notebook/Codon Bias Graph.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import argparse\n",
    "import os\n",
    "import json\n",
    "import logging\n",
    "import string\n",
    "import collections\n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sqlalchemy import create_engine\n",
    "import networkx as nx\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "if os.getcwd().endswith('notebook'):\n",
    "    os.chdir('..')\n",
    "\n",
    "from rna_learn.alphabet import CODON_REDUNDANCY\n",
    "from rna_learn.codon_bias.trna_ai import compute_trna_ai_weights\n",
    "from rna_learn.codon_bias.graph import load_codon_bias\n",
    "from rna_learn.codon_bias.distance import compute_codon_bias_distance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "sns.set(palette='colorblind', font_scale=1.3)\n",
    "palette = sns.color_palette()\n",
    "logging.basicConfig(level=logging.INFO, format=\"%(asctime)s (%(levelname)s) %(message)s\")\n",
    "logger = logging.getLogger(__name__)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "db_path = os.path.join(os.getcwd(), 'data/db/seq.db')\n",
    "engine = create_engine(f'sqlite+pysqlite:///{db_path}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "codon_to_aa_table = { \n",
    "    'ATA':'I', 'ATC':'I', 'ATT':'I', 'ATG':'M', \n",
    "    'ACA':'T', 'ACC':'T', 'ACG':'T', 'ACT':'T', \n",
    "    'AAC':'N', 'AAT':'N', 'AAA':'K', 'AAG':'K', \n",
    "    'AGC':'S', 'AGT':'S', 'AGA':'R', 'AGG':'R',                  \n",
    "    'CTA':'L', 'CTC':'L', 'CTG':'L', 'CTT':'L', \n",
    "    'CCA':'P', 'CCC':'P', 'CCG':'P', 'CCT':'P', \n",
    "    'CAC':'H', 'CAT':'H', 'CAA':'Q', 'CAG':'Q', \n",
    "    'CGA':'R', 'CGC':'R', 'CGG':'R', 'CGT':'R', \n",
    "    'GTA':'V', 'GTC':'V', 'GTG':'V', 'GTT':'V', \n",
    "    'GCA':'A', 'GCC':'A', 'GCG':'A', 'GCT':'A', \n",
    "    'GAC':'D', 'GAT':'D', 'GAA':'E', 'GAG':'E', \n",
    "    'GGA':'G', 'GGC':'G', 'GGG':'G', 'GGT':'G', \n",
    "    'TCA':'S', 'TCC':'S', 'TCG':'S', 'TCT':'S', \n",
    "    'TTC':'F', 'TTT':'F', 'TTA':'L', 'TTG':'L', \n",
    "    'TAC':'Y', 'TAT':'Y', 'TAA':'*', 'TAG':'*', \n",
    "    'TGC':'C', 'TGT':'C', 'TGA':'*', 'TGG':'W', \n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'I': ['ATA', 'ATC', 'ATT'],\n",
       " 'M': ['ATG'],\n",
       " 'T': ['ACA', 'ACC', 'ACG', 'ACT'],\n",
       " 'N': ['AAC', 'AAT'],\n",
       " 'K': ['AAA', 'AAG'],\n",
       " 'S': ['AGC', 'AGT', 'TCA', 'TCC', 'TCG', 'TCT'],\n",
       " 'R': ['AGA', 'AGG', 'CGA', 'CGC', 'CGG', 'CGT'],\n",
       " 'L': ['CTA', 'CTC', 'CTG', 'CTT', 'TTA', 'TTG'],\n",
       " 'P': ['CCA', 'CCC', 'CCG', 'CCT'],\n",
       " 'H': ['CAC', 'CAT'],\n",
       " 'Q': ['CAA', 'CAG'],\n",
       " 'V': ['GTA', 'GTC', 'GTG', 'GTT'],\n",
       " 'A': ['GCA', 'GCC', 'GCG', 'GCT'],\n",
       " 'D': ['GAC', 'GAT'],\n",
       " 'E': ['GAA', 'GAG'],\n",
       " 'G': ['GGA', 'GGC', 'GGG', 'GGT'],\n",
       " 'F': ['TTC', 'TTT'],\n",
       " 'Y': ['TAC', 'TAT'],\n",
       " '*': ['TAA', 'TAG', 'TGA'],\n",
       " 'C': ['TGC', 'TGT'],\n",
       " 'W': ['TGG']}"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "aa_to_codons = collections.defaultdict(list)\n",
    "for codon, aa in codon_to_aa_table.items():\n",
    "    aa_to_codons[aa].append(codon)\n",
    "aa_to_codons = dict(aa_to_codons)\n",
    "aa_to_codons"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load graph"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 335 ms, sys: 50.1 ms, total: 385 ms\n",
      "Wall time: 387 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "graph_path = os.path.join(os.getcwd(), 'data/codon_bias_graph.gpickle')\n",
    "graph = nx.read_gpickle(graph_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nx.number_connected_components(graph)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2672"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(nx.node_connected_component(graph, 2336))"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {},
   "source": [
    "graph.nodes[562]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load species Ids"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "species_query = \"\"\"\n",
