3.analysis/3.t_tests.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "sys.path.insert(0, \"../utils/\")\n",
    "from data_loader import load_data\n",
    "from scipy.stats import ttest_ind \n",
    "from scipy.stats import f_oneway\n",
    "import pathlib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "latent_dir = pathlib.Path(\"../2.train-VAE/results/latent_df.parquet\").resolve()\n",
    "latent_df = pd.read_parquet(latent_dir)\n",
    "metadata_dir = pathlib.Path(\".././0.data-download/data/metadata_df.parquet\").resolve()\n",
    "metadata_df = pd.read_parquet(metadata_dir)\n",
    "data_dir = pathlib.Path(\"../0.data-download/data/\").resolve()\n",
    "model_df, dependency_df = load_data(data_dir, adult_or_pediatric=\"all\")\n",
    "gsea_results_dir = pathlib.Path(\"../3.analysis/results/combined_gsea_results.parquet.gz\").resolve()\n",
    "gsea_results_df = pd.read_parquet(gsea_results_dir)\n",
    "significant_gsea_dir = pathlib.Path(\"../3.analysis/results/significant_gsea_results.parquet.gz\")\n",
    "significant_gsea_df = pd.read_parquet(significant_gsea_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Creating categorized lists of sample IDs used in BVAE training\n",
    "# note that 10 of the 912 used samples have Unknown Sex\n",
    "\n",
    "ped_ids = metadata_df.query(\"AgeCategory == 'Pediatric'\").ModelID.tolist()\n",
    "adult_ids = metadata_df.query(\"AgeCategory == 'Adult'\").ModelID.tolist()\n",
    "male_ids = metadata_df.query(\"Sex == 'Male'\").ModelID.tolist()\n",
    "female_ids = metadata_df.query(\"Sex == 'Female'\").ModelID.tolist()\n",
    "ped_male_ids = metadata_df.query(\"AgeCategory == 'Pediatric'\").query(\"Sex == 'Male'\").ModelID.tolist()\n",
    "adult_male_ids = metadata_df.query(\"AgeCategory == 'Adult'\").query(\"Sex == 'Male'\").ModelID.tolist()\n",
    "ped_female_ids = metadata_df.query(\"AgeCategory == 'Pediatric'\").query(\"Sex == 'Female'\").ModelID.tolist()\n",
    "adult_female_ids = metadata_df.query(\"AgeCategory == 'Adult'\").query(\"Sex == 'Female'\").ModelID.tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Generating latent dataframes for each category and dropping the id column to prep for t tests\n",
    "def generate_latent_df(latent_df, category_ids):\n",
    "    \"\"\"\n",
    "    Generate a latent DataFrame filtered by category IDs and drop the 'ModelID' column.\n",
    "\n",
    "    Parameters:\n",
    "    latent_df (pd.DataFrame): The original DataFrame containing latent variables.\n",
    "    category_ids (list): The list of ModelIDs to filter the DataFrame by.\n",
    "\n",
    "    Returns:\n",
    "    pd.DataFrame: The filtered DataFrame with the 'ModelID' column dropped.\n",
    "    \"\"\"\n",
    "    filtered_df = latent_df[latent_df['ModelID'].isin(category_ids)].copy()\n",
    "    filtered_df_float = filtered_df.drop(columns=[\"ModelID\"])\n",
    "    filtered_df_float.reset_index(drop=True, inplace=True)\n",
    "    return filtered_df, filtered_df_float\n",
    "\n",
    "\n",
    "# Usage for each category\n",
    "adult_latent_df, adult_latent_df_float = generate_latent_df(latent_df, adult_ids)\n",
    "ped_latent_df, ped_latent_df_float = generate_latent_df(latent_df, ped_ids)\n",
    "male_latent_df, male_latent_df_float = generate_latent_df(latent_df, male_ids)\n",
    "female_latent_df, female_latent_df_float = generate_latent_df(latent_df, female_ids)\n",
    "ped_male_latent_df, ped_male_latent_df_float = generate_latent_df(latent_df, ped_male_ids)\n",
    "adult_male_latent_df, adult_male_latent_df_float = generate_latent_df(latent_df, adult_male_ids)\n",
    "ped_female_latent_df, ped_female_latent_df_float = generate_latent_df(latent_df, ped_female_ids)\n",
    "adult_female_latent_df, adult_female_latent_df_float = generate_latent_df(latent_df, adult_female_ids)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(33, 4)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>t_stat</th>\n",
       "      <th>p_value</th>\n",
       "      <th>comparison</th>\n",
       "      <th>latent_feature</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-3.218127</td>\n",
       "      <td>1.333718e-03</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-5.107409</td>\n",
       "      <td>3.941735e-07</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2.121312</td>\n",
       "      <td>3.415234e-02</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-2.344316</td>\n",
       "      <td>1.926592e-02</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1.278132</td>\n",
       "      <td>2.015132e-01</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.418699</td>\n",
       "      <td>6.755301e-01</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.418699</td>\n",
       "      <td>6.755301e-01</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1.033402</td>\n",
       "      <td>3.016773e-01</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.418699</td>\n",
       "      <td>6.755302e-01</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.418699</td>\n",
       "      <td>6.755301e-01</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>-4.261201</td>\n",
       "      <td>2.234542e-05</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.418699</td>\n",
       "      <td>6.755301e-01</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>-4.041902</td>\n",
       "      <td>5.728084e-05</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>-9.486921</td>\n",
       "      <td>1.827865e-20</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>-6.114014</td>\n",
       "      <td>1.413787e-09</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>-1.441360</td>\n",
       "      <td>1.498104e-01</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>-4.173298</td>\n",
       "      <td>3.275839e-05</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>-2.666039</td>\n",
       "      <td>7.804609e-03</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>-8.084050</td>\n",
       "      <td>1.884506e-15</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>-6.600415</td>\n",
       "      <td>6.785679e-11</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.418699</td>\n",
       "      <td>6.755301e-01</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>22</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>-4.472322</td>\n",
       "      <td>8.666170e-06</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>-3.965499</td>\n",
       "      <td>7.870679e-05</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>-6.820863</td>\n",
       "      <td>1.603158e-11</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>-3.776063</td>\n",
       "      <td>1.691588e-04</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>-2.461114</td>\n",
       "      <td>1.402637e-02</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>-4.796875</td>\n",
       "      <td>1.868623e-06</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>0.782335</td>\n",
       "      <td>4.342114e-01</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>31</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>-4.838553</td>\n",
       "      <td>1.524005e-06</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>-6.065182</td>\n",
       "      <td>1.896237e-09</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>-6.116637</td>\n",
       "      <td>1.391584e-09</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>34</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>1.226049</td>\n",
       "      <td>2.204820e-01</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>35</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>-5.741979</td>\n",
       "      <td>1.256397e-08</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>36</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      t_stat       p_value          comparison  latent_feature\n",
       "0  -3.218127  1.333718e-03  Adult vs Pediatric               1\n",
       "2  -5.107409  3.941735e-07  Adult vs Pediatric               3\n",
       "3   2.121312  3.415234e-02  Adult vs Pediatric               4\n",
       "4  -2.344316  1.926592e-02  Adult vs Pediatric               5\n",
       "5   1.278132  2.015132e-01  Adult vs Pediatric               6\n",
       "6   0.418699  6.755301e-01  Adult vs Pediatric               7\n",
       "7   0.418699  6.755301e-01  Adult vs Pediatric               8\n",
       "8   1.033402  3.016773e-01  Adult vs Pediatric               9\n",
       "9   0.418699  6.755302e-01  Adult vs Pediatric              10\n",
       "10  0.418699  6.755301e-01  Adult vs Pediatric              11\n",
       "11 -4.261201  2.234542e-05  Adult vs Pediatric              12\n",
       "12  0.418699  6.755301e-01  Adult vs Pediatric              13\n",
       "13 -4.041902  5.728084e-05  Adult vs Pediatric              14\n",
       "14 -9.486921  1.827865e-20  Adult vs Pediatric              15\n",
       "15 -6.114014  1.413787e-09  Adult vs Pediatric              16\n",
       "16 -1.441360  1.498104e-01  Adult vs Pediatric              17\n",
       "17 -4.173298  3.275839e-05  Adult vs Pediatric              18\n",
       "18 -2.666039  7.804609e-03  Adult vs Pediatric              19\n",
       "19 -8.084050  1.884506e-15  Adult vs Pediatric              20\n",
       "20 -6.600415  6.785679e-11  Adult vs Pediatric              21\n",
       "21  0.418699  6.755301e-01  Adult vs Pediatric              22\n",
       "22 -4.472322  8.666170e-06  Adult vs Pediatric              23\n",
       "23 -3.965499  7.870679e-05  Adult vs Pediatric              24\n",
       "24 -6.820863  1.603158e-11  Adult vs Pediatric              25\n",
       "26 -3.776063  1.691588e-04  Adult vs Pediatric              27\n",
       "27 -2.461114  1.402637e-02  Adult vs Pediatric              28\n",
       "29 -4.796875  1.868623e-06  Adult vs Pediatric              30\n",
       "30  0.782335  4.342114e-01  Adult vs Pediatric              31\n",
       "31 -4.838553  1.524005e-06  Adult vs Pediatric              32\n",
       "32 -6.065182  1.896237e-09  Adult vs Pediatric              33\n",
       "33 -6.116637  1.391584e-09  Adult vs Pediatric              34\n",
       "34  1.226049  2.204820e-01  Adult vs Pediatric              35\n",
