example.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# PerturbationX\n",
    "PerturbationX is a package for analyzing causal networks in combination with gene expression data. It is based on the [TopoNPA](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4227138/) algorithm. It was developed as part of a Master's thesis at the University of Ljubljana, Faculty of Computer and Information Science and in collaboration with the National Institute of Biology. This notebook demonstrates the use of PerturbationX for analyzing a small causal network.\n",
    "\n",
    "## Installation\n",
    "The package can be installed from PyPI or directly from GitHub. It requires a Python version of 3.10 or newer. It is based on NetworkX and pandas and requires Cytoscape for visualization. The latter can be downloaded from [here](https://cytoscape.org/download.html).\n",
    "\n",
    "```bash\n",
    "python -m pip install perturbationx # PyPI\n",
    "python -m pip install git+https://github.com/mikethenut/perturbationx # GitHub\n",
    "```\n",
    "\n",
    "## Network import\n",
    "For demonstration purposes, a small example network with accompanying datasets is included in the package. The network core is based on the human Osmotic Stress network from the [Causal Biological Networks database](https://www.causalbionet.com/), available under Network ID [osmotic_stress_2.0_hs](https://www.causalbionet.com/Home/NetworkVisualization/networkGraph?ID=osmotic_stress_2.0_hs). All non-causal edges were removed and the largest connected component was then selected. The network boundary and datasets were generated synthetically and do not correspond to any real biological system."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "is_executing": true
   },
   "outputs": [],
   "source": [
    "import perturbationx as px\n",
    "from perturbationx import CausalNetwork\n",
    "\n",
    "core_network_file = px.resources.get_osmotic_stress_core_path() # path to the network core\n",
    "boundary_network_file = px.resources.get_osmotic_stress_boundary_path() # path to the network boundary\n",
    "osmotic_stress_network = CausalNetwork.from_tsv(core_network_file, edge_type=\"core\")\n",
    "osmotic_stress_network.add_edges_from_tsv(boundary_network_file, edge_type=\"boundary\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this example, we imported core and boundary edges separately. The files contain three columns, with a header row of 'subject', 'object', and 'relation'. Adding a 'type' column would allow us to concatenate the files and import them in one step. When the input file specifies valid edge types in this way, the argument passed to the import function is ignored.\n",
    "\n",
    "A CausalNetwork can also be instantiated from CSV files (from_csv), other delimited files (from_dsv), Cytoscape json files (from_cyjs_json), NetworkX graphs (from_networkx), and pandas DataFrames (from_pandas). Refer to the documentation for more information.\n",
    "\n",
    "## Network manipulation\n",
    "We can now inspect the network. We will first look at the number of nodes and edges."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of nodes: 131, 0 of which are core nodes.\n",
      "Number of edges: 436, 40 of which are core edges.\n"
     ]
    }
   ],
   "source": [
    "node_count = osmotic_stress_network.number_of_nodes()\n",
    "core_node_count = osmotic_stress_network.number_of_nodes(typ=\"core\")\n",
    "edge_count = osmotic_stress_network.number_of_edges()\n",
    "core_edge_count = osmotic_stress_network.number_of_edges(typ=\"core\")\n",
    "\n",
    "print(f\"Number of nodes: {node_count}, {core_node_count} of which are core nodes.\")\n",
    "print(f\"Number of edges: {edge_count}, {core_edge_count} of which are core edges.\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Wait, why does the network report zero core nodes? This is because node types are inferred dynamically during computation. While the network contains 131 nodes, none of them are currently assigned to the core or boundary. In order to see what the algorithm will do, we can use the following command."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-11-05 21:15:50,120 INFO -- core edges: 40\n",
      "2023-11-05 21:15:50,122 INFO -- boundary edges: 396\n",
      "2023-11-05 21:15:50,123 INFO -- core nodes: 31\n",
      "2023-11-05 21:15:50,125 INFO -- outer boundary nodes: 100\n",
      "2023-11-05 21:15:50,126 INFO -- core nodes with boundary edges (inner boundary): 10\n",
      "The network contains 31 core nodes and 100 boundary nodes.\n"
     ]
    }
   ],
   "source": [
    "osmotic_stress_network.infer_graph_attributes(inplace=True)\n",
    "\n",
    "# Recompute node counts\n",
    "core_node_count = osmotic_stress_network.number_of_nodes(typ=\"core\")\n",
    "boundary_node_count = osmotic_stress_network.number_of_nodes(typ=\"boundary\")\n",
    "\n",
    "print(f\"The network contains {core_node_count} core nodes and {boundary_node_count} boundary nodes.\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can see, the nodes are now correctly assigned. We obtained the same statistics from internal logs. Logging is enabled by default and can be disabled by setting `verbose=False`.\n",
    "\n",
    "Next, let's try adding a new edge to the network."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "New edge: ('a(CHEBI:\"hydrogen peroxide\")', 'a(SCHEM:Dinoprostone)', {'relation': '1.0', 'type': 'infer'})\n"
     ]
    }
   ],
   "source": [
    "src, trg = 'a(CHEBI:\"hydrogen peroxide\")', 'a(SCHEM:Dinoprostone)'\n",
    "osmotic_stress_network.add_edge(src, trg, '1.0')\n",
    "\n",
    "for edge in osmotic_stress_network.edges():\n",
    "    v1, v2, data = edge\n",
    "    if v1 == src and v2 == trg:\n",
    "        print(\"New edge: \" + str(edge))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The edge was successfully added, however its type was set to 'infer'. Let's specify the edge's type to prevent any errors during computation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Updated edge: ('a(CHEBI:\"hydrogen peroxide\")', 'a(SCHEM:Dinoprostone)', {'relation': '1.0', 'type': 'core'})\n"
     ]
    }
   ],
   "source": [
    "osmotic_stress_network.modify_edge(src, trg, typ='core')\n",
    "\n",
    "for edge in osmotic_stress_network.edges():\n",
    "    v1, v2, data = edge\n",
