n"
     ]
    }
   ],
   "source": [
    "w_cc_dataframes = []\n",
    "\n",
    "for i in range(0, len(branchAwindows)):\n",
    "    dataframe = pd.read_csv(f'{PATH_TO_ABS_MATRICES}win{i}_abs.csv')\n",
    "    dataframe = filter_df(dataframe, 0.5) # filter out all values below 0.5\n",
    "    \n",
    "    A = branchAwindows[i].copy()\n",
    "    B = branchBwindows[i].copy() \n",
    "    \n",
    "    rows = []\n",
    "    \n",
    "    for index, row in dataframe.iterrows():\n",
    "        tf = row['TF']\n",
    "        gene = row['GENE']\n",
    "        w = row['abs']\n",
    "        wA = astropy.stats.biweight_midcorrelation(A.obs_vector(tf), A.obs_vector(gene))\n",
    "        wB = astropy.stats.biweight_midcorrelation(B.obs_vector(tf), B.obs_vector(gene))\n",
    "        cc_bicor = abs(wA-wB)\n",
    "        wA = np.corrcoef(A.obs_vector(tf), A.obs_vector(gene))[0, 1]\n",
    "        wB = np.corrcoef(B.obs_vector(tf), B.obs_vector(gene))[0, 1]\n",
    "        cc_coef = abs(wA-wB)\n",
    "\n",
    "        \n",
    "        rows.append([tf, gene, w, cc_coef, cc_bicor])\n",
    "    w_cc_dataframes.append(pd.DataFrame(rows, columns=['TF', 'GENE', 'w_abs', 'cc_coef_abs', 'cc_bicor_abs']))    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "d2f1e5d3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TF</th>\n",
       "      <th>GENE</th>\n",
       "      <th>w_abs</th>\n",
       "      <th>cc_coef_abs</th>\n",
       "      <th>cc_bicor_abs</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>1810037I17Rik</td>\n",
       "      <td>0.759367</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>2010111I01Rik</td>\n",
       "      <td>0.643845</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>2210016F16Rik</td>\n",
       "      <td>2.120124</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>2310011J03Rik</td>\n",
       "      <td>3.651726</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>Actr2</td>\n",
       "      <td>3.651726</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>8505</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp326</td>\n",
       "      <td>3.701298</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8506</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp407</td>\n",
       "      <td>4.984579</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8507</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp512</td>\n",
       "      <td>0.733906</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8508</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zmynd8</td>\n",
       "      <td>1.499678</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8509</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zranb2</td>\n",
       "      <td>2.455203</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8510 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        TF           GENE     w_abs  cc_coef_abs  cc_bicor_abs\n",
       "0     Ezh2  1810037I17Rik  0.759367          0.0           0.0\n",
       "1     Ezh2  2010111I01Rik  0.643845          0.0           0.0\n",
       "2     Ezh2  2210016F16Rik  2.120124          0.0           0.0\n",
       "3     Ezh2  2310011J03Rik  3.651726          0.0           0.0\n",
       "4     Ezh2          Actr2  3.651726          0.0           0.0\n",
       "...    ...            ...       ...          ...           ...\n",
       "8505  Rest         Zfp326  3.701298          0.0           0.0\n",
       "8506  Rest         Zfp407  4.984579          0.0           0.0\n",
       "8507  Rest         Zfp512  0.733906          0.0           0.0\n",
       "8508  Rest         Zmynd8  1.499678          0.0           0.0\n",
       "8509  Rest         Zranb2  2.455203          0.0           0.0\n",
       "\n",
       "[8510 rows x 5 columns]"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "w_cc_dataframes[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "1b6983f6",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TF</th>\n",
       "      <th>GENE</th>\n",
       "      <th>w_abs</th>\n",
       "      <th>cc_coef_abs</th>\n",
       "      <th>cc_bicor_abs</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>1810037I17Rik</td>\n",
       "      <td>1.800194</td>\n",
       "      <td>0.073319</td>\n",
       "      <td>0.064313</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>2210016F16Rik</td>\n",
       "      <td>3.325726</td>\n",
       "      <td>0.161181</td>\n",
       "      <td>0.164022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>3830406C13Rik</td>\n",
       "      <td>0.729930</td>\n",
       "      <td>0.029488</td>\n",
       "      <td>0.009544</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>4833420G17Rik</td>\n",
       "      <td>2.802151</td>\n",
       "      <td>0.114864</td>\n",
       "      <td>0.105545</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>4931406P16Rik</td>\n",
       "      <td>0.839976</td>\n",
       "      <td>0.016909</td>\n",
       "      <td>0.009813</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>11755</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zc3h7a</td>\n",
       "      <td>2.047629</td>\n",
       "      <td>0.107217</td>\n",
       "      <td>0.114232</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11756</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zcchc6</td>\n",
       "      <td>1.112263</td>\n",
       "      <td>0.075051</td>\n",
       "      <td>0.052891</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11757</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zcchc7</td>\n",
       "      <td>0.947237</td>\n",
       "      <td>0.166832</td>\n",
       "      <td>0.134063</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11758</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp110</td>\n",
       "      <td>1.704517</td>\n",
       "      <td>0.180745</td>\n",
       "      <td>0.152108</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11759</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp518a</td>\n",
       "      <td>2.545763</td>\n",
       "      <td>0.163253</td>\n",
       "      <td>0.112958</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>11760 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         TF           GENE     w_abs  cc_coef_abs  cc_bicor_abs\n",
       "0      Ezh2  1810037I17Rik  1.800194     0.073319      0.064313\n",
       "1      Ezh2  2210016F16Rik  3.325726     0.161181      0.164022\n",
       "2      Ezh2  3830406C13Rik  0.729930     0.029488      0.009544\n",
       "3      Ezh2  4833420G17Rik  2.802151     0.114864      0.105545\n",
       "4      Ezh2  4931406P16Rik  0.839976     0.016909      0.009813\n",
       "...     ...            ...       ...          ...           ...\n",
       "11755  Rest         Zc3h7a  2.047629     0.107217      0.114232\n",
       "11756  Rest         Zcchc6  1.112263     0.075051      0.052891\n",
       "11757  Rest         Zcchc7  0.947237     0.166832      0.134063\n",
       "11758  Rest         Zfp110  1.704517     0.180745      0.152108\n",
       "11759  Rest        Zfp518a  2.545763     0.163253      0.112958\n",
       "\n",
       "[11760 rows x 5 columns]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "w_cc_dataframes[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "9ed01430",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TF</th>\n",
       "      <th>GENE</th>\n",
       "      <th>w_abs</th>\n",
       "      <th>cc_coef_abs</th>\n",
       "      <th>cc_bicor_abs</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>1810058I24Rik</td>\n",
       "      <td>2.071438</td>\n",
       "      <td>0.410831</td>\n",
       "      <td>0.343949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>2410015M20Rik</td>\n",
       "      <td>2.838728</td>\n",
       "      <td>0.483066</td>\n",
       "      <td>0.366533</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>3830406C13Rik</td>\n",
       "      <td>2.071438</td>\n",
       "      <td>0.203650</td>\n",
       "      <td>0.208148</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>4833420G17Rik</td>\n",
       "      <td>1.095240</td>\n",
       "      <td>0.192126</td>\n",
       "      <td>0.155420</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>4921524J17Rik</td>\n",
       "      <td>2.865398</td>\n",
       "      <td>0.816723</td>\n",
       "      <td>0.698781</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>12076</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zcchc7</td>\n",
       "      <td>0.698076</td>\n",
       "      <td>0.385435</td>\n",
       "      <td>0.245508</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12077</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp207</td>\n",
       "      <td>2.790385</td>\n",
       "      <td>0.160590</td>\n",
       "      <td>0.269295</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12078</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp407</td>\n",
       "      <td>2.739675</td>\n",
       "      <td>0.478757</td>\n",
       "      <td>0.390642</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12079</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp518a</td>\n",
       "      <td>1.521264</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12080</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zmynd8</td>\n",
       "      <td>3.134357</td>\n",
       "      <td>0.059448</td>\n",
       "      <td>0.173098</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>12081 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         TF           GENE     w_abs  cc_coef_abs  cc_bicor_abs\n",
       "0      Ezh2  1810058I24Rik  2.071438     0.410831      0.343949\n",
       "1      Ezh2  2410015M20Rik  2.838728     0.483066      0.366533\n",
       "2      Ezh2  3830406C13Rik  2.071438     0.203650      0.208148\n",
       "3      Ezh2  4833420G17Rik  1.095240     0.192126      0.155420\n",
       "4      Ezh2  4921524J17Rik  2.865398     0.816723      0.698781\n",
       "...     ...            ...       ...          ...           ...\n",
       "12076  Rest         Zcchc7  0.698076     0.385435      0.245508\n",
       "12077  Rest         Zfp207  2.790385     0.160590      0.269295\n",
       "12078  Rest         Zfp407  2.739675     0.478757      0.390642\n",
       "12079  Rest        Zfp518a  1.521264          NaN           NaN\n",
       "12080  Rest         Zmynd8  3.134357     0.059448      0.173098\n",
       "\n",
       "[12081 rows x 5 columns]"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "w_cc_dataframes[2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "982092b6",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TF</th>\n",
       "      <th>GENE</th>\n",
       "      <th>w_abs</th>\n",
       "      <th>cc_coef_abs</th>\n",
       "      <th>cc_bicor_abs</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>4833420G17Rik</td>\n",
       "      <td>4.151750</td>\n",
       "      <td>0.169465</td>\n",
       "      <td>0.155415</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>Ablim1</td>\n",
       "      <td>4.325007</td>\n",
       "      <td>0.228184</td>\n",
       "      <td>0.258895</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>Acap1</td>\n",
       "      <td>2.208852</td>\n",
       "      <td>0.176180</td>\n",
       "      <td>0.158432</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>Acot13</td>\n",
       "      <td>1.434090</td>\n",
       "      <td>0.156252</td>\n",
       "      <td>0.076892</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>Adam19</td>\n",
       "      <td>2.973806</td>\n",
       "      <td>0.018931</td>\n",
       "      <td>0.005074</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>12417</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zbtb1</td>\n",
       "      <td>7.573699</td>\n",
       "      <td>0.242525</td>\n",
       "      <td>0.198932</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12418</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp518a</td>\n",
       "      <td>3.991905</td>\n",
       "      <td>0.172934</td>\n",
       "      <td>0.126091</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12419</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp707</td>\n",
       "      <td>0.763305</td>\n",
       "      <td>0.145737</td>\n",
       "      <td>0.107849</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12420</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfx</td>\n",
       "      <td>1.284463</td>\n",
       "      <td>0.193232</td>\n",
       "      <td>0.075004</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12421</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zmynd8</td>\n",
       "      <td>1.866047</td>\n",
       "      <td>0.065771</td>\n",
       "      <td>0.033780</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>12422 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         TF           GENE     w_abs  cc_coef_abs  cc_bicor_abs\n",
       "0      Ezh2  4833420G17Rik  4.151750     0.169465      0.155415\n",
       "1      Ezh2         Ablim1  4.325007     0.228184      0.258895\n",
       "2      Ezh2          Acap1  2.208852     0.176180      0.158432\n",
       "3      Ezh2         Acot13  1.434090     0.156252      0.076892\n",
       "4      Ezh2         Adam19  2.973806     0.018931      0.005074\n",
       "...     ...            ...       ...          ...           ...\n",
       "12417  Rest          Zbtb1  7.573699     0.242525      0.198932\n",
       "12418  Rest        Zfp518a  3.991905     0.172934      0.126091\n",
       "12419  Rest         Zfp707  0.763305     0.145737      0.107849\n",
       "12420  Rest            Zfx  1.284463     0.193232      0.075004\n",
       "12421  Rest         Zmynd8  1.866047     0.065771      0.033780\n",
       "\n",
       "[12422 rows x 5 columns]"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "w_cc_dataframes[3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "d1cf66e3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TF</th>\n",
       "      <th>GENE</th>\n",
       "      <th>w_abs</th>\n",
       "      <th>cc_coef_abs</th>\n",
       "      <th>cc_bicor_abs</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>0610030E20Rik</td>\n",
       "      <td>3.792624</td>\n",
       "      <td>0.169295</td>\n",
       "      <td>0.142573</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>1810058I24Rik</td>\n",
       "      <td>3.968208</td>\n",
       "      <td>0.195791</td>\n",
       "      <td>0.185573</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>2210016F16Rik</td>\n",
       "      <td>2.470725</td>\n",
       "      <td>0.084101</td>\n",
       "      <td>0.096360</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>2410015M20Rik</td>\n",
       "      <td>3.968208</td>\n",
       "      <td>0.544187</td>\n",
       "      <td>0.297660</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>Abi1</td>\n",
       "      <td>3.908071</td>\n",
       "      <td>0.185623</td>\n",
       "      <td>0.272632</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>12570</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp518a</td>\n",
       "      <td>2.484976</td>\n",
       "      <td>0.109616</td>\n",
       "      <td>0.142073</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12571</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp707</td>\n",
       "      <td>3.353105</td>\n",
       "      <td>0.014814</td>\n",
       "      <td>0.055117</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12572</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfx</td>\n",
       "      <td>0.528062</td>\n",
       "      <td>0.171519</td>\n",
       "      <td>0.230108</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12573</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zmynd11</td>\n",
       "      <td>1.311668</td>\n",
       "      <td>0.141789</td>\n",
       "      <td>0.153280</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12574</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zmynd8</td>\n",
       "      <td>1.924801</td>\n",
       "      <td>0.359896</td>\n",
       "      <td>0.256965</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>12575 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         TF           GENE     w_abs  cc_coef_abs  cc_bicor_abs\n",
       "0      Ezh2  0610030E20Rik  3.792624     0.169295      0.142573\n",
       "1      Ezh2  1810058I24Rik  3.968208     0.195791      0.185573\n",
       "2      Ezh2  2210016F16Rik  2.470725     0.084101      0.096360\n",
       "3      Ezh2  2410015M20Rik  3.968208     0.544187      0.297660\n",
       "4      Ezh2           Abi1  3.908071     0.185623      0.272632\n",
       "...     ...            ...       ...          ...           ...\n",
       "12570  Rest        Zfp518a  2.484976     0.109616      0.142073\n",
       "12571  Rest         Zfp707  3.353105     0.014814      0.055117\n",
       "12572  Rest            Zfx  0.528062     0.171519      0.230108\n",
       "12573  Rest        Zmynd11  1.311668     0.141789      0.153280\n",
       "12574  Rest         Zmynd8  1.924801     0.359896      0.256965\n",
       "\n",
       "[12575 rows x 5 columns]"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "w_cc_dataframes[4]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9404501e",
   "metadata": {},
   "source": [
    "Write result dataframes to csv files:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "ea538e37",
   "metadata": {},
   "outputs": [],
   "source": [
    "PATH_TO_W_CC_DFS = 'data_files/correlation_abs/'\n",
    "for i in range(0,len(w_cc_dataframes)):\n",
    "    df = w_cc_dataframes[i]\n",
    "    df.to_csv(f'{PATH_TO_W_CC_DFS}w{i}_SCGRN_correlation.csv', index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5caa03fb",
   "metadata": {},
   "source": [
    "Plot distribution of Abs[w] and Abs[cc]:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "c6e18a90",
   "metadata": {},
   "outputs": [],
   "source": [
    "w_vals = []\n",
    "cc_coef_vals = []\n",
    "cc_bicor_vals = []\n",
    "\n",
    "for i in range(len(w_cc_dataframes)):\n",
    "    w_vals.append([float(x) for x in w_cc_dataframes[i]['w_abs']])\n",
    "    cc_coef_vals.append( [float(x) for x in w_cc_dataframes[i]['cc_coef_abs']])\n",
    "    cc_bicor_vals.append( [float(x) for x in w_cc_dataframes[i]['cc_bicor_abs']])\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "368a4e12",
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot(w_vals, cc_coef_vals, cc_bicor_vals, title):\n",
    "    fig, ax = plt.subplots(3,1, figsize =(10,8), sharey=True )\n",
    "    ax[0].hist(w_vals, bins=100, label='w_abs')\n",
    "    ax[0].set_title('Abs[w]')\n",
    "    ax[1].hist(cc_coef_vals, bins=100, label='cc_coef_abs')\n",
    "    ax[1].set_title('Abs[cc_coef]')\n",
    "    ax[2].hist(cc_bicor_vals, bins=100, label='cc_bicor_abs')\n",
    "    ax[2].set_title('Abs[cc_bicor]')\n",
    "    fig.suptitle(title, fontsize=16)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "5bf3472e",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i in range (0, len(w_vals)):\n",
    "    plot(w_vals[i], cc_coef_vals[i], cc_bicor_vals[i], f'window {i}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "404985d1",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "08bbb714",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a897bfec",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "a9682711",
   "metadata": {},
   "source": [
    "### 6. Determine which correlation method to use"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "6532ca67",
   "metadata": {},
   "outputs": [],
   "source": [
    "# method to compare the ranks when sorting by column 1 and by comlumn 2\n",
    "def compare_ranks(dataframe, column1, column2):\n",
    "    sums = []\n",
    "    for tf in set(dataframe['TF']):\n",
    "        df = dataframe[dataframe['TF']==tf]\n",
    "        df1 = df.sort_values(column1)\n",
    "        ranks1 = list(df1.index)\n",
    "        df2 = df.sort_values(column2)\n",
    "        ranks2 = list(df2.index)\n",
    "        \n",
    "        sum = 0\n",
    "        for i in range(len(df)):\n",
    "            sum = sum + abs(ranks1[i]-ranks2[i])\n",
    "        sum = sum/len(ranks1)\n",
    "        sums.append(sum)\n",
    "    \n",
    "    return sums"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "2352c508",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[2.0,\n",
       " 11.743589743589743,\n",
       " 9.235294117647058,\n",
       " 93.05384615384615,\n",
       " 23.81081081081081,\n",
       " 3.7333333333333334,\n",
       " 56.31515151515151,\n",
       " 52.32258064516129,\n",
       " 75.93636363636364,\n",
       " 40.39344262295082,\n",
       " 2.4285714285714284,\n",
       " 7.7,\n",
       " 114.98780487804878,\n",
       " 4.166666666666667,\n",
       " 2.6666666666666665,\n",
       " 20.46875,\n",
       " 5.818181818181818,\n",
       " 1.6,\n",
       " 118.97450424929178,\n",
       " 15.047619047619047,\n",
       " 106.06430868167203,\n",
       " 19.6875,\n",
       " 158.869918699187,\n",
       " 5.6923076923076925,\n",
       " 4.666666666666667,\n",
       " 2.857142857142857,\n",
       " 4.142857142857143,\n",
       " 86.47509578544062,\n",
       " 6.705882352941177,\n",
       " 168.98585858585858,\n",
       " 13.116279069767442,\n",
       " 3.5,\n",
       " 26.714285714285715,\n",
       " 18.166666666666668,\n",
       " 68.53398058252426,\n",
       " 13.090909090909092,\n",
       " 15.5,\n",
       " 51.77922077922078,\n",
       " 118.14899713467048,\n",
       " 211.98742138364778,\n",
       " 10.64864864864865,\n",
       " 23.37313432835821,\n",
       " 184.97826086956522,\n",
       " 17.818181818181817,\n",
       " 3.142857142857143,\n",
       " 124.33333333333333,\n",
       " 139.3148614609572,\n",
       " 28.86315789473684,\n",
       " 115.10843373493977,\n",
       " 78.81967213114754,\n",
       " 143.33333333333334,\n",
       " 161.9205020920502,\n",
       " 9.8125,\n",
       " 54.82142857142857,\n",
       " 172.41980198019803,\n",
       " 0.0,\n",
       " 148.22685185185185,\n",
       " 58.566265060240966,\n",
       " 152.7922077922078,\n",
       " 2.3333333333333335,\n",
       " 9.741935483870968,\n",
       " 171.62226640159045,\n",
       " 39.163636363636364,\n",
       " 20.215384615384615,\n",
       " 140.24824355971896,\n",
       " 101.86292834890966,\n",
       " 62.03125,\n",
       " 2.0,\n",
       " 3.7777777777777777,\n",
       " 85.45255474452554,\n",