       "35 -5.741979  1.256397e-08  Adult vs Pediatric              36"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# t tests comparing adult vs ped for each latent dimension\n",
    "\n",
    "t_test_adult_vs_ped = ttest_ind(adult_latent_df_float, ped_latent_df_float)\n",
    "t_test_adult_vs_ped = pd.DataFrame(t_test_adult_vs_ped).T\n",
    "t_test_adult_vs_ped.columns = [\"t_stat\", \"p_value\"]\n",
    "t_test_adult_vs_ped['comparison'] = 'Adult vs Pediatric'\n",
    "t_test_adult_vs_ped['latent_feature'] = t_test_adult_vs_ped.index + 1\n",
    "# Remove rows with NaN values\n",
    "t_test_adult_vs_ped = t_test_adult_vs_ped.dropna()\n",
    "print(t_test_adult_vs_ped.shape)\n",
    "\n",
    "t_test_adult_vs_ped.head(50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(33, 4)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>t_stat</th>\n",
       "      <th>p_value</th>\n",
       "      <th>comparison</th>\n",
       "      <th>latent_feature</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.512374</td>\n",
       "      <td>0.130773</td>\n",
       "      <td>Male vs Female</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.730208</td>\n",
       "      <td>0.083919</td>\n",
       "      <td>Male vs Female</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.090873</td>\n",
       "      <td>0.927613</td>\n",
       "      <td>Male vs Female</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.416794</td>\n",
       "      <td>0.156872</td>\n",
       "      <td>Male vs Female</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.349296</td>\n",
       "      <td>0.726945</td>\n",
       "      <td>Male vs Female</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     t_stat   p_value      comparison  latent_feature\n",
       "0  1.512374  0.130773  Male vs Female               1\n",
       "2  1.730208  0.083919  Male vs Female               3\n",
       "3  0.090873  0.927613  Male vs Female               4\n",
       "4  1.416794  0.156872  Male vs Female               5\n",
       "5  0.349296  0.726945  Male vs Female               6"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# t tests comparing male vs female for each latent dimension\n",
    "\n",
    "t_test_male_vs_female = ttest_ind(male_latent_df_float, female_latent_df_float)\n",
    "t_test_male_vs_female = pd.DataFrame(t_test_male_vs_female).T\n",
    "t_test_male_vs_female.columns = [\"t_stat\", \"p_value\"]\n",
    "t_test_male_vs_female['comparison'] = 'Male vs Female'\n",
    "t_test_male_vs_female['latent_feature'] = t_test_male_vs_female.index + 1\n",
    "# Remove rows with NaN values\n",
    "t_test_male_vs_female = t_test_male_vs_female.dropna()\n",
    "print(t_test_male_vs_female.shape)\n",
    "\n",
    "t_test_male_vs_female.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(31, 4)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>t_stat</th>\n",
       "      <th>p_value</th>\n",
       "      <th>comparison</th>\n",
       "      <th>latent_feature</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.801326</td>\n",
       "      <td>0.072211</td>\n",
       "      <td>Adult Male vs Pediatric Male</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3.084566</td>\n",
       "      <td>0.002143</td>\n",
       "      <td>Adult Male vs Pediatric Male</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-1.600661</td>\n",
       "      <td>0.110039</td>\n",
       "      <td>Adult Male vs Pediatric Male</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.170902</td>\n",
       "      <td>0.242156</td>\n",
       "      <td>Adult Male vs Pediatric Male</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>-0.595032</td>\n",
       "      <td>0.552072</td>\n",
       "      <td>Adult Male vs Pediatric Male</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     t_stat   p_value                    comparison  latent_feature\n",
       "0  1.801326  0.072211  Adult Male vs Pediatric Male               1\n",
       "2  3.084566  0.002143  Adult Male vs Pediatric Male               3\n",
       "3 -1.600661  0.110039  Adult Male vs Pediatric Male               4\n",
       "4  1.170902  0.242156  Adult Male vs Pediatric Male               5\n",
       "5 -0.595032  0.552072  Adult Male vs Pediatric Male               6"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# t tests comparing adult male vs ped male for each latent dimension\n",
    "\n",
    "t_test_adult_male_vs_ped_male = ttest_ind(ped_male_latent_df_float, adult_male_latent_df_float)\n",
    "t_test_adult_male_vs_ped_male = pd.DataFrame(t_test_adult_male_vs_ped_male).T\n",
    "t_test_adult_male_vs_ped_male.columns = [\"t_stat\", \"p_value\"]\n",
    "t_test_adult_male_vs_ped_male['comparison'] = 'Adult Male vs Pediatric Male'\n",
    "t_test_adult_male_vs_ped_male['latent_feature'] = t_test_adult_male_vs_ped_male.index + 1\n",
    "# Remove rows with NaN values\n",
    "t_test_adult_male_vs_ped_male = t_test_adult_male_vs_ped_male.dropna()\n",
    "print(t_test_adult_male_vs_ped_male.shape)\n",
    "\n",
    "t_test_adult_male_vs_ped_male.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(29, 4)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>t_stat</th>\n",
       "      <th>p_value</th>\n",
       "      <th>comparison</th>\n",
       "      <th>latent_feature</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2.826528</td>\n",
       "      <td>0.004938</td>\n",
       "      <td>Adult Female vs Pediatric Female</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4.123023</td>\n",
       "      <td>0.000045</td>\n",
       "      <td>Adult Female vs Pediatric Female</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-1.147304</td>\n",
       "      <td>0.251932</td>\n",
       "      <td>Adult Female vs Pediatric Female</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2.252685</td>\n",
       "      <td>0.024812</td>\n",
       "      <td>Adult Female vs Pediatric Female</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>-1.150781</td>\n",
       "      <td>0.250500</td>\n",
       "      <td>Adult Female vs Pediatric Female</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     t_stat   p_value                        comparison  latent_feature\n",
       "0  2.826528  0.004938  Adult Female vs Pediatric Female               1\n",
       "2  4.123023  0.000045  Adult Female vs Pediatric Female               3\n",
       "3 -1.147304  0.251932  Adult Female vs Pediatric Female               4\n",
       "4  2.252685  0.024812  Adult Female vs Pediatric Female               5\n",
       "5 -1.150781  0.250500  Adult Female vs Pediatric Female               6"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# t tests comparing adult female vs ped female for each latent dimension\n",
    "\n",
    "t_test_adult_female_vs_ped_female = ttest_ind(ped_female_latent_df_float, adult_female_latent_df_float)\n",
    "t_test_adult_female_vs_ped_female = pd.DataFrame(t_test_adult_female_vs_ped_female).T\n",
    "t_test_adult_female_vs_ped_female.columns = [\"t_stat\", \"p_value\"]\n",
    "t_test_adult_female_vs_ped_female['comparison'] = 'Adult Female vs Pediatric Female'\n",
    "t_test_adult_female_vs_ped_female['latent_feature'] = t_test_adult_female_vs_ped_female.index + 1\n",
    "# Remove rows with NaN values\n",
    "t_test_adult_female_vs_ped_female = t_test_adult_female_vs_ped_female.dropna()\n",
    "print(t_test_adult_female_vs_ped_female.shape)\n",
    "\n",
    "t_test_adult_female_vs_ped_female.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(25, 4)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>t_stat</th>\n",
       "      <th>p_value</th>\n",
       "      <th>comparison</th>\n",
       "      <th>latent_feature</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.173129</td>\n",
       "      <td>0.862813</td>\n",
       "      <td>Pediatric Male vs Pediatric Female</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.478101</td>\n",
       "      <td>0.633363</td>\n",
       "      <td>Pediatric Male vs Pediatric Female</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.175723</td>\n",
       "      <td>0.860779</td>\n",
       "      <td>Pediatric Male vs Pediatric Female</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.373064</td>\n",
       "      <td>0.709695</td>\n",
       "      <td>Pediatric Male vs Pediatric Female</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>-0.889634</td>\n",
       "      <td>0.375269</td>\n",
       "      <td>Pediatric Male vs Pediatric Female</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     t_stat   p_value                          comparison  latent_feature\n",
       "0  0.173129  0.862813  Pediatric Male vs Pediatric Female               1\n",
       "2  0.478101  0.633363  Pediatric Male vs Pediatric Female               3\n",
       "3  0.175723  0.860779  Pediatric Male vs Pediatric Female               4\n",
       "4  0.373064  0.709695  Pediatric Male vs Pediatric Female               5\n",
       "5 -0.889634  0.375269  Pediatric Male vs Pediatric Female               6"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# t tests comparing ped male vs ped female for each latent dimension\n",
    "\n",
    "t_test_ped_male_vs_ped_female = ttest_ind(ped_female_latent_df_float, ped_male_latent_df_float)\n",
    "t_test_ped_male_vs_ped_female = pd.DataFrame(t_test_ped_male_vs_ped_female).T\n",
    "t_test_ped_male_vs_ped_female.columns = [\"t_stat\", \"p_value\"]\n",
    "t_test_ped_male_vs_ped_female['comparison'] = 'Pediatric Male vs Pediatric Female'\n",
    "t_test_ped_male_vs_ped_female['latent_feature'] = t_test_ped_male_vs_ped_female.index + 1\n",
    "# Remove rows with NaN values\n",
    "t_test_ped_male_vs_ped_female = t_test_ped_male_vs_ped_female.dropna()\n",
    "print(t_test_ped_male_vs_ped_female.shape)\n",
    "\n",
    "t_test_ped_male_vs_ped_female.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(33, 4)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>t_stat</th>\n",
       "      <th>p_value</th>\n",
       "      <th>comparison</th>\n",
       "      <th>latent_feature</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-1.734332</td>\n",
       "      <td>0.083239</td>\n",