    "    if v1 == src and v2 == trg:\n",
    "        print(\"Updated edge: \" + str(edge))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We specified the edge relation as `1.0`. This is the causal relationship between vertices and is used to assign edge weights during computation. By default, numeric parsing is enabled to allow setting weights directly. The default settings also map `increases`, `directlyIncreases`, `->` and `=>` to a weight of 1.0, and `decreases`, `directlyDecreases`, `-|` and `=|` to a weight of -1.0. Different syntax can be used, however the mapping must be specified by correctly configuring the `RelationTranslator` class and passing it to the causal network. Unexpected relations will be mapped to a weight of zero, indicating no causal relationship. Refer to the documentation for more information.\n",
    "\n",
    "## Computing the NPA score\n",
    "\n",
    "Let's remove the extra edge and import our test datasets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "osmotic_stress_network.remove_edge(src, trg)\n",
    "\n",
    "import pandas as pd\n",
    "\n",
    "datasets = dict()\n",
    "idx = 1\n",
    "for dataset_path in px.resources.get_osmotic_stress_dataset_paths():\n",
    "    data_name = \"test_%d\" % idx\n",
    "    datasets[data_name] = pd.read_table(dataset_path)\n",
    "    idx += 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-16T14:36:32.941001Z",
     "start_time": "2023-09-16T14:36:32.907189800Z"
    }
   },
   "source": [
    "The implemention expects datasets to be passed as a dictionary of pandas DataFrames. Each DataFrame should contain at least the columns `nodeID` and `logFC`, however `t` or `stderr` are also required for computing statistics. We will now pass our dataset map to the network and begin computation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-16T14:36:33.585461200Z",
     "start_time": "2023-09-16T14:36:33.567800200Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-11-05 21:15:53,385 INFO -- PREPROCESSING NETWORK\n",
      "2023-11-05 21:15:53,394 INFO -- core edges: 40\n",
      "2023-11-05 21:15:53,395 INFO -- boundary edges: 396\n",
      "2023-11-05 21:15:53,395 INFO -- core nodes: 31\n",
      "2023-11-05 21:15:53,396 INFO -- outer boundary nodes: 100\n",
      "2023-11-05 21:15:53,397 INFO -- core nodes with boundary edges (inner boundary): 10\n",
      "2023-11-05 21:15:53,996 INFO -- COMPUTING NPA FOR DATASET 'test_1'\n",
      "2023-11-05 21:15:54,000 INFO -- boundary nodes matched with dataset: 100\n",
      "2023-11-05 21:15:54,001 INFO -- boundary edges remaining: 385\n",
      "2023-11-05 21:15:54,002 INFO -- core nodes with boundary edges remaining: 7\n",
      "2023-11-05 21:15:55,414 INFO -- COMPUTING NPA FOR DATASET 'test_2'\n",
      "2023-11-05 21:15:55,417 INFO -- boundary nodes matched with dataset: 100\n",
      "2023-11-05 21:15:55,418 INFO -- boundary edges remaining: 385\n",
      "2023-11-05 21:15:55,419 INFO -- core nodes with boundary edges remaining: 7\n",
      "2023-11-05 21:15:56,839 INFO -- COMPUTING NPA FOR DATASET 'test_3'\n",
      "2023-11-05 21:15:56,843 INFO -- boundary nodes matched with dataset: 100\n",
      "2023-11-05 21:15:56,844 INFO -- boundary edges remaining: 385\n",
      "2023-11-05 21:15:56,844 INFO -- core nodes with boundary edges remaining: 7\n"
     ]
    }
   ],
   "source": [
    "test_results = osmotic_stress_network.toponpa(datasets)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Results contain four types of information: global, node, distribution, and metadata. Let's first take a look at global and node information."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-16T12:49:11.983597400Z",
     "start_time": "2023-09-16T12:49:11.935721700Z"
    },
    "tags": []
   },
   "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>NPA</th>\n",
       "      <th>variance</th>\n",
       "      <th>ci_lower</th>\n",
       "      <th>ci_upper</th>\n",
       "      <th>k2_value</th>\n",
       "      <th>o_value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>test_1</th>\n",
       "      <td>0.005144</td>\n",
       "      <td>0.000568</td>\n",
       "      <td>-0.041569</td>\n",
       "      <td>0.051856</td>\n",
       "      <td>0.002</td>\n",
       "      <td>0.112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_2</th>\n",
       "      <td>0.117399</td>\n",
       "      <td>0.016193</td>\n",
       "      <td>-0.132009</td>\n",
       "      <td>0.366806</td>\n",
       "      <td>0.002</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>test_3</th>\n",
       "      <td>1.750560</td>\n",
       "      <td>8.072700</td>\n",
       "      <td>-3.818187</td>\n",
       "      <td>7.319307</td>\n",
       "      <td>0.006</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             NPA  variance  ci_lower  ci_upper  k2_value  o_value\n",
       "test_1  0.005144  0.000568 -0.041569  0.051856     0.002    0.112\n",
       "test_2  0.117399  0.016193 -0.132009  0.366806     0.002    0.000\n",
       "test_3  1.750560  8.072700 -3.818187  7.319307     0.006    0.000"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['contribution', 'coefficient', 'variance', 'ci_lower', 'ci_upper', 'p_value']\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>attr</th>\n",
       "      <th>contribution</th>\n",
       "      <th>coefficient</th>\n",
       "      <th>variance</th>\n",
       "      <th>ci_lower</th>\n",
       "      <th>ci_upper</th>\n",
       "      <th>p_value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>tscript(p(HGNC:HIF1A))</th>\n",
       "      <td>0.026643</td>\n",
       "      <td>-0.682930</td>\n",
       "      <td>0.273868</td>\n",
       "      <td>-1.708626</td>\n",
       "      <td>0.342767</td>\n",
       "      <td>0.095949</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>p(HGNC:AQP5)</th>\n",
       "      <td>0.038163</td>\n",
       "      <td>-0.667370</td>\n",
       "      <td>0.255508</td>\n",
       "      <td>-1.658088</td>\n",
       "      <td>0.323348</td>\n",
       "      <td>0.093371</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>kin(p(HGNC:EGFR))</th>\n",
       "      <td>0.016734</td>\n",
       "      <td>-0.541236</td>\n",
       "      <td>0.213143</td>\n",
       "      <td>-1.446100</td>\n",
       "      <td>0.363628</td>\n",