       " 170.94208494208493,\n",
       " 18.793650793650794]"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = w_cc_dataframes[4]\n",
    "sums_coef = compare_ranks(df, 'w_abs', 'cc_coef_abs')\n",
    "sums_coef"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "ec41e6b4",
   "metadata": {},
   "outputs": [],
   "source": [
    "sums_bicor = compare_ranks(df, 'w_abs', 'cc_bicor_abs')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "3c9cc869",
   "metadata": {},
   "outputs": [],
   "source": [
    "coef_diff = 0\n",
    "bicor_diff = 0\n",
    "for i in range(len(sums_coef)):\n",
    "    coef_diff += sums_coef[i]\n",
    "    bicor_diff += sums_bicor[i]\n",
    "    \n",
    "coef_diff = coef_diff/len(sums_coef)\n",
    "bicor_diff = bicor_diff/len(sums_coef)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "e2c2a93b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(58.660652147613696, 58.69767694738248)"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "coef_diff, bicor_diff"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "a743d3ea",
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib.pyplot import figure"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "b55058b1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f37f8296828>"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = [x for x in range(0,72)]\n",
    "\n",
    "figure(figsize=(15, 6))\n",
    "\n",
    "plt.plot(x, (sums_coef), 'bo', label='coef')\n",
    "plt.plot(x, (sums_bicor), 'rx', label='bicor')\n",
    "plt.legend()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "australian-script",
   "metadata": {},
   "source": [
    "I will use the pearson correlation coefficient but still they don't differ very much."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "choice-brick",
   "metadata": {},
   "source": [
    "### 7. Try to evaluate significance of correlation results with permutation test for p-value (not practicable yet)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f8826891",
   "metadata": {},
   "source": [
    "#### 7.1 Randomize expression matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "ba9bf123",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy.random as rd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "f6f47f14",
   "metadata": {},
   "outputs": [],
   "source": [
    "branchAwindows = []\n",
    "branchBwindows = []\n",
    "\n",
    "A_windows_files=(glob.glob(f'{PATH_TO_WINDOW_ANNDATA}*A.h5ad'))\n",
    "B_windows_files=(glob.glob(f'{PATH_TO_WINDOW_ANNDATA}*B.h5ad'))\n",
    "A_windows_files = sorted(A_windows_files)\n",
    "B_windows_files = sorted(B_windows_files)\n",
    "\n",
    "for i in range(len(A_windows_files)):\n",
    "    branchAwindows.append(sc.read_h5ad(A_windows_files[i]))\n",
    "    branchBwindows.append(sc.read_h5ad(B_windows_files[i]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "9c1eea9d",
   "metadata": {},
   "outputs": [],
   "source": [
    "def shuffle2d(arr2d):\n",
    "    \"\"\"Shuffes entries of 2-d array arr2d, preserving shape.\"\"\"\n",
    "    list2d = arr2d.tolist()\n",
    "    reshape = []\n",
    "    data = []\n",
    "    iend = 0\n",
    "    for row in list2d:\n",
    "        data.extend(row)\n",
    "        istart, iend = iend, iend+len(row)\n",
    "        reshape.append((istart, iend))\n",
    "    rd.shuffle(data)\n",
    "    return np.array([data[istart:iend] for (istart,iend) in reshape])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a676ef2a",
   "metadata": {},
   "source": [
    "Example:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "a19404a4",
   "metadata": {},
   "outputs": [],
   "source": [
    "matrix = branchAwindows[3].X\n",
    "shuffled_matrix = shuffle2d(matrix)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "1f6a7249",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([[-0.29072848, -0.3315403 , -0.47442967, ..., -0.31714624,\n",
       "         -0.33347058, -0.31607842],\n",
       "        [-0.29072848, -0.3315403 , -0.47442967, ..., -0.31714624,\n",
       "         -0.33347058, -0.31607842],\n",
       "        [-0.29072848,  4.056121  , -0.47442967, ..., -0.31714624,\n",
       "         -0.33347058, -0.31607842],\n",
       "        ...,\n",
       "        [-0.29072848, -0.3315403 , -0.2978558 , ..., -0.31714624,\n",
       "         -0.33347058,  2.1098306 ],\n",
       "        [-0.29072848, -0.3315403 , -0.47442967, ..., -0.31714624,\n",
       "         -0.33347058, -0.31607842],\n",
       "        [-0.29072848, -0.3315403 , -0.47442967, ..., -0.31714624,\n",
       "         -0.33347058, -0.31607842]], dtype=float32),\n",
       " array([[-0.20748407, -0.2685321 , -0.25669026, ..., -0.21526991,\n",
       "         -0.40584263, -0.37261307],\n",
       "        [-0.27228117,  1.15971673, -0.48232967, ..., -0.34254035,\n",
       "         -0.31028366, -0.45893404],\n",
       "        [-0.41818815, -0.26284239,  1.2942667 , ..., -0.60374743,\n",
       "         -0.28113917, -0.17949069],\n",
       "        ...,\n",
       "        [-0.45723256,  3.78234363, -0.46871236, ..., -0.90658909,\n",
       "         -0.31836531, -0.28898853],\n",
       "        [-0.49720114, -0.18701635, -0.28226253, ..., -0.20586516,\n",
       "         -0.26464128, -0.70579433],\n",
       "        [-0.29247862, -0.28196168, -0.21671177, ..., -0.41747174,\n",
       "         -0.38989735, -0.25232235]]))"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "matrix, shuffled_matrix"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "484c6013",
   "metadata": {},
   "source": [
    "Compute correlation values for 100 randomized matrices and append them to a dataframe with correlation values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "39abf2f2",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "random matrix 0\n",
      "random matrix 1\n",
      "random matrix 2\n",
      "random matrix 3\n",
      "random matrix 4\n",
      "random matrix 5\n",
      "random matrix 6\n",
      "random matrix 7\n",
      "random matrix 8\n",
      "random matrix 9\n",
      "random matrix 10\n",
      "random matrix 11\n",
      "random matrix 12\n",
      "random matrix 13\n",
      "random matrix 14\n",
      "random matrix 15\n",
      "random matrix 16\n",
      "random matrix 17\n",
      "random matrix 18\n",
      "random matrix 19\n",
      "random matrix 20\n",
      "random matrix 21\n",
      "random matrix 22\n",
      "random matrix 23\n",
      "random matrix 24\n",
      "random matrix 25\n",
      "random matrix 26\n",
      "random matrix 27\n",
      "random matrix 28\n",
      "random matrix 29\n",
      "random matrix 30\n",
      "random matrix 31\n",
      "random matrix 32\n",
      "random matrix 33\n",
      "random matrix 34\n",
      "random matrix 35\n",
      "random matrix 36\n",
      "random matrix 37\n",
      "random matrix 38\n",
      "random matrix 39\n",
      "random matrix 40\n",
      "random matrix 41\n",
      "random matrix 42\n",
      "random matrix 43\n",
      "random matrix 44\n",
      "random matrix 45\n",
      "random matrix 46\n",
      "random matrix 47\n",
      "random matrix 48\n",
      "random matrix 49\n",
      "random matrix 50\n",
      "random matrix 51\n",
      "random matrix 52\n",
      "random matrix 53\n",
      "random matrix 54\n",
      "random matrix 55\n",
      "random matrix 56\n",
      "random matrix 57\n",
      "random matrix 58\n",
      "random matrix 59\n",
      "random matrix 60\n",
      "random matrix 61\n",
      "random matrix 62\n",
      "random matrix 63\n",
      "random matrix 64\n",
      "random matrix 65\n",
      "random matrix 66\n",
      "random matrix 67\n",
      "random matrix 68\n",
      "random matrix 69\n",
      "random matrix 70\n",
      "random matrix 71\n",
      "random matrix 72\n",
      "random matrix 73\n",
      "random matrix 74\n",
      "random matrix 75\n",
      "random matrix 76\n",
      "random matrix 77\n",
      "random matrix 78\n",
      "random matrix 79\n",
      "random matrix 80\n",
      "random matrix 81\n",
      "random matrix 82\n",
      "random matrix 83\n",
      "random matrix 84\n",
      "random matrix 85\n",
      "random matrix 86\n",
      "random matrix 87\n",
      "random matrix 88\n",
      "random matrix 89\n",
      "random matrix 90\n",
      "random matrix 91\n",
      "random matrix 92\n",
      "random matrix 93\n",
      "random matrix 94\n",
      "random matrix 95\n",
      "random matrix 96\n",
      "random matrix 97\n",
      "random matrix 98\n",
      "random matrix 99\n"
     ]
    }
   ],
   "source": [
    "# for window 0\n",
    "matrixA = branchAwindows[0].X\n",
    "matrixB = branchBwindows[0].X\n",
    "matA_copy = branchAwindows[0].copy()\n",
    "matB_copy = branchBwindows[0].copy()\n",
    "dataframe = pd.read_csv(f'data_files/correlation_abs/w0_SCGRN_correlation.csv')\n",
    "out_dataframe = dataframe\n",
    "\n",
    "for i in range(100):\n",
    "    print(f'random matrix {i}')\n",
    "    shuffled_matA = shuffle2d(matrixA)\n",
    "    shuffled_matB = shuffle2d(matrixB)\n",
    "    matA_copy.X = shuffled_matA\n",
    "    matB_copy.X = shuffled_matB\n",
    "    corr_col = []\n",
    "    \n",
    "    for index, row in dataframe.iterrows():\n",
    "        tf = row['TF']\n",
    "        gene = row['GENE']\n",
    "        wA = np.corrcoef(matA_copy.obs_vector(tf), matA_copy.obs_vector(gene))[0, 1]\n",
    "        wB = np.corrcoef(matB_copy.obs_vector(tf), matB_copy.obs_vector(gene))[0, 1]\n",
    "        cc_coef = abs(wA-wB)\n",
    "        corr_col.append(cc_coef)\n",
    "    \n",
    "    out_dataframe[f'random {i}'] = corr_col"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "a5652bc9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TF</th>\n",
       "      <th>GENE</th>\n",
       "      <th>w_abs</th>\n",
       "      <th>cc_coef_abs</th>\n",
       "      <th>cc_bicor_abs</th>\n",
       "      <th>random 0</th>\n",
       "      <th>random 1</th>\n",
       "      <th>random 2</th>\n",
       "      <th>random 3</th>\n",
       "      <th>random 4</th>\n",
       "      <th>...</th>\n",
       "      <th>random 90</th>\n",
       "      <th>random 91</th>\n",
       "      <th>random 92</th>\n",
       "      <th>random 93</th>\n",
       "      <th>random 94</th>\n",
       "      <th>random 95</th>\n",
       "      <th>random 96</th>\n",
       "      <th>random 97</th>\n",
       "      <th>random 98</th>\n",
       "      <th>random 99</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>1810037I17Rik</td>\n",
       "      <td>0.759367</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.141573</td>\n",
       "      <td>0.187443</td>\n",
       "      <td>0.140750</td>\n",
       "      <td>0.090845</td>\n",
       "      <td>0.170943</td>\n",
       "      <td>...</td>\n",
       "      <td>0.044857</td>\n",
       "      <td>0.050462</td>\n",
       "      <td>0.101005</td>\n",
       "      <td>0.079727</td>\n",
       "      <td>0.104379</td>\n",
       "      <td>0.066838</td>\n",
       "      <td>0.149059</td>\n",
       "      <td>0.108076</td>\n",
       "      <td>0.198730</td>\n",
       "      <td>0.071676</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>2010111I01Rik</td>\n",
       "      <td>0.643845</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.247837</td>\n",
       "      <td>0.074895</td>\n",
       "      <td>0.056414</td>\n",
       "      <td>0.086139</td>\n",
       "      <td>0.034460</td>\n",
       "      <td>...</td>\n",
       "      <td>0.031962</td>\n",
       "      <td>0.171828</td>\n",
       "      <td>0.073368</td>\n",
       "      <td>0.135006</td>\n",
       "      <td>0.052079</td>\n",
       "      <td>0.080099</td>\n",
       "      <td>0.190555</td>\n",
       "      <td>0.050349</td>\n",
       "      <td>0.004757</td>\n",
       "      <td>0.016654</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>2210016F16Rik</td>\n",
       "      <td>2.120124</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.164798</td>\n",
       "      <td>0.075472</td>\n",
       "      <td>0.116873</td>\n",
       "      <td>0.015009</td>\n",
       "      <td>0.043209</td>\n",
       "      <td>...</td>\n",
       "      <td>0.080394</td>\n",
       "      <td>0.036877</td>\n",
       "      <td>0.014747</td>\n",
       "      <td>0.163786</td>\n",
       "      <td>0.028343</td>\n",
       "      <td>0.116505</td>\n",
       "      <td>0.032195</td>\n",
       "      <td>0.017193</td>\n",
       "      <td>0.042952</td>\n",
       "      <td>0.070206</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>2310011J03Rik</td>\n",
       "      <td>3.651726</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.082150</td>\n",
       "      <td>0.082227</td>\n",
       "      <td>0.160136</td>\n",
       "      <td>0.140024</td>\n",
       "      <td>0.037751</td>\n",
       "      <td>...</td>\n",
       "      <td>0.014172</td>\n",
       "      <td>0.014732</td>\n",
       "      <td>0.029537</td>\n",
       "      <td>0.030489</td>\n",
       "      <td>0.064550</td>\n",
       "      <td>0.017432</td>\n",
       "      <td>0.044642</td>\n",
       "      <td>0.217021</td>\n",
       "      <td>0.084769</td>\n",
       "      <td>0.016223</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>Actr2</td>\n",
       "      <td>3.651726</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.098336</td>\n",
       "      <td>0.134584</td>\n",
       "      <td>0.040926</td>\n",
       "      <td>0.008271</td>\n",
       "      <td>0.014974</td>\n",
       "      <td>...</td>\n",
       "      <td>0.068731</td>\n",
       "      <td>0.005026</td>\n",
       "      <td>0.080777</td>\n",
       "      <td>0.179003</td>\n",
       "      <td>0.101708</td>\n",
       "      <td>0.011542</td>\n",
       "      <td>0.007658</td>\n",
       "      <td>0.020221</td>\n",
       "      <td>0.053523</td>\n",
       "      <td>0.345400</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8505</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp326</td>\n",
       "      <td>3.701298</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.065161</td>\n",
       "      <td>0.053761</td>\n",
       "      <td>0.113183</td>\n",
       "      <td>0.037799</td>\n",
       "      <td>0.043411</td>\n",
       "      <td>...</td>\n",
       "      <td>0.159267</td>\n",
       "      <td>0.123309</td>\n",
       "      <td>0.194488</td>\n",
       "      <td>0.112962</td>\n",
       "      <td>0.035218</td>\n",
       "      <td>0.005911</td>\n",
       "      <td>0.022111</td>\n",
       "      <td>0.084952</td>\n",
       "      <td>0.030184</td>\n",
       "      <td>0.007544</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8506</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp407</td>\n",
       "      <td>4.984579</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.066704</td>\n",
       "      <td>0.118158</td>\n",
       "      <td>0.053080</td>\n",
       "      <td>0.025651</td>\n",
       "      <td>0.096090</td>\n",
       "      <td>...</td>\n",
       "      <td>0.123877</td>\n",
       "      <td>0.135158</td>\n",
       "      <td>0.010307</td>\n",
       "      <td>0.019190</td>\n",
       "      <td>0.034601</td>\n",
       "      <td>0.119896</td>\n",
       "      <td>0.050266</td>\n",
       "      <td>0.078878</td>\n",
       "      <td>0.141204</td>\n",
       "      <td>0.001037</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8507</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp512</td>\n",
       "      <td>0.733906</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.064585</td>\n",
       "      <td>0.001528</td>\n",
       "      <td>0.042059</td>\n",
       "      <td>0.188767</td>\n",
       "      <td>0.162882</td>\n",
       "      <td>...</td>\n",
       "      <td>0.115525</td>\n",
       "      <td>0.010437</td>\n",
       "      <td>0.072259</td>\n",
       "      <td>0.005476</td>\n",
       "      <td>0.020251</td>\n",
       "      <td>0.095369</td>\n",
       "      <td>0.098025</td>\n",
       "      <td>0.015364</td>\n",
       "      <td>0.150075</td>\n",
       "      <td>0.061293</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8508</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zmynd8</td>\n",
       "      <td>1.499678</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.019550</td>\n",
       "      <td>0.035188</td>\n",
       "      <td>0.002834</td>\n",
       "      <td>0.079891</td>\n",
       "      <td>0.129816</td>\n",
       "      <td>...</td>\n",
       "      <td>0.002579</td>\n",
       "      <td>0.115711</td>\n",
       "      <td>0.019467</td>\n",
       "      <td>0.068151</td>\n",
       "      <td>0.022730</td>\n",
       "      <td>0.083652</td>\n",
       "      <td>0.024051</td>\n",
       "      <td>0.018295</td>\n",
       "      <td>0.020815</td>\n",
       "      <td>0.103962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8509</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zranb2</td>\n",
       "      <td>2.455203</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.167527</td>\n",
       "      <td>0.108769</td>\n",
       "      <td>0.199451</td>\n",
       "      <td>0.110591</td>\n",
       "      <td>0.112195</td>\n",
       "      <td>...</td>\n",
       "      <td>0.077486</td>\n",
       "      <td>0.009475</td>\n",
       "      <td>0.027129</td>\n",
       "      <td>0.039880</td>\n",
       "      <td>0.003872</td>\n",
       "      <td>0.021762</td>\n",
       "      <td>0.064292</td>\n",
       "      <td>0.197679</td>\n",
       "      <td>0.025672</td>\n",
       "      <td>0.027155</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8510 rows × 105 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        TF           GENE     w_abs  cc_coef_abs  cc_bicor_abs  random 0  \\\n",
       "0     Ezh2  1810037I17Rik  0.759367          0.0           0.0  0.141573   \n",
       "1     Ezh2  2010111I01Rik  0.643845          0.0           0.0  0.247837   \n",
       "2     Ezh2  2210016F16Rik  2.120124          0.0           0.0  0.164798   \n",
       "3     Ezh2  2310011J03Rik  3.651726          0.0           0.0  0.082150   \n",
       "4     Ezh2          Actr2  3.651726          0.0           0.0  0.098336   \n",
       "...    ...            ...       ...          ...           ...       ...   \n",
       "8505  Rest         Zfp326  3.701298          0.0           0.0  0.065161   \n",
       "8506  Rest         Zfp407  4.984579          0.0           0.0  0.066704   \n",
       "8507  Rest         Zfp512  0.733906          0.0           0.0  0.064585   \n",
       "8508  Rest         Zmynd8  1.499678          0.0           0.0  0.019550   \n",
       "8509  Rest         Zranb2  2.455203          0.0           0.0  0.167527   \n",
       "\n",
       "      random 1  random 2  random 3  random 4  ...  random 90  random 91  \\\n",
       "0     0.187443  0.140750  0.090845  0.170943  ...   0.044857   0.050462   \n",
       "1     0.074895  0.056414  0.086139  0.034460  ...   0.031962   0.171828   \n",
       "2     0.075472  0.116873  0.015009  0.043209  ...   0.080394   0.036877   \n",
       "3     0.082227  0.160136  0.140024  0.037751  ...   0.014172   0.014732   \n",
       "4     0.134584  0.040926  0.008271  0.014974  ...   0.068731   0.005026   \n",
       "...        ...       ...       ...       ...  ...        ...        ...   \n",
       "8505  0.053761  0.113183  0.037799  0.043411  ...   0.159267   0.123309   \n",
       "8506  0.118158  0.053080  0.025651  0.096090  ...   0.123877   0.135158   \n",
       "8507  0.001528  0.042059  0.188767  0.162882  ...   0.115525   0.010437   \n",
       "8508  0.035188  0.002834  0.079891  0.129816  ...   0.002579   0.115711   \n",
       "8509  0.108769  0.199451  0.110591  0.112195  ...   0.077486   0.009475   \n",
       "\n",
       "      random 92  random 93  random 94  random 95  random 96  random 97  \\\n",
       "0      0.101005   0.079727   0.104379   0.066838   0.149059   0.108076   \n",
       "1      0.073368   0.135006   0.052079   0.080099   0.190555   0.050349   \n",
       "2      0.014747   0.163786   0.028343   0.116505   0.032195   0.017193   \n",
       "3      0.029537   0.030489   0.064550   0.017432   0.044642   0.217021   \n",
       "4      0.080777   0.179003   0.101708   0.011542   0.007658   0.020221   \n",
       "...         ...        ...        ...        ...        ...        ...   \n",
       "8505   0.194488   0.112962   0.035218   0.005911   0.022111   0.084952   \n",
       "8506   0.010307   0.019190   0.034601   0.119896   0.050266   0.078878   \n",
       "8507   0.072259   0.005476   0.020251   0.095369   0.098025   0.015364   \n",
       "8508   0.019467   0.068151   0.022730   0.083652   0.024051   0.018295   \n",
       "8509   0.027129   0.039880   0.003872   0.021762   0.064292   0.197679   \n",
       "\n",
       "      random 98  random 99  \n",
       "0      0.198730   0.071676  \n",
       "1      0.004757   0.016654  \n",
       "2      0.042952   0.070206  \n",
       "3      0.084769   0.016223  \n",
       "4      0.053523   0.345400  \n",
       "...         ...        ...  \n",
       "8505   0.030184   0.007544  \n",
       "8506   0.141204   0.001037  \n",
       "8507   0.150075   0.061293  \n",
       "8508   0.020815   0.103962  \n",
       "8509   0.025672   0.027155  \n",
       "\n",
       "[8510 rows x 105 columns]"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "out_dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "715c31c5",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "random matrix 0\n",
      "random matrix 1\n",
      "random matrix 2\n",
      "random matrix 3\n",
      "random matrix 4\n",
      "random matrix 5\n",
      "random matrix 6\n",
      "random matrix 7\n",
      "random matrix 8\n",
      "random matrix 9\n",
      "random matrix 10\n",
      "random matrix 11\n",
      "random matrix 12\n",
      "random matrix 13\n",
      "random matrix 14\n",
      "random matrix 15\n",
      "random matrix 16\n",
      "random matrix 17\n",
      "random matrix 18\n",
      "random matrix 19\n",
      "random matrix 20\n",
      "random matrix 21\n",
      "random matrix 22\n",
      "random matrix 23\n",
      "random matrix 24\n",
      "random matrix 25\n",
      "random matrix 26\n",
      "random matrix 27\n",
      "random matrix 28\n",
      "random matrix 29\n",
      "random matrix 30\n",
      "random matrix 31\n",
      "random matrix 32\n",
      "random matrix 33\n",
      "random matrix 34\n",