       "      <td>Adult Male vs Adult Female</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-2.009238</td>\n",
       "      <td>0.044843</td>\n",
       "      <td>Adult Male vs Adult Female</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-0.176256</td>\n",
       "      <td>0.860137</td>\n",
       "      <td>Adult Male vs Adult Female</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-1.646554</td>\n",
       "      <td>0.100037</td>\n",
       "      <td>Adult Male vs Adult Female</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>-0.133334</td>\n",
       "      <td>0.893963</td>\n",
       "      <td>Adult Male vs Adult Female</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     t_stat   p_value                  comparison  latent_feature\n",
       "0 -1.734332  0.083239  Adult Male vs Adult Female               1\n",
       "2 -2.009238  0.044843  Adult Male vs Adult Female               3\n",
       "3 -0.176256  0.860137  Adult Male vs Adult Female               4\n",
       "4 -1.646554  0.100037  Adult Male vs Adult Female               5\n",
       "5 -0.133334  0.893963  Adult Male vs Adult Female               6"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# t tests comparing adult male vs adult female for each latent dimension\n",
    "\n",
    "t_test_adult_male_vs_adult_female = ttest_ind(adult_female_latent_df_float, adult_male_latent_df_float)\n",
    "t_test_adult_male_vs_adult_female = pd.DataFrame(t_test_adult_male_vs_adult_female).T\n",
    "t_test_adult_male_vs_adult_female.columns = [\"t_stat\", \"p_value\"]\n",
    "t_test_adult_male_vs_adult_female['comparison'] = 'Adult Male vs Adult Female'\n",
    "t_test_adult_male_vs_adult_female['latent_feature'] = t_test_adult_male_vs_adult_female.index + 1\n",
    "# Remove rows with NaN values\n",
    "t_test_adult_male_vs_adult_female = t_test_adult_male_vs_adult_female.dropna()\n",
    "print(t_test_adult_male_vs_adult_female.shape)\n",
    "\n",
    "t_test_adult_male_vs_adult_female.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "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>t_stat</th>\n",
       "      <th>p_value</th>\n",
       "      <th>comparison</th>\n",
       "      <th>latent_feature</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>-9.486921</td>\n",
       "      <td>1.827865e-20</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>-8.084050</td>\n",
       "      <td>1.884506e-15</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>106</th>\n",
       "      <td>7.336605</td>\n",
       "      <td>1.201889e-12</td>\n",
       "      <td>Adult Female vs Pediatric Female</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>-6.820863</td>\n",
       "      <td>1.603158e-11</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>-6.600415</td>\n",
       "      <td>6.785679e-11</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>140</th>\n",
       "      <td>0.122658</td>\n",
       "      <td>9.025627e-01</td>\n",
       "      <td>Pediatric Male vs Pediatric Female</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>0.090873</td>\n",
       "      <td>9.276128e-01</td>\n",
       "      <td>Male vs Female</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>64</th>\n",
       "      <td>0.084749</td>\n",
       "      <td>9.324789e-01</td>\n",
       "      <td>Male vs Female</td>\n",
       "      <td>35</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>138</th>\n",
       "      <td>0.080982</td>\n",
       "      <td>9.355782e-01</td>\n",
       "      <td>Pediatric Male vs Pediatric Female</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>182</th>\n",
       "      <td>-0.005472</td>\n",
       "      <td>9.956357e-01</td>\n",
       "      <td>Adult Male vs Adult Female</td>\n",
       "      <td>35</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>184 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       t_stat       p_value                          comparison  \\\n",
       "13  -9.486921  1.827865e-20                  Adult vs Pediatric   \n",
       "18  -8.084050  1.884506e-15                  Adult vs Pediatric   \n",
       "106  7.336605  1.201889e-12    Adult Female vs Pediatric Female   \n",
       "23  -6.820863  1.603158e-11                  Adult vs Pediatric   \n",
       "19  -6.600415  6.785679e-11                  Adult vs Pediatric   \n",
       "..        ...           ...                                 ...   \n",
       "140  0.122658  9.025627e-01  Pediatric Male vs Pediatric Female   \n",
       "35   0.090873  9.276128e-01                      Male vs Female   \n",
       "64   0.084749  9.324789e-01                      Male vs Female   \n",
       "138  0.080982  9.355782e-01  Pediatric Male vs Pediatric Female   \n",
       "182 -0.005472  9.956357e-01          Adult Male vs Adult Female   \n",
       "\n",
       "     latent_feature  \n",
       "13               15  \n",
       "18               20  \n",
       "106              15  \n",
       "23               25  \n",
       "19               21  \n",
       "..              ...  \n",
       "140              23  \n",
       "35                4  \n",
       "64               35  \n",
       "138              20  \n",
       "182              35  \n",
       "\n",
       "[184 rows x 4 columns]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Combining and saving t test results\n",
    "t_test_results_df = pd.concat([\n",
    "    t_test_adult_vs_ped, \n",
    "    t_test_male_vs_female, \n",
    "    t_test_adult_male_vs_ped_male, \n",
    "    t_test_adult_female_vs_ped_female, \n",
    "    t_test_ped_male_vs_ped_female, \n",
    "    t_test_adult_male_vs_adult_female\n",
    "]).reset_index(drop=True)\n",
    "t_test_results_dir = pathlib.Path(\"./results/t_test_results.tsv\")\n",
    "t_test_results_df.to_parquet(t_test_results_dir)\n",
    "\n",
    "# sort to show most significant p-values\n",
    "t_test_results_df.sort_values(by='p_value', ascending = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/juliacurd/anaconda3/envs/gene_dependency_representations/lib/python3.12/site-packages/scipy/stats/_axis_nan_policy.py:605: ConstantInputWarning: Each of the input arrays is constant; the F statistic is not defined or infinite\n",
      "  res = hypotest_fun_out(*samples, axis=axis, **kwds)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>f_stat</th>\n",
       "      <th>p_value</th>\n",
       "      <th>latent_feature</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>4.304991</td>\n",
       "      <td>5.009211e-03</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>9.917823</td>\n",
       "      <td>1.927772e-06</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.305123</td>\n",
       "      <td>2.714218e-01</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2.819108</td>\n",
       "      <td>3.801336e-02</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.622503</td>\n",
       "      <td>6.005391e-01</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.351409</td>\n",
       "      <td>7.881409e-01</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.351410</td>\n",
       "      <td>7.881405e-01</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>2.375147</td>\n",
       "      <td>6.870333e-02</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.563613</td>\n",
       "      <td>6.391101e-01</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.351410</td>\n",
       "      <td>7.881407e-01</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>6.891740</td>\n",
       "      <td>1.360873e-04</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.351410</td>\n",
       "      <td>7.881405e-01</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>6.246126</td>\n",
       "      <td>3.363315e-04</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>30.093961</td>\n",
       "      <td>1.387354e-18</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>13.225361</td>\n",
       "      <td>1.835405e-08</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>1.201460</td>\n",
       "      <td>3.080948e-01</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>6.954200</td>\n",
       "      <td>1.246682e-04</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>3.520247</td>\n",
       "      <td>1.470903e-02</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>22.296928</td>\n",
       "      <td>5.926836e-14</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>16.280155</td>\n",
       "      <td>2.535182e-10</td>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>0.563614</td>\n",
       "      <td>6.391097e-01</td>\n",
       "      <td>22</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>7.554415</td>\n",
       "      <td>5.366294e-05</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>5.984939</td>\n",
       "      <td>4.846715e-04</td>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>17.138916</td>\n",
       "      <td>7.637740e-11</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>6.372055</td>\n",
       "      <td>2.819658e-04</td>\n",
       "      <td>27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>3.853585</td>\n",
       "      <td>9.322380e-03</td>\n",
       "      <td>28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>9.014319</td>\n",
       "      <td>6.880753e-06</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>1.816591</td>\n",
       "      <td>1.424569e-01</td>\n",
       "      <td>31</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>9.342065</td>\n",
       "      <td>4.337161e-06</td>\n",
       "      <td>32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>13.510820</td>\n",
       "      <td>1.229089e-08</td>\n",
       "      <td>33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>13.705157</td>\n",
       "      <td>9.355415e-09</td>\n",
       "      <td>34</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>0.456759</td>\n",
       "      <td>7.125801e-01</td>\n",
       "      <td>35</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>13.185820</td>\n",
       "      <td>1.940259e-08</td>\n",
       "      <td>36</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       f_stat       p_value  latent_feature\n",
       "0    4.304991  5.009211e-03               1\n",