       "      <td>0.120532</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cat(p(HGNC:PTGS2))</th>\n",
       "      <td>0.038208</td>\n",
       "      <td>-0.871322</td>\n",
       "      <td>0.503980</td>\n",
       "      <td>-2.262731</td>\n",
       "      <td>0.520087</td>\n",
       "      <td>0.109844</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>p(HGNC:EGFR,pmod(P,Y))</th>\n",
       "      <td>0.016734</td>\n",
       "      <td>-0.541236</td>\n",
       "      <td>0.213143</td>\n",
       "      <td>-1.446100</td>\n",
       "      <td>0.363628</td>\n",
       "      <td>0.120532</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>a(SCHEM:Dinoprostone)</th>\n",
       "      <td>0.048062</td>\n",
       "      <td>-0.748938</td>\n",
       "      <td>0.268949</td>\n",
       "      <td>-1.765380</td>\n",
       "      <td>0.267504</td>\n",
       "      <td>0.074349</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>p(HGNC:HSPA1A)</th>\n",
       "      <td>0.013488</td>\n",
       "      <td>-0.687185</td>\n",
       "      <td>0.441027</td>\n",
       "      <td>-1.988795</td>\n",
       "      <td>0.614424</td>\n",
       "      <td>0.150390</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tscript(complex(SCOMP:\"AP-1 Complex\"))</th>\n",
       "      <td>0.011957</td>\n",
       "      <td>-0.633816</td>\n",
       "      <td>0.627025</td>\n",
       "      <td>-2.185813</td>\n",
       "      <td>0.918180</td>\n",
       "      <td>0.211732</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>kin(p(SFAM:\"PRKA Family\"))</th>\n",
       "      <td>0.057843</td>\n",
       "      <td>-0.711537</td>\n",
       "      <td>0.270254</td>\n",
       "      <td>-1.730444</td>\n",
       "      <td>0.307369</td>\n",
       "      <td>0.085545</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>p(HGNC:EGF)</th>\n",
       "      <td>0.028964</td>\n",
       "      <td>-0.503502</td>\n",
       "      <td>0.200832</td>\n",
       "      <td>-1.381845</td>\n",
       "      <td>0.374841</td>\n",
       "      <td>0.130606</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>p(HGNC:HSPA1B)</th>\n",
       "      <td>0.013488</td>\n",
       "      <td>-0.687185</td>\n",
       "      <td>0.441027</td>\n",
       "      <td>-1.988795</td>\n",
       "      <td>0.614424</td>\n",
       "      <td>0.150390</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>p(HGNC:SLC6A12)</th>\n",
       "      <td>0.031004</td>\n",
       "      <td>-0.758725</td>\n",
       "      <td>0.892567</td>\n",
       "      <td>-2.610416</td>\n",
       "      <td>1.092967</td>\n",
       "      <td>0.210962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>kin(p(SFAM:\"MAPK Erk1/2 Family\"))</th>\n",
       "      <td>0.060674</td>\n",
       "      <td>-0.651810</td>\n",
       "      <td>0.238653</td>\n",
       "      <td>-1.609295</td>\n",
       "      <td>0.305674</td>\n",
       "      <td>0.091061</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>p(HGNC:NRG1)</th>\n",
       "      <td>0.032483</td>\n",
       "      <td>-0.795125</td>\n",
       "      <td>0.396958</td>\n",
       "      <td>-2.029992</td>\n",
       "      <td>0.439742</td>\n",
       "      <td>0.103472</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>kin(p(HGNC:ATM))</th>\n",
       "      <td>0.049270</td>\n",
       "      <td>-0.656697</td>\n",
       "      <td>0.455038</td>\n",
       "      <td>-1.978821</td>\n",
       "      <td>0.665426</td>\n",
       "      <td>0.165149</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>p(HGNC:AQP3)</th>\n",
       "      <td>0.012135</td>\n",
       "      <td>-0.651810</td>\n",
       "      <td>0.238653</td>\n",
       "      <td>-1.609295</td>\n",
       "      <td>0.305674</td>\n",
       "      <td>0.091061</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>kin(p(HGNC:MAPK14))</th>\n",
       "      <td>0.027421</td>\n",
       "      <td>-0.692837</td>\n",
       "      <td>0.354713</td>\n",
       "      <td>-1.860148</td>\n",
       "      <td>0.474474</td>\n",
       "      <td>0.122353</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>a(SCHEM:Calcium)</th>\n",
       "      <td>0.012721</td>\n",
       "      <td>0.667370</td>\n",
       "      <td>0.255508</td>\n",
       "      <td>-0.323348</td>\n",
       "      <td>1.658088</td>\n",
       "      <td>0.093371</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>p(HGNC:SLC5A3)</th>\n",
       "      <td>0.013488</td>\n",
       "      <td>-0.687185</td>\n",
       "      <td>0.441027</td>\n",
       "      <td>-1.988795</td>\n",
       "      <td>0.614424</td>\n",
       "      <td>0.150390</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tscript(p(HGNC:NFAT5))</th>\n",
       "      <td>0.134878</td>\n",
       "      <td>-0.687185</td>\n",
       "      <td>0.441027</td>\n",
       "      <td>-1.988795</td>\n",
       "      <td>0.614424</td>\n",
       "      <td>0.150390</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cat(p(HGNC:ERBB3))</th>\n",
       "      <td>0.021159</td>\n",
       "      <td>-0.571884</td>\n",
       "      <td>0.170333</td>\n",
       "      <td>-1.380789</td>\n",
       "      <td>0.237021</td>\n",
       "      <td>0.082924</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>a(CHEBI:\"hydrogen peroxide\")</th>\n",
       "      <td>0.057445</td>\n",
       "      <td>-0.578969</td>\n",
       "      <td>0.228553</td>\n",
       "      <td>-1.515975</td>\n",
       "      <td>0.358036</td>\n",
       "      <td>0.112938</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>p(HGNC:AQP1)</th>\n",
       "      <td>0.025605</td>\n",
       "      <td>-0.669498</td>\n",
       "      <td>0.319636</td>\n",
       "      <td>-1.777589</td>\n",
       "      <td>0.438594</td>\n",
       "      <td>0.118169</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>bp(GOBP:\"hyperosmotic response\")</th>\n",
       "      <td>0.083611</td>\n",
       "      <td>-0.698490</td>\n",
       "      <td>0.293734</td>\n",
       "      <td>-1.760736</td>\n",
       "      <td>0.363757</td>\n",
       "      <td>0.098735</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>p(HGNC:AKR1B1)</th>\n",
       "      <td>0.013488</td>\n",
       "      <td>-0.687185</td>\n",
       "      <td>0.441027</td>\n",