      "random matrix 35\n",
      "random matrix 36\n",
      "random matrix 37\n",
      "random matrix 38\n",
      "random matrix 39\n",
      "random matrix 40\n",
      "random matrix 41\n",
      "random matrix 42\n",
      "random matrix 43\n",
      "random matrix 44\n",
      "random matrix 45\n",
      "random matrix 46\n",
      "random matrix 47\n",
      "random matrix 48\n",
      "random matrix 49\n",
      "random matrix 50\n",
      "random matrix 51\n",
      "random matrix 52\n",
      "random matrix 53\n",
      "random matrix 54\n",
      "random matrix 55\n",
      "random matrix 56\n",
      "random matrix 57\n",
      "random matrix 58\n",
      "random matrix 59\n",
      "random matrix 60\n",
      "random matrix 61\n",
      "random matrix 62\n",
      "random matrix 63\n",
      "random matrix 64\n",
      "random matrix 65\n",
      "random matrix 66\n",
      "random matrix 67\n",
      "random matrix 68\n",
      "random matrix 69\n",
      "random matrix 70\n",
      "random matrix 71\n",
      "random matrix 72\n",
      "random matrix 73\n",
      "random matrix 74\n",
      "random matrix 75\n",
      "random matrix 76\n",
      "random matrix 77\n",
      "random matrix 78\n",
      "random matrix 79\n",
      "random matrix 80\n",
      "random matrix 81\n",
      "random matrix 82\n",
      "random matrix 83\n",
      "random matrix 84\n",
      "random matrix 85\n",
      "random matrix 86\n",
      "random matrix 87\n",
      "random matrix 88\n",
      "random matrix 89\n",
      "random matrix 90\n",
      "random matrix 91\n",
      "random matrix 92\n",
      "random matrix 93\n",
      "random matrix 94\n",
      "random matrix 95\n",
      "random matrix 96\n",
      "random matrix 97\n",
      "random matrix 98\n",
      "random matrix 99\n"
     ]
    }
   ],
   "source": [
    "# for window 1\n",
    "matrixA = branchAwindows[1].X\n",
    "matrixB = branchBwindows[1].X\n",
    "matA_copy = branchAwindows[1].copy()\n",
    "matB_copy = branchBwindows[1].copy()\n",
    "dataframe = pd.read_csv(f'data_files/correlation_abs/w1_SCGRN_correlation.csv')\n",
    "out_dataframe2 = dataframe\n",
    "\n",
    "for i in range(100):\n",
    "    print(f'random matrix {i}')\n",
    "    shuffled_matA = shuffle2d(matrixA)\n",
    "    shuffled_matB = shuffle2d(matrixB)\n",
    "    matA_copy.X = shuffled_matA\n",
    "    matB_copy.X = shuffled_matB\n",
    "    corr_col = []\n",
    "    \n",
    "    for index, row in dataframe.iterrows():\n",
    "        tf = row['TF']\n",
    "        gene = row['GENE']\n",
    "        wA = np.corrcoef(matA_copy.obs_vector(tf), matA_copy.obs_vector(gene))[0, 1]\n",
    "        wB = np.corrcoef(matB_copy.obs_vector(tf), matB_copy.obs_vector(gene))[0, 1]\n",
    "        cc_coef = abs(wA-wB)\n",
    "        corr_col.append(cc_coef)\n",
    "    \n",
    "    out_dataframe2[f'random {i}'] = corr_col"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "5424925b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TF</th>\n",
       "      <th>GENE</th>\n",
       "      <th>w_abs</th>\n",
       "      <th>cc_coef_abs</th>\n",
       "      <th>cc_bicor_abs</th>\n",
       "      <th>random 0</th>\n",
       "      <th>random 1</th>\n",
       "      <th>random 2</th>\n",
       "      <th>random 3</th>\n",
       "      <th>random 4</th>\n",
       "      <th>...</th>\n",
       "      <th>random 90</th>\n",
       "      <th>random 91</th>\n",
       "      <th>random 92</th>\n",
       "      <th>random 93</th>\n",
       "      <th>random 94</th>\n",
       "      <th>random 95</th>\n",
       "      <th>random 96</th>\n",
       "      <th>random 97</th>\n",
       "      <th>random 98</th>\n",
       "      <th>random 99</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>1810037I17Rik</td>\n",
       "      <td>1.800194</td>\n",
       "      <td>0.073319</td>\n",
       "      <td>0.064313</td>\n",
       "      <td>0.113287</td>\n",
       "      <td>0.063105</td>\n",
       "      <td>0.150345</td>\n",
       "      <td>0.346625</td>\n",
       "      <td>0.059212</td>\n",
       "      <td>...</td>\n",
       "      <td>0.062333</td>\n",
       "      <td>0.148554</td>\n",
       "      <td>0.040941</td>\n",
       "      <td>0.081292</td>\n",
       "      <td>0.238616</td>\n",
       "      <td>0.154349</td>\n",
       "      <td>0.036505</td>\n",
       "      <td>0.072601</td>\n",
       "      <td>0.169645</td>\n",
       "      <td>0.152825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>2210016F16Rik</td>\n",
       "      <td>3.325726</td>\n",
       "      <td>0.161181</td>\n",
       "      <td>0.164022</td>\n",
       "      <td>0.066083</td>\n",
       "      <td>0.094788</td>\n",
       "      <td>0.294349</td>\n",
       "      <td>0.149165</td>\n",
       "      <td>0.046935</td>\n",
       "      <td>...</td>\n",
       "      <td>0.169953</td>\n",
       "      <td>0.055911</td>\n",
       "      <td>0.100949</td>\n",
       "      <td>0.070887</td>\n",
       "      <td>0.060781</td>\n",
       "      <td>0.037781</td>\n",
       "      <td>0.172746</td>\n",
       "      <td>0.041630</td>\n",
       "      <td>0.112643</td>\n",
       "      <td>0.216104</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>3830406C13Rik</td>\n",
       "      <td>0.729930</td>\n",
       "      <td>0.029488</td>\n",
       "      <td>0.009544</td>\n",
       "      <td>0.074445</td>\n",
       "      <td>0.106442</td>\n",
       "      <td>0.012774</td>\n",
       "      <td>0.096644</td>\n",
       "      <td>0.292691</td>\n",
       "      <td>...</td>\n",
       "      <td>0.054524</td>\n",
       "      <td>0.001573</td>\n",
       "      <td>0.070519</td>\n",
       "      <td>0.136322</td>\n",
       "      <td>0.073395</td>\n",
       "      <td>0.126074</td>\n",
       "      <td>0.010329</td>\n",
       "      <td>0.183610</td>\n",
       "      <td>0.058618</td>\n",
       "      <td>0.021069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>4833420G17Rik</td>\n",
       "      <td>2.802151</td>\n",
       "      <td>0.114864</td>\n",
       "      <td>0.105545</td>\n",
       "      <td>0.052703</td>\n",
       "      <td>0.358186</td>\n",
       "      <td>0.170185</td>\n",
       "      <td>0.188190</td>\n",
       "      <td>0.083525</td>\n",
       "      <td>...</td>\n",
       "      <td>0.032972</td>\n",
       "      <td>0.055658</td>\n",
       "      <td>0.010152</td>\n",
       "      <td>0.067152</td>\n",
       "      <td>0.017488</td>\n",
       "      <td>0.058235</td>\n",
       "      <td>0.045138</td>\n",
       "      <td>0.159930</td>\n",
       "      <td>0.055771</td>\n",
       "      <td>0.041404</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Ezh2</td>\n",
       "      <td>4931406P16Rik</td>\n",
       "      <td>0.839976</td>\n",
       "      <td>0.016909</td>\n",
       "      <td>0.009813</td>\n",
       "      <td>0.126701</td>\n",
       "      <td>0.338071</td>\n",
       "      <td>0.012385</td>\n",
       "      <td>0.032387</td>\n",
       "      <td>0.101930</td>\n",
       "      <td>...</td>\n",
       "      <td>0.045997</td>\n",
       "      <td>0.121840</td>\n",
       "      <td>0.157710</td>\n",
       "      <td>0.021286</td>\n",
       "      <td>0.048446</td>\n",
       "      <td>0.115952</td>\n",
       "      <td>0.154234</td>\n",
       "      <td>0.134438</td>\n",
       "      <td>0.039177</td>\n",
       "      <td>0.123738</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11755</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zc3h7a</td>\n",
       "      <td>2.047629</td>\n",
       "      <td>0.107217</td>\n",
       "      <td>0.114232</td>\n",
       "      <td>0.418721</td>\n",
       "      <td>0.129951</td>\n",
       "      <td>0.004503</td>\n",
       "      <td>0.095438</td>\n",
       "      <td>0.011929</td>\n",
       "      <td>...</td>\n",
       "      <td>0.086167</td>\n",
       "      <td>0.063235</td>\n",
       "      <td>0.020508</td>\n",
       "      <td>0.224237</td>\n",
       "      <td>0.067980</td>\n",
       "      <td>0.104951</td>\n",
       "      <td>0.056078</td>\n",
       "      <td>0.123117</td>\n",
       "      <td>0.168167</td>\n",
       "      <td>0.029980</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11756</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zcchc6</td>\n",
       "      <td>1.112263</td>\n",
       "      <td>0.075051</td>\n",
       "      <td>0.052891</td>\n",
       "      <td>0.042276</td>\n",
       "      <td>0.115894</td>\n",
       "      <td>0.076807</td>\n",
       "      <td>0.142061</td>\n",
       "      <td>0.041933</td>\n",
       "      <td>...</td>\n",
       "      <td>0.323494</td>\n",
       "      <td>0.124991</td>\n",
       "      <td>0.190114</td>\n",
       "      <td>0.118167</td>\n",
       "      <td>0.036154</td>\n",
       "      <td>0.056776</td>\n",
       "      <td>0.004326</td>\n",
       "      <td>0.015165</td>\n",
       "      <td>0.031823</td>\n",
       "      <td>0.071005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11757</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zcchc7</td>\n",
       "      <td>0.947237</td>\n",
       "      <td>0.166832</td>\n",
       "      <td>0.134063</td>\n",
       "      <td>0.050270</td>\n",
       "      <td>0.023260</td>\n",
       "      <td>0.052951</td>\n",
       "      <td>0.422522</td>\n",
       "      <td>0.199744</td>\n",
       "      <td>...</td>\n",
       "      <td>0.088980</td>\n",
       "      <td>0.017932</td>\n",
       "      <td>0.074360</td>\n",
       "      <td>0.245744</td>\n",
       "      <td>0.015430</td>\n",
       "      <td>0.184772</td>\n",
       "      <td>0.031428</td>\n",
       "      <td>0.344297</td>\n",
       "      <td>0.117000</td>\n",
       "      <td>0.127827</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11758</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp110</td>\n",
       "      <td>1.704517</td>\n",
       "      <td>0.180745</td>\n",
       "      <td>0.152108</td>\n",
       "      <td>0.165838</td>\n",
       "      <td>0.158022</td>\n",
       "      <td>0.040166</td>\n",
       "      <td>0.035489</td>\n",
       "      <td>0.005364</td>\n",
       "      <td>...</td>\n",
       "      <td>0.029971</td>\n",
       "      <td>0.098692</td>\n",
       "      <td>0.002740</td>\n",
       "      <td>0.093915</td>\n",
       "      <td>0.033433</td>\n",
       "      <td>0.148712</td>\n",
       "      <td>0.169541</td>\n",
       "      <td>0.142538</td>\n",
       "      <td>0.074442</td>\n",
       "      <td>0.056037</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11759</th>\n",
       "      <td>Rest</td>\n",
       "      <td>Zfp518a</td>\n",
       "      <td>2.545763</td>\n",
       "      <td>0.163253</td>\n",
       "      <td>0.112958</td>\n",
       "      <td>0.068075</td>\n",
       "      <td>0.155856</td>\n",
       "      <td>0.149806</td>\n",
       "      <td>0.018999</td>\n",
       "      <td>0.170297</td>\n",
       "      <td>...</td>\n",
       "      <td>0.074240</td>\n",
       "      <td>0.056931</td>\n",
       "      <td>0.090652</td>\n",
       "      <td>0.088780</td>\n",
       "      <td>0.022553</td>\n",
       "      <td>0.338248</td>\n",
       "      <td>0.092220</td>\n",
       "      <td>0.127783</td>\n",
       "      <td>0.003189</td>\n",
       "      <td>0.204747</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>11760 rows × 105 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         TF           GENE     w_abs  cc_coef_abs  cc_bicor_abs  random 0  \\\n",
       "0      Ezh2  1810037I17Rik  1.800194     0.073319      0.064313  0.113287   \n",
       "1      Ezh2  2210016F16Rik  3.325726     0.161181      0.164022  0.066083   \n",
       "2      Ezh2  3830406C13Rik  0.729930     0.029488      0.009544  0.074445   \n",
       "3      Ezh2  4833420G17Rik  2.802151     0.114864      0.105545  0.052703   \n",
       "4      Ezh2  4931406P16Rik  0.839976     0.016909      0.009813  0.126701   \n",
       "...     ...            ...       ...          ...           ...       ...   \n",
       "11755  Rest         Zc3h7a  2.047629     0.107217      0.114232  0.418721   \n",
       "11756  Rest         Zcchc6  1.112263     0.075051      0.052891  0.042276   \n",
       "11757  Rest         Zcchc7  0.947237     0.166832      0.134063  0.050270   \n",
       "11758  Rest         Zfp110  1.704517     0.180745      0.152108  0.165838   \n",
       "11759  Rest        Zfp518a  2.545763     0.163253      0.112958  0.068075   \n",
       "\n",
       "       random 1  random 2  random 3  random 4  ...  random 90  random 91  \\\n",
       "0      0.063105  0.150345  0.346625  0.059212  ...   0.062333   0.148554   \n",
       "1      0.094788  0.294349  0.149165  0.046935  ...   0.169953   0.055911   \n",
       "2      0.106442  0.012774  0.096644  0.292691  ...   0.054524   0.001573   \n",
       "3      0.358186  0.170185  0.188190  0.083525  ...   0.032972   0.055658   \n",
       "4      0.338071  0.012385  0.032387  0.101930  ...   0.045997   0.121840   \n",
       "...         ...       ...       ...       ...  ...        ...        ...   \n",
       "11755  0.129951  0.004503  0.095438  0.011929  ...   0.086167   0.063235   \n",
       "11756  0.115894  0.076807  0.142061  0.041933  ...   0.323494   0.124991   \n",
       "11757  0.023260  0.052951  0.422522  0.199744  ...   0.088980   0.017932   \n",
       "11758  0.158022  0.040166  0.035489  0.005364  ...   0.029971   0.098692   \n",
       "11759  0.155856  0.149806  0.018999  0.170297  ...   0.074240   0.056931   \n",
       "\n",
       "       random 92  random 93  random 94  random 95  random 96  random 97  \\\n",
       "0       0.040941   0.081292   0.238616   0.154349   0.036505   0.072601   \n",
       "1       0.100949   0.070887   0.060781   0.037781   0.172746   0.041630   \n",
       "2       0.070519   0.136322   0.073395   0.126074   0.010329   0.183610   \n",
       "3       0.010152   0.067152   0.017488   0.058235   0.045138   0.159930   \n",
       "4       0.157710   0.021286   0.048446   0.115952   0.154234   0.134438   \n",
       "...          ...        ...        ...        ...        ...        ...   \n",
       "11755   0.020508   0.224237   0.067980   0.104951   0.056078   0.123117   \n",
       "11756   0.190114   0.118167   0.036154   0.056776   0.004326   0.015165   \n",
       "11757   0.074360   0.245744   0.015430   0.184772   0.031428   0.344297   \n",
       "11758   0.002740   0.093915   0.033433   0.148712   0.169541   0.142538   \n",
       "11759   0.090652   0.088780   0.022553   0.338248   0.092220   0.127783   \n",
       "\n",
       "       random 98  random 99  \n",
       "0       0.169645   0.152825  \n",
       "1       0.112643   0.216104  \n",
       "2       0.058618   0.021069  \n",
       "3       0.055771   0.041404  \n",
       "4       0.039177   0.123738  \n",
       "...          ...        ...  \n",
       "11755   0.168167   0.029980  \n",
       "11756   0.031823   0.071005  \n",
       "11757   0.117000   0.127827  \n",
       "11758   0.074442   0.056037  \n",
       "11759   0.003189   0.204747  \n",
       "\n",
       "[11760 rows x 105 columns]"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "out_dataframe2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "7a5b066e",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "random matrix 0\n",
      "random matrix 1\n",
      "random matrix 2\n",
      "random matrix 3\n",
      "random matrix 4\n",
      "random matrix 5\n",
      "random matrix 6\n",
      "random matrix 7\n",
      "random matrix 8\n",
      "random matrix 9\n",
      "random matrix 10\n",
      "random matrix 11\n",
      "random matrix 12\n",
      "random matrix 13\n",
      "random matrix 14\n",
      "random matrix 15\n",
      "random matrix 16\n",
      "random matrix 17\n",
      "random matrix 18\n",
      "random matrix 19\n",
      "random matrix 20\n",
      "random matrix 21\n",
      "random matrix 22\n",
      "random matrix 23\n",
      "random matrix 24\n",
      "random matrix 25\n",
      "random matrix 26\n",
      "random matrix 27\n",
      "random matrix 28\n",
      "random matrix 29\n",
      "random matrix 30\n",
      "random matrix 31\n",
      "random matrix 32\n",
      "random matrix 33\n",
      "random matrix 34\n",
      "random matrix 35\n",
      "random matrix 36\n",
      "random matrix 37\n",
      "random matrix 38\n",
      "random matrix 39\n",
      "random matrix 40\n",
      "random matrix 41\n",
      "random matrix 42\n",
      "random matrix 43\n",
      "random matrix 44\n",
      "random matrix 45\n",
      "random matrix 46\n",
      "random matrix 47\n",
      "random matrix 48\n",
      "random matrix 49\n",
      "random matrix 50\n",
      "random matrix 51\n",
      "random matrix 52\n",
      "random matrix 53\n",
      "random matrix 54\n",
      "random matrix 55\n",
      "random matrix 56\n",
      "random matrix 57\n",
      "random matrix 58\n",
      "random matrix 59\n",
      "random matrix 60\n",
      "random matrix 61\n",
      "random matrix 62\n",
      "random matrix 63\n",
      "random matrix 64\n",
      "random matrix 65\n",
      "random matrix 66\n",
      "random matrix 67\n",
      "random matrix 68\n",
      "random matrix 69\n",
      "random matrix 70\n",
      "random matrix 71\n",
      "random matrix 72\n",
      "random matrix 73\n",
      "random matrix 74\n",
      "random matrix 75\n",
      "random matrix 76\n",
      "random matrix 77\n",
      "random matrix 78\n",
      "random matrix 79\n",
      "random matrix 80\n",
      "random matrix 81\n",
      "random matrix 82\n",
      "random matrix 83\n",
      "random matrix 84\n",
      "random matrix 85\n",
      "random matrix 86\n",
      "random matrix 87\n",
      "random matrix 88\n",
      "random matrix 89\n",
      "random matrix 90\n",
      "random matrix 91\n",
      "random matrix 92\n",
      "random matrix 93\n",
      "random matrix 94\n",
      "random matrix 95\n",
      "random matrix 96\n",
      "random matrix 97\n",
      "random matrix 98\n",
      "random matrix 99\n",
      "window 0 done\n",
      "random matrix 0\n",
      "random matrix 1\n",
      "random matrix 2\n",
      "random matrix 3\n",
      "random matrix 4\n",
      "random matrix 5\n",
      "random matrix 6\n",
      "random matrix 7\n",
      "random matrix 8\n",
      "random matrix 9\n",
      "random matrix 10\n",
      "random matrix 11\n",
      "random matrix 12\n",
      "random matrix 13\n",
      "random matrix 14\n",
      "random matrix 15\n",
      "random matrix 16\n",
      "random matrix 17\n",
      "random matrix 18\n",
      "random matrix 19\n",
      "random matrix 20\n",
      "random matrix 21\n",
      "random matrix 22\n",
      "random matrix 23\n",
      "random matrix 24\n",
      "random matrix 25\n",
      "random matrix 26\n",
      "random matrix 27\n",
      "random matrix 28\n",
      "random matrix 29\n",
      "random matrix 30\n",
      "random matrix 31\n",
      "random matrix 32\n",
      "random matrix 33\n",
      "random matrix 34\n",
      "random matrix 35\n",
      "random matrix 36\n",
      "random matrix 37\n",
      "random matrix 38\n",
      "random matrix 39\n",
      "random matrix 40\n",
      "random matrix 41\n",
      "random matrix 42\n",
      "random matrix 43\n",
      "random matrix 44\n",
      "random matrix 45\n",
      "random matrix 46\n",
      "random matrix 47\n",
      "random matrix 48\n",
      "random matrix 49\n",
      "random matrix 50\n",
      "random matrix 51\n",
      "random matrix 52\n",
      "random matrix 53\n",
      "random matrix 54\n",
      "random matrix 55\n",
      "random matrix 56\n",
      "random matrix 57\n",
      "random matrix 58\n",
      "random matrix 59\n",
      "random matrix 60\n",
      "random matrix 61\n",
      "random matrix 62\n",
      "random matrix 63\n",
      "random matrix 64\n",
      "random matrix 65\n",
      "random matrix 66\n",
      "random matrix 67\n",
      "random matrix 68\n",
      "random matrix 69\n",
      "random matrix 70\n",
      "random matrix 71\n",
      "random matrix 72\n",
      "random matrix 73\n",
      "random matrix 74\n",
      "random matrix 75\n",
      "random matrix 76\n",
      "random matrix 77\n",
      "random matrix 78\n",
      "random matrix 79\n",
      "random matrix 80\n",
      "random matrix 81\n",
      "random matrix 82\n",
      "random matrix 83\n",
      "random matrix 84\n",
      "random matrix 85\n",
      "random matrix 86\n",
      "random matrix 87\n",
      "random matrix 88\n",
      "random matrix 89\n",
      "random matrix 90\n",
      "random matrix 91\n",
      "random matrix 92\n",
      "random matrix 93\n",
      "random matrix 94\n",
      "random matrix 95\n",
      "random matrix 96\n",
      "random matrix 97\n",
      "random matrix 98\n",
      "random matrix 99\n",
      "window 1 done\n",
      "random matrix 0\n",
      "random matrix 1\n",
      "random matrix 2\n",
      "random matrix 3\n",
      "random matrix 4\n",
      "random matrix 5\n",
      "random matrix 6\n",
      "random matrix 7\n",
      "random matrix 8\n",
      "random matrix 9\n",
      "random matrix 10\n",
      "random matrix 11\n",
      "random matrix 12\n",
      "random matrix 13\n",
      "random matrix 14\n",
      "random matrix 15\n",
      "random matrix 16\n",
      "random matrix 17\n",
      "random matrix 18\n",
      "random matrix 19\n",
      "random matrix 20\n",
      "random matrix 21\n",
      "random matrix 22\n",
      "random matrix 23\n",
      "random matrix 24\n",
      "random matrix 25\n",
      "random matrix 26\n",
      "random matrix 27\n",
      "random matrix 28\n",
      "random matrix 29\n",
      "random matrix 30\n",
      "random matrix 31\n",
      "random matrix 32\n",
      "random matrix 33\n",
      "random matrix 34\n",
      "random matrix 35\n",
      "random matrix 36\n",
      "random matrix 37\n",
      "random matrix 38\n",
      "random matrix 39\n",
      "random matrix 40\n",
      "random matrix 41\n",
      "random matrix 42\n",
      "random matrix 43\n",
      "random matrix 44\n",
      "random matrix 45\n",
      "random matrix 46\n",
      "random matrix 47\n",
      "random matrix 48\n",
      "random matrix 49\n",
      "random matrix 50\n",
      "random matrix 51\n",
      "random matrix 52\n",
      "random matrix 53\n",
      "random matrix 54\n",
      "random matrix 55\n",
      "random matrix 56\n",