       "2    9.917823  1.927772e-06               3\n",
       "3    1.305123  2.714218e-01               4\n",
       "4    2.819108  3.801336e-02               5\n",
       "5    0.622503  6.005391e-01               6\n",
       "6    0.351409  7.881409e-01               7\n",
       "7    0.351410  7.881405e-01               8\n",
       "8    2.375147  6.870333e-02               9\n",
       "9    0.563613  6.391101e-01              10\n",
       "10   0.351410  7.881407e-01              11\n",
       "11   6.891740  1.360873e-04              12\n",
       "12   0.351410  7.881405e-01              13\n",
       "13   6.246126  3.363315e-04              14\n",
       "14  30.093961  1.387354e-18              15\n",
       "15  13.225361  1.835405e-08              16\n",
       "16   1.201460  3.080948e-01              17\n",
       "17   6.954200  1.246682e-04              18\n",
       "18   3.520247  1.470903e-02              19\n",
       "19  22.296928  5.926836e-14              20\n",
       "20  16.280155  2.535182e-10              21\n",
       "21   0.563614  6.391097e-01              22\n",
       "22   7.554415  5.366294e-05              23\n",
       "23   5.984939  4.846715e-04              24\n",
       "24  17.138916  7.637740e-11              25\n",
       "26   6.372055  2.819658e-04              27\n",
       "27   3.853585  9.322380e-03              28\n",
       "29   9.014319  6.880753e-06              30\n",
       "30   1.816591  1.424569e-01              31\n",
       "31   9.342065  4.337161e-06              32\n",
       "32  13.510820  1.229089e-08              33\n",
       "33  13.705157  9.355415e-09              34\n",
       "34   0.456759  7.125801e-01              35\n",
       "35  13.185820  1.940259e-08              36"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# ANOVA Testing\n",
    "f_statistic, p_value = f_oneway(adult_male_latent_df_float, ped_male_latent_df_float, adult_female_latent_df_float, ped_female_latent_df_float)\n",
    "anova_df = pd.DataFrame({'f_stat': f_statistic.tolist(), 'p_value': p_value.tolist()})\n",
    "anova_df['latent_feature'] = anova_df.index + 1\n",
    "# Remove rows with NaN values\n",
    "anova_df = anova_df.dropna()\n",
    "anova_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Lung cancer in adults vs neuroblastoma in peds comparison\n",
    "\n",
    "# Extract Neuroblastoma model IDs from the model dataframe\n",
    "NB_ids = model_df.query(\"OncotreePrimaryDisease == 'Neuroblastoma'\").ModelID.tolist()\n",
    "\n",
    "# Create a copy of the pediatric latent space dataframe\n",
    "ped_NB_latent_df = ped_latent_df.copy()\n",
    "\n",
    "# Filter the pediatric latent dataframe to include only Neuroblastoma models\n",
    "for index, row in ped_NB_latent_df.iterrows():\n",
    "    if row['ModelID'] not in NB_ids:\n",
    "        ped_NB_latent_df.drop(index, inplace=True)\n",
    "\n",
    "# Drop the 'ModelID' column from the filtered pediatric latent dataframe\n",
    "# and reset the index\n",
    "ped_NB_latent_float_df = ped_NB_latent_df.drop(columns=[\"ModelID\"])\n",
    "ped_NB_latent_float_df.reset_index(drop=True, inplace=True)\n",
    "\n",
    "# Extract Non-Small Cell Lung Cancer (NSCLC) model IDs from the model dataframe\n",
    "LC_ids = model_df.query(\"OncotreePrimaryDisease == 'Non-Small Cell Lung Cancer'\").ModelID.tolist()\n",
    "\n",
    "# Create a copy of the adult latent space dataframe\n",
    "adult_LC_latent_df = adult_latent_df.copy()\n",
    "\n",
    "# Filter the adult latent dataframe to include only NSCLC models\n",
    "for index, row in adult_LC_latent_df.iterrows():\n",
    "    if row['ModelID'] not in LC_ids:\n",
    "        adult_LC_latent_df.drop(index, inplace=True)\n",
    "\n",
    "# Drop the 'ModelID' column from the filtered adult latent dataframe\n",
    "# and reset the index\n",
    "adult_LC_latent_float_df = adult_LC_latent_df.drop(columns=[\"ModelID\"])\n",
    "adult_LC_latent_float_df.reset_index(drop=True, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(24, 4)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>t_stat</th>\n",
       "      <th>p_value</th>\n",
       "      <th>comparison</th>\n",
       "      <th>latent_feature</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>-9.156894</td>\n",
       "      <td>2.724268e-15</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>-7.072663</td>\n",
       "      <td>1.348224e-10</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>-5.599585</td>\n",
       "      <td>1.528553e-07</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>-5.462841</td>\n",
       "      <td>2.822401e-07</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>-5.404335</td>\n",
       "      <td>3.660317e-07</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>-4.977234</td>\n",
       "      <td>2.330309e-06</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>-4.970053</td>\n",
       "      <td>2.402213e-06</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>-4.749113</td>\n",
       "      <td>6.042632e-06</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>34</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>-3.962528</td>\n",
       "      <td>1.302539e-04</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-3.958313</td>\n",
       "      <td>1.322873e-04</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>-3.919951</td>\n",
       "      <td>1.522374e-04</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>-2.818412</td>\n",
       "      <td>5.699367e-03</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>-2.462109</td>\n",
       "      <td>1.532239e-02</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>-2.345417</td>\n",
       "      <td>2.075011e-02</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>-2.318796</td>\n",
       "      <td>2.220421e-02</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>-1.900919</td>\n",
       "      <td>5.986003e-02</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>-1.856479</td>\n",
       "      <td>6.599007e-02</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>-1.681487</td>\n",
       "      <td>9.543023e-02</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-1.602900</td>\n",
       "      <td>1.117485e-01</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>-0.749436</td>\n",
       "      <td>4.551517e-01</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>-0.744265</td>\n",
       "      <td>4.582611e-01</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.721099</td>\n",
       "      <td>4.723382e-01</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>0.659798</td>\n",
       "      <td>5.107262e-01</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>35</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.422667</td>\n",
       "      <td>6.733412e-01</td>\n",
       "      <td>Lung Cancer vs Neuroblastoma</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      t_stat       p_value                    comparison  latent_feature\n",
       "14 -9.156894  2.724268e-15  Lung Cancer vs Neuroblastoma              15\n",
       "19 -7.072663  1.348224e-10  Lung Cancer vs Neuroblastoma              20\n",
       "24 -5.599585  1.528553e-07  Lung Cancer vs Neuroblastoma              25\n",
       "15 -5.462841  2.822401e-07  Lung Cancer vs Neuroblastoma              16\n",
       "29 -5.404335  3.660317e-07  Lung Cancer vs Neuroblastoma              30\n",
       "32 -4.977234  2.330309e-06  Lung Cancer vs Neuroblastoma              33\n",
       "20 -4.970053  2.402213e-06  Lung Cancer vs Neuroblastoma              21\n",
       "33 -4.749113  6.042632e-06  Lung Cancer vs Neuroblastoma              34\n",
       "31 -3.962528  1.302539e-04  Lung Cancer vs Neuroblastoma              32\n",
       "2  -3.958313  1.322873e-04  Lung Cancer vs Neuroblastoma               3\n",
       "35 -3.919951  1.522374e-04  Lung Cancer vs Neuroblastoma              36\n",
       "22 -2.818412  5.699367e-03  Lung Cancer vs Neuroblastoma              23\n",
       "17 -2.462109  1.532239e-02  Lung Cancer vs Neuroblastoma              18\n",
       "23 -2.345417  2.075011e-02  Lung Cancer vs Neuroblastoma              24\n",
       "27 -2.318796  2.220421e-02  Lung Cancer vs Neuroblastoma              28\n",
       "11 -1.900919  5.986003e-02  Lung Cancer vs Neuroblastoma              12\n",
       "13 -1.856479  6.599007e-02  Lung Cancer vs Neuroblastoma              14\n",
       "26 -1.681487  9.543023e-02  Lung Cancer vs Neuroblastoma              27\n",
       "4  -1.602900  1.117485e-01  Lung Cancer vs Neuroblastoma               5\n",
       "18 -0.749436  4.551517e-01  Lung Cancer vs Neuroblastoma              19\n",
       "16 -0.744265  4.582611e-01  Lung Cancer vs Neuroblastoma              17\n",
       "0  -0.721099  4.723382e-01  Lung Cancer vs Neuroblastoma               1\n",
       "34  0.659798  5.107262e-01  Lung Cancer vs Neuroblastoma              35\n",
       "3   0.422667  6.733412e-01  Lung Cancer vs Neuroblastoma               4"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# t tests comparing Lung Cancer in adult vs Neuroblastoma in ped for each latent dimension\n",
    "\n",
    "t_test_diff_adult_vs_ped = ttest_ind(adult_LC_latent_float_df, ped_NB_latent_float_df)\n",
    "t_test_diff_adult_vs_ped = pd.DataFrame(t_test_diff_adult_vs_ped).T\n",
    "t_test_diff_adult_vs_ped.columns = [\"t_stat\", \"p_value\"]\n",
    "t_test_diff_adult_vs_ped['comparison'] = 'Lung Cancer vs Neuroblastoma'\n",
    "t_test_diff_adult_vs_ped['latent_feature'] = t_test_diff_adult_vs_ped.index + 1\n",
    "# Remove rows with NaN values\n",
    "t_test_diff_adult_vs_ped = t_test_diff_adult_vs_ped.dropna()\n",
    "print(t_test_diff_adult_vs_ped.shape)\n",
    "\n",
    "t_test_diff_adult_vs_ped.sort_values(by = 'p_value', ascending= True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'Acute Myeloid Leukemia', 'Ewing Sarcoma', 'Mature B-Cell Neoplasms'}"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Obtaining shared cancer types in ped and adult\n",
    "adult_types = model_df.query(\"AgeCategory == 'Adult'\").OncotreePrimaryDisease.tolist()\n",
    "adult_types = [x for x in adult_types if adult_types.count(x) >= 5]\n",
    "adult_types = list(set(adult_types))\n",