       "      <td>-1.988795</td>\n",
       "      <td>0.614424</td>\n",
       "      <td>0.150390</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>kin(complex(SCOMP:\"AMP Activated Protein Kinase Complex\"))</th>\n",
       "      <td>0.014461</td>\n",
       "      <td>0.711537</td>\n",
       "      <td>0.270254</td>\n",
       "      <td>-0.307369</td>\n",
       "      <td>1.730444</td>\n",
       "      <td>0.085545</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>p(HGNC:CFTR,pmod(P,S,768))</th>\n",
       "      <td>0.028921</td>\n",
       "      <td>0.711537</td>\n",
       "      <td>0.270254</td>\n",
       "      <td>-0.307369</td>\n",
       "      <td>1.730444</td>\n",
       "      <td>0.085545</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cat(p(HGNC:PTGES2))</th>\n",
       "      <td>0.025182</td>\n",
       "      <td>-0.663954</td>\n",
       "      <td>0.230599</td>\n",
       "      <td>-1.605142</td>\n",
       "      <td>0.277235</td>\n",
       "      <td>0.083388</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tport(p(HGNC:CFTR))</th>\n",
       "      <td>0.028921</td>\n",
       "      <td>-0.711537</td>\n",
       "      <td>0.270254</td>\n",
       "      <td>-1.730444</td>\n",
       "      <td>0.307369</td>\n",
       "      <td>0.085545</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cat(p(HGNC:PARP1))</th>\n",
       "      <td>0.007847</td>\n",
       "      <td>0.482390</td>\n",
       "      <td>0.382737</td>\n",
       "      <td>-0.730157</td>\n",
       "      <td>1.694937</td>\n",
       "      <td>0.217773</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>p(HGNC:PLCG1,pmod(P,Y))</th>\n",
       "      <td>0.009001</td>\n",
       "      <td>-0.352568</td>\n",
       "      <td>0.182584</td>\n",
       "      <td>-1.190057</td>\n",
       "      <td>0.484922</td>\n",
       "      <td>0.204655</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "attr                                                contribution  coefficient  \\\n",
       "tscript(p(HGNC:HIF1A))                                  0.026643    -0.682930   \n",
       "p(HGNC:AQP5)                                            0.038163    -0.667370   \n",
       "kin(p(HGNC:EGFR))                                       0.016734    -0.541236   \n",
       "cat(p(HGNC:PTGS2))                                      0.038208    -0.871322   \n",
       "p(HGNC:EGFR,pmod(P,Y))                                  0.016734    -0.541236   \n",
       "a(SCHEM:Dinoprostone)                                   0.048062    -0.748938   \n",
       "p(HGNC:HSPA1A)                                          0.013488    -0.687185   \n",
       "tscript(complex(SCOMP:\"AP-1 Complex\"))                  0.011957    -0.633816   \n",
       "kin(p(SFAM:\"PRKA Family\"))                              0.057843    -0.711537   \n",
       "p(HGNC:EGF)                                             0.028964    -0.503502   \n",
       "p(HGNC:HSPA1B)                                          0.013488    -0.687185   \n",
       "p(HGNC:SLC6A12)                                         0.031004    -0.758725   \n",
       "kin(p(SFAM:\"MAPK Erk1/2 Family\"))                       0.060674    -0.651810   \n",
       "p(HGNC:NRG1)                                            0.032483    -0.795125   \n",
       "kin(p(HGNC:ATM))                                        0.049270    -0.656697   \n",
       "p(HGNC:AQP3)                                            0.012135    -0.651810   \n",
       "kin(p(HGNC:MAPK14))                                     0.027421    -0.692837   \n",
       "a(SCHEM:Calcium)                                        0.012721     0.667370   \n",
       "p(HGNC:SLC5A3)                                          0.013488    -0.687185   \n",
       "tscript(p(HGNC:NFAT5))                                  0.134878    -0.687185   \n",
       "cat(p(HGNC:ERBB3))                                      0.021159    -0.571884   \n",
       "a(CHEBI:\"hydrogen peroxide\")                            0.057445    -0.578969   \n",
       "p(HGNC:AQP1)                                            0.025605    -0.669498   \n",
       "bp(GOBP:\"hyperosmotic response\")                        0.083611    -0.698490   \n",
       "p(HGNC:AKR1B1)                                          0.013488    -0.687185   \n",
       "kin(complex(SCOMP:\"AMP Activated Protein Kinase...      0.014461     0.711537   \n",
       "p(HGNC:CFTR,pmod(P,S,768))                              0.028921     0.711537   \n",
       "cat(p(HGNC:PTGES2))                                     0.025182    -0.663954   \n",
       "tport(p(HGNC:CFTR))                                     0.028921    -0.711537   \n",
       "cat(p(HGNC:PARP1))                                      0.007847     0.482390   \n",
       "p(HGNC:PLCG1,pmod(P,Y))                                 0.009001    -0.352568   \n",
       "\n",
       "attr                                                variance  ci_lower  \\\n",
       "tscript(p(HGNC:HIF1A))                              0.273868 -1.708626   \n",
       "p(HGNC:AQP5)                                        0.255508 -1.658088   \n",
       "kin(p(HGNC:EGFR))                                   0.213143 -1.446100   \n",
       "cat(p(HGNC:PTGS2))                                  0.503980 -2.262731   \n",
       "p(HGNC:EGFR,pmod(P,Y))                              0.213143 -1.446100   \n",
       "a(SCHEM:Dinoprostone)                               0.268949 -1.765380   \n",
       "p(HGNC:HSPA1A)                                      0.441027 -1.988795   \n",
       "tscript(complex(SCOMP:\"AP-1 Complex\"))              0.627025 -2.185813   \n",
       "kin(p(SFAM:\"PRKA Family\"))                          0.270254 -1.730444   \n",
       "p(HGNC:EGF)                                         0.200832 -1.381845   \n",
       "p(HGNC:HSPA1B)                                      0.441027 -1.988795   \n",
       "p(HGNC:SLC6A12)                                     0.892567 -2.610416   \n",
       "kin(p(SFAM:\"MAPK Erk1/2 Family\"))                   0.238653 -1.609295   \n",
       "p(HGNC:NRG1)                                        0.396958 -2.029992   \n",
       "kin(p(HGNC:ATM))                                    0.455038 -1.978821   \n",
       "p(HGNC:AQP3)                                        0.238653 -1.609295   \n",