      "random matrix 57\n",
      "random matrix 58\n",
      "random matrix 59\n",
      "random matrix 60\n",
      "random matrix 61\n",
      "random matrix 62\n",
      "random matrix 63\n",
      "random matrix 64\n",
      "random matrix 65\n",
      "random matrix 66\n",
      "random matrix 67\n",
      "random matrix 68\n",
      "random matrix 69\n",
      "random matrix 70\n",
      "random matrix 71\n",
      "random matrix 72\n",
      "random matrix 73\n",
      "random matrix 74\n",
      "random matrix 75\n",
      "random matrix 76\n",
      "random matrix 77\n",
      "random matrix 78\n",
      "random matrix 79\n",
      "random matrix 80\n",
      "random matrix 81\n",
      "random matrix 82\n",
      "random matrix 83\n",
      "random matrix 84\n",
      "random matrix 85\n",
      "random matrix 86\n",
      "random matrix 87\n",
      "random matrix 88\n",
      "random matrix 89\n",
      "random matrix 90\n",
      "random matrix 91\n",
      "random matrix 92\n",
      "random matrix 93\n",
      "random matrix 94\n",
      "random matrix 95\n",
      "random matrix 96\n",
      "random matrix 97\n",
      "random matrix 98\n",
      "random matrix 99\n",
      "window 2 done\n",
      "random matrix 0\n",
      "random matrix 1\n",
      "random matrix 2\n",
      "random matrix 3\n",
      "random matrix 4\n",
      "random matrix 5\n",
      "random matrix 6\n",
      "random matrix 7\n",
      "random matrix 8\n",
      "random matrix 9\n",
      "random matrix 10\n",
      "random matrix 11\n",
      "random matrix 12\n",
      "random matrix 13\n",
      "random matrix 14\n",
      "random matrix 15\n",
      "random matrix 16\n",
      "random matrix 17\n",
      "random matrix 18\n",
      "random matrix 19\n",
      "random matrix 20\n",
      "random matrix 21\n",
      "random matrix 22\n",
      "random matrix 23\n",
      "random matrix 24\n",
      "random matrix 25\n",
      "random matrix 26\n",
      "random matrix 27\n",
      "random matrix 28\n",
      "random matrix 29\n",
      "random matrix 30\n",
      "random matrix 31\n",
      "random matrix 32\n",
      "random matrix 33\n",
      "random matrix 34\n",
      "random matrix 35\n",
      "random matrix 36\n",
      "random matrix 37\n",
      "random matrix 38\n",
      "random matrix 39\n",
      "random matrix 40\n",
      "random matrix 41\n",
      "random matrix 42\n",
      "random matrix 43\n",
      "random matrix 44\n",
      "random matrix 45\n",
      "random matrix 46\n",
      "random matrix 47\n",
      "random matrix 48\n",
      "random matrix 49\n",
      "random matrix 50\n",
      "random matrix 51\n",
      "random matrix 52\n",
      "random matrix 53\n",
      "random matrix 54\n",
      "random matrix 55\n",
      "random matrix 56\n",
      "random matrix 57\n",
      "random matrix 58\n",
      "random matrix 59\n",
      "random matrix 60\n",
      "random matrix 61\n",
      "random matrix 62\n",
      "random matrix 63\n",
      "random matrix 64\n",
      "random matrix 65\n",
      "random matrix 66\n",
      "random matrix 67\n",
      "random matrix 68\n",
      "random matrix 69\n",
      "random matrix 70\n",
      "random matrix 71\n",
      "random matrix 72\n",
      "random matrix 73\n",
      "random matrix 74\n",
      "random matrix 75\n",
      "random matrix 76\n",
      "random matrix 77\n",
      "random matrix 78\n",
      "random matrix 79\n",
      "random matrix 80\n",
      "random matrix 81\n",
      "random matrix 82\n",
      "random matrix 83\n",
      "random matrix 84\n",
      "random matrix 85\n",
      "random matrix 86\n",
      "random matrix 87\n",
      "random matrix 88\n",
      "random matrix 89\n",
      "random matrix 90\n",
      "random matrix 91\n",
      "random matrix 92\n",
      "random matrix 93\n",
      "random matrix 94\n",
      "random matrix 95\n",
      "random matrix 96\n",
      "random matrix 97\n",
      "random matrix 98\n",
      "random matrix 99\n",
      "window 3 done\n",
      "random matrix 0\n",
      "random matrix 1\n",
      "random matrix 2\n",
      "random matrix 3\n",
      "random matrix 4\n",
      "random matrix 5\n",
      "random matrix 6\n",
      "random matrix 7\n",
      "random matrix 8\n",
      "random matrix 9\n",
      "random matrix 10\n",
      "random matrix 11\n",
      "random matrix 12\n",
      "random matrix 13\n",
      "random matrix 14\n",
      "random matrix 15\n",
      "random matrix 16\n",
      "random matrix 17\n",
      "random matrix 18\n",
      "random matrix 19\n",
      "random matrix 20\n",
      "random matrix 21\n",
      "random matrix 22\n",
      "random matrix 23\n",
      "random matrix 24\n",
      "random matrix 25\n",
      "random matrix 26\n",
      "random matrix 27\n",
      "random matrix 28\n",
      "random matrix 29\n",
      "random matrix 30\n",
      "random matrix 31\n",
      "random matrix 32\n",
      "random matrix 33\n",
      "random matrix 34\n",
      "random matrix 35\n",
      "random matrix 36\n",
      "random matrix 37\n",
      "random matrix 38\n",
      "random matrix 39\n",
      "random matrix 40\n",
      "random matrix 41\n",
      "random matrix 42\n",
      "random matrix 43\n",
      "random matrix 44\n",
      "random matrix 45\n",
      "random matrix 46\n",
      "random matrix 47\n",
      "random matrix 48\n",
      "random matrix 49\n",
      "random matrix 50\n",
      "random matrix 51\n",
      "random matrix 52\n",
      "random matrix 53\n",
      "random matrix 54\n",
      "random matrix 55\n",
      "random matrix 56\n",
      "random matrix 57\n",
      "random matrix 58\n",
      "random matrix 59\n",
      "random matrix 60\n",
      "random matrix 61\n",
      "random matrix 62\n",
      "random matrix 63\n",
      "random matrix 64\n",
      "random matrix 65\n",
      "random matrix 66\n",
      "random matrix 67\n",
      "random matrix 68\n",
      "random matrix 69\n",
      "random matrix 70\n",
      "random matrix 71\n",
      "random matrix 72\n",
      "random matrix 73\n",
      "random matrix 74\n",
      "random matrix 75\n",
      "random matrix 76\n",
      "random matrix 77\n",
      "random matrix 78\n",
      "random matrix 79\n",
      "random matrix 80\n",
      "random matrix 81\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "random matrix 82\n",
      "random matrix 83\n",
      "random matrix 84\n",
      "random matrix 85\n",
      "random matrix 86\n",
      "random matrix 87\n",
      "random matrix 88\n",
      "random matrix 89\n",
      "random matrix 90\n",
      "random matrix 91\n",
      "random matrix 92\n",
      "random matrix 93\n",
      "random matrix 94\n",
      "random matrix 95\n",
      "random matrix 96\n",
      "random matrix 97\n",
      "random matrix 98\n",
      "random matrix 99\n",
      "window 4 done\n"
     ]
    }
   ],
   "source": [
    "# for all windows\n",
    "\n",
    "out_dfs = []\n",
    "for j in range(len(branchAwindows)):\n",
    "    matrixA = branchAwindows[j].X\n",
    "    matrixB = branchBwindows[j].X\n",
    "    matA_copy = branchAwindows[j].copy()\n",
    "    matB_copy = branchBwindows[j].copy()\n",
    "    dataframe = pd.read_csv(f'data_files/correlation_abs/w{j}_SCGRN_correlation.csv')\n",
    "    out_dataframe = dataframe\n",
    "\n",
    "    for i in range(100):\n",
    "        print(f'random matrix {i}')\n",
    "        shuffled_matA = shuffle2d(matrixA)\n",
    "        shuffled_matB = shuffle2d(matrixB)\n",
    "        matA_copy.X = shuffled_matA\n",
    "        matB_copy.X = shuffled_matB\n",
    "        corr_col = []\n",
    "\n",
    "        for index, row in dataframe.iterrows():\n",
    "            tf = row['TF']\n",
    "            gene = row['GENE']\n",
    "            wA = np.corrcoef(matA_copy.obs_vector(tf), matA_copy.obs_vector(gene))[0, 1]\n",
    "            wB = np.corrcoef(matB_copy.obs_vector(tf), matB_copy.obs_vector(gene))[0, 1]\n",
    "            cc_coef = abs(wA-wB)\n",
    "            corr_col.append(cc_coef)\n",
    "\n",
    "        out_dataframe[f'random {i}'] = corr_col\n",
    "    \n",
    "    out_dfs.append(out_dataframe)\n",
    "    print(f'window {j} done')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "dc31e372",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(        TF           GENE     w_abs  cc_coef_abs  cc_bicor_abs  random 0  \\\n",
       " 0     Ezh2  1810037I17Rik  0.759367          0.0           0.0  0.117190   \n",
       " 1     Ezh2  2010111I01Rik  0.643845          0.0           0.0  0.040679   \n",
       " 2     Ezh2  2210016F16Rik  2.120124          0.0           0.0  0.030796   \n",
       " 3     Ezh2  2310011J03Rik  3.651726          0.0           0.0  0.022006   \n",
       " 4     Ezh2          Actr2  3.651726          0.0           0.0  0.029727   \n",
       " ...    ...            ...       ...          ...           ...       ...   \n",
       " 8505  Rest         Zfp326  3.701298          0.0           0.0  0.058648   \n",
       " 8506  Rest         Zfp407  4.984579          0.0           0.0  0.069716   \n",
       " 8507  Rest         Zfp512  0.733906          0.0           0.0  0.094928   \n",
       " 8508  Rest         Zmynd8  1.499678          0.0           0.0  0.047972   \n",
       " 8509  Rest         Zranb2  2.455203          0.0           0.0  0.191484   \n",
       " \n",
       "       random 1  random 2  random 3  random 4  ...  random 90  random 91  \\\n",
       " 0     0.008381  0.137376  0.295552  0.027401  ...   0.122538   0.073748   \n",
       " 1     0.122186  0.079715  0.018592  0.000315  ...   0.035985   0.007977   \n",
       " 2     0.049158  0.002331  0.074769  0.074345  ...   0.043631   0.249424   \n",
       " 3     0.055461  0.090605  0.043595  0.104574  ...   0.076915   0.045999   \n",
       " 4     0.147113  0.013007  0.162027  0.083111  ...   0.083424   0.207534   \n",
       " ...        ...       ...       ...       ...  ...        ...        ...   \n",
       " 8505  0.157934  0.047592  0.237340  0.160418  ...   0.069494   0.234134   \n",
       " 8506  0.002252  0.013635  0.157513  0.009579  ...   0.175664   0.134645   \n",
       " 8507  0.006057  0.149677  0.161842  0.040137  ...   0.103963   0.102691   \n",
       " 8508  0.089609  0.082630  0.051533  0.167510  ...   0.102008   0.050717   \n",
       " 8509  0.111286  0.016009  0.061877  0.073682  ...   0.021027   0.085877   \n",
       " \n",
       "       random 92  random 93  random 94  random 95  random 96  random 97  \\\n",
       " 0      0.012190   0.071153   0.024947   0.062418   0.048792   0.138379   \n",
       " 1      0.039463   0.053757   0.059524   0.029516   0.114197   0.245190   \n",
       " 2      0.014841   0.025687   0.135378   0.022549   0.065014   0.203037   \n",
       " 3      0.028987   0.092291   0.050110   0.099907   0.019067   0.061746   \n",
       " 4      0.182115   0.169060   0.071160   0.068533   0.027143   0.157203   \n",
       " ...         ...        ...        ...        ...        ...        ...   \n",
       " 8505   0.040576   0.121518   0.214849   0.083143   0.079454   0.096100   \n",
       " 8506   0.079065   0.243314   0.019916   0.030224   0.156092   0.016312   \n",
       " 8507   0.017823   0.018946   0.050601   0.048373   0.028980   0.084750   \n",
       " 8508   0.200402   0.139791   0.059757   0.014037   0.000156   0.143845   \n",
       " 8509   0.132109   0.020098   0.003823   0.022736   0.065386   0.088029   \n",
       " \n",
       "       random 98  random 99  \n",
       " 0      0.073791   0.082000  \n",
       " 1      0.162109   0.028524  \n",
       " 2      0.012494   0.059925  \n",
       " 3      0.108812   0.029707  \n",
       " 4      0.008809   0.316191  \n",
       " ...         ...        ...  \n",
       " 8505   0.131553   0.004032  \n",
       " 8506   0.000802   0.055606  \n",
       " 8507   0.075521   0.154883  \n",
       " 8508   0.204062   0.103100  \n",
       " 8509   0.000327   0.126868  \n",
       " \n",
       " [8510 rows x 105 columns],\n",
       "          TF           GENE     w_abs  cc_coef_abs  cc_bicor_abs  random 0  \\\n",
       " 0      Ezh2  1810037I17Rik  1.800194     0.073319      0.064313  0.117234   \n",
       " 1      Ezh2  2210016F16Rik  3.325726     0.161181      0.164022  0.068830   \n",
       " 2      Ezh2  3830406C13Rik  0.729930     0.029488      0.009544  0.047546   \n",
       " 3      Ezh2  4833420G17Rik  2.802151     0.114864      0.105545  0.105914   \n",
       " 4      Ezh2  4931406P16Rik  0.839976     0.016909      0.009813  0.163563   \n",
       " ...     ...            ...       ...          ...           ...       ...   \n",
       " 11755  Rest         Zc3h7a  2.047629     0.107217      0.114232  0.049452   \n",
       " 11756  Rest         Zcchc6  1.112263     0.075051      0.052891  0.236270   \n",
       " 11757  Rest         Zcchc7  0.947237     0.166832      0.134063  0.083346   \n",
       " 11758  Rest         Zfp110  1.704517     0.180745      0.152108  0.004053   \n",
       " 11759  Rest        Zfp518a  2.545763     0.163253      0.112958  0.101765   \n",
       " \n",
       "        random 1  random 2  random 3  random 4  ...  random 90  random 91  \\\n",
       " 0      0.236452  0.119975  0.135797  0.168429  ...   0.145857   0.206808   \n",
       " 1      0.138806  0.068374  0.267927  0.132772  ...   0.113956   0.092969   \n",
       " 2      0.121034  0.077867  0.413963  0.256024  ...   0.272415   0.210711   \n",
       " 3      0.004423  0.166924  0.166732  0.188280  ...   0.075276   0.067068   \n",
       " 4      0.090155  0.249392  0.120613  0.126571  ...   0.011637   0.230733   \n",
       " ...         ...       ...       ...       ...  ...        ...        ...   \n",
       " 11755  0.067045  0.170373  0.116647  0.045902  ...   0.140505   0.022819   \n",
       " 11756  0.185725  0.126458  0.131193  0.207078  ...   0.116759   0.314962   \n",
       " 11757  0.111989  0.035020  0.028500  0.030770  ...   0.261372   0.027192   \n",
       " 11758  0.113604  0.011535  0.043448  0.134354  ...   0.148098   0.327943   \n",
       " 11759  0.060153  0.028037  0.005410  0.002772  ...   0.070027   0.140118   \n",
       " \n",
       "        random 92  random 93  random 94  random 95  random 96  random 97  \\\n",
       " 0       0.061214   0.177129   0.055458   0.050373   0.000298   0.155045   \n",
       " 1       0.127534   0.058744   0.057160   0.016782   0.094312   0.089481   \n",
       " 2       0.010406   0.139999   0.086655   0.054836   0.049291   0.214383   \n",
       " 3       0.176415   0.089019   0.338223   0.198425   0.156882   0.052082   \n",
       " 4       0.164735   0.009645   0.246852   0.022893   0.147279   0.088065   \n",
       " ...          ...        ...        ...        ...        ...        ...   \n",
       " 11755   0.000905   0.153060   0.220337   0.080527   0.236605   0.339985   \n",
       " 11756   0.122658   0.035047   0.088496   0.178868   0.203837   0.119189   \n",
       " 11757   0.080852   0.133415   0.006967   0.316641   0.277054   0.088903   \n",
       " 11758   0.048912   0.240867   0.101451   0.128246   0.084605   0.060286   \n",
       " 11759   0.031759   0.008811   0.078586   0.277254   0.166329   0.101621   \n",
       " \n",
       "        random 98  random 99  \n",
       " 0       0.062923   0.145870  \n",
       " 1       0.198125   0.121974  \n",
       " 2       0.081054   0.046188  \n",
       " 3       0.024814   0.007664  \n",
       " 4       0.087526   0.196377  \n",
       " ...          ...        ...  \n",
       " 11755   0.055238   0.007157  \n",
       " 11756   0.006320   0.119132  \n",
       " 11757   0.314621   0.069100  \n",
       " 11758   0.142394   0.049191  \n",
       " 11759   0.062484   0.136490  \n",
       " \n",
       " [11760 rows x 105 columns])"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "out_dfs[0], out_dfs[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "06b36a71",
   "metadata": {},
   "outputs": [],
   "source": [
    "PATH_TO_W_CC_DFS = 'data_files/permutation_correlation/'\n",
    "for i in range(0,len(out_dfs)):\n",
    "    df = out_dfs[i]\n",
    "    df.to_csv(f'{PATH_TO_W_CC_DFS}w{i}_SCGRN_correlation_perm.csv', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d36f6ae4",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# for all windows - 10000 permutations\n",
    "\n",
    "out_dfs_10000 = []\n",
    "for j in range(len(branchAwindows)):\n",
    "    matrixA = branchAwindows[j].X\n",
    "    matrixB = branchBwindows[j].X\n",
    "    matA_copy = branchAwindows[j].copy()\n",
    "    matB_copy = branchBwindows[j].copy()\n",
    "    dataframe = pd.read_csv(f'data_files/correlation_abs/w{j}_SCGRN_correlation.csv')\n",
    "    out_dataframe = dataframe\n",
    "\n",
    "    for i in range(10000):\n",
    "        print(f'random matrix {i}')\n",
    "        shuffled_matA = shuffle2d(matrixA)\n",
    "        shuffled_matB = shuffle2d(matrixB)\n",
    "        matA_copy.X = shuffled_matA\n",
    "        matB_copy.X = shuffled_matB\n",
    "        corr_col = []\n",
    "\n",
    "        for index, row in dataframe.iterrows():\n",
    "            tf = row['TF']\n",
    "            gene = row['GENE']\n",
    "            wA = np.corrcoef(matA_copy.obs_vector(tf), matA_copy.obs_vector(gene))[0, 1]\n",
    "            wB = np.corrcoef(matB_copy.obs_vector(tf), matB_copy.obs_vector(gene))[0, 1]\n",
    "            cc_coef = abs(wA-wB)\n",
    "            corr_col.append(cc_coef)\n",
    "\n",
    "        out_dataframe[f'random {i}'] = corr_col\n",
    "    \n",
    "    out_dfs_10000.append(out_dataframe)\n",
    "    out_dataframe.to_csv(f'{PATH_TO_W_CC_DFS}w{i}_SCGRN_correlation_perm_10000.csv', index=False)\n",
    "    print(f'window {j} done')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0fc486ed",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "matrixA = branchAwindows[0].X\n",
    "matrixB = branchBwindows[0].X\n",
    "matA_copy = branchAwindows[0].copy()\n",
    "matB_copy = branchBwindows[0].copy()\n",
    "dataframe = pd.read_csv(f'data_files/correlation_abs/w0_SCGRN_correlation.csv')\n",
    "out_dataframe = dataframe\n",
    "\n",
    "random_dict = {}\n",
    "\n",
    "for i in range(10000):\n",
    "    print(f'random matrix {i}')\n",
    "    shuffled_matA = shuffle2d(matrixA)\n",
    "    shuffled_matB = shuffle2d(matrixB)\n",
    "    matA_copy.X = shuffled_matA\n",
    "    matB_copy.X = shuffled_matB\n",
    "    corr_col = []\n",
    "\n",
    "    for index, row in dataframe.iterrows():\n",
    "        tf = row['TF']\n",
    "        gene = row['GENE']\n",
    "        wA = np.corrcoef(matA_copy.obs_vector(tf), matA_copy.obs_vector(gene))[0, 1]\n",
    "        wB = np.corrcoef(matB_copy.obs_vector(tf), matB_copy.obs_vector(gene))[0, 1]\n",
    "        cc_coef = abs(wA-wB)\n",
    "        corr_col.append(cc_coef)\n",
    "\n",
    "    # out_dataframe[f'random {i}'] = corr_col\n",
    "    # random_dict[i]=corr_col\n",
    "\n",
    "# out_dfs_10000.append(out_dataframe)\n",
    "# out_dataframe.to_csv(f'{PATH_TO_W_CC_DFS}w{0}_SCGRN_correlation_perm_10000.csv', index=False)\n",
    "# print(f'window {0} done')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "014b30f3",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_p_value(matA, matB, tf, gene, permutations, true_val):\n",
    "    values = []\n",
    "    for i in range(permutations):\n",
    "        shuffled_matA = shuffle2d(matA.X)\n",
    "        shuffled_matB = shuffle2d(matB.X)\n",
    "        matA.X = shuffled_matA\n",
    "        matB.X = shuffled_matB\n",
    "        wA = np.corrcoef(matA.obs_vector(tf), matA.obs_vector(gene))[0, 1]\n",
    "        wB = np.corrcoef(matB.obs_vector(tf), matB.obs_vector(gene))[0, 1]\n",
    "        cc_coef = abs(wA-wB)\n",
    "        values.append(cc_coef)\n",
    "    \n",
    "    counter = 0\n",
    "    for value in values:\n",
    "        if value > true_val:\n",
    "            counter += 1\n",
    "    p = counter/permutations\n",
    "    print(p)\n",
    "    return p"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "6b1c165a",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "startet window 3\n",
      "0.1\n",
      "0.0\n",
      "0.04\n",
      "0.11\n",
      "0.83\n",
      "0.79\n",
      "0.0\n",
      "0.7\n",
      "0.3\n",
      "0.04\n",
      "0.23\n",
      "0.47\n",
      "0.71\n",
      "0.05\n",
      "0.18\n",
      "0.78\n",
      "0.06\n",
      "0.27\n",
      "0.89\n",
      "0.38\n",
      "0.08\n",
      "0.9\n",
      "0.56\n",
      "0.07\n",
      "0.01\n",
      "0.03\n",
      "0.06\n",
      "0.85\n",
      "0.09\n",
      "0.02\n",
      "0.04\n",
      "0.25\n",
      "0.26\n",
      "0.49\n",
      "0.1\n",
      "0.17\n",
      "0.39\n",
      "0.36\n",
      "0.03\n",
      "0.03\n",
      "0.27\n",
      "0.07\n",
      "0.2\n",
      "0.11\n",
      "0.74\n",
      "0.93\n",
      "0.12\n",
      "0.49\n",
      "0.53\n",
      "0.55\n",
      "0.13\n",
      "0.0\n",
      "0.22\n",
      "0.35\n",
      "0.17\n",
      "0.27\n",
      "0.74\n",
      "0.1\n",
      "0.14\n",
      "0.22\n",
      "0.05\n",
      "0.15\n",
      "0.4\n",
      "0.53\n",
      "0.85\n",
      "0.0\n",
      "0.25\n",
      "0.19\n",
      "0.63\n",
      "0.41\n",
      "0.2\n",
      "0.38\n",
      "0.33\n",
      "0.51\n",
      "0.18\n",
      "0.12\n",
      "0.17\n",
      "0.88\n",
      "0.4\n",
      "0.73\n",
      "0.81\n",
      "0.0\n",
      "0.99\n",
      "0.5\n",
      "0.07\n",
      "0.08\n",
      "0.18\n",
      "0.39\n",
      "0.19\n",
      "0.09\n",
      "0.32\n",
      "0.1\n",
      "0.67\n",
      "0.13\n",
      "0.03\n",
      "0.16\n",
      "0.14\n",
      "0.52\n",
      "0.29\n",
      "0.41\n",
      "0.34\n",
      "0.0\n",
      "0.13\n",
      "0.1\n",
      "0.27\n",
      "0.84\n",
      "0.16\n",
      "0.83\n",
      "0.67\n",
      "0.11\n",
      "0.45\n",
      "0.73\n",
      "0.1\n",
      "0.11\n",
      "0.03\n",
      "0.69\n",
      "0.54\n",
      "0.33\n",
      "0.35\n",
      "0.0\n",
      "0.21\n",
      "0.18\n",
      "0.49\n",
      "0.8\n",
      "0.14\n",
      "0.51\n",
      "0.21\n",
      "0.96\n",
      "0.89\n",
      "0.0\n",
      "0.09\n",
      "0.07\n",
      "0.26\n",
      "0.26\n",
      "0.18\n",
      "0.31\n",
      "0.24\n",
      "0.13\n",
      "0.66\n",
      "0.75\n",
      "0.04\n",
      "0.34\n",
      "0.23\n",
      "0.06\n",
      "0.01\n",
      "0.27\n",
      "0.16\n",
      "0.03\n",
      "0.37\n",
      "0.05\n",
      "0.15\n",
      "0.01\n",
      "0.03\n",