    "\n",
    "ped_types = model_df.query(\"AgeCategory == 'Pediatric'\").OncotreePrimaryDisease.tolist()\n",
    "ped_types = [x for x in ped_types if ped_types.count(x) >= 5]\n",
    "ped_types = list(set(ped_types))\n",
    "\n",
    "all_types = set(adult_types) | set(ped_types)\n",
    "\n",
    "shared_types = set(adult_types) & set(ped_types)\n",
    "shared_types"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "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>t_stat</th>\n",
       "      <th>p_value</th>\n",
       "      <th>comparison</th>\n",
       "      <th>cancer_type</th>\n",
       "      <th>latent_feature</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>101</th>\n",
       "      <td>1.939462</td>\n",
       "      <td>0.064287</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>Acute Myeloid Leukemia</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>1.933352</td>\n",
       "      <td>0.065073</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>Acute Myeloid Leukemia</td>\n",
       "      <td>28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>103</th>\n",
       "      <td>1.477448</td>\n",
       "      <td>0.152559</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>Acute Myeloid Leukemia</td>\n",
       "      <td>32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>1.338527</td>\n",
       "      <td>0.195037</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>Ewing Sarcoma</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>74</th>\n",
       "      <td>1.330467</td>\n",
       "      <td>0.195869</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>Acute Myeloid Leukemia</td>\n",
       "      <td>3</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>62</th>\n",
       "      <td>-0.113085</td>\n",
       "      <td>0.910434</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>Mature B-Cell Neoplasms</td>\n",
       "      <td>27</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>107</th>\n",
       "      <td>0.106392</td>\n",
       "      <td>0.916156</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>Acute Myeloid Leukemia</td>\n",
       "      <td>36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52</th>\n",
       "      <td>-0.027256</td>\n",
       "      <td>0.978369</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>Mature B-Cell Neoplasms</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-0.008869</td>\n",
       "      <td>0.993007</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>Ewing Sarcoma</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>-0.004827</td>\n",
       "      <td>0.996195</td>\n",
       "      <td>Adult vs Pediatric</td>\n",
       "      <td>Ewing Sarcoma</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>68 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       t_stat   p_value          comparison              cancer_type  \\\n",
       "101  1.939462  0.064287  Adult vs Pediatric   Acute Myeloid Leukemia   \n",
       "99   1.933352  0.065073  Adult vs Pediatric   Acute Myeloid Leukemia   \n",
       "103  1.477448  0.152559  Adult vs Pediatric   Acute Myeloid Leukemia   \n",
       "15   1.338527  0.195037  Adult vs Pediatric            Ewing Sarcoma   \n",
       "74   1.330467  0.195869  Adult vs Pediatric   Acute Myeloid Leukemia   \n",
       "..        ...       ...                 ...                      ...   \n",
       "62  -0.113085  0.910434  Adult vs Pediatric  Mature B-Cell Neoplasms   \n",
       "107  0.106392  0.916156  Adult vs Pediatric   Acute Myeloid Leukemia   \n",
       "52  -0.027256  0.978369  Adult vs Pediatric  Mature B-Cell Neoplasms   \n",
       "4   -0.008869  0.993007  Adult vs Pediatric            Ewing Sarcoma   \n",
       "19  -0.004827  0.996195  Adult vs Pediatric            Ewing Sarcoma   \n",
       "\n",
       "     latent_feature  \n",
       "101              30  \n",
       "99               28  \n",
       "103              32  \n",
       "15               16  \n",
       "74                3  \n",
       "..              ...  \n",
       "62               27  \n",
       "107              36  \n",
       "52               17  \n",
       "4                 5  \n",
       "19               20  \n",
       "\n",
       "[68 rows x 5 columns]"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Comparing the shared cancer types\n",
    "# Initialize a list to store the t-test results for each shared cancer type\n",
    "comp_dfs = []\n",
    "\n",
    "# Iterate over each cancer type shared between adult and pediatric datasets\n",
    "for cancer_type in shared_types:\n",
    "    # Extract the model IDs for the current cancer type from the model dataframe\n",
    "    type_ids = model_df.query(\"OncotreePrimaryDisease == \" + \"'\" + cancer_type + \"'\").ModelID.tolist()\n",
    "\n",
    "    # Filter the pediatric latent dataframe to include only models of the current cancer type\n",
    "    ped_type_latent_df = ped_latent_df.copy()\n",
    "    ped_type_latent_df = ped_latent_df[ped_latent_df['ModelID'].isin(type_ids)].copy()\n",
    "\n",
    "    # Drop the 'ModelID' column and reset the index in the pediatric latent dataframe\n",
    "    ped_type_latent_float_df = ped_type_latent_df.drop(columns=[\"ModelID\"])\n",
    "    ped_type_latent_float_df.reset_index(drop=True, inplace=True)\n",
    "\n",
    "    # Filter the adult latent dataframe to include only models of the current cancer type\n",
    "    adult_type_latent_df = adult_latent_df.copy()\n",
    "    adult_type_latent_df = adult_latent_df[adult_latent_df['ModelID'].isin(type_ids)].copy()\n",
    "\n",
    "    # Drop the 'ModelID' column and reset the index in the adult latent dataframe\n",
    "    adult_type_latent_float_df = adult_type_latent_df.drop(columns=[\"ModelID\"])\n",
    "    adult_type_latent_float_df.reset_index(drop=True, inplace=True)\n",
    "\n",
    "    # Perform a t-test comparing the latent features between adult and pediatric datasets\n",
    "    t_test_type_adult_vs_ped = ttest_ind(adult_type_latent_float_df, ped_type_latent_float_df)\n",
    "\n",
    "    # Convert the t-test results to a DataFrame and add additional information\n",
    "    t_test_type_adult_vs_ped = pd.DataFrame(t_test_type_adult_vs_ped).T\n",
    "    t_test_type_adult_vs_ped.columns = [\"t_stat\", \"p_value\"]\n",
    "    t_test_type_adult_vs_ped['comparison'] = 'Adult vs Pediatric'\n",
    "    t_test_type_adult_vs_ped['cancer_type'] = cancer_type\n",
    "    t_test_type_adult_vs_ped['latent_feature'] = t_test_type_adult_vs_ped.index + 1\n",
    "\n",
    "    # Append the t-test results for the current cancer type to the list\n",
    "    comp_dfs.append(t_test_type_adult_vs_ped)\n",
    "\n",
    "# Concatenate all the t-test results into a single DataFrame\n",
    "t_test_type_results_df = pd.concat(comp_dfs).reset_index(drop=True)\n",
    "\n",
    "# Remove any rows with NaN values from the results\n",
    "t_test_type_results_df = t_test_type_results_df.dropna()\n",
    "\n",
    "# Sort the t-test results by the p-value in ascending order\n",
    "t_test_type_results_df.sort_values(by='p_value', ascending=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "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>Pathway</th>\n",
       "      <th>F-statistic</th>\n",
       "      <th>p-value</th>\n",
       "      <th>Higher in</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>222</th>\n",
       "      <td>Metabolism R-HSA-1430728</td>\n",
       "      <td>14.444604</td>\n",
       "      <td>0.000482</td>\n",
       "      <td>Adult</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>Chaperonin-mediated Protein Folding R-HSA-390466</td>\n",
       "      <td>9.192164</td>\n",
       "      <td>0.004251</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Metabolism Of Nucleotides R-HSA-15869</td>\n",
       "      <td>9.002977</td>\n",
       "      <td>0.004624</td>\n",
       "      <td>Adult</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>723</th>\n",
       "      <td>Mitochondrial Translation Termination R-HSA-54...</td>\n",
       "      <td>8.652187</td>\n",
       "      <td>0.005410</td>\n",
       "      <td>Adult</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>451</th>\n",
       "      <td>Gluconeogenesis R-HSA-70263</td>\n",
       "      <td>8.275562</td>\n",
       "      <td>0.006415</td>\n",
       "      <td>Adult</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>247</th>\n",
       "      <td>Post-translational Protein Phosphorylation R-H...</td>\n",
       "      <td>8.155295</td>\n",
       "      <td>0.006776</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>248</th>\n",
       "      <td>Regulation Of IGF Transport And Uptake By IGFB...</td>\n",
       "      <td>8.155295</td>\n",
       "      <td>0.006776</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49</th>\n",
       "      <td>Metabolism Of Porphyrins R-HSA-189445</td>\n",
       "      <td>8.024915</td>\n",
       "      <td>0.007193</td>\n",
       "      <td>Adult</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>Heme Biosynthesis R-HSA-189451</td>\n",
       "      <td>8.024914</td>\n",
       "      <td>0.007193</td>\n",
       "      <td>Adult</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>147</th>\n",
       "      <td>Cooperation Of PDCL (PhLP1) And TRiC/CCT In G-...</td>\n",
       "      <td>7.778841</td>\n",
       "      <td>0.008055</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>773</th>\n",
       "      <td>Mitochondrial Translation R-HSA-5368287</td>\n",
       "      <td>7.778092</td>\n",
       "      <td>0.008058</td>\n",
       "      <td>Adult</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>233</th>\n",
       "      <td>PTK6 Regulates RHO GTPases, RAS GTPase And MAP...</td>\n",
       "      <td>7.708078</td>\n",
       "      <td>0.008323</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>851</th>\n",
       "      <td>Cristae Formation R-HSA-8949613</td>\n",
       "      <td>7.590092</td>\n",
       "      <td>0.008791</td>\n",
       "      <td>Adult</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>987</th>\n",
       "      <td>Establishment Of Sister Chromatid Cohesion R-H...</td>\n",