       "kin(p(HGNC:MAPK14))                                 0.354713 -1.860148   \n",
       "a(SCHEM:Calcium)                                    0.255508 -0.323348   \n",
       "p(HGNC:SLC5A3)                                      0.441027 -1.988795   \n",
       "tscript(p(HGNC:NFAT5))                              0.441027 -1.988795   \n",
       "cat(p(HGNC:ERBB3))                                  0.170333 -1.380789   \n",
       "a(CHEBI:\"hydrogen peroxide\")                        0.228553 -1.515975   \n",
       "p(HGNC:AQP1)                                        0.319636 -1.777589   \n",
       "bp(GOBP:\"hyperosmotic response\")                    0.293734 -1.760736   \n",
       "p(HGNC:AKR1B1)                                      0.441027 -1.988795   \n",
       "kin(complex(SCOMP:\"AMP Activated Protein Kinase...  0.270254 -0.307369   \n",
       "p(HGNC:CFTR,pmod(P,S,768))                          0.270254 -0.307369   \n",
       "cat(p(HGNC:PTGES2))                                 0.230599 -1.605142   \n",
       "tport(p(HGNC:CFTR))                                 0.270254 -1.730444   \n",
       "cat(p(HGNC:PARP1))                                  0.382737 -0.730157   \n",
       "p(HGNC:PLCG1,pmod(P,Y))                             0.182584 -1.190057   \n",
       "\n",
       "attr                                                ci_upper   p_value  \n",
       "tscript(p(HGNC:HIF1A))                              0.342767  0.095949  \n",
       "p(HGNC:AQP5)                                        0.323348  0.093371  \n",
       "kin(p(HGNC:EGFR))                                   0.363628  0.120532  \n",
       "cat(p(HGNC:PTGS2))                                  0.520087  0.109844  \n",
       "p(HGNC:EGFR,pmod(P,Y))                              0.363628  0.120532  \n",
       "a(SCHEM:Dinoprostone)                               0.267504  0.074349  \n",
       "p(HGNC:HSPA1A)                                      0.614424  0.150390  \n",
       "tscript(complex(SCOMP:\"AP-1 Complex\"))              0.918180  0.211732  \n",
       "kin(p(SFAM:\"PRKA Family\"))                          0.307369  0.085545  \n",
       "p(HGNC:EGF)                                         0.374841  0.130606  \n",
       "p(HGNC:HSPA1B)                                      0.614424  0.150390  \n",
       "p(HGNC:SLC6A12)                                     1.092967  0.210962  \n",
       "kin(p(SFAM:\"MAPK Erk1/2 Family\"))                   0.305674  0.091061  \n",
       "p(HGNC:NRG1)                                        0.439742  0.103472  \n",
       "kin(p(HGNC:ATM))                                    0.665426  0.165149  \n",
       "p(HGNC:AQP3)                                        0.305674  0.091061  \n",
       "kin(p(HGNC:MAPK14))                                 0.474474  0.122353  \n",
       "a(SCHEM:Calcium)                                    1.658088  0.093371  \n",
       "p(HGNC:SLC5A3)                                      0.614424  0.150390  \n",
       "tscript(p(HGNC:NFAT5))                              0.614424  0.150390  \n",
       "cat(p(HGNC:ERBB3))                                  0.237021  0.082924  \n",
       "a(CHEBI:\"hydrogen peroxide\")                        0.358036  0.112938  \n",
       "p(HGNC:AQP1)                                        0.438594  0.118169  \n",
       "bp(GOBP:\"hyperosmotic response\")                    0.363757  0.098735  \n",
       "p(HGNC:AKR1B1)                                      0.614424  0.150390  \n",
       "kin(complex(SCOMP:\"AMP Activated Protein Kinase...  1.730444  0.085545  \n",
       "p(HGNC:CFTR,pmod(P,S,768))                          1.730444  0.085545  \n",
       "cat(p(HGNC:PTGES2))                                 0.277235  0.083388  \n",
       "tport(p(HGNC:CFTR))                                 0.307369  0.085545  \n",
       "cat(p(HGNC:PARP1))                                  1.694937  0.217773  \n",
       "p(HGNC:PLCG1,pmod(P,Y))                             0.484922  0.204655  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(test_results.global_info())\n",
    "\n",
    "print(test_results.node_attributes())\n",
    "display(test_results.node_info(\"test_3\"))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The datasets all have nonsensical variance due to their synthetic nature. Regardless, we can see that the NPA scores vary significantly. For the dataset with the lowest score, the permutation tests produced p-values above `0.05`, indicating that the result is not significant.\n",
    "\n",
    "Node information is accessed by specifying either the name of a dataset or a node attribute. The TopoNPA algorithm assigns six attributes per dataset: contribution, coefficient, variance, lower and upper confidence interval limits, and p-values. The last four are all derived from standard error and are thus unrealistic in this example. Coefficients are counterparts to the `logFC` values we provided for boundary nodes. Contributions are estimates of how important a node is for the total NPA score. Since we have a lot of nodes, we might want to look at only the ones with the highest contribution."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-16T12:49:17.302414300Z",
     "start_time": "2023-09-16T12:49:17.254953200Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Leading nodes for both datasets: {'kin(p(HGNC:ATM))', 'kin(p(SFAM:\"PRKA Family\"))', 'p(HGNC:AQP1)', 'kin(p(HGNC:MAPK14))', 'tscript(p(HGNC:NFAT5))', 'bp(GOBP:\"hyperosmotic response\")', 'kin(p(SFAM:\"MAPK Erk1/2 Family\"))', 'p(HGNC:AQP5)', 'tscript(p(HGNC:HIF1A))', 'p(HGNC:CFTR,pmod(P,S,768))', 'p(HGNC:SLC6A12)', 'tport(p(HGNC:CFTR))', 'a(SCHEM:Dinoprostone)', 'p(HGNC:NRG1)', 'a(CHEBI:\"hydrogen peroxide\")'}\n",
      "Leading nodes for only dataset 2: {'cat(p(HGNC:ERBB3))'}\n",
      "Leading nodes for only dataset 3: {'cat(p(HGNC:PTGS2))', 'cat(p(HGNC:PTGES2))', 'p(HGNC:EGF)'}\n"
     ]
    }
   ],
   "source": [
    "leading_nodes_2 = test_results.get_leading_nodes(\"test_2\")\n",
    "leading_nodes_3 = test_results.get_leading_nodes(\"test_3\")\n",
    "\n",
    "print(\"Leading nodes for both datasets: \" + str(leading_nodes_2.intersection(leading_nodes_3)))\n",
    "print(\"Leading nodes for only dataset 2: \" + str(leading_nodes_2.difference(leading_nodes_3)))\n",