      "0.03\n",
      "0.08\n",
      "0.43\n",
      "0.04\n",
      "0.16\n",
      "0.18\n",
      "0.23\n",
      "0.28\n",
      "0.79\n",
      "0.15\n",
      "0.65\n",
      "0.37\n",
      "0.72\n",
      "0.07\n",
      "0.98\n",
      "0.09\n",
      "0.01\n",
      "0.04\n",
      "0.34\n",
      "0.47\n",
      "0.5\n",
      "0.5\n",
      "0.38\n",
      "0.05\n",
      "0.71\n",
      "0.23\n",
      "0.04\n",
      "0.06\n",
      "0.04\n",
      "0.12\n",
      "0.06\n",
      "0.03\n",
      "0.19\n",
      "0.48\n",
      "0.11\n",
      "0.33\n",
      "0.77\n",
      "0.26\n",
      "0.26\n",
      "0.23\n",
      "0.21\n",
      "0.53\n",
      "0.4\n",
      "0.27\n",
      "0.38\n",
      "0.0\n",
      "0.18\n",
      "0.11\n",
      "0.34\n",
      "0.21\n",
      "0.45\n",
      "0.05\n",
      "0.25\n",
      "0.96\n",
      "0.02\n",
      "0.13\n",
      "0.47\n",
      "0.55\n",
      "0.52\n",
      "0.16\n",
      "0.07\n",
      "0.11\n",
      "0.67\n",
      "0.2\n",
      "0.63\n",
      "0.17\n",
      "0.16\n",
      "0.27\n",
      "0.29\n",
      "0.06\n",
      "0.04\n",
      "0.13\n",
      "0.25\n",
      "0.63\n",
      "0.07\n",
      "0.21\n",
      "0.15\n",
      "0.49\n",
      "0.17\n",
      "0.01\n",
      "0.16\n",
      "0.35\n",
      "0.29\n",
      "0.02\n",
      "0.04\n",
      "0.02\n",
      "0.05\n",
      "0.26\n",
      "0.12\n",
      "0.55\n",
      "0.6\n",
      "0.73\n",
      "0.45\n",
      "0.81\n",
      "0.04\n",
      "0.24\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-64-fa0b07dddf2b>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      9\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     10\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mindex\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrow\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdataframe\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0miterrows\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 11\u001b[0;31m     \u001b[0mp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mget_p_value\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmatrixA\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmatrixB\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrow\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'TF'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrow\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'GENE'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrow\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'cc_coef_abs'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     12\u001b[0m     \u001b[0mp_vals\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     13\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m<ipython-input-62-db50ea994e29>\u001b[0m in \u001b[0;36mget_p_value\u001b[0;34m(matA, matB, tf, gene, permutations, true_val)\u001b[0m\n\u001b[1;32m      2\u001b[0m     \u001b[0mvalues\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m     \u001b[0;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpermutations\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m         \u001b[0mshuffled_matA\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mshuffle2d\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmatA\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      5\u001b[0m         \u001b[0mshuffled_matB\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mshuffle2d\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmatB\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      6\u001b[0m         \u001b[0mmatA\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mX\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mshuffled_matA\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m<ipython-input-52-67f2e11fe3ba>\u001b[0m in \u001b[0;36mshuffle2d\u001b[0;34m(arr2d)\u001b[0m\n\u001b[1;32m     10\u001b[0m         \u001b[0mreshape\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mistart\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0miend\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     11\u001b[0m     \u001b[0mrd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshuffle\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 12\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mistart\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0miend\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mistart\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0miend\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mreshape\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m<ipython-input-52-67f2e11fe3ba>\u001b[0m in \u001b[0;36m<listcomp>\u001b[0;34m(.0)\u001b[0m\n\u001b[1;32m     10\u001b[0m         \u001b[0mreshape\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mistart\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0miend\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     11\u001b[0m     \u001b[0mrd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshuffle\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 12\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mistart\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0miend\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mistart\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0miend\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mreshape\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "print(f'startet window {3}')\n",
    "matrixA = branchAwindows[3]\n",
    "matrixB = branchBwindows[3]\n",
    "\n",
    "p_vals = []\n",
    "\n",
    "dataframe = pd.read_csv(f'data_files/correlation_abs/w3_SCGRN_correlation.csv')\n",
    "out_df = dataframe\n",
    "\n",
    "for index, row in dataframe.iterrows():\n",
    "    p = get_p_value(matrixA, matrixB, row['TF'], row['GENE'], 10000, row['cc_coef_abs'])\n",
    "    p_vals.append(p)\n",
    "\n",
    "out_df['p-value'] = p_vals\n",
    "# out_df.to_csv(f'data_files/correlation_abs/w{i}_SCGRN_correlation_perm_10000.csv', index=False)\n",
    "print(f'finished window {3}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "pretty-vulnerability",
   "metadata": {},
   "source": [
    "#### 7.2 trying to make the shuffling and correlation computation faster"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "51590d9f",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import scanpy as sc\n",
    "import anndata\n",
    "import glob\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "39d49778",
   "metadata": {},
   "outputs": [],
   "source": [
    "import multiprocessing\n",
    "import concurrent.futures\n",
    "import time\n",
    "import numpy.random as rd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "69b23f4d",
   "metadata": {},
   "outputs": [],
   "source": [
    "branchAwindows = []\n",
    "branchBwindows = []\n",
    "\n",
    "A_windows_files=(glob.glob(f'data_files/windows_anndata/*A.h5ad'))\n",
    "B_windows_files=(glob.glob(f'data_files/windows_anndata/*B.h5ad'))\n",
    "A_windows_files = sorted(A_windows_files)\n",
    "B_windows_files = sorted(B_windows_files)\n",
    "\n",
    "for i in range(len(A_windows_files)):\n",
    "    branchAwindows.append(sc.read_h5ad(A_windows_files[i]))\n",
    "    branchBwindows.append(sc.read_h5ad(B_windows_files[i]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "id": "e5317449",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_shuffled_corrcoef(matA, matB, tf, gene):\n",
    "    \n",
    "    # starting time\n",
    "    start = time.time()\n",
    "    \n",
    "    shuffled_matA = shuffle2d_2(matA.X)\n",
    "    shuffled_matB = shuffle2d_2(matB.X)\n",
    "    \n",
    "    # end time\n",
    "    end = time.time()\n",
    "    # total time taken\n",
    "    print(f\"Run time for shuffling is {end - start}\")\n",
    "    \n",
    "    # starting time\n",
    "    start2 = time.time()\n",
    "    \n",
    "    matA.X = shuffled_matA\n",
    "    matB.X = shuffled_matB\n",
    "    wA = np.corrcoef(matA.obs_vector(tf), matA.obs_vector(gene))[0, 1]\n",
    "    wB = np.corrcoef(matB.obs_vector(tf), matB.obs_vector(gene))[0, 1]\n",
    "    cc_coef = abs(wA-wB)\n",
    "    \n",
    "    # end time\n",
    "    end2 = time.time()\n",
    "    # total time taken\n",
    "    print(f\"Run time for corrcoef is {end2 - start2}\")\n",
    "    \n",
    "    return cc_coef"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "id": "c19ee64b",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_p_value_2(matA, matB, tf, gene, permutations, true_val):\n",
    "   \n",
    "    # starting time\n",
    "    start = time.time()\n",
    "\n",
    "    values = []\n",
    "        \n",
    "    with concurrent.futures.ProcessPoolExecutor() as executor:\n",
    "        tasks = [executor.submit(get_shuffled_corrcoef, matA, matB, tf, gene) for _ in range(permutations)]\n",
    "        for task in concurrent.futures.as_completed(tasks):\n",
    "            values.append(task.result())\n",
    "    \n",
    "    counter = 0\n",
    "    for value in values:\n",
    "        if value > true_val:\n",
    "            counter += 1\n",
    "    p = counter/permutations\n",
    "    print(p)\n",
    "    # end time\n",
    "    end = time.time()\n",
    "    # total time taken\n",
    "    print(f\"Run time for P-value is {end - start}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "id": "97731084",
   "metadata": {},
   "outputs": [],
   "source": [
    "def shuffle2d(arr2d):\n",
    "    \"\"\"Shuffes entries of 2-d array arr2d, preserving shape.\"\"\"\n",
    "    list2d = arr2d.tolist()\n",
    "    reshape = []\n",
    "    data = []\n",
    "    iend = 0\n",
    "    for row in list2d:\n",
    "        data.extend(row)\n",
    "        istart, iend = iend, iend+len(row)\n",
    "        reshape.append((istart, iend))\n",
    "    rd.shuffle(data)\n",
    "    return np.array([data[istart:iend] for (istart,iend) in reshape])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "id": "70977779",
   "metadata": {},
   "outputs": [],
   "source": [
    "# def shuffle2d_test(arr2d):\n",
    "#    shapee = arr2d.shape\n",
    "#    print(shapee)\n",
    "#    to1d = shapee[0]*shapee[1]\n",
    "#    print(to1d)\n",
    "#    arr2d = arr2d.reshape(to1d,)\n",
    "#    np.random.shuffle(arr2d)\n",
    "#    arr2d = arr2d.reshape(shapee)\n",
    "#    print(arr2d.shape)\n",
    "#    return "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "id": "documentary-vegetable",
   "metadata": {},
   "outputs": [],
   "source": [
    "# method to shuffle gene expression matrix (as 2d array)\n",
    "def shuffle2d_2(arr2d):\n",
    "    shapee = arr2d.shape           # save shape of original matrix\n",
    "    to1d = shapee[0]*shapee[1]     # get size of matrix as 1d array\n",
    "    arr2d = arr2d.reshape(to1d,)   # reshape matrix into 1d array\n",
    "    np.random.shuffle(arr2d)       # shuffle array\n",
    "    arr2d = arr2d.reshape(shapee)  # reshape to original form -> now shuffled\n",
    "    return arr2d"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "id": "4ae0f1cb",
   "metadata": {},
   "outputs": [],
   "source": [
    "# shuffle2d_test(matA.X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "id": "2b07d5ab",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Run time for shuffling is 0.010926008224487305\n",
      "Run time for corrcoef is 0.002603769302368164\n",
      "1.0\n",
      "Run time for P-value is 3.261845827102661\n"
     ]
    }
   ],
   "source": [
    "#test\n",
    "matA = branchAwindows[1]\n",
    "matB = branchBwindows[1]\n",
    "get_p_value_2(matA, matB, 'Rest', 'Zcchc6', 1, 0.12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "a116273a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "48.710385958353676"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(0.29226231575012207*10000)/60"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "bed53cf2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Run time for shuffling is 0.011691808700561523\n",
      "\n",
      "\n",
      "Run time for corrcoef is 0.008134603500366211Run time for shuffling is 0.017081499099731445Run time for corrcoef is 0.003803253173828125\n",
      "Run time for shuffling is 0.017433643341064453\n",
      "Run time for corrcoef is 0.003842592239379883\n",
      "Run time for shuffling is 0.011986255645751953\n",
      "Run time for corrcoef is 0.0029740333557128906\n",
      "Run time for shuffling is 0.01213526725769043\n",
      "Run time for corrcoef is 0.002951383590698242\n",
      "Run time for shuffling is 0.012038707733154297\n",
      "Run time for corrcoef is 0.002782106399536133\n",
      "Run time for shuffling is 0.013411760330200195\n",
      "Run time for corrcoef is 0.0031762123107910156\n",
      "Run time for shuffling is 0.0179443359375\n",
      "Run time for corrcoef is 0.003835439682006836\n",
      "Run time for shuffling is 0.017591238021850586\n",
      "Run time for corrcoef is 0.003959178924560547\n",
      "Run time for shuffling is 0.015452146530151367\n",
      "Run time for corrcoef is 0.003937244415283203\n",
      "Run time for shuffling is 0.014491558074951172\n",
      "Run time for corrcoef is 0.002983570098876953\n",
      "Run time for shuffling is 0.019179344177246094\n",
      "Run time for shuffling is 0.012751340866088867\n",
      "Run time for corrcoef is 0.003925323486328125\n",
      "Run time for corrcoef is 0.002956390380859375\n",
      "Run time for shuffling is 0.018749237060546875\n",
      "Run time for corrcoef is 0.003857851028442383Run time for shuffling is 0.012131929397583008\n",
      "\n",
      "Run time for corrcoef is 0.0028433799743652344\n",
      "Run time for shuffling is 0.01176595687866211\n",
      "Run time for corrcoef is 0.0027670860290527344\n",
      "Run time for shuffling is 0.017511367797851562\n",
      "Run time for corrcoef is 0.0037717819213867188\n",
      "Run time for shuffling is 0.01740431785583496\n",
      "Run time for corrcoef is 0.0037965774536132812\n",
      "Run time for shuffling is 0.01772904396057129\n",
      "Run time for corrcoef is 0.0037941932678222656\n",
      "Run time for shuffling is 0.019353389739990234\n",
      "Run time for corrcoef is 0.0039904117584228516\n",
      "Run time for shuffling is 0.021353483200073242\n",
      "Run time for shuffling is 0.015954256057739258Run time for corrcoef is 0.003947734832763672\n",
      "\n",
      "Run time for corrcoef is 0.0031452178955078125\n",
      "Run time for shuffling is 0.019145727157592773\n",
      "Run time for corrcoef is 0.003974199295043945\n",
      "Run time for shuffling is 0.018463134765625\n",
      "Run time for corrcoef is 0.0038309097290039062Run time for shuffling is 0.025487661361694336\n",
      "\n",
      "Run time for corrcoef is 0.004082441329956055\n",
      "Run time for shuffling is 0.018907546997070312\n",
      "Run time for corrcoef is 0.00376129150390625\n",
      "Run time for shuffling is 0.012912750244140625\n",
      "Run time for corrcoef is 0.0037245750427246094\n",
      "Run time for shuffling is 0.012877225875854492\n",
      "Run time for corrcoef is 0.0028581619262695312\n",
      "Run time for shuffling is 0.018041133880615234\n",
      "Run time for corrcoef is 0.0037009716033935547\n",
      "Run time for shuffling is 0.012505054473876953\n",
      "Run time for corrcoef is 0.002874612808227539\n",
      "Run time for shuffling is 0.01855301856994629\n",
      "Run time for corrcoef is 0.003698110580444336\n",
      "Run time for shuffling is 0.012314081192016602\n",
      "Run time for corrcoef is 0.0028955936431884766\n",
      "Run time for shuffling is 0.01300501823425293\n",
      "Run time for corrcoef is 0.0027871131896972656\n",
      "Run time for shuffling is 0.018144607543945312\n",
      "Run time for corrcoef is 0.003762960433959961\n",
      "Run time for shuffling is 0.017052412033081055\n",
      "Run time for corrcoef is 0.003705263137817383\n",
      "Run time for shuffling is 0.02653670310974121Run time for shuffling is 0.012506246566772461\n",
      "\n",
      "Run time for corrcoef is 0.0029888153076171875Run time for corrcoef is 0.0035462379455566406\n",
      "\n",
      "Run time for shuffling is 0.021706342697143555\n",
      "Run time for corrcoef is 0.003913402557373047\n",
      "Run time for shuffling is 0.01279139518737793\n",
      "Run time for corrcoef is 0.0028972625732421875\n",
      "Run time for shuffling is 0.018641233444213867\n",
      "Run time for corrcoef is 0.0037622451782226562\n",
      "Run time for shuffling is 0.018208980560302734\n",
      "Run time for corrcoef is 0.0037539005279541016\n",
      "Run time for shuffling is 0.018760204315185547\n",
      "Run time for corrcoef is 0.0038652420043945312\n",
      "Run time for shuffling is 0.0184171199798584\n",
      "Run time for corrcoef is 0.0038132667541503906\n",
      "Run time for shuffling is 0.013018369674682617\n",
      "Run time for corrcoef is 0.003199338912963867\n",
      "Run time for shuffling is 0.018582820892333984\n",
      "Run time for corrcoef is 0.0038564205169677734\n",
      "Run time for shuffling is 0.01935887336730957\n",
      "Run time for corrcoef is 0.0035600662231445312\n",
      "Run time for shuffling is 0.013240575790405273\n",
      "Run time for corrcoef is 0.0028841495513916016\n",
      "Run time for shuffling is 0.020730257034301758\n",
      "Run time for corrcoef is 0.0035104751586914062\n",
      "Run time for shuffling is 0.017112255096435547\n",
      "Run time for corrcoef is 0.0029375553131103516\n",
      "Run time for shuffling is 0.025381088256835938\n",
      "Run time for shuffling is 0.012632608413696289\n",
      "Run time for corrcoef is 0.002990245819091797\n",
      "Run time for corrcoef is 0.002817869186401367\n",
      "Run time for shuffling is 0.018610715866088867\n",
      "Run time for corrcoef is 0.0038499832153320312\n",
      "Run time for shuffling is 0.01847052574157715\n",
      "Run time for shuffling is 0.012449264526367188\n",
      "Run time for corrcoef is 0.0038175582885742188\n",
      "Run time for corrcoef is 0.0027675628662109375\n",
      "Run time for shuffling is 0.01222848892211914\n",
      "Run time for corrcoef is 0.0028650760650634766\n",
      "Run time for shuffling is 0.014055252075195312Run time for shuffling is 0.02715444564819336\n",
      "\n",
      "Run time for corrcoef is 0.002921581268310547\n",
      "Run time for corrcoef is 0.004004001617431641\n",
      "Run time for shuffling is 0.012454986572265625\n",
      "Run time for corrcoef is 0.0027658939361572266\n",
      "Run time for shuffling is 0.012174844741821289\n",
      "Run time for corrcoef is 0.002811908721923828\n",
      "Run time for shuffling is 0.017464637756347656\n",
      "Run time for shuffling is 0.011586189270019531\n",
      "Run time for corrcoef is 0.0037360191345214844\n",
      "Run time for corrcoef is 0.0025892257690429688\n",
      "Run time for shuffling is 0.017507314682006836\n",
      "Run time for corrcoef is 0.003760099411010742\n",
      "Run time for shuffling is 0.015545368194580078\n",
      "Run time for corrcoef is 0.0038444995880126953\n",
      "Run time for shuffling is 0.011590719223022461\n",
      "Run time for corrcoef is 0.0027952194213867188\n",
      "Run time for shuffling is 0.011936187744140625\n",
      "Run time for corrcoef is 0.002775430679321289\n",
      "Run time for shuffling is 0.027790546417236328\n",
      "Run time for shuffling is 0.01737833023071289\n",
      "Run time for shuffling is 0.011714458465576172Run time for corrcoef is 0.003700733184814453\n",
      "\n",
      "Run time for corrcoef is 0.0037384033203125\n",
      "Run time for corrcoef is 0.0026254653930664062\n",
      "Run time for shuffling is 0.017495155334472656\n",
      "Run time for corrcoef is 0.0037794113159179688\n",
      "Run time for shuffling is 0.011629581451416016\n",
      "Run time for corrcoef is 0.002816438674926758\n",
      "Run time for shuffling is 0.011638164520263672\n",
      "Run time for corrcoef is 0.0026941299438476562\n",
      "Run time for shuffling is 0.01161646842956543\n",
      "Run time for corrcoef is 0.0026884078979492188\n",
      "Run time for shuffling is 0.014078617095947266\n",
      "Run time for corrcoef is 0.0030143260955810547\n",
      "Run time for shuffling is 0.012021541595458984\n",
      "Run time for corrcoef is 0.002828359603881836\n",
      "Run time for shuffling is 0.012039899826049805\n",
      "Run time for corrcoef is 0.0029172897338867188\n",
      "Run time for shuffling is 0.01438760757446289\n",
      "Run time for corrcoef is 0.0028123855590820312\n",
      "Run time for shuffling is 0.01819777488708496\n",
      "Run time for shuffling is 0.011920928955078125\n",
      "Run time for corrcoef is 0.0037920475006103516\n",
      "Run time for corrcoef is 0.002591848373413086\n",
      "Run time for shuffling is 0.013688325881958008\n",
      "Run time for corrcoef is 0.002870798110961914\n",
      "Run time for shuffling is 0.023429155349731445\n",
      "Run time for shuffling is 0.014276981353759766Run time for shuffling is 0.026463031768798828Run time for shuffling is 0.027064085006713867\n",
      "\n",
      "\n",
      "Run time for corrcoef is 0.004049062728881836Run time for corrcoef is 0.0038056373596191406\n",
      "Run time for corrcoef is 0.0033118724822998047\n",
      "Run time for corrcoef is 0.003777742385864258\n",
      "\n",
      "Run time for shuffling is 0.016382932662963867\n",
      "Run time for corrcoef is 0.0040187835693359375\n",
      "Run time for shuffling is 0.014765739440917969Run time for shuffling is 0.023456573486328125\n",
      "\n",
      "Run time for corrcoef is 0.003480195999145508Run time for corrcoef is 0.003721952438354492\n",
      "\n",
      "Run time for shuffling is 0.02325749397277832Run time for shuffling is 0.01477956771850586\n",
      "\n",
      "Run time for corrcoef is 0.002678394317626953\n",
      "Run time for corrcoef is 0.005698442459106445\n",
      "Run time for shuffling is 0.015144824981689453\n",
      "Run time for corrcoef is 0.0036995410919189453\n",
      "Run time for shuffling is 0.014220237731933594\n",
      "Run time for corrcoef is 0.0029366016387939453\n",
      "Run time for shuffling is 0.016412019729614258\n",
      "Run time for corrcoef is 0.0036706924438476562\n",
      "Run time for shuffling is 0.013912439346313477\n",
      "Run time for corrcoef is 0.0036623477935791016\n",
      "Run time for shuffling is 0.014065980911254883\n",
      "Run time for corrcoef is 0.0036339759826660156\n",
      "Run time for shuffling is 0.013786792755126953\n",