       "      <td>7.426942</td>\n",
       "      <td>0.009486</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>113</th>\n",
       "      <td>Nuclear Envelope Breakdown R-HSA-2980766</td>\n",
       "      <td>7.391553</td>\n",
       "      <td>0.009644</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>198</th>\n",
       "      <td>Depolymerisation Of Nuclear Lamina R-HSA-4419969</td>\n",
       "      <td>7.391039</td>\n",
       "      <td>0.009646</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>250</th>\n",
       "      <td>Signaling By FGFR4 In Disease R-HSA-5655291</td>\n",
       "      <td>7.358090</td>\n",
       "      <td>0.009796</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>311</th>\n",
       "      <td>Signaling By FGFR3 In Disease R-HSA-5655332</td>\n",
       "      <td>7.353371</td>\n",
       "      <td>0.009818</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>470</th>\n",
       "      <td>Mitochondrial Translation Elongation R-HSA-538...</td>\n",
       "      <td>7.322628</td>\n",
       "      <td>0.009960</td>\n",
       "      <td>Adult</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>Protein Folding R-HSA-391251</td>\n",
       "      <td>7.275299</td>\n",
       "      <td>0.010184</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>420</th>\n",
       "      <td>Sulfur Amino Acid Metabolism R-HSA-1614635</td>\n",
       "      <td>7.182403</td>\n",
       "      <td>0.010638</td>\n",
       "      <td>Adult</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70</th>\n",
       "      <td>Mitotic Prophase R-HSA-68875</td>\n",
       "      <td>6.785066</td>\n",
       "      <td>0.012843</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>437</th>\n",
       "      <td>Signaling By FGFR2 In Disease R-HSA-5655253</td>\n",
       "      <td>6.771385</td>\n",
       "      <td>0.012928</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>293</th>\n",
       "      <td>Mitochondrial Biogenesis R-HSA-1592230</td>\n",
       "      <td>6.750306</td>\n",
       "      <td>0.013058</td>\n",
       "      <td>Adult</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72</th>\n",
       "      <td>RAC1 GTPase Cycle R-HSA-9013149</td>\n",
       "      <td>6.723484</td>\n",
       "      <td>0.013227</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>294</th>\n",
       "      <td>P75 NTR Receptor-Mediated Signaling R-HSA-193704</td>\n",
       "      <td>6.677718</td>\n",
       "      <td>0.013520</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>Cooperation Of Prefoldin And TriC/CCT In Actin...</td>\n",
       "      <td>6.566420</td>\n",
       "      <td>0.014262</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>261</th>\n",
       "      <td>PI-3K cascade:FGFR1 R-HSA-5654689</td>\n",
       "      <td>6.538412</td>\n",
       "      <td>0.014455</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>419</th>\n",
       "      <td>NRAGE Signals Death Thru JNK R-HSA-193648</td>\n",
       "      <td>6.456695</td>\n",
       "      <td>0.015036</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56</th>\n",
       "      <td>NCAM Signaling For Neurite Out-Growth R-HSA-37...</td>\n",
       "      <td>6.215993</td>\n",
       "      <td>0.016898</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52</th>\n",
       "      <td>Citric Acid Cycle (TCA Cycle) R-HSA-71403</td>\n",
       "      <td>6.212264</td>\n",
       "      <td>0.016928</td>\n",
       "      <td>Adult</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>815</th>\n",
       "      <td>TP53 Regulates Transcription Of Additional Cel...</td>\n",
       "      <td>6.136642</td>\n",
       "      <td>0.017565</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>380</th>\n",
       "      <td>Post-translational Modification: Synthesis Of ...</td>\n",
       "      <td>6.135750</td>\n",
       "      <td>0.017572</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>243</th>\n",
       "      <td>VLDLR Internalisation And Degradation R-HSA-88...</td>\n",
       "      <td>6.086234</td>\n",
       "      <td>0.018003</td>\n",
       "      <td>Adult</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>Purine Ribonucleoside Monophosphate Biosynthes...</td>\n",
       "      <td>6.081600</td>\n",
       "      <td>0.018044</td>\n",
       "      <td>Adult</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58</th>\n",
       "      <td>Platelet Aggregation (Plug Formation) R-HSA-76009</td>\n",
       "      <td>6.063546</td>\n",
       "      <td>0.018204</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57</th>\n",
       "      <td>Integrin Signaling R-HSA-354192</td>\n",
       "      <td>6.063546</td>\n",
       "      <td>0.018204</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>215</th>\n",
       "      <td>G Alpha (Q) Signaling Events R-HSA-416476</td>\n",
       "      <td>5.983603</td>\n",
       "      <td>0.018932</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>Prefoldin Mediated Transfer Of Substrate To CC...</td>\n",
       "      <td>5.980471</td>\n",
       "      <td>0.018961</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>688</th>\n",
       "      <td>PI3K/AKT Signaling In Cancer R-HSA-2219528</td>\n",
       "      <td>5.950496</td>\n",
       "      <td>0.019243</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>926</th>\n",
       "      <td>TP53 Regulates Transcription Of Cell Death Gen...</td>\n",
       "      <td>5.942924</td>\n",
       "      <td>0.019315</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>614</th>\n",
       "      <td>Deadenylation Of mRNA R-HSA-429947</td>\n",
       "      <td>5.940704</td>\n",
       "      <td>0.019336</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62</th>\n",
       "      <td>Glyoxylate Metabolism And Glycine Degradation ...</td>\n",
       "      <td>5.929499</td>\n",
       "      <td>0.019443</td>\n",
       "      <td>Adult</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>Nucleotide Biosynthesis R-HSA-8956320</td>\n",
       "      <td>5.824697</td>\n",
       "      <td>0.020475</td>\n",
       "      <td>Adult</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>366</th>\n",
       "      <td>PI-3K cascade:FGFR3 R-HSA-5654710</td>\n",
       "      <td>5.820442</td>\n",
       "      <td>0.020518</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>367</th>\n",
       "      <td>PI-3K cascade:FGFR4 R-HSA-5654720</td>\n",
       "      <td>5.820442</td>\n",
       "      <td>0.020518</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>368</th>\n",
       "      <td>PI-3K cascade:FGFR2 R-HSA-5654695</td>\n",
       "      <td>5.820442</td>\n",
       "      <td>0.020518</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>330</th>\n",
       "      <td>Cell Death Signaling Via NRAGE, NRIF And NADE ...</td>\n",
       "      <td>5.817368</td>\n",
       "      <td>0.020549</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>234</th>\n",
       "      <td>Cargo Trafficking To Periciliary Membrane R-HS...</td>\n",
       "      <td>5.800607</td>\n",
       "      <td>0.020720</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>152</th>\n",
       "      <td>FRS-mediated FGFR1 Signaling R-HSA-5654693</td>\n",
       "      <td>5.762524</td>\n",
       "      <td>0.021114</td>\n",
       "      <td>Pediatric</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                               Pathway  F-statistic   p-value  \\\n",
       "222                           Metabolism R-HSA-1430728    14.444604  0.000482   \n",
       "31    Chaperonin-mediated Protein Folding R-HSA-390466     9.192164  0.004251   \n",
       "0                Metabolism Of Nucleotides R-HSA-15869     9.002977  0.004624   \n",
       "723  Mitochondrial Translation Termination R-HSA-54...     8.652187  0.005410   \n",
       "451                        Gluconeogenesis R-HSA-70263     8.275562  0.006415   \n",
       "247  Post-translational Protein Phosphorylation R-H...     8.155295  0.006776   \n",
       "248  Regulation Of IGF Transport And Uptake By IGFB...     8.155295  0.006776   \n",
       "49               Metabolism Of Porphyrins R-HSA-189445     8.024915  0.007193   \n",
       "50                      Heme Biosynthesis R-HSA-189451     8.024914  0.007193   \n",
       "147  Cooperation Of PDCL (PhLP1) And TRiC/CCT In G-...     7.778841  0.008055   \n",
       "773            Mitochondrial Translation R-HSA-5368287     7.778092  0.008058   \n",
       "233  PTK6 Regulates RHO GTPases, RAS GTPase And MAP...     7.708078  0.008323   \n",
       "851                    Cristae Formation R-HSA-8949613     7.590092  0.008791   \n",
       "987  Establishment Of Sister Chromatid Cohesion R-H...     7.426942  0.009486   \n",
       "113           Nuclear Envelope Breakdown R-HSA-2980766     7.391553  0.009644   \n",
       "198   Depolymerisation Of Nuclear Lamina R-HSA-4419969     7.391039  0.009646   \n",
       "250        Signaling By FGFR4 In Disease R-HSA-5655291     7.358090  0.009796   \n",
       "311        Signaling By FGFR3 In Disease R-HSA-5655332     7.353371  0.009818   \n",
       "470  Mitochondrial Translation Elongation R-HSA-538...     7.322628  0.009960   \n",
       "24                        Protein Folding R-HSA-391251     7.275299  0.010184   \n",
       "420         Sulfur Amino Acid Metabolism R-HSA-1614635     7.182403  0.010638   \n",
       "70                        Mitotic Prophase R-HSA-68875     6.785066  0.012843   \n",
       "437        Signaling By FGFR2 In Disease R-HSA-5655253     6.771385  0.012928   \n",
       "293             Mitochondrial Biogenesis R-HSA-1592230     6.750306  0.013058   \n",
       "72                     RAC1 GTPase Cycle R-HSA-9013149     6.723484  0.013227   \n",
       "294   P75 NTR Receptor-Mediated Signaling R-HSA-193704     6.677718  0.013520   \n",
       "36   Cooperation Of Prefoldin And TriC/CCT In Actin...     6.566420  0.014262   \n",
       "261                  PI-3K cascade:FGFR1 R-HSA-5654689     6.538412  0.014455   \n",
       "419          NRAGE Signals Death Thru JNK R-HSA-193648     6.456695  0.015036   \n",
       "56   NCAM Signaling For Neurite Out-Growth R-HSA-37...     6.215993  0.016898   \n",