    "print(\"Leading nodes for only dataset 3: \" + str(leading_nodes_3.difference(leading_nodes_2)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Comparing leading nodes between the two datasets with significant results, we can see that there is a high degree of conformity. While this is not guaranteed (especially when using synthetic data), it is likely that the same nodes will show up across different datasets.\n",
    "\n",
    "Next, let's take a look at permutation test distributions."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-16T12:49:39.196088400Z",
     "start_time": "2023-09-16T12:49:39.183227100Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['k2', 'o']\n"
     ]
    }
   ],
   "source": [
    "print(test_results.distributions())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By default, two permutation tests are computed. The `O` permutation test shuffles boundary nodes and tests the significance of their ordering. The `K2` permutation test shuffles core edges according to the degree sequence and tests the importance of the core network structure. A `K1` test is also available, which shuffles edges completely at random. The tests can be further adjusted via optional arguments. Refer to the documentation for more information."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 640x480 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x1200 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "array([<Axes: ylabel='test_1'>, <Axes: ylabel='test_2'>,\n",
       "       <Axes: ylabel='test_3'>], dtype=object)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_results.plot_distribution('o')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-16T12:49:42.717727Z",
     "start_time": "2023-09-16T12:49:42.425775100Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 640x480 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x1200 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "array([<Axes: ylabel='test_1'>, <Axes: ylabel='test_2'>,\n",
       "       <Axes: ylabel='test_3'>], dtype=object)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_results.plot_distribution('k2')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Finally, we can inspect the metadata. Information about the programming environment, package version, network and dataset information is all added automatically. Additional metadata can be assigned to a CausalNetwork through the metadata attribute, and it will appear here prepended with `network_`.\n",
    "\n",
    "Results can be exported to a JSON file using the `to_json()` method."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'missing_value_pruning_mode': 'nullify',\n",
       " 'opposing_value_pruning_mode': None,\n",
       " 'opposing_value_minimum_amplitude': 1.0,\n",
       " 'boundary_edge_minimum': 6,\n",
       " 'exact_boundary_outdegree': True,\n",
       " 'alpha': 0.95,\n",
       " 'full_core_permutation': True,\n",
       " 'permutation_iterations': 500,\n",
       " 'permutation_rate': 1.0,\n",
       " 'seed': None,\n",
       " 'datetime_utc': '2023-11-05T20:15:58.155957',\n",
       " 'python_implementation': 'CPython',\n",
       " 'python_version': '3.10.0',\n",
       " 'system_name': 'Windows',\n",
       " 'system_release': '10',\n",
       " 'system_version': '10.0.19045',\n",
       " 'network_core_edges': 40,\n",
       " 'network_boundary_edges': 396,\n",
       " 'network_core_nodes': 31,\n",
       " 'network_outer_boundary_nodes': 100,\n",
       " 'network_inner_boundary_nodes': 10,\n",
       " 'network_title': 'osmotic_stress_core',\n",
       " 'network_collection': 'resources',\n",
       " 'dataset_test_1': {'matched_outer_boundary_nodes': 100,\n",
       "  'matched_boundary_edges': 385,\n",
       "  'matched_inner_boundary_nodes': 7},\n",
       " 'network_dataset_test_1': {'matched_outer_boundary_nodes': 100,\n",
       "  'matched_boundary_edges': 385,\n",
       "  'matched_inner_boundary_nodes': 7},\n",
       " 'dataset_test_2': {'matched_outer_boundary_nodes': 100,\n",
       "  'matched_boundary_edges': 385,\n",
       "  'matched_inner_boundary_nodes': 7},\n",
       " 'network_dataset_test_2': {'matched_outer_boundary_nodes': 100,\n",
       "  'matched_boundary_edges': 385,\n",
       "  'matched_inner_boundary_nodes': 7},\n",
       " 'dataset_test_3': {'matched_outer_boundary_nodes': 100,\n",
       "  'matched_boundary_edges': 385,\n",
       "  'matched_inner_boundary_nodes': 7},\n",
       " 'network_dataset_test_3': {'matched_outer_boundary_nodes': 100,\n",
       "  'matched_boundary_edges': 385,\n",
       "  'matched_inner_boundary_nodes': 7},\n",
       " 'perturbationx_version': '1.0.1',\n",
       " 'numpy_version': '1.26.1',\n",
       " 'networkx_version': '3.2.1',\n",
       " 'pandas_version': '2.1.2',\n",
       " 'scipy_version': '1.11.3',\n",
       " 'matplotlib_version': '3.8.1',\n",
       " 'seaborn_version': '0.13.0',\n",
       " 'py4cytoscape_version': '1.8.0'}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "display(test_results.metadata())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Network visualization\n",
    "\n",
    "The last feature that the result object offers is Cytoscape integration through the py4cytoscape package. Let's import the package and verify that Cytoscape is running."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-16T13:49:15.986119200Z",
     "start_time": "2023-09-16T13:49:15.949123600Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-11-05 21:16:05,373 DEBUG -- Calling cytoscape_version_info()\n",
      "2023-11-05 21:16:05,374 DEBUG -- ǀCalling cyrest_get('version', base_url='http://127.0.0.1:1234/v1')\n",
      "2023-11-05 21:16:05,376 DEBUG -- Attempting to direct connect to Cytoscape on http://127.0.0.1:1234/v1\n",
      "2023-11-05 21:16:05,470 DEBUG -- Detected py4cytoscape running on Cytoscape workstation\n",
      "2023-11-05 21:16:05,472 DEBUG -- ǀHTTP GET(http://127.0.0.1:1234/v1/version)\n",
      "2023-11-05 21:16:05,488 DEBUG -- ǀOK[200], content: {\"apiVersion\":\"v1\",\"cytoscapeVersion\":\"3.10.1\"}\n",
      "2023-11-05 21:16:05,490 DEBUG -- ǀReturning 'cyrest_get': {'apiVersion': 'v1', 'cytoscapeVersion': '3.10.1'}\n",
      "2023-11-05 21:16:05,493 DEBUG -- Returning 'cytoscape_version_info': {'apiVersion': 'v1', 'cytoscapeVersion': '3.10.1', 'automationAPIVersion': '1.8.0', 'py4cytoscapeVersion': '1.8.0'}\n",