      "Run time for corrcoef is 0.003596782684326172\n",
      "Run time for shuffling is 0.023116111755371094Run time for shuffling is 0.014435529708862305\n",
      "\n",
      "Run time for corrcoef is 0.0035560131072998047\n",
      "Run time for corrcoef is 0.005995035171508789\n",
      "Run time for shuffling is 0.01428079605102539\n",
      "Run time for corrcoef is 0.003597259521484375\n",
      "Run time for shuffling is 0.022617816925048828\n",
      "Run time for corrcoef is 0.0035512447357177734\n",
      "Run time for shuffling is 0.022081375122070312Run time for shuffling is 0.014710426330566406\n",
      "\n",
      "Run time for corrcoef is 0.0036368370056152344\n",
      "Run time for corrcoef is 0.005998134613037109\n",
      "Run time for shuffling is 0.014589071273803711\n",
      "Run time for corrcoef is 0.004111528396606445\n",
      "Run time for shuffling is 0.020039081573486328\n",
      "Run time for shuffling is 0.011718273162841797\n",
      "Run time for corrcoef is 0.004525423049926758\n",
      "Run time for corrcoef is 0.0025298595428466797\n",
      "Run time for shuffling is 0.01778721809387207\n",
      "Run time for corrcoef is 0.0037665367126464844Run time for shuffling is 0.018340110778808594\n",
      "\n",
      "Run time for corrcoef is 0.0037415027618408203\n",
      "Run time for shuffling is 0.011949300765991211\n",
      "Run time for corrcoef is 0.002685546875\n",
      "Run time for shuffling is 0.01765275001525879\n",
      "Run time for corrcoef is 0.003809690475463867\n",
      "Run time for shuffling is 0.017554283142089844\n",
      "Run time for corrcoef is 0.0036935806274414062\n",
      "Run time for shuffling is 0.011800765991210938\n",
      "Run time for corrcoef is 0.0027360916137695312\n",
      "Run time for shuffling is 0.011827707290649414\n",
      "Run time for corrcoef is 0.002758026123046875\n",
      "Run time for shuffling is 0.017412900924682617\n",
      "Run time for corrcoef is 0.0037841796875\n",
      "Run time for shuffling is 0.017469167709350586\n",
      "Run time for corrcoef is 0.0035157203674316406\n",
      "Run time for shuffling is 0.014522552490234375\n",
      "Run time for corrcoef is 0.0031652450561523438\n",
      "Run time for shuffling is 0.022882461547851562\n",
      "Run time for corrcoef is 0.0028650760650634766\n",
      "Run time for shuffling is 0.017525196075439453\n",
      "Run time for corrcoef is 0.002828359603881836\n",
      "Run time for shuffling is 0.017536401748657227\n",
      "Run time for corrcoef is 0.003958463668823242\n",
      "Run time for shuffling is 0.018396854400634766\n",
      "Run time for corrcoef is 0.0037610530853271484\n",
      "Run time for shuffling is 0.012024879455566406\n",
      "Run time for corrcoef is 0.002841949462890625\n",
      "Run time for shuffling is 0.018454551696777344\n",
      "Run time for corrcoef is 0.0034422874450683594\n",
      "Run time for shuffling is 0.01840519905090332\n",
      "Run time for corrcoef is 0.003426074981689453\n",
      "Run time for shuffling is 0.017614364624023438\n",
      "Run time for corrcoef is 0.0037772655487060547\n",
      "Run time for shuffling is 0.017626523971557617\n",
      "Run time for corrcoef is 0.0036995410919189453\n",
      "Run time for shuffling is 0.017522811889648438\n",
      "Run time for corrcoef is 0.003790616989135742\n",
      "Run time for shuffling is 0.01801753044128418\n",
      "Run time for corrcoef is 0.0035636425018310547\n",
      "Run time for shuffling is 0.012122869491577148\n",
      "Run time for corrcoef is 0.002779245376586914\n",
      "Run time for shuffling is 0.012372255325317383\n",
      "Run time for corrcoef is 0.002802133560180664\n",
      "Run time for shuffling is 0.018074989318847656\n",
      "Run time for corrcoef is 0.003841400146484375\n",
      "Run time for shuffling is 0.01249384880065918\n",
      "Run time for corrcoef is 0.0029845237731933594\n",
      "Run time for shuffling is 0.0224912166595459\n",
      "Run time for shuffling is 0.012996912002563477\n",
      "Run time for corrcoef is 0.0035965442657470703Run time for corrcoef is 0.0026373863220214844\n",
      "\n",
      "Run time for shuffling is 0.018413066864013672Run time for shuffling is 0.012427091598510742\n",
      "\n",
      "Run time for shuffling is 0.01904773712158203\n",
      "Run time for corrcoef is 0.003837108612060547Run time for corrcoef is 0.003936290740966797\n",
      "\n",
      "Run time for corrcoef is 0.003774404525756836\n",
      "Run time for shuffling is 0.01853179931640625\n",
      "Run time for corrcoef is 0.0037851333618164062\n",
      "Run time for shuffling is 0.012524843215942383Run time for shuffling is 0.018424034118652344\n",
      "\n",
      "Run time for corrcoef is 0.002766132354736328\n",
      "Run time for corrcoef is 0.003393888473510742\n",
      "Run time for shuffling is 0.017897605895996094\n",
      "Run time for corrcoef is 0.003840923309326172Run time for shuffling is 0.020284414291381836\n",
      "\n",
      "Run time for corrcoef is 0.0030879974365234375\n",
      "Run time for shuffling is 0.013570785522460938\n",
      "Run time for corrcoef is 0.0041103363037109375\n",
      "Run time for shuffling is 0.015604496002197266\n",
      "Run time for corrcoef is 0.003120899200439453\n",
      "Run time for shuffling is 0.014635801315307617\n",
      "Run time for corrcoef is 0.004099369049072266\n",
      "Run time for shuffling is 0.015472412109375\n",
      "Run time for corrcoef is 0.0029783248901367188\n",
      "Run time for shuffling is 0.013806581497192383\n",
      "Run time for corrcoef is 0.003155231475830078\n",
      "Run time for shuffling is 0.01536417007446289\n",
      "Run time for corrcoef is 0.003930330276489258\n",
      "Run time for shuffling is 0.014579057693481445\n",
      "Run time for corrcoef is 0.004277706146240234\n",
      "Run time for shuffling is 0.011709451675415039\n",
      "Run time for shuffling is 0.0206756591796875Run time for corrcoef is 0.002705097198486328\n",
      "\n",
      "Run time for corrcoef is 0.006124019622802734\n",
      "Run time for shuffling is 0.017478227615356445\n",
      "Run time for corrcoef is 0.003725767135620117\n",
      "Run time for shuffling is 0.011704683303833008\n",
      "Run time for corrcoef is 0.002778768539428711\n",
      "Run time for shuffling is 0.01420903205871582\n",
      "Run time for corrcoef is 0.002769947052001953\n",
      "Run time for shuffling is 0.012461185455322266\n",
      "Run time for corrcoef is 0.002755403518676758\n",
      "Run time for shuffling is 0.012393474578857422\n",
      "Run time for corrcoef is 0.0028655529022216797\n",
      "Run time for shuffling is 0.018391847610473633\n",
      "Run time for corrcoef is 0.003737926483154297\n",
      "Run time for shuffling is 0.018629789352416992\n",
      "Run time for corrcoef is 0.0035054683685302734\n",
      "Run time for shuffling is 0.019113779067993164\n",
      "Run time for corrcoef is 0.0031614303588867188\n",
      "Run time for shuffling is 0.021396875381469727\n",
      "Run time for corrcoef is 0.004118680953979492\n",
      "Run time for shuffling is 0.014600038528442383\n",
      "Run time for corrcoef is 0.004445791244506836\n",
      "Run time for shuffling is 0.01697826385498047\n",
      "Run time for corrcoef is 0.004075050354003906\n",
      "Run time for shuffling is 0.01781773567199707Run time for shuffling is 0.011495113372802734\n",
      "\n",
      "Run time for corrcoef is 0.002588987350463867Run time for corrcoef is 0.003813028335571289\n",
      "\n",
      "Run time for shuffling is 0.017386198043823242\n",
      "Run time for corrcoef is 0.0020051002502441406Run time for shuffling is 0.017559528350830078\n",
      "Run time for shuffling is 0.012446880340576172\n",
      "Run time for corrcoef is 0.0021080970764160156Run time for corrcoef is 0.0015177726745605469\n",
      "\n",
      "Run time for shuffling is 0.018078088760375977\n",
      "\n",
      "Run time for corrcoef is 0.0018718242645263672Run time for shuffling is 0.01231527328491211\n",
      "Run time for shuffling is 0.012758731842041016\n",
      "\n",
      "Run time for corrcoef is 0.0012431144714355469Run time for corrcoef is 0.002000093460083008\n",
      "\n",
      "Run time for shuffling is 0.018193721771240234\n",
      "Run time for corrcoef is 0.001957416534423828Run time for shuffling is 0.0151214599609375\n",
      "\n",
      "Run time for corrcoef is 0.001293182373046875Run time for shuffling is 0.012766361236572266Run time for shuffling is 0.012562990188598633\n",
      "\n",
      "\n",
      "Run time for corrcoef is 0.0013341903686523438Run time for corrcoef is 0.0012962818145751953\n",
      "\n",
      "Run time for shuffling is 0.012521982192993164\n",
      "Run time for corrcoef is 0.001270294189453125\n",
      "Run time for shuffling is 0.01788949966430664\n",
      "Run time for corrcoef is 0.0018038749694824219Run time for shuffling is 0.017620563507080078Run time for shuffling is 0.011034488677978516\n",
      "\n",
      "\n",
      "Run time for corrcoef is 0.0017366409301757812Run time for corrcoef is 0.0020322799682617188\n",
      "\n",
      "Run time for shuffling is 0.011912345886230469Run time for shuffling is 0.011675357818603516\n",
      "\n",
      "Run time for corrcoef is 0.0012297630310058594\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Run time for corrcoef is 0.002294301986694336\n",
      "Run time for shuffling is 0.017579078674316406\n",
      "Run time for corrcoef is 0.0018773078918457031Run time for shuffling is 0.01190495491027832\n",
      "Run time for corrcoef is 0.001203775405883789\n",
      "Run time for shuffling is 0.017632484436035156\n",
      "\n",
      "Run time for shuffling is 0.01183319091796875Run time for corrcoef is 0.0016703605651855469\n",
      "\n",
      "Run time for corrcoef is 0.0016341209411621094\n",
      "Run time for shuffling is 0.017911672592163086\n",
      "Run time for corrcoef is 0.001773834228515625Run time for shuffling is 0.012092351913452148\n",
      "Run time for corrcoef is 0.001163482666015625Run time for shuffling is 0.018133878707885742Run time for shuffling is 0.012005090713500977\n",
      "\n",
      "\n",
      "\n",
      "Run time for corrcoef is 0.0018389225006103516Run time for corrcoef is 0.0011782646179199219\n",
      "\n",
      "Run time for shuffling is 0.017767906188964844\n",
      "Run time for corrcoef is 0.0017917156219482422\n",
      "Run time for shuffling is 0.017855167388916016\n",
      "Run time for shuffling is 0.011635780334472656Run time for corrcoef is 0.0018157958984375\n",
      "Run time for shuffling is 0.017593860626220703\n",
      "\n",
      "Run time for corrcoef is 0.0011987686157226562\n",
      "Run time for corrcoef is 0.002980947494506836Run time for shuffling is 0.011922836303710938\n",
      "\n",
      "Run time for corrcoef is 0.0012202262878417969\n",
      "Run time for shuffling is 0.01779961585998535\n",
      "Run time for shuffling is 0.011571645736694336Run time for corrcoef is 0.0018305778503417969\n",
      "Run time for corrcoef is 0.0011949539184570312\n",
      "\n",
      "Run time for shuffling is 0.011793136596679688\n",
      "Run time for corrcoef is 0.0012021064758300781Run time for shuffling is 0.0177156925201416\n",
      "\n",
      "Run time for corrcoef is 0.001790761947631836Run time for shuffling is 0.017709970474243164\n",
      "Run time for corrcoef is 0.0017735958099365234Run time for shuffling is 0.012360572814941406\n",
      "Run time for shuffling is 0.01838064193725586\n",
      "\n",
      "Run time for corrcoef is 0.0018448829650878906\n",
      "\n",
      "Run time for corrcoef is 0.001922607421875Run time for shuffling is 0.012453317642211914\n",
      "\n",
      "Run time for shuffling is 0.018348217010498047\n",
      "Run time for corrcoef is 0.0029129981994628906Run time for shuffling is 0.012318134307861328\n",
      "Run time for corrcoef is 0.0018568038940429688\n",
      "\n",
      "Run time for corrcoef is 0.0023338794708251953\n",
      "Run time for shuffling is 0.01260685920715332\n",
      "Run time for shuffling is 0.018238544464111328\n",
      "Run time for corrcoef is 0.0012891292572021484\n",
      "Run time for corrcoef is 0.0018482208251953125\n",
      "Run time for shuffling is 0.012573719024658203\n",
      "Run time for shuffling is 0.018550872802734375\n",
      "Run time for corrcoef is 0.001306772232055664\n",
      "Run time for corrcoef is 0.0017023086547851562\n",
      "Run time for shuffling is 0.017907142639160156\n",
      "Run time for corrcoef is 0.0018451213836669922\n",
      "Run time for shuffling is 0.018105745315551758\n",
      "Run time for shuffling is 0.01193094253540039\n",
      "Run time for corrcoef is 0.0019173622131347656\n",
      "Run time for corrcoef is 0.0012180805206298828\n",
      "Run time for shuffling is 0.01266026496887207\n",
      "Run time for corrcoef is 0.0011813640594482422\n",
      "Run time for shuffling is 0.017421483993530273\n",
      "Run time for corrcoef is 0.001814126968383789Run time for shuffling is 0.011468887329101562\n",
      "Run time for corrcoef is 0.0011870861053466797\n",
      "Run time for shuffling is 0.01734614372253418\n",
      "\n",
      "Run time for corrcoef is 0.0017576217651367188Run time for shuffling is 0.011445760726928711\n",
      "\n",
      "Run time for corrcoef is 0.0011606216430664062\n",
      "Run time for shuffling is 0.012075662612915039\n",
      "Run time for corrcoef is 0.0012023448944091797\n",
      "Run time for shuffling is 0.011390209197998047\n",
      "Run time for corrcoef is 0.001171112060546875\n",
      "Run time for shuffling is 0.020360708236694336\n",
      "Run time for shuffling is 0.011442899703979492Run time for shuffling is 0.011851310729980469\n",
      "\n",
      "Run time for corrcoef is 0.0012407302856445312Run time for corrcoef is 0.0012364387512207031Run time for corrcoef is 0.0017464160919189453\n",
      "\n",
      "Run time for shuffling is 0.011825799942016602\n",
      "\n",
      "Run time for shuffling is 0.011936187744140625\n",
      "Run time for corrcoef is 0.0012023448944091797Run time for corrcoef is 0.001157999038696289\n",
      "\n",
      "Run time for shuffling is 0.011815786361694336\n",
      "Run time for corrcoef is 0.0012705326080322266\n",
      "Run time for shuffling is 0.011865377426147461\n",
      "Run time for corrcoef is 0.0012276172637939453\n",
      "Run time for shuffling is 0.01750946044921875Run time for shuffling is 0.011928081512451172\n",
      "\n",
      "Run time for shuffling is 0.011600494384765625Run time for corrcoef is 0.0011873245239257812\n",
      "\n",
      "Run time for corrcoef is 0.0011749267578125Run time for shuffling is 0.011558771133422852\n",
      "\n",
      "Run time for corrcoef is 0.0011763572692871094\n",
      "Run time for corrcoef is 0.0017304420471191406\n",
      "Run time for shuffling is 0.017644882202148438\n",
      "Run time for shuffling is 0.011907339096069336Run time for corrcoef is 0.0018496513366699219\n",
      "\n",
      "Run time for shuffling is 0.01776123046875Run time for corrcoef is 0.0011775493621826172\n",
      "Run time for corrcoef is 0.001833200454711914\n",
      "Run time for shuffling is 0.011342048645019531\n",
      "Run time for corrcoef is 0.0011715888977050781\n",
      "\n",
      "Run time for shuffling is 0.011882543563842773\n",
      "Run time for corrcoef is 0.001237630844116211Run time for shuffling is 0.020977020263671875\n",
      "\n",
      "Run time for corrcoef is 0.0018963813781738281\n",
      "Run time for shuffling is 0.012379169464111328Run time for shuffling is 0.01209259033203125\n",
      "\n",
      "Run time for corrcoef is 0.001216888427734375Run time for corrcoef is 0.0011854171752929688\n",
      "\n",
      "Run time for shuffling is 0.012578010559082031\n",
      "Run time for shuffling is 0.01780247688293457\n",
      "Run time for corrcoef is 0.001302957534790039\n",
      "Run time for corrcoef is 0.0017876625061035156\n",
      "Run time for shuffling is 0.01184391975402832\n",
      "Run time for corrcoef is 0.0011811256408691406\n",
      "Run time for shuffling is 0.019173622131347656\n",
      "Run time for shuffling is 0.012070417404174805Run time for corrcoef is 0.0017616748809814453\n",
      "\n",
      "Run time for corrcoef is 0.0011677742004394531\n",
      "Run time for shuffling is 0.012188434600830078\n",
      "Run time for corrcoef is 0.0012028217315673828\n",
      "Run time for shuffling is 0.012994766235351562\n",
      "Run time for shuffling is 0.01223134994506836\n",
      "Run time for corrcoef is 0.0012786388397216797\n",
      "Run time for corrcoef is 0.0029621124267578125\n",
      "Run time for shuffling is 0.01221323013305664\n",
      "Run time for corrcoef is 0.0013630390167236328\n",
      "Run time for shuffling is 0.02035355567932129\n",
      "Run time for shuffling is 0.011919736862182617\n",
      "Run time for corrcoef is 0.0018930435180664062Run time for corrcoef is 0.001285552978515625\n",
      "\n",
      "Run time for shuffling is 0.011822223663330078\n",
      "Run time for corrcoef is 0.0012753009796142578\n",
      "Run time for shuffling is 0.012059211730957031\n",
      "Run time for shuffling is 0.017345428466796875\n",
      "Run time for shuffling is 0.011564493179321289\n",
      "Run time for corrcoef is 0.001251220703125\n",
      "Run time for corrcoef is 0.0018219947814941406Run time for corrcoef is 0.0012328624725341797\n",
      "\n",
      "Run time for shuffling is 0.015074729919433594Run time for shuffling is 0.01191854476928711\n",
      "\n",
      "Run time for corrcoef is 0.0012209415435791016Run time for corrcoef is 0.0012023448944091797\n",
      "\n",
      "Run time for shuffling is 0.01216578483581543Run time for shuffling is 0.017311573028564453\n",
      "\n",
      "Run time for corrcoef is 0.0018694400787353516Run time for corrcoef is 0.0028972625732421875Run time for shuffling is 0.01211690902709961\n",
      "\n",
      "\n",
      "Run time for corrcoef is 0.002908468246459961Run time for shuffling is 0.011431694030761719\n",
      "Run time for corrcoef is 0.0012199878692626953\n",
      "\n",
      "Run time for shuffling is 0.0123443603515625\n",
      "Run time for corrcoef is 0.0013480186462402344\n",
      "Run time for shuffling is 0.012563228607177734\n",
      "Run time for corrcoef is 0.0012774467468261719Run time for shuffling is 0.011847257614135742\n",
      "\n",
      "Run time for corrcoef is 0.0012717247009277344\n",
      "Run time for shuffling is 0.012007713317871094\n",
      "Run time for corrcoef is 0.001287698745727539\n",
      "Run time for shuffling is 0.012232065200805664Run time for shuffling is 0.017531871795654297\n",
      "\n",
      "Run time for corrcoef is 0.0016033649444580078Run time for corrcoef is 0.003007650375366211Run time for shuffling is 0.012018918991088867\n",
      "\n",
      "\n",
      "Run time for corrcoef is 0.0014050006866455078\n",
      "Run time for shuffling is 0.012075424194335938\n",
      "Run time for corrcoef is 0.0012073516845703125\n",
      "Run time for shuffling is 0.012200355529785156\n",
      "Run time for corrcoef is 0.0012543201446533203\n",
      "Run time for shuffling is 0.011529922485351562\n",
      "Run time for corrcoef is 0.0011589527130126953\n",
      "Run time for shuffling is 0.013032674789428711\n",
      "Run time for shuffling is 0.011913776397705078\n",
      "Run time for corrcoef is 0.0012063980102539062Run time for corrcoef is 0.0011608600616455078\n",
      "\n",
      "Run time for shuffling is 0.013209819793701172\n",
      "Run time for corrcoef is 0.001219034194946289Run time for shuffling is 0.012061119079589844\n",
      "\n",
      "Run time for corrcoef is 0.0014331340789794922\n",
      "Run time for shuffling is 0.012431621551513672\n",
      "Run time for corrcoef is 0.0012662410736083984Run time for shuffling is 0.018512487411499023\n",
      "\n",
      "Run time for corrcoef is 0.0018796920776367188Run time for shuffling is 0.012526512145996094\n",
      "Run time for corrcoef is 0.0011649131774902344\n",
      "\n",
      "Run time for shuffling is 0.012326478958129883\n",
      "Run time for corrcoef is 0.0012297630310058594Run time for shuffling is 0.012555837631225586\n",
      "\n",
      "Run time for corrcoef is 0.0011951923370361328\n",
      "Run time for shuffling is 0.012633085250854492\n",
      "Run time for corrcoef is 0.0011696815490722656\n",
      "Run time for shuffling is 0.013416290283203125\n",
      "Run time for corrcoef is 0.001220703125Run time for shuffling is 0.01235342025756836\n",
      "Run time for corrcoef is 0.001203775405883789Run time for shuffling is 0.012436389923095703\n",
      "\n",
      "Run time for corrcoef is 0.0012009143829345703\n",
      "\n",
      "Run time for shuffling is 0.012655973434448242\n",
      "Run time for corrcoef is 0.0011637210845947266\n",
      "Run time for shuffling is 0.012423038482666016\n",
      "Run time for shuffling is 0.01858997344970703Run time for corrcoef is 0.001260519027709961\n",
      "\n",
      "Run time for corrcoef is 0.0018494129180908203Run time for shuffling is 0.012563705444335938\n",
      "\n",
      "Run time for corrcoef is 0.001199960708618164\n",
      "Run time for shuffling is 0.012464523315429688\n",
      "Run time for corrcoef is 0.0011963844299316406\n",
      "Run time for shuffling is 0.012671232223510742\n",
      "Run time for corrcoef is 0.001219034194946289\n",
      "Run time for shuffling is 0.012567996978759766\n",
      "Run time for corrcoef is 0.0012249946594238281Run time for shuffling is 0.01268625259399414\n",
      "Run time for corrcoef is 0.0012249946594238281\n",
      "\n",
      "Run time for shuffling is 0.012605428695678711\n",
      "Run time for corrcoef is 0.001207590103149414\n",
      "Run time for shuffling is 0.012570381164550781\n",
      "Run time for corrcoef is 0.0012373924255371094\n",
      "Run time for shuffling is 0.012607336044311523\n",
      "Run time for shuffling is 0.018674612045288086Run time for corrcoef is 0.001256704330444336\n",
      "\n",
      "Run time for corrcoef is 0.0018486976623535156Run time for shuffling is 0.012380123138427734\n",
      "Run time for corrcoef is 0.00115966796875\n",
      "\n",
      "Run time for shuffling is 0.012573719024658203\n",
      "Run time for corrcoef is 0.0012202262878417969\n",
      "Run time for shuffling is 0.01303553581237793\n",
      "Run time for corrcoef is 0.0012462139129638672\n",