       "52           Citric Acid Cycle (TCA Cycle) R-HSA-71403     6.212264  0.016928   \n",
       "815  TP53 Regulates Transcription Of Additional Cel...     6.136642  0.017565   \n",
       "380  Post-translational Modification: Synthesis Of ...     6.135750  0.017572   \n",
       "243  VLDLR Internalisation And Degradation R-HSA-88...     6.086234  0.018003   \n",
       "15   Purine Ribonucleoside Monophosphate Biosynthes...     6.081600  0.018044   \n",
       "58   Platelet Aggregation (Plug Formation) R-HSA-76009     6.063546  0.018204   \n",
       "57                     Integrin Signaling R-HSA-354192     6.063546  0.018204   \n",
       "215          G Alpha (Q) Signaling Events R-HSA-416476     5.983603  0.018932   \n",
       "44   Prefoldin Mediated Transfer Of Substrate To CC...     5.980471  0.018961   \n",
       "688         PI3K/AKT Signaling In Cancer R-HSA-2219528     5.950496  0.019243   \n",
       "926  TP53 Regulates Transcription Of Cell Death Gen...     5.942924  0.019315   \n",
       "614                 Deadenylation Of mRNA R-HSA-429947     5.940704  0.019336   \n",
       "62   Glyoxylate Metabolism And Glycine Degradation ...     5.929499  0.019443   \n",
       "11               Nucleotide Biosynthesis R-HSA-8956320     5.824697  0.020475   \n",
       "366                  PI-3K cascade:FGFR3 R-HSA-5654710     5.820442  0.020518   \n",
       "367                  PI-3K cascade:FGFR4 R-HSA-5654720     5.820442  0.020518   \n",
       "368                  PI-3K cascade:FGFR2 R-HSA-5654695     5.820442  0.020518   \n",
       "330  Cell Death Signaling Via NRAGE, NRIF And NADE ...     5.817368  0.020549   \n",
       "234  Cargo Trafficking To Periciliary Membrane R-HS...     5.800607  0.020720   \n",
       "152         FRS-mediated FGFR1 Signaling R-HSA-5654693     5.762524  0.021114   \n",
       "\n",
       "     Higher in  \n",
       "222      Adult  \n",
       "31   Pediatric  \n",
       "0        Adult  \n",
       "723      Adult  \n",
       "451      Adult  \n",
       "247  Pediatric  \n",
       "248  Pediatric  \n",
       "49       Adult  \n",
       "50       Adult  \n",
       "147  Pediatric  \n",
       "773      Adult  \n",
       "233  Pediatric  \n",
       "851      Adult  \n",
       "987  Pediatric  \n",
       "113  Pediatric  \n",
       "198  Pediatric  \n",
       "250  Pediatric  \n",
       "311  Pediatric  \n",
       "470      Adult  \n",
       "24   Pediatric  \n",
       "420      Adult  \n",
       "70   Pediatric  \n",
       "437  Pediatric  \n",
       "293      Adult  \n",
       "72   Pediatric  \n",
       "294  Pediatric  \n",
       "36   Pediatric  \n",
       "261  Pediatric  \n",
       "419  Pediatric  \n",
       "56   Pediatric  \n",
       "52       Adult  \n",
       "815  Pediatric  \n",
       "380  Pediatric  \n",
       "243      Adult  \n",
       "15       Adult  \n",
       "58   Pediatric  \n",
       "57   Pediatric  \n",
       "215  Pediatric  \n",
       "44   Pediatric  \n",
       "688  Pediatric  \n",
       "926  Pediatric  \n",
       "614  Pediatric  \n",
       "62       Adult  \n",
       "11       Adult  \n",
       "366  Pediatric  \n",
       "367  Pediatric  \n",
       "368  Pediatric  \n",
       "330  Pediatric  \n",
       "234  Pediatric  \n",
       "152  Pediatric  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Prepare a DataFrame to store ANOVA results for multiple pathways\n",
    "anova_results = []\n",
    "\n",
    "# Add \"z_\" prefix to the latent dimensions in the t-test DataFrame\n",
    "t_test_adult_vs_ped['z_dim'] = 'z_' + t_test_adult_vs_ped['latent_feature'].astype(str)\n",
    "t_test_adult_vs_ped['group'] = t_test_adult_vs_ped['t_stat'].apply(lambda x: 'Adult' if x > 0 else 'Pediatric')\n",
    "\n",
    "# Filter significant latent features\n",
    "significant_latent_features = t_test_adult_vs_ped[t_test_adult_vs_ped['p_value'] < 0.05]\n",
    "\n",
    "# Add the 'z_dim' column if not already added\n",
    "significant_latent_features.loc[:, 'z_dim'] = 'z_' + significant_latent_features['latent_feature'].astype(str)\n",
    "\n",
    "\n",
    "# Loop through all unique pathways (Terms) in the GSEA DataFrame\n",
    "for pathway in gsea_results_df['Term'].unique():\n",
    "    # Filter for the pathway\n",
    "    filtered_gsea_df = gsea_results_df[gsea_results_df['Term'] == pathway]\n",
    "    \n",
    "    # Further filter by significant latent features\n",
    "    filtered_gsea_df = filtered_gsea_df[filtered_gsea_df['z_dim'].isin(significant_latent_features['z_dim'])]\n",
    "    \n",
    "    # Merge GSEA DataFrame with t-test DataFrame to get group information\n",
    "    merged_df = pd.merge(filtered_gsea_df, significant_latent_features[['z_dim', 'group']], on='z_dim', how='inner')\n",
    "\n",
    "    # Group by 'group' and collect ES values\n",
    "    grouped_data = merged_df.groupby('group')['es'].apply(list)\n",
    "\n",
    "    # Ensure we have data for both groups\n",
    "    if len(grouped_data) == 2 and all(len(vals) > 1 for vals in grouped_data):\n",
    "        # Perform ANOVA\n",
    "        f_statistic, p_value = f_oneway(*grouped_data)\n",
    "\n",
    "        # Determine which group has higher enrichment score\n",
    "        adult_mean = np.mean(grouped_data['Adult'])\n",
    "        pediatric_mean = np.mean(grouped_data['Pediatric'])\n",
    "        higher_group = 'Adult' if adult_mean > pediatric_mean else 'Pediatric'\n",
    "\n",
    "        # Store results\n",
    "        anova_results.append({\n",
    "            'Pathway': pathway,\n",
    "            'F-statistic': f_statistic,\n",
    "            'p-value': p_value,\n",
    "            'Higher in': higher_group\n",
    "        })\n",
    "\n",
    "# Convert the results to a DataFrame\n",
    "anova_results_df = pd.DataFrame(anova_results)\n",
    "\n",
    "# Apply a significance threshold (e.g., p < 0.05)\n",
    "significant_anova_results_df = anova_results_df[anova_results_df['p-value'] < 0.05]\n",
    "\n",
    "anova_dir = pathlib.Path(\"./results/anova_results.csv\")\n",
    "significant_anova_results_df.to_csv(anova_dir)\n",
    "\n",
    "# Display significant pathways\n",
    "significant_anova_results_df.sort_values(by='F-statistic', key=abs, ascending = False).head(50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_3697599/2057681926.py:9: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  significant_anova_results_df.rename(columns={'Pathway': 'Term'}, inplace=True)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "# Define cut-offs\n",
    "lfc_cutoff = 0.584\n",
    "fdr_cutoff = 0.25\n",
    "\n",
    "# Merge GSEA results with ANOVA results based on the 'Pathway' column\n",
    "significant_anova_results_df.rename(columns={'Pathway': 'Term'}, inplace=True)\n",
    "merged_df = significant_gsea_df.merge(significant_anova_results_df[['Term', 'Higher in', 'F-statistic']], on='Term', how='left')\n",
    "\n",
    "# Define masks for adult and pediatric pathways\n",
    "adult_mask = merged_df['Higher in'] == 'Adult'\n",
    "pediatric_mask = merged_df['Higher in'] == 'Pediatric'\n",
    "\n",
    "# Sort by the absolute value of NES and select the top 5 adult and top 5 pediatric pathways\n",
    "top_adult = merged_df[adult_mask].loc[merged_df[adult_mask]['nes'].abs().nlargest(1).index]\n",
    "top_pediatric = merged_df[pediatric_mask].loc[merged_df[pediatric_mask]['nes'].abs().nlargest(1).index]\n",
    "\n",
    "# Combine the top pathways\n",
    "top_pathways_df = pd.concat([top_adult, top_pediatric])\n",
    "\n",
    "# Create the plot\n",
    "plt.figure(figsize=(12, 8))\n",
    "\n",
    "# Plot all pathways in grey\n",
    "plt.scatter(x=gsea_results_df['es'], \n",
    "            y=gsea_results_df['fdr'].apply(lambda x: -np.log10(x)), \n",
    "            s=10, color='grey', label='All pathways')\n",
    "\n",
    "# Plot top adult pathways in red\n",
    "plt.scatter(x=top_adult['es'], \n",
    "            y=top_adult['fdr'].apply(lambda x: -np.log10(x)), \n",
    "            s=10, color='red', label='Top Adult pathway')\n",
    "\n",
    "# Plot top pediatric pathways in blue\n",
    "plt.scatter(x=top_pediatric['es'], \n",
    "            y=top_pediatric['fdr'].apply(lambda x: -np.log10(x)), \n",
    "            s=10, color='blue', label='Top Pediatric pathway')\n",
    "\n",
    "# Add LFC and FDR cut-off lines\n",
    "plt.axhline(y=-np.log10(fdr_cutoff), color='r', linestyle='--', linewidth=1)\n",
    "plt.axvline(x=lfc_cutoff, color='g', linestyle='--', linewidth=1)\n",
    "plt.axvline(x=-lfc_cutoff, color='g', linestyle='--', linewidth=1)\n",
    "\n",
    "# Add labels for the top pathways\n",
    "\n",
    "for i in top_pathways_df.index:\n",
    "    es_value = top_pathways_df.at[i, 'es']\n",
    "    fdr_value = top_pathways_df.at[i, 'fdr']\n",
    "    pathway_label = top_pathways_df.at[i, 'Term']\n",
    "    \n",
    "    # Add the label for each point\n",
    "    plt.text((es_value), (-np.log10(fdr_value)), pathway_label, fontsize=8, ha='right' if es_value > 0 else 'left')\n",
    "\n",
    "\n",
    "# Label axes and add title\n",
    "plt.xlabel('log2 Fold Change (ES)')\n",
    "plt.ylabel('-log10(fdr)')\n",
    "plt.ylim(0, 20)\n",
    "plt.title('Gene Set Enrichment Analysis')\n",
    "\n",
    "# Add legend\n",
    "plt.legend()\n",
    "\n",
    "# Save the figure\n",
    "gsea_save_path = pathlib.Path(\"../1.data-exploration/figures/gsea_top_labeled.png\")\n",
    "plt.savefig(gsea_save_path, bbox_inches=\"tight\", dpi=600)\n",
    "\n",
    "# Show the plot\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[21], line 40\u001b[0m\n\u001b[1;32m     37\u001b[0m \u001b[38;5;66;03m# Loop through all unique pathways (Terms) in the GSEA DataFrame\u001b[39;00m\n\u001b[1;32m     38\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m pathway \u001b[38;5;129;01min\u001b[39;00m gsea_results_df[\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mTerm\u001b[39m\u001b[38;5;124m'\u001b[39m]\u001b[38;5;241m.\u001b[39munique():\n\u001b[1;32m     39\u001b[0m     \u001b[38;5;66;03m# Filter for the pathway\u001b[39;00m\n\u001b[0;32m---> 40\u001b[0m     filtered_gsea_df \u001b[38;5;241m=\u001b[39m gsea_results_df[\u001b[43mgsea_results_df\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[38;5;124;43mTerm\u001b[39;49m\u001b[38;5;124;43m'\u001b[39;49m\u001b[43m]\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m==\u001b[39;49m\u001b[43m \u001b[49m\u001b[43mpathway\u001b[49m]\n\u001b[1;32m     42\u001b[0m     \u001b[38;5;66;03m# Further filter by significant latent features\u001b[39;00m\n\u001b[1;32m     43\u001b[0m     filtered_gsea_df \u001b[38;5;241m=\u001b[39m filtered_gsea_df[filtered_gsea_df[\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mz_dim\u001b[39m\u001b[38;5;124m'\u001b[39m]\u001b[38;5;241m.\u001b[39misin(significant_latent_features[\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mz_dim\u001b[39m\u001b[38;5;124m'\u001b[39m])]\n",