      "2023-11-05 21:16:05,495 DEBUG -- --------------------\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'apiVersion': 'v1',\n",
       " 'cytoscapeVersion': '3.10.1',\n",
       " 'automationAPIVersion': '1.8.0',\n",
       " 'py4cytoscapeVersion': '1.8.0'}"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import py4cytoscape as p4c\n",
    "p4c.cytoscape_version_info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If no errors were returned, we can display our results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-16T15:02:27.350350100Z",
     "start_time": "2023-09-16T15:02:25.521963800Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Applying default style...\n",
      "Applying preferred layout\n"
     ]
    }
   ],
   "source": [
    "cynet = test_results.display_network()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The method returns an NPAResultDisplay instance, which we can use to further adjust the display. For example, we can color the nodes using any node attribute and any dataset. Let's show node contributions for the `test_3` dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "cynet_suid = cynet.color_nodes(\"test_3\", \"contribution\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "While coloring nodes is helpful for identifying key parts of the network, it is insufficient when viewing large networks. For this reason, two methods were implemented for visualizing leading nodes The first is `highlight_leading_nodes`. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "cynet_suid = cynet.highlight_leading_nodes(\"test_3\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "All leading nodes have now been highlighted with an emboldened node border. To include related edges and edges we can specify optional arguments. Let's try it on the second method, `extract_leading_nodes`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "cynet_suid = cynet.extract_leading_nodes(\"test_3\", include_shortest_paths=\"directed\", path_length_tolerance=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By default, this method performs extraction in place by hiding all unselected nodes and edges. We have included shortest directed paths between each pair of leading nodes, with a path length tolerance of $1$. The length tolerance selects also paths that are up to $1$ edge longer than the shortest path. It is also possible to select undirected shortest paths, however this usually results in a smaller subgraph, as the edges do not need to be directed in the same way.\n",
    "\n",
    "The other way to include related nodes is by selecting neighbors. Let's use out-of-place extraction for that."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "subnet_suid = cynet.extract_leading_nodes(\"test_3\", include_neighbors=1, neighborhood_type=\"union\", inplace=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can see, the entire core network was selected, as every node is at most one edge away from a leading node. If we instead chose a neighborhood type of intersection, the nodes would have to be first neighbors of all leading nodes to be included. It is also possible to combine display of neighborhoods and paths. In this case, the selected nodes and edges are combined.\n",
    "\n",
    "The extracted subnetwork inherits the network style of the original. We can always reset the display to clear coloring, highlights and invisibility."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "128"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cynet.reset_display(reset_color=True, reset_highlight=True, reset_visibility=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Advanced usage\n",
    "\n",
    "The package offers some additional options when it comes to network analysis. Once imported, parts of the core network can be randomly wired or rewired. Let's try adding two random edges between a couple of nodes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-11-05 21:16:33,404 INFO -- WIRING EDGES\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[('p(HGNC:EGFR,pmod(P,Y))', 'p(HGNC:HSPA1A)', -1, 'core'),\n",
       " ('p(HGNC:EGFR,pmod(P,Y))', 'p(HGNC:HSPA1A)', 1, 'core')]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "[('a(SCHEM:Calcium)', 'tport(p(HGNC:CFTR))', -1, 'core'),\n",
       " ('p(HGNC:EGFR,pmod(P,Y))', 'a(SCHEM:Calcium)', -1, 'core')]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "[('p(HGNC:EGFR,pmod(P,Y))', 'tport(p(HGNC:CFTR))', -1, 'core'),\n",
       " ('p(HGNC:HSPA1A)', 'a(SCHEM:Calcium)', 1, 'core')]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "[('a(SCHEM:Calcium)', 'p(HGNC:EGFR,pmod(P,Y))', 1, 'core'),\n",
       " ('p(HGNC:EGFR,pmod(P,Y))', 'tport(p(HGNC:CFTR))', 1, 'core')]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "[('a(SCHEM:Calcium)', 'tport(p(HGNC:CFTR))', 1, 'core'),\n",
       " ('p(HGNC:HSPA1A)', 'p(HGNC:EGFR,pmod(P,Y))', -1, 'core')]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "modifications = osmotic_stress_network.wire_edges(\n",
    "    ['tport(p(HGNC:CFTR))', 'p(HGNC:EGFR,pmod(P,Y))', 'a(SCHEM:Calcium)', 'p(HGNC:HSPA1A)'], \n",
    "    number_of_edges=2, iterations=5, edge_relations=[1, -1], datasets=None\n",
    ")\n",
    "for modification in modifications:\n",
    "    display(modification)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If we don't pass in any datasets, we will only recieve lists of modifications. These can be applied to the network using the `modify_network` method. Modifications can also be constructed and passed in by hand. If a relation of an edge is set to None, the edge will instead be deleted from the network.\n",
    "\n",
    "If we pass in our test datasets, we will also receive an evaluation of the permutations using the TopoNPA algorithm."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-11-05 21:16:34,625 INFO -- WIRING EDGES\n",
      "2023-11-05 21:16:34,642 INFO -- PREPROCESSING NETWORK\n",
      "2023-11-05 21:16:34,685 INFO -- COMPUTING NPA FOR DATASET 'test_1'\n",
      "2023-11-05 21:16:34,691 INFO -- boundary nodes matched with dataset: 100\n",
      "2023-11-05 21:16:34,692 INFO -- boundary edges remaining: 385\n",
      "2023-11-05 21:16:34,693 INFO -- core nodes with boundary edges remaining: 7\n",
      "2023-11-05 21:16:34,720 INFO -- COMPUTING NPA FOR DATASET 'test_2'\n",
      "2023-11-05 21:16:34,725 INFO -- boundary nodes matched with dataset: 100\n",