      "Run time for shuffling is 0.012399435043334961\n",
      "Run time for shuffling is 0.01856374740600586\n",
      "Run time for corrcoef is 0.0012040138244628906Run time for corrcoef is 0.0016796588897705078\n",
      "\n",
      "Run time for shuffling is 0.01261281967163086\n",
      "Run time for corrcoef is 0.0012080669403076172Run time for shuffling is 0.012305736541748047\n",
      "\n",
      "Run time for corrcoef is 0.0011703968048095703\n",
      "Run time for shuffling is 0.01564478874206543\n",
      "Run time for shuffling is 0.012128829956054688\n",
      "Run time for corrcoef is 0.0016188621520996094Run time for corrcoef is 0.0011708736419677734\n",
      "\n",
      "Run time for shuffling is 0.012880325317382812\n",
      "Run time for shuffling is 0.018377065658569336\n",
      "Run time for corrcoef is 0.0012786388397216797Run time for corrcoef is 0.0017774105072021484\n",
      "Run time for shuffling is 0.012292623519897461\n",
      "\n",
      "Run time for corrcoef is 0.0011680126190185547\n",
      "Run time for shuffling is 0.012395381927490234\n",
      "Run time for corrcoef is 0.0011970996856689453\n",
      "Run time for shuffling is 0.012636899948120117\n",
      "Run time for corrcoef is 0.0011365413665771484\n",
      "Run time for shuffling is 0.011964797973632812\n",
      "Run time for corrcoef is 0.0011186599731445312\n",
      "Run time for shuffling is 0.012152433395385742\n",
      "Run time for corrcoef is 0.0011708736419677734\n",
      "Run time for shuffling is 0.011665821075439453\n",
      "Run time for corrcoef is 0.0011343955993652344\n",
      "Run time for shuffling is 0.01595449447631836\n",
      "Run time for corrcoef is 0.0012111663818359375Run time for shuffling is 0.012324333190917969\n",
      "\n",
      "Run time for corrcoef is 0.0012083053588867188Run time for shuffling is 0.01235508918762207\n",
      "\n",
      "Run time for corrcoef is 0.0011806488037109375\n",
      "Run time for shuffling is 0.01210784912109375\n",
      "Run time for corrcoef is 0.0011870861053466797\n",
      "Run time for shuffling is 0.012111186981201172\n",
      "Run time for corrcoef is 0.0012180805206298828\n",
      "Run time for shuffling is 0.012188434600830078\n",
      "Run time for corrcoef is 0.001252889633178711Run time for shuffling is 0.017673969268798828\n",
      "\n",
      "Run time for corrcoef is 0.0018515586853027344Run time for shuffling is 0.011945247650146484\n",
      "Run time for corrcoef is 0.001149892807006836\n",
      "\n",
      "Run time for shuffling is 0.012804746627807617Run time for shuffling is 0.018256187438964844\n",
      "Run time for corrcoef is 0.0019485950469970703\n",
      "Run time for shuffling is 0.017609357833862305\n",
      "\n",
      "Run time for corrcoef is 0.002393484115600586Run time for corrcoef is 0.0018463134765625Run time for shuffling is 0.01725029945373535\n",
      "\n",
      "Run time for corrcoef is 0.0032918453216552734Run time for shuffling is 0.017880678176879883\n",
      "\n",
      "Run time for shuffling is 0.01193380355834961\n",
      "\n",
      "Run time for corrcoef is 0.0012664794921875Run time for corrcoef is 0.0017809867858886719Run time for shuffling is 0.011619091033935547\n",
      "\n",
      "\n",
      "Run time for corrcoef is 0.0013446807861328125\n",
      "Run time for shuffling is 0.013986825942993164\n",
      "Run time for corrcoef is 0.0012969970703125\n",
      "Run time for shuffling is 0.01737380027770996\n",
      "Run time for shuffling is 0.011889457702636719Run time for corrcoef is 0.0017218589782714844\n",
      "\n",
      "Run time for corrcoef is 0.0012962818145751953\n",
      "Run time for shuffling is 0.012076616287231445\n",
      "Run time for corrcoef is 0.0012667179107666016Run time for shuffling is 0.011846303939819336\n",
      "Run time for corrcoef is 0.001276254653930664\n",
      "\n",
      "Run time for shuffling is 0.014227867126464844\n",
      "Run time for corrcoef is 0.0019249916076660156Run time for shuffling is 0.012209892272949219\n",
      "\n",
      "Run time for corrcoef is 0.0013358592987060547\n",
      "Run time for shuffling is 0.012347936630249023\n",
      "Run time for shuffling is 0.012268781661987305\n",
      "Run time for corrcoef is 0.0013725757598876953Run time for corrcoef is 0.0012950897216796875\n",
      "\n",
      "Run time for shuffling is 0.012125253677368164\n",
      "Run time for corrcoef is 0.0013475418090820312\n",
      "Run time for shuffling is 0.018761634826660156Run time for shuffling is 0.01736617088317871\n",
      "\n",
      "Run time for shuffling is 0.012077808380126953\n",
      "Run time for corrcoef is 0.0020322799682617188\n",
      "Run time for corrcoef is 0.002101421356201172\n",
      "Run time for corrcoef is 0.0012764930725097656\n",
      "Run time for shuffling is 0.012543916702270508\n",
      "Run time for corrcoef is 0.0013027191162109375\n",
      "Run time for shuffling is 0.012694597244262695\n",
      "Run time for corrcoef is 0.0013015270233154297\n",
      "Run time for shuffling is 0.014142990112304688\n",
      "Run time for corrcoef is 0.0017313957214355469\n",
      "Run time for shuffling is 0.011729955673217773\n",
      "Run time for corrcoef is 0.0013091564178466797\n",
      "Run time for shuffling is 0.011862993240356445\n",
      "Run time for corrcoef is 0.0012938976287841797\n",
      "Run time for shuffling is 0.015004396438598633\n",
      "Run time for corrcoef is 0.0012409687042236328\n",
      "Run time for shuffling is 0.01147007942199707\n",
      "Run time for corrcoef is 0.00119781494140625\n",
      "Run time for shuffling is 0.011506080627441406\n",
      "Run time for corrcoef is 0.0012247562408447266\n",
      "Run time for shuffling is 0.011575460433959961\n",
      "Run time for corrcoef is 0.001598358154296875\n",
      "Run time for shuffling is 0.012304067611694336\n",
      "Run time for corrcoef is 0.0012505054473876953\n",
      "Run time for shuffling is 0.011516332626342773\n",
      "Run time for shuffling is 0.017259597778320312\n",
      "Run time for corrcoef is 0.0013148784637451172Run time for shuffling is 0.017421960830688477\n",
      "\n",
      "Run time for corrcoef is 0.001859426498413086\n",
      "Run time for corrcoef is 0.0031061172485351562\n",
      "Run time for shuffling is 0.017007827758789062\n",
      "Run time for shuffling is 0.014088153839111328\n",
      "Run time for corrcoef is 0.002185344696044922\n",
      "Run time for corrcoef is 0.0015718936920166016\n",
      "Run time for shuffling is 0.011711359024047852\n",
      "Run time for corrcoef is 0.001318216323852539\n",
      "Run time for shuffling is 0.017482280731201172\n",
      "Run time for corrcoef is 0.0018916130065917969\n",
      "Run time for shuffling is 0.011471271514892578\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Run time for corrcoef is 0.0012614727020263672\n",
      "Run time for shuffling is 0.017295360565185547\n",
      "Run time for shuffling is 0.011939048767089844Run time for corrcoef is 0.001947164535522461\n",
      "\n",
      "Run time for corrcoef is 0.0012357234954833984\n",
      "Run time for shuffling is 0.01681971549987793\n",
      "Run time for corrcoef is 0.0021643638610839844Run time for shuffling is 0.011676788330078125\n",
      "\n",
      "Run time for corrcoef is 0.0018169879913330078Run time for shuffling is 0.011772632598876953\n",
      "\n",
      "Run time for corrcoef is 0.0012671947479248047\n",
      "Run time for shuffling is 0.011618375778198242\n",
      "Run time for corrcoef is 0.0012242794036865234\n",
      "Run time for shuffling is 0.01697850227355957\n",
      "Run time for shuffling is 0.011021852493286133\n",
      "Run time for corrcoef is 0.0018682479858398438\n",
      "Run time for corrcoef is 0.0011835098266601562\n",
      "Run time for shuffling is 0.014101028442382812\n",
      "Run time for corrcoef is 0.0018053054809570312\n",
      "Run time for shuffling is 0.014087677001953125\n",
      "Run time for shuffling is 0.01732611656188965\n",
      "Run time for corrcoef is 0.0013003349304199219\n",
      "Run time for corrcoef is 0.0018422603607177734\n",
      "Run time for shuffling is 0.011619806289672852\n",
      "Run time for corrcoef is 0.0012633800506591797\n",
      "Run time for shuffling is 0.017215728759765625\n",
      "Run time for shuffling is 0.012188196182250977\n",
      "Run time for corrcoef is 0.0019812583923339844Run time for corrcoef is 0.0012431144714355469\n",
      "\n",
      "Run time for shuffling is 0.011676311492919922Run time for shuffling is 0.011662721633911133\n",
      "\n",
      "Run time for corrcoef is 0.0012860298156738281Run time for corrcoef is 0.0016765594482421875Run time for shuffling is 0.014649391174316406\n",
      "\n",
      "\n",
      "Run time for corrcoef is 0.0017499923706054688\n",
      "Run time for shuffling is 0.011923789978027344\n",
      "Run time for corrcoef is 0.0012552738189697266Run time for shuffling is 0.016654491424560547\n",
      "\n",
      "Run time for shuffling is 0.01170802116394043Run time for corrcoef is 0.001840829849243164\n",
      "\n",
      "Run time for corrcoef is 0.0012297630310058594Run time for shuffling is 0.020150423049926758\n",
      "\n",
      "Run time for corrcoef is 0.0014503002166748047\n",
      "Run time for shuffling is 0.011520862579345703\n",
      "Run time for corrcoef is 0.0012781620025634766\n",
      "Run time for shuffling is 0.02022528648376465\n",
      "Run time for corrcoef is 0.0012402534484863281\n",
      "Run time for shuffling is 0.011412858963012695Run time for shuffling is 0.011606454849243164\n",
      "\n",
      "Run time for corrcoef is 0.001211404800415039Run time for corrcoef is 0.0012044906616210938\n",
      "\n",
      "Run time for shuffling is 0.011975765228271484\n",
      "Run time for corrcoef is 0.001272439956665039\n",
      "Run time for shuffling is 0.011757850646972656\n",
      "Run time for corrcoef is 0.001262664794921875\n",
      "Run time for shuffling is 0.017792224884033203\n",
      "Run time for corrcoef is 0.0017702579498291016\n",
      "Run time for shuffling is 0.012189626693725586\n",
      "Run time for corrcoef is 0.0012104511260986328\n",
      "Run time for shuffling is 0.0171968936920166\n",
      "Run time for shuffling is 0.018187284469604492Run time for corrcoef is 0.0018510818481445312\n",
      "\n",
      "Run time for corrcoef is 0.0018033981323242188\n",
      "Run time for shuffling is 0.011881351470947266\n",
      "Run time for corrcoef is 0.0012443065643310547\n",
      "Run time for shuffling is 0.01156473159790039\n",
      "Run time for corrcoef is 0.0011968612670898438Run time for shuffling is 0.012878179550170898\n",
      "\n",
      "Run time for shuffling is 0.011987924575805664Run time for corrcoef is 0.0012264251708984375\n",
      "\n",
      "Run time for corrcoef is 0.0011737346649169922\n",
      "Run time for shuffling is 0.01192617416381836\n",
      "Run time for shuffling is 0.017832040786743164Run time for corrcoef is 0.0012042522430419922\n",
      "\n",
      "Run time for corrcoef is 0.0018143653869628906\n",
      "Run time for shuffling is 0.012327194213867188\n",
      "Run time for shuffling is 0.012065410614013672Run time for corrcoef is 0.0013616085052490234\n",
      "\n",
      "Run time for corrcoef is 0.0015642642974853516\n",
      "Run time for shuffling is 0.01219034194946289\n",
      "Run time for shuffling is 0.018176555633544922\n",
      "Run time for corrcoef is 0.0013499259948730469\n",
      "Run time for corrcoef is 0.001903533935546875\n",
      "Run time for shuffling is 0.012521505355834961\n",
      "Run time for corrcoef is 0.0013356208801269531\n",
      "Run time for shuffling is 0.012363672256469727\n",
      "Run time for corrcoef is 0.0012755393981933594\n",
      "Run time for shuffling is 0.012152910232543945\n",
      "Run time for corrcoef is 0.001295328140258789\n",
      "Run time for shuffling is 0.01312398910522461\n",
      "Run time for corrcoef is 0.0013415813446044922\n",
      "Run time for shuffling is 0.013022661209106445Run time for shuffling is 0.012655019760131836\n",
      "\n",
      "Run time for corrcoef is 0.002048969268798828\n",
      "Run time for corrcoef is 0.0019314289093017578\n",
      "Run time for shuffling is 0.01879596710205078\n",
      "Run time for corrcoef is 0.0020558834075927734\n",
      "Run time for shuffling is 0.019316911697387695\n",
      "Run time for corrcoef is 0.002054452896118164\n",
      "Run time for shuffling is 0.01961827278137207\n",
      "Run time for corrcoef is 0.002084016799926758Run time for shuffling is 0.01608586311340332\n",
      "\n",
      "Run time for corrcoef is 0.0014009475708007812\n",
      "Run time for shuffling is 0.019416093826293945\n",
      "Run time for corrcoef is 0.0020699501037597656\n",
      "Run time for shuffling is 0.022021770477294922\n",
      "Run time for corrcoef is 0.0021142959594726562Run time for shuffling is 0.019095420837402344\n",
      "\n",
      "Run time for corrcoef is 0.002012968063354492\n",
      "Run time for shuffling is 0.019792795181274414\n",
      "Run time for corrcoef is 0.0020563602447509766\n",
      "Run time for shuffling is 0.01927495002746582\n",
      "Run time for shuffling is 0.01594853401184082\n",
      "Run time for corrcoef is 0.0020334720611572266\n",
      "Run time for corrcoef is 0.0017063617706298828\n",
      "Run time for shuffling is 0.01395726203918457\n",
      "Run time for corrcoef is 0.0014622211456298828\n",
      "Run time for shuffling is 0.012635231018066406\n",
      "Run time for corrcoef is 0.0013899803161621094\n",
      "Run time for shuffling is 0.018891334533691406\n",
      "Run time for corrcoef is 0.001966714859008789\n",
      "Run time for shuffling is 0.017633676528930664Run time for shuffling is 0.017808198928833008\n",
      "\n",
      "Run time for corrcoef is 0.0020024776458740234Run time for corrcoef is 0.001913309097290039\n",
      "\n",
      "Run time for shuffling is 0.011799335479736328\n",
      "Run time for corrcoef is 0.0013539791107177734\n",
      "Run time for shuffling is 0.017486095428466797\n",
      "Run time for shuffling is 0.01216268539428711\n",
      "Run time for corrcoef is 0.002053499221801758Run time for corrcoef is 0.0013530254364013672\n",
      "\n",
      "Run time for shuffling is 0.01835012435913086\n",
      "Run time for corrcoef is 0.0020589828491210938\n",
      "Run time for shuffling is 0.018581867218017578\n",
      "Run time for corrcoef is 0.0020046234130859375\n",
      "Run time for shuffling is 0.01453852653503418\n",
      "Run time for corrcoef is 0.0019366741180419922\n",
      "Run time for shuffling is 0.01883101463317871\n",
      "Run time for corrcoef is 0.0020360946655273438Run time for shuffling is 0.012597322463989258\n",
      "\n",
      "Run time for corrcoef is 0.0016324520111083984\n",
      "Run time for shuffling is 0.019032716751098633\n",
      "Run time for corrcoef is 0.0024013519287109375\n",
      "Run time for shuffling is 0.019459962844848633\n",
      "Run time for corrcoef is 0.0021195411682128906\n",
      "Run time for shuffling is 0.01336216926574707\n",
      "Run time for corrcoef is 0.0014028549194335938\n",
      "Run time for shuffling is 0.019639253616333008\n",
      "Run time for corrcoef is 0.0019876956939697266\n",
      "Run time for shuffling is 0.019771814346313477\n",
      "Run time for corrcoef is 0.002002716064453125\n",
      "Run time for shuffling is 0.02020740509033203\n",
      "Run time for shuffling is 0.016376018524169922\n",
      "Run time for corrcoef is 0.0019125938415527344\n",
      "Run time for corrcoef is 0.001783132553100586\n",
      "Run time for shuffling is 0.012835502624511719\n",
      "Run time for corrcoef is 0.001241922378540039\n",
      "Run time for shuffling is 0.01947188377380371\n",
      "Run time for shuffling is 0.012422323226928711\n",
      "Run time for corrcoef is 0.001847982406616211\n",
      "Run time for corrcoef is 0.0012259483337402344\n",
      "Run time for shuffling is 0.017951488494873047\n",
      "Run time for corrcoef is 0.0019559860229492188\n",
      "Run time for shuffling is 0.01605224609375\n",
      "Run time for corrcoef is 0.0013415813446044922\n",
      "Run time for shuffling is 0.019206523895263672\n",
      "Run time for shuffling is 0.01291513442993164\n",
      "Run time for corrcoef is 0.0020270347595214844Run time for corrcoef is 0.0013463497161865234\n",
      "\n",
      "Run time for shuffling is 0.019150257110595703\n",
      "Run time for corrcoef is 0.0019669532775878906\n",
      "Run time for shuffling is 0.019321203231811523\n",
      "Run time for shuffling is 0.01668095588684082\n",
      "Run time for corrcoef is 0.0019259452819824219\n",
      "Run time for corrcoef is 0.0018215179443359375\n",
      "Run time for shuffling is 0.019016027450561523\n",
      "Run time for corrcoef is 0.0019469261169433594\n",
      "Run time for shuffling is 0.01347804069519043\n",
      "Run time for shuffling is 0.018557310104370117\n",
      "Run time for corrcoef is 0.002058744430541992\n",
      "Run time for corrcoef is 0.0019366741180419922\n",
      "Run time for shuffling is 0.014184951782226562\n",
      "Run time for corrcoef is 0.0013630390167236328\n",
      "Run time for shuffling is 0.019130945205688477\n",
      "Run time for shuffling is 0.01883101463317871\n",
      "Run time for corrcoef is 0.002023935317993164\n",
      "Run time for corrcoef is 0.002156496047973633\n",
      "Run time for shuffling is 0.01728987693786621\n",
      "Run time for corrcoef is 0.0025489330291748047\n",
      "Run time for shuffling is 0.018758058547973633\n",
      "Run time for shuffling is 0.0161440372467041\n",
      "Run time for corrcoef is 0.0019350051879882812\n",
      "Run time for corrcoef is 0.0017566680908203125\n",
      "Run time for shuffling is 0.01152658462524414\n",
      "Run time for corrcoef is 0.0011746883392333984\n",
      "Run time for shuffling is 0.0178220272064209\n",
      "Run time for corrcoef is 0.0017850399017333984\n",
      "Run time for shuffling is 0.01491403579711914\n",
      "Run time for corrcoef is 0.0018587112426757812\n",
      "Run time for shuffling is 0.017153024673461914\n",
      "Run time for shuffling is 0.012184381484985352Run time for corrcoef is 0.0018618106842041016\n",
      "\n",
      "Run time for corrcoef is 0.001336812973022461\n",
      "Run time for shuffling is 0.017635583877563477\n",
      "Run time for shuffling is 0.017844676971435547\n",
      "Run time for corrcoef is 0.0017681121826171875\n",
      "Run time for corrcoef is 0.00179290771484375Run time for shuffling is 0.01417994499206543\n",
      "\n",
      "Run time for corrcoef is 0.0011968612670898438\n",
      "Run time for shuffling is 0.01572275161743164\n",
      "Run time for corrcoef is 0.0017056465148925781\n",
      "Run time for shuffling is 0.01327204704284668\n",
      "Run time for corrcoef is 0.001300811767578125\n",
      "Run time for shuffling is 0.019350528717041016\n",
      "Run time for corrcoef is 0.001794576644897461\n",
      "Run time for shuffling is 0.019304275512695312\n",
      "Run time for corrcoef is 0.0017862319946289062\n",
      "Run time for shuffling is 0.018319368362426758\n",
      "Run time for shuffling is 0.01785445213317871Run time for corrcoef is 0.002004384994506836\n",
      "\n",
      "Run time for corrcoef is 0.0017735958099365234Run time for shuffling is 0.013610124588012695\n",
      "\n",
      "Run time for corrcoef is 0.0014126300811767578\n",
      "Run time for shuffling is 0.014902114868164062\n",
      "Run time for corrcoef is 0.0013332366943359375\n",
      "Run time for shuffling is 0.019504070281982422\n",
      "Run time for corrcoef is 0.0017633438110351562\n",
      "Run time for shuffling is 0.019100427627563477\n",
      "Run time for shuffling is 0.0118408203125\n",
      "Run time for corrcoef is 0.0018208026885986328\n",
      "Run time for corrcoef is 0.0012295246124267578\n",
      "Run time for shuffling is 0.017617225646972656\n",
      "Run time for corrcoef is 0.001855611801147461\n",
      "Run time for shuffling is 0.019553661346435547\n",
      "Run time for corrcoef is 0.002063751220703125\n",
      "Run time for shuffling is 0.019322633743286133\n",
      "Run time for corrcoef is 0.002004861831665039\n",
      "Run time for shuffling is 0.0194704532623291\n",
      "Run time for corrcoef is 0.0019867420196533203\n",
      "Run time for shuffling is 0.018727779388427734\n",
      "Run time for corrcoef is 0.0019249916076660156\n",
      "Run time for shuffling is 0.01960444450378418\n",
      "Run time for shuffling is 0.020166873931884766\n",
      "Run time for corrcoef is 0.0019850730895996094\n",
      "Run time for corrcoef is 0.0018002986907958984\n",
      "Run time for shuffling is 0.019479990005493164\n",
      "Run time for corrcoef is 0.0013527870178222656\n",
      "Run time for shuffling is 0.020319700241088867\n",
      "Run time for corrcoef is 0.0019683837890625\n",
      "Run time for shuffling is 0.020481586456298828\n",
      "Run time for shuffling is 0.019504070281982422\n",
      "Run time for corrcoef is 0.0019326210021972656\n",
      "Run time for corrcoef is 0.001943349838256836\n",
      "Run time for shuffling is 0.01940631866455078\n",
      "Run time for corrcoef is 0.0019059181213378906Run time for shuffling is 0.018551111221313477\n",
      "\n",
      "Run time for corrcoef is 0.0021152496337890625\n",
      "Run time for shuffling is 0.019242048263549805\n",
      "Run time for corrcoef is 0.0018470287322998047\n",
      "Run time for shuffling is 0.01417684555053711\n",
      "Run time for corrcoef is 0.0012104511260986328Run time for shuffling is 0.019443750381469727\n",
      "Run time for corrcoef is 0.0017948150634765625\n",
      "Run time for shuffling is 0.019345760345458984\n",
      "\n",
      "Run time for corrcoef is 0.0017020702362060547\n",
      "Run time for shuffling is 0.018735170364379883\n",
      "Run time for corrcoef is 0.0017044544219970703\n",
      "Run time for shuffling is 0.016594886779785156\n",
      "Run time for corrcoef is 0.0016269683837890625\n",
      "Run time for shuffling is 0.012121200561523438\n",
      "Run time for corrcoef is 0.0012946128845214844\n",
      "Run time for shuffling is 0.01843404769897461\n",
      "Run time for corrcoef is 0.0019218921661376953\n",
      "Run time for shuffling is 0.013464212417602539\n",
      "Run time for corrcoef is 0.0019953250885009766\n",
      "Run time for shuffling is 0.01844477653503418\n",