      "File \u001b[0;32m~/anaconda3/envs/gene_dependency_representations/lib/python3.12/site-packages/pandas/core/ops/common.py:76\u001b[0m, in \u001b[0;36m_unpack_zerodim_and_defer.<locals>.new_method\u001b[0;34m(self, other)\u001b[0m\n\u001b[1;32m     72\u001b[0m             \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28mNotImplemented\u001b[39m\n\u001b[1;32m     74\u001b[0m other \u001b[38;5;241m=\u001b[39m item_from_zerodim(other)\n\u001b[0;32m---> 76\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mmethod\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mother\u001b[49m\u001b[43m)\u001b[49m\n",
      "File \u001b[0;32m~/anaconda3/envs/gene_dependency_representations/lib/python3.12/site-packages/pandas/core/arraylike.py:40\u001b[0m, in \u001b[0;36mOpsMixin.__eq__\u001b[0;34m(self, other)\u001b[0m\n\u001b[1;32m     38\u001b[0m \u001b[38;5;129m@unpack_zerodim_and_defer\u001b[39m(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124m__eq__\u001b[39m\u001b[38;5;124m\"\u001b[39m)\n\u001b[1;32m     39\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21m__eq__\u001b[39m(\u001b[38;5;28mself\u001b[39m, other):\n\u001b[0;32m---> 40\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_cmp_method\u001b[49m\u001b[43m(\u001b[49m\u001b[43mother\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43moperator\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43meq\u001b[49m\u001b[43m)\u001b[49m\n",
      "File \u001b[0;32m~/anaconda3/envs/gene_dependency_representations/lib/python3.12/site-packages/pandas/core/series.py:6119\u001b[0m, in \u001b[0;36mSeries._cmp_method\u001b[0;34m(self, other, op)\u001b[0m\n\u001b[1;32m   6116\u001b[0m lvalues \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_values\n\u001b[1;32m   6117\u001b[0m rvalues \u001b[38;5;241m=\u001b[39m extract_array(other, extract_numpy\u001b[38;5;241m=\u001b[39m\u001b[38;5;28;01mTrue\u001b[39;00m, extract_range\u001b[38;5;241m=\u001b[39m\u001b[38;5;28;01mTrue\u001b[39;00m)\n\u001b[0;32m-> 6119\u001b[0m res_values \u001b[38;5;241m=\u001b[39m \u001b[43mops\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mcomparison_op\u001b[49m\u001b[43m(\u001b[49m\u001b[43mlvalues\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mrvalues\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mop\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m   6121\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_construct_result(res_values, name\u001b[38;5;241m=\u001b[39mres_name)\n",
      "File \u001b[0;32m~/anaconda3/envs/gene_dependency_representations/lib/python3.12/site-packages/pandas/core/ops/array_ops.py:344\u001b[0m, in \u001b[0;36mcomparison_op\u001b[0;34m(left, right, op)\u001b[0m\n\u001b[1;32m    341\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m invalid_comparison(lvalues, rvalues, op)\n\u001b[1;32m    343\u001b[0m \u001b[38;5;28;01melif\u001b[39;00m lvalues\u001b[38;5;241m.\u001b[39mdtype \u001b[38;5;241m==\u001b[39m \u001b[38;5;28mobject\u001b[39m \u001b[38;5;129;01mor\u001b[39;00m \u001b[38;5;28misinstance\u001b[39m(rvalues, \u001b[38;5;28mstr\u001b[39m):\n\u001b[0;32m--> 344\u001b[0m     res_values \u001b[38;5;241m=\u001b[39m \u001b[43mcomp_method_OBJECT_ARRAY\u001b[49m\u001b[43m(\u001b[49m\u001b[43mop\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mlvalues\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mrvalues\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    346\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[1;32m    347\u001b[0m     res_values \u001b[38;5;241m=\u001b[39m _na_arithmetic_op(lvalues, rvalues, op, is_cmp\u001b[38;5;241m=\u001b[39m\u001b[38;5;28;01mTrue\u001b[39;00m)\n",
      "File \u001b[0;32m~/anaconda3/envs/gene_dependency_representations/lib/python3.12/site-packages/pandas/core/ops/array_ops.py:129\u001b[0m, in \u001b[0;36mcomp_method_OBJECT_ARRAY\u001b[0;34m(op, x, y)\u001b[0m\n\u001b[1;32m    127\u001b[0m     result \u001b[38;5;241m=\u001b[39m libops\u001b[38;5;241m.\u001b[39mvec_compare(x\u001b[38;5;241m.\u001b[39mravel(), y\u001b[38;5;241m.\u001b[39mravel(), op)\n\u001b[1;32m    128\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[0;32m--> 129\u001b[0m     result \u001b[38;5;241m=\u001b[39m \u001b[43mlibops\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mscalar_compare\u001b[49m\u001b[43m(\u001b[49m\u001b[43mx\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mravel\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43my\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mop\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m    130\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m result\u001b[38;5;241m.\u001b[39mreshape(x\u001b[38;5;241m.\u001b[39mshape)\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "# Prepare a DataFrame to store ANOVA results for multiple pathways\n",
    "anova_results = []\n",
    "\n",
    "# Loop through each OncotreePrimaryDisease type in the dataset\n",
    "for disease in model_df['OncotreePrimaryDisease'].unique():\n",
    "    # Extract model IDs for the current disease\n",
    "    disease_ids = model_df.query(f\"OncotreePrimaryDisease == '{disease}'\").ModelID.tolist()\n",
    "\n",
    "    # Create a copy of the latent space dataframe (either adult or pediatric as applicable)\n",
    "    latent_df = latent_df.copy()  # Use the appropriate dataframe here (adult or pediatric)\n",
    "\n",
    "    # Filter the latent dataframe to include only the current disease models\n",
    "    disease_latent_df = latent_df[latent_df['ModelID'].isin(disease_ids)]\n",
    "\n",
    "    # Drop the 'ModelID' column and reset the index\n",
    "    disease_latent_float_df = disease_latent_df.drop(columns=[\"ModelID\"]).reset_index(drop=True)\n",
    "\n",
    "    # Extract model IDs for the rest of the dataset\n",
    "    other_ids = model_df.query(f\"OncotreePrimaryDisease != '{disease}'\").ModelID.tolist()\n",
    "\n",
    "    # Filter the latent dataframe to include only the models not in the current disease\n",
    "    other_latent_df = latent_df[latent_df['ModelID'].isin(other_ids)]\n",
    "    other_latent_float_df = other_latent_df.drop(columns=[\"ModelID\"]).reset_index(drop=True)\n",
    "\n",
    "    # Perform t-tests comparing current disease vs the rest for each latent dimension\n",
    "    t_test_results = ttest_ind(disease_latent_float_df, other_latent_float_df)\n",
    "    t_test_results_df = pd.DataFrame(t_test_results).T\n",
    "    t_test_results_df.columns = [\"t_stat\", \"p_value\"]\n",
    "    t_test_results_df['comparison'] = f\"{disease} vs Rest\"\n",
    "    t_test_results_df['latent_feature'] = t_test_results_df.index + 1\n",
    "    t_test_results_df['z_dim'] = 'z_' + t_test_results_df['latent_feature'].astype(str)\n",
    "    t_test_results_df['group'] = t_test_results_df['t_stat'].apply(lambda x: disease if x > 0 else 'Rest')\n",
    "\n",
    "    # Filter significant latent features\n",
    "    significant_latent_features = t_test_results_df[t_test_results_df['p_value'] < 0.05]\n",
    "\n",
    "    # Loop through all unique pathways (Terms) in the GSEA DataFrame\n",
    "    for pathway in gsea_results_df['Term'].unique():\n",
    "        # Filter for the pathway\n",
    "        filtered_gsea_df = gsea_results_df[gsea_results_df['Term'] == pathway]\n",
    "        \n",
    "        # Further filter by significant latent features\n",
    "        filtered_gsea_df = filtered_gsea_df[filtered_gsea_df['z_dim'].isin(significant_latent_features['z_dim'])]\n",
    "        \n",
    "        # Merge GSEA DataFrame with t-test DataFrame to get group information\n",
    "        merged_df = pd.merge(filtered_gsea_df, significant_latent_features[['z_dim', 'group']], on='z_dim', how='inner')\n",
    "\n",
    "        # Group by 'group' and collect ES values\n",
    "        grouped_data = merged_df.groupby('group')['es'].apply(list)\n",
    "\n",
    "        # Ensure we have data for both groups\n",
    "        if len(grouped_data) == 2 and all(len(vals) > 1 for vals in grouped_data):\n",
    "            # Perform ANOVA\n",
    "            f_statistic, p_value = f_oneway(*grouped_data)\n",
    "\n",
    "            # Determine which group has higher enrichment score\n",
    "            disease_mean = np.mean(grouped_data[disease])\n",
    "            rest_mean = np.mean(grouped_data['Rest'])\n",
    "            higher_group = disease if disease_mean > rest_mean else 'Other Types'\n",
    "\n",
    "            # Store results\n",
    "            anova_results.append({\n",
    "                'Cancer Type': disease,\n",
    "                'Pathway': pathway,\n",
    "                'F-statistic': f_statistic,\n",
    "                'p-value': p_value,\n",
    "                'Higher in': higher_group\n",
    "            })\n",
    "\n",
    "# Convert the results to a DataFrame\n",
    "anova_results_df = pd.DataFrame(anova_results)\n",
    "\n",
    "# Apply a significance threshold (e.g., p < 0.05)\n",
    "significant_anova_results_df = anova_results_df[anova_results_df['p-value'] < 0.05]\n",
    "\n",
    "# Save the results to a CSV file\n",
    "anova_dir = pathlib.Path(\"./results/cancer_type_anova_results.csv\")\n",
    "significant_anova_results_df.to_csv(anova_dir, index=False)\n",
    "\n",
    "# Display significant pathways\n",
    "significant_anova_results_df.sort_values(by='F-statistic', key=abs, ascending=False).head(50)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "base",
   "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": "3.12.5"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}