      "2023-11-05 21:16:34,726 INFO -- boundary edges remaining: 385\n",
      "2023-11-05 21:16:34,727 INFO -- core nodes with boundary edges remaining: 7\n",
      "2023-11-05 21:16:34,743 INFO -- COMPUTING NPA FOR DATASET 'test_3'\n",
      "2023-11-05 21:16:34,747 INFO -- boundary nodes matched with dataset: 100\n",
      "2023-11-05 21:16:34,748 INFO -- boundary edges remaining: 385\n",
      "2023-11-05 21:16:34,749 INFO -- core nodes with boundary edges remaining: 7\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[('tport(p(HGNC:CFTR))', 'a(SCHEM:Calcium)', -1, 'core'),\n",
       " ('p(HGNC:EGFR,pmod(P,Y))', 'a(SCHEM:Calcium)', 1, 'core')]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "{'test_1': 0.002935597323214852,\n",
       " 'test_2': 0.07395491316870083,\n",
       " 'test_3': 0.974211131537148}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "[('a(SCHEM:Calcium)', 'tport(p(HGNC:CFTR))', -1, 'core'),\n",
       " ('p(HGNC:EGFR,pmod(P,Y))', 'a(SCHEM:Calcium)', -1, 'core')]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "{'test_1': 0.005267391568917243,\n",
       " 'test_2': 0.11608037044920416,\n",
       " 'test_3': 1.8019464239281209}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "[('p(HGNC:EGFR,pmod(P,Y))', 'p(HGNC:HSPA1A)', 1, 'core'),\n",
       " ('a(SCHEM:Calcium)', 'p(HGNC:EGFR,pmod(P,Y))', -1, 'core')]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "{'test_1': 0.005261593247009843,\n",
       " 'test_2': 0.11592317474799156,\n",
       " 'test_3': 1.8102068444031594}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "[('a(SCHEM:Calcium)', 'p(HGNC:EGFR,pmod(P,Y))', 1, 'core'),\n",
       " ('a(SCHEM:Calcium)', 'p(HGNC:EGFR,pmod(P,Y))', -1, 'core')]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "{'test_1': 0.005153815038587717,\n",
       " 'test_2': 0.11423531161319098,\n",
       " 'test_3': 1.7711877984533957}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "[('p(HGNC:HSPA1A)', 'tport(p(HGNC:CFTR))', -1, 'core'),\n",
       " ('tport(p(HGNC:CFTR))', 'p(HGNC:HSPA1A)', 1, 'core')]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "{'test_1': 0.005266004590349019,\n",
       " 'test_2': 0.12090783645597367,\n",
       " 'test_3': 1.7959738118234712}"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "results = osmotic_stress_network.wire_edges(\n",
    "    ['tport(p(HGNC:CFTR))', 'p(HGNC:EGFR,pmod(P,Y))', 'a(SCHEM:Calcium)', 'p(HGNC:HSPA1A)'], \n",
    "    number_of_edges=2, iterations=5, edge_relations=[1, -1], datasets=datasets\n",
    ")\n",
    "\n",
    "for modification, scores in results:\n",
    "    display(modification)\n",
    "    display(scores)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A number of optional arguments are available to adjust evaluation, both in edge wiring and standard computation. We briefly describe them here and explain the logic behind them. An in-depth evaluation was performed as part of the Master's thesis \"Integration of gene expression data with causal networks\", which can be found in the root of the project's repository.\n",
    "\n",
    "### Missing value pruning mode\n",
    "When values aren't passed for boundary nodes, their associated edges must be removed. We have two options when doing this: we can either first remove the edge weights for missing edges and then balance, or we can first balance the weights and then remove the edges. The first option can be selected by passing \"remove\". It indicates that the edge was somehow incorrect and should not have been included. The second option is the default and can be selected by passing \"nullify\". It indicates that the edge should be included, but cannot be processed in the current computation.\n",
    "\n",
    "### Boundary edge minimum\n",
    "Although weights are expected for boundary edges, they are not used directly. Instead, the absolute weights of all boundary edges that share a source core node are first balanced, so that they sum to $1$. This ensures that boundary edges do not overtake the core edges in terms of importance for the NPA score. However, that can still occur if a core node only has a few boundary edges, resulting in large weights. This is why a boundary edge minimum is applied, and the boundary edges of any core node that are fewer in number are removed. Specifically, they are removed at the same time as boundary weights are balanced. If the number of edges falls below the minimum afterwards, during edge nullification, the remaining edges are not removed. The parameter defaults to $6$ and can be set to $0$ to remove this behaviour altogether.\n",
    "\n",
    "### Exact boundary outdegree\n",
    "When we use nullification to remove edges, the absolute weights of boundary edges for a core node will no longer sum to $1$. In this case, we can choose to use the exact boundary outdegrees of core nodes when computing the NPA score, which is the default behavior. Alternatively, we may choose to use binary outdegrees, which are $1$ for any node with at least one boundary outdegree, and $0$ otherwise. The second option was selected as the default behavior in the R NPA package.\n",
    "\n",
    "### Opposing value pruning mode\n",
    "Sometimes, we may encounted data that causally opposes our biological understanding. For example, we may select three regulators **A**, **B**, and **C**, which all promote the expression of **X**, yet we observe a decrease of **X** in our measurements. Should the decrease still be processed by those edges? If we would like to avoid this, we can use an opposing value pruning mode of either \"remove\" or \"nullify\", similar to the removal of edges for missing values. By default, this behaviour is disabled with a value of `None`.\n",
    "\n",
    "### Opposing value minimum amplitude\n",
    "Although many measurements may oppose our causal edges, most of them are likely to be insignificant. In this case, the insignificance of the measurement is likely more important than its causality. Following this logic, only measurements with a large enough absolute value should be considered as causally opposing and removed. This parameter defaults to a value of $1$, and is disregarded when opposing value pruning is not used."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.10.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}