      "Run time for corrcoef is 0.0018677711486816406\n",
      "Run time for shuffling is 0.013980865478515625\n",
      "Run time for corrcoef is 0.0019643306732177734\n",
      "Run time for shuffling is 0.019727706909179688\n",
      "Run time for corrcoef is 0.0016908645629882812\n",
      "Run time for shuffling is 0.017438173294067383\n",
      "Run time for corrcoef is 0.001226186752319336\n",
      "Run time for shuffling is 0.01634073257446289\n",
      "Run time for corrcoef is 0.0017142295837402344\n",
      "Run time for shuffling is 0.018558740615844727\n",
      "Run time for shuffling is 0.015171527862548828\n",
      "Run time for corrcoef is 0.001886606216430664\n",
      "Run time for corrcoef is 0.001825571060180664\n",
      "Run time for shuffling is 0.01856398582458496Run time for shuffling is 0.013372659683227539\n",
      "\n",
      "Run time for corrcoef is 0.001795053482055664Run time for corrcoef is 0.0018596649169921875\n",
      "\n",
      "Run time for shuffling is 0.01350712776184082\n",
      "Run time for corrcoef is 0.0014700889587402344\n",
      "Run time for shuffling is 0.012779951095581055\n",
      "Run time for corrcoef is 0.0012750625610351562\n",
      "Run time for shuffling is 0.018630027770996094\n",
      "Run time for corrcoef is 0.0019130706787109375\n",
      "Run time for shuffling is 0.01306295394897461\n",
      "Run time for shuffling is 0.018656015396118164\n",
      "Run time for corrcoef is 0.001744985580444336Run time for shuffling is 0.012226581573486328\n",
      "Run time for corrcoef is 0.0012638568878173828\n",
      "Run time for corrcoef is 0.0014774799346923828\n",
      "\n",
      "Run time for shuffling is 0.01305532455444336Run time for shuffling is 0.018525123596191406\n",
      "\n",
      "Run time for corrcoef is 0.0012738704681396484Run time for corrcoef is 0.00176239013671875\n",
      "\n",
      "Run time for shuffling is 0.013006925582885742\n",
      "Run time for corrcoef is 0.0012710094451904297\n",
      "Run time for shuffling is 0.012593507766723633\n",
      "Run time for corrcoef is 0.0017580986022949219\n",
      "Run time for shuffling is 0.012543916702270508\n",
      "Run time for corrcoef is 0.001199483871459961\n",
      "Run time for shuffling is 0.019441604614257812\n",
      "Run time for corrcoef is 0.001813650131225586"
     ]
    }
   ],
   "source": [
    "matA = branchAwindows[1]\n",
    "matB = branchBwindows[1]\n",
    "get_p_value_2(matA, matB, 'Rest', 'Zcchc6', 10000, 0.12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "852fb4c9",
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "%%timeit\n",
    "get_p_value_2(matA, matB, 'Rest', 'Zcchc6', 1000, 0.12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "19cb0423",
   "metadata": {},
   "outputs": [],
   "source": [
    "%%timeit\n",
    "get_p_value_2(matA, matB, 'Rest', 'Zcchc6', 10000, 0.12)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "buried-sodium",
   "metadata": {},
   "source": [
    "more timing test below (s. TIMING)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "617d91c1",
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in range(2,len(branchAwindows)):\n",
    "    print(f'startet window {i}')\n",
    "    matrixA = branchAwindows[i]\n",
    "    matrixB = branchBwindows[i]\n",
    "\n",
    "    p_vals = []\n",
    "\n",
    "    dataframe = pd.read_csv(f'data_files/correlation_abs/w{i}_SCGRN_correlation.csv')\n",
    "    out_df = dataframe\n",
    "\n",
    "    processes = []\n",
    "    # q = Queue()\n",
    "    for index, row in dataframe.iterrows():\n",
    "        process = multiprocessing.Process(target=get_p_value_q, args = [matrixA, matrixB, row['TF'], row['GENE'], 5, row['cc_coef_abs']])\n",
    "        processes.append(process)\n",
    "        process.start()\n",
    "        \n",
    "    for process in processes:\n",
    "        process.join()\n",
    "    \n",
    "    out_df['p-value'] = p_vals\n",
    "    # out_df.to_csv(f'data_files/correlation_abs/w{i}_SCGRN_correlation_perm_10.csv', index=False)\n",
    "    print(f'finished window {i}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "412db201",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "inclusive-elevation",
   "metadata": {},
   "source": [
    "#### Using NaNCorrMp.calculate() from https://github.com/bukson/nancorrmp"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "significant-absence",
   "metadata": {},
   "source": [
    "Testing NaNCorrMp.calculate() first version: dataframe for each tf and regulated genes as input"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "optimum-reach",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_dataframes(window_df):\n",
    "    dataframes = []\n",
    "    tfs = set(window_df['TF'])\n",
    "    for tf in tfs:\n",
    "        df = window_df[window_df['TF'] == tf]\n",
    "        dataframes.append(df)\n",
    "    return dataframes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "corrected-neighbor",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_anndatas(anndata_window, dfs):\n",
    "    anndatas = {}\n",
    "    for df in dfs:\n",
    "        genes = list(df['GENE'])\n",
    "        tf = list(df['TF'])[0]\n",
    "        genes.insert(0, tf)\n",
    "        adata = anndata_window[:,list(genes)]\n",
    "        anndatas[tf] = adata\n",
    "    return anndatas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "republican-summer",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_correlations(anndatasA, anndatasB):\n",
    "    correlation_abs = {}\n",
    "    \n",
    "    tfs = set(anndatasA.keys()).union(set(anndatasB.keys()))\n",
    "    for tf in tfs:\n",
    "        # assuming tf appears in both branches\n",
    "        if tf in anndatasA.keys() and tf in anndatasB.keys():\n",
    "            start = time.time()\n",
    "            cA = NaNCorrMp.calculate(anndatasA[tf].to_df(), n_jobs=20)\n",
    "            cB = NaNCorrMp.calculate(anndatasB[tf].to_df(), n_jobs=20)\n",
    "            end = time.time()\n",
    "            # print(f'time for computing correlation for both matrices: {end-start}')\n",
    "            \n",
    "            corrA = cA.iloc[0]\n",
    "            corrB = cB.iloc[0]\n",
    "            \n",
    "            genes = set(corrA.index).union(set(corrB.index))\n",
    "            genes = list(genes)\n",
    "            \n",
    "            corrs = []\n",
    "            for gene in genes:\n",
    "                if not gene in corrA.index:\n",
    "                    ccA = 0.0\n",
    "                    ccB = corrB[gene]\n",
    "                elif not gene in corrB.index:\n",
    "                    ccB = 0.0\n",
    "                    ccA = corrB[gene]\n",
    "                else:\n",
    "                    ccA = corrA[gene]\n",
    "                    ccB = corrB[gene]\n",
    "                corrs.append(abs(ccA-ccB))     # get the absolute value between branchA and branchB correlation\n",
    "            \n",
    "            df = pd.DataFrame([corrs], columns = genes)\n",
    "            # print(f'genes: {genes}')\n",
    "            # print(f'dataframe with correlations: {df}')\n",
    "            correlation_abs[tf]=(df.iloc[0]) # get only first row -> only correlation between tf and gene (not gene-gene)\n",
    "                \n",
    "    return correlation_abs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "invalid-wesley",
   "metadata": {},
   "outputs": [],
   "source": [
    "a1 = branchAwindows[1]\n",
    "b1 = branchBwindows[1]\n",
    "df1 = pd.read_csv(f'data_files/correlation_abs/w1_SCGRN_correlation.csv') # test dataframe (window 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "better-questionnaire",
   "metadata": {},
   "outputs": [],
   "source": [
    "dfs = get_dataframes(df1)\n",
    "dfs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "narrative-current",
   "metadata": {},
   "outputs": [],
   "source": [
    "ada1 = get_anndatas(a1, dfs)\n",
    "adb1 = get_anndatas(b1, dfs)\n",
    "ada1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "upper-happening",
   "metadata": {},
   "outputs": [],
   "source": [
    "corr = get_correlations(ada1, adb1)\n",
    "print(corr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "parliamentary-boards",
   "metadata": {},
   "outputs": [],
   "source": [
    "print(corr['Trp53'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "piano-plaintiff",
   "metadata": {},
   "source": [
    "timing test for one run:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "continent-copying",
   "metadata": {},
   "outputs": [],
   "source": [
    "df1 = pd.read_csv(f'data_files/correlation_abs/w1_SCGRN_correlation.csv') # test dataframe (window 1)\n",
    "a1 = branchAwindows[1].copy()\n",
    "b1 = branchBwindows[1].copy()\n",
    "\n",
    "dfs = get_dataframes(df1)      # slice original dataframe into dataframes for each tf\n",
    "\n",
    "start = time.time()\n",
    "\n",
    "shuffle2d(a1.X)\n",
    "shuffle2d(b1.X)\n",
    "print('shuffling done')\n",
    "\n",
    "ada1 = get_anndatas(a1, dfs)   # get anndata objects with only tf and its genes for each tf\n",
    "adb1 = get_anndatas(b1, dfs)\n",
    "\n",
    "print('started computing correlation')\n",
    "corr = get_correlations(ada1, adb1)   # get list of dfs with absolute correlation values for each tf\n",
    "print('correlation computed')\n",
    "end = time.time()\n",
    "\n",
    "print(f'runtime for one iteration/ permutation: {end-start}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "loose-decision",
   "metadata": {},
   "source": [
    "Testing NaNCorrMp.calculate() second version: whole dataframe as input"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dimensional-privilege",
   "metadata": {},
   "outputs": [],
   "source": [
    "def correlation(inputlist):\n",
    "    df_a = inputlist[0]\n",
    "    n_jobs1 = inputlist[1]\n",
    "    cc = NaNCorrMp.calculate(df_a, n_jobs = n_jobs1)\n",
    "    return cc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "sunset-artwork",
   "metadata": {},
   "outputs": [],
   "source": [
    "input_list = [[a1.to_df(), 25],[b1.to_df(), 25]]\n",
    "res = []\n",
    "\n",
    "start = time.perf_counter()\n",
    "with concurrent.futures.ProcessPoolExecutor() as executor:\n",
    "    results = executor.map(correlation, input_list)\n",
    "    for r in results:\n",
    "        res.append(r)\n",
    "end = time.perf_counter()\n",
    "print(f'Finished in {round(end - start, 2)} s')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "intermediate-crest",
   "metadata": {},
   "outputs": [],
   "source": [
    "df1 = pd.read_csv(f'data_files/correlation_abs/w1_SCGRN_correlation.csv') # test dataframe (window 1)\n",
    "a1 = branchAwindows[1].copy()\n",
    "b1 = branchBwindows[1].copy()\n",
    "\n",
    "counts_larger_than_cc = []\n",
    "for j in range(len(df1)):            # fill list with 0 for each tf-gene pair\n",
    "    counts_larger_than_cc.append(0)\n",
    "\n",
    "# start one permutation\n",
    "# - shuffle matrices\n",
    "# - compute correlation in matrix\n",
    "# - compare true and random values\n",
    "# start = time.time()\n",
    "\n",
    "def function1(num):\n",
    "    # print('start one permutation')\n",
    "    start1 = time.time()\n",
    "    for i in range(num):\n",
    "        start_sh = time.time()\n",
    "        shuffle2d(a1.X)\n",
    "        shuffle2d(b1.X)\n",
    "        end_sh = time.time()\n",
    "        # print(f'shuffling done in {end_sh - start_sh}')\n",
    "\n",
    "        start_cc = time.time()\n",
    "\n",
    "\n",
    "        input_list = [[a1.to_df(), 25],[b1.to_df(), 25]]\n",
    "        res = []\n",
    "\n",
    "        start_cc = time.perf_counter()\n",
    "        with concurrent.futures.ProcessPoolExecutor() as executor:\n",
    "            results = executor.map(correlation, input_list)\n",
    "            for r in results:\n",
    "                res.append(r)\n",
    "        end_cc = time.perf_counter()\n",
    "        # print(f'Finished in {round(end_cc - start_cc, 2)} s')\n",
    "\n",
    "        cA = res[0]\n",
    "        cB = res[1]\n",
    "\n",
    "        #cA = NaNCorrMp.calculate(a1.to_df(), n_jobs=30)\n",
    "        end_cc = time.time()\n",
    "        #cB = NaNCorrMp.calculate(b1.to_df(), n_jobs=20)\n",
    "        # end_cc = time.time()\n",
    "        # print(f'computing correlation done in {end_cc-start_cc}')\n",
    "\n",
    "        for index, row in df1.iterrows():\n",
    "            tf = row['TF']\n",
    "            gene = row['GENE']\n",
    "            true_cc = row['cc_coef_abs']\n",
    "            wA = cA.at[tf, gene]\n",
    "            wB = cB.at[tf, gene]\n",
    "\n",
    "            if abs(wA-wB) > true_cc:            # compare random correlation to true correlation and count +1 if larger\n",
    "                counts_larger_than_cc[index] += 1   \n",
    "\n",
    "        end = time.time()\n",
    "        # print(f'time for one permutation: {end-start}')\n",
    "\n",
    "    end1 = time.time()\n",
    "    print(f'time for {num} permutation: {end1-start1}')\n",
    "    return counts_larger_than_cc\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "civilian-authorization",
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'function1' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-2-c128703f372e>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mcount_list\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfunction1\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m5\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'function1' is not defined"
     ]
    }
   ],
   "source": [
    "count_list = function1(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "common-monkey",
   "metadata": {},
   "outputs": [],
   "source": [
    "start2 = time.perf_counter()\n",
    "res_lst = []\n",
    "with concurrent.futures.ProcessPoolExecutor() as executor:\n",
    "    results = executor.map(function1, [5, 5, 5, 5])    # number of processes !!! (40)\n",
    "    for r in results:\n",
    "        res_lst.append(r)\n",
    "end2 = time.perf_counter()\n",
    "print(f'Finished in {round(end2 - start2, 2)} s')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "finished-honduras",
   "metadata": {},
   "outputs": [],
   "source": [
    "count_list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "unknown-indicator",
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'cA' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-3-7ccc20c1c643>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mcA\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'cA' is not defined"
     ]
    }
   ],
   "source": [
    "cA"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "adjusted-location",
   "metadata": {},
   "source": [
    "Better but still too time consuming!"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fifth-doctrine",
   "metadata": {},
   "source": [
    "Some more timing tests:\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "manual-green",
   "metadata": {},
   "outputs": [],
   "source": [
    "a1 = branchAwindows[1].copy()\n",
    "a1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ongoing-sunday",
   "metadata": {},
   "outputs": [],
   "source": [
    "# test\n",
    "test = branchBwindows[1].copy()\n",
    "Xb = test.obs_vector('Nrf1')\n",
    "Yb = test.obs_vector('Psmb1')\n",
    "w2 = np.corrcoef(Xb, Yb)[0, 1]\n",
    "w2\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "binding-makeup",
   "metadata": {},
   "outputs": [],
   "source": [
    "type(Xb)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "healthy-strike",
   "metadata": {},
   "outputs": [],
   "source": [
    "Xb.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "coupled-bikini",
   "metadata": {},
   "outputs": [],
   "source": [
    "random_dataframe = pd.DataFrame(np.random.rand(100, 4))\n",
    "# corr = NaNCorrMp.calculate(random_dataframe)\n",
    "random_dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "disabled-possibility",
   "metadata": {},
   "outputs": [],
   "source": [
    "corr = NaNCorrMp.calculate(random_dataframe)\n",
    "corr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "intellectual-margin",
   "metadata": {},
   "outputs": [],
   "source": [
    "w2 = np.corrcoef(random_dataframe[0], random_dataframe[1])\n",
    "w2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "encouraging-thong",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "injured-tamil",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "transparent-field",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "quarterly-biology",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "resident-puppy",
   "metadata": {},
   "source": [
    "#### TIMING\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "awful-retrieval",
   "metadata": {},
   "outputs": [],
   "source": [
    "window1_A = branchAwindows[1].copy()\n",
    "\n",
    "start = time.time()\n",
    "shuffle2d(window1_A.X)\n",
    "end = time.time()\n",
    "\n",
    "print(f'runtime for shuffling one matrix: {end-start} seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "funded-leadership",
   "metadata": {},
   "outputs": [],
   "source": [
    "window1_A = branchAwindows[1].copy()\n",
    "\n",
    "start = time.time()\n",
    "shuffle2d(window1_A.X)\n",
    "end = time.time()\n",
    "\n",
    "print(f'runtime for shuffling one matrix: {end-start} seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "minus-immigration",
   "metadata": {},
   "outputs": [],
   "source": [
    "window1_A = branchAwindows[1].copy()\n",
    "window1_B = branchBwindows[1].copy()\n",
    "dataframe = pd.read_csv(f'data_files/correlation_abs/test1.csv') # test tf-gene pair (Ezh2 <-> 1810037I17Rik)\n",
    "\n",
    "start = time.time()\n",
    "get_counts(dataframe, window1_A, window1_B, 100)\n",
    "end = time.time()\n",
    "\n",
    "print(f'runtime for 100 permutation for one tf-gene pair: {end-start} seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "still-mambo",
   "metadata": {},
   "outputs": [],
   "source": [
    "window1_A = branchAwindows[1].copy()\n",
    "window1_B = branchBwindows[1].copy()\n",
    "dataframe = pd.read_csv(f'data_files/correlation_abs/test1.csv') # test tf-gene pair (Ezh2 <-> 1810037I17Rik)\n",
    "\n",
    "start = time.time()\n",
    "get_counts(dataframe, window1_A, window1_B, 1000)\n",
    "end = time.time()\n",
    "\n",
    "print(f'runtime for 1000 permutation for one tf-gene pair: {end-start} seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "written-samba",
   "metadata": {},
   "outputs": [],
   "source": [
    "window1_A = branchAwindows[1].copy()\n",
    "window1_B = branchBwindows[1].copy()\n",
    "dataframe = pd.read_csv(f'data_files/correlation_abs/test1.csv') # test tf-gene pair (Ezh2 <-> 1810037I17Rik)\n",
    "\n",
    "start = time.time()\n",
    "get_counts(dataframe, window1_A, window1_B, 10000)\n",
    "end = time.time()\n",
    "\n",
    "print(f'runtime for 10000 permutation for one tf-gene pair: {end-start} seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "jewish-float",
   "metadata": {},
   "outputs": [],
   "source": [
    "window1_A = branchAwindows[1].copy()\n",
    "window1_B = branchBwindows[1].copy()\n",
    "dataframe = pd.read_csv(f'data_files/correlation_abs/test10.csv') # test for first 10 tf-gene pairs in window 1\n",
    "\n",
    "start = time.time()\n",
    "get_counts(dataframe, window1_A, window1_B, 10000)\n",
    "end = time.time()\n",
    "\n",
    "print(f'runtime for 10000 permutation for 10 tf-gene pairs: {end-start} seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "detected-dominican",
   "metadata": {},
   "outputs": [],
   "source": [
    "window1_A = branchAwindows[1].copy()\n",
    "window1_B = branchBwindows[1].copy()\n",
    "dataframe = pd.read_csv(f'data_files/correlation_abs/test100.csv') # test for first 100 tf-gene pairs in window 1\n",
    "\n",
    "start = time.time()\n",
    "get_counts(dataframe, window1_A, window1_B, 10000)\n",
    "end = time.time()\n",
    "\n",
    "print(f'runtime for 10000 permutation for 100 tf-gene pairs: {end-start} seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "respected-wonder",
   "metadata": {},
   "outputs": [],
   "source": [
    "window1_A = branchAwindows[1].copy()\n",
    "window1_B = branchBwindows[1].copy()\n",
    "dataframe = pd.read_csv(f'data_files/correlation_abs/test100.csv') # test for first 100 tf-gene pairs in window 1\n",
    "\n",
    "start = time.time()\n",
    "get_counts(dataframe, window1_A, window1_B, 1)\n",
    "end = time.time()\n",
    "\n",
    "print(f'runtime one permutation for 100 tf-gene pairs: {end-start} seconds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "polyphonic-meeting",
   "metadata": {},
   "outputs": [],
   "source": [
    "window1_A = branchAwindows[1].copy()\n",
    "window1_B = branchBwindows[1].copy()\n",
    "dataframe = pd.read_csv(f'data_files/correlation_abs/w1_SCGRN_correlation.csv') # test with tf-gene pairs from window 1\n",
    "\n",
    "start = time.time()\n",
    "get_counts(dataframe, window1_A, window1_B, 1)\n",
    "end = time.time()\n",
    "\n",
    "print(f'runtime one permutation for ~11000 tf-gene pairs: {end-start} seconds')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "twelve-majority",
   "metadata": {},
   "source": [
    "testing on smaller dataframes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "constitutional-sense",
   "metadata": {},
   "outputs": [],
   "source": [
    "# test 1\n",
    "dataframe = pd.read_csv(f'data_files/correlation_abs/test100.csv') # test dataframe (first 100 of window 1)\n",
    "test100_counts = get_counts(dataframe, branchAwindows[1].copy(), branchBwindows[1].copy(), 1000)\n",
    "print(test100_counts)\n",
    "p_values = get_p_values(test100_counts, 1000)\n",
    "print(p_values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "sunrise-edmonton",
   "metadata": {},
   "outputs": [],
   "source": [
    "# test 2\n",
    "dataframe = pd.read_csv(f'data_files/correlation_abs/test200_300.csv') # test dataframe (entries 200-299 of window 1)\n",
    "test200_300_counts = get_counts(dataframe, branchAwindows[1].copy(), branchBwindows[1].copy(), 1000)\n",
    "p_values = get_p_values(test200_300_counts, 1000)\n",
    "print(p_values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bright-exemption",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "smart-candy",
   "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": "3.6.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}