    "select species_taxid, species, phylum, superkingdom, growth_tmp from species_traits\n",
    "\"\"\"\n",
    "species_df = pd.read_sql(species_query, engine)\n",
    "species_taxids = species_df['species_taxid'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>species_taxid</th>\n",
       "      <th>species</th>\n",
       "      <th>phylum</th>\n",
       "      <th>superkingdom</th>\n",
       "      <th>growth_tmp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>25</td>\n",
       "      <td>Shewanella hanedai</td>\n",
       "      <td>Proteobacteria</td>\n",
       "      <td>Bacteria</td>\n",
       "      <td>16.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>102</td>\n",
       "      <td>Polaribacter glomeratus</td>\n",
       "      <td>Bacteroidetes</td>\n",
       "      <td>Bacteria</td>\n",
       "      <td>15.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>141</th>\n",
       "      <td>531</td>\n",
       "      <td>Polaribacter irgensii</td>\n",
       "      <td>Bacteroidetes</td>\n",
       "      <td>Bacteria</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>184</th>\n",
       "      <td>659</td>\n",
       "      <td>Photobacterium phosphoreum</td>\n",
       "      <td>Proteobacteria</td>\n",
       "      <td>Bacteria</td>\n",
       "      <td>19.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>194</th>\n",
       "      <td>688</td>\n",
       "      <td>Aliivibrio logei</td>\n",
       "      <td>Proteobacteria</td>\n",
       "      <td>Bacteria</td>\n",
       "      <td>16.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2548</th>\n",
       "      <td>1217908</td>\n",
       "      <td>Octadecabacter antarcticus</td>\n",
       "      <td>Proteobacteria</td>\n",
       "      <td>Bacteria</td>\n",
       "      <td>4.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2621</th>\n",
       "      <td>1424659</td>\n",
       "      <td>Cryobacterium flavum</td>\n",
       "      <td>Actinobacteria</td>\n",
       "      <td>Bacteria</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2622</th>\n",
       "      <td>1424661</td>\n",
       "      <td>Cryobacterium luteum</td>\n",
       "      <td>Actinobacteria</td>\n",
       "      <td>Bacteria</td>\n",
       "      <td>10.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2667</th>\n",
       "      <td>1706046</td>\n",
       "      <td>Sulfurirhabdus autotrophica</td>\n",
       "      <td>Proteobacteria</td>\n",
       "      <td>Bacteria</td>\n",
       "      <td>15.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2677</th>\n",
       "      <td>1835254</td>\n",
       "      <td>Polynucleobacter duraquae</td>\n",
       "      <td>Proteobacteria</td>\n",
       "      <td>Bacteria</td>\n",
       "      <td>18.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>89 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      species_taxid                      species          phylum superkingdom  \\\n",
       "6                25           Shewanella hanedai  Proteobacteria     Bacteria   \n",
       "20              102      Polaribacter glomeratus   Bacteroidetes     Bacteria   \n",
       "141             531        Polaribacter irgensii   Bacteroidetes     Bacteria   \n",
       "184             659   Photobacterium phosphoreum  Proteobacteria     Bacteria   \n",
       "194             688             Aliivibrio logei  Proteobacteria     Bacteria   \n",
       "...             ...                          ...             ...          ...   \n",
       "2548        1217908   Octadecabacter antarcticus  Proteobacteria     Bacteria   \n",
       "2621        1424659         Cryobacterium flavum  Actinobacteria     Bacteria   \n",
       "2622        1424661         Cryobacterium luteum  Actinobacteria     Bacteria   \n",
       "2667        1706046  Sulfurirhabdus autotrophica  Proteobacteria     Bacteria   \n",
       "2677        1835254    Polynucleobacter duraquae  Proteobacteria     Bacteria   \n",
       "\n",
       "      growth_tmp  \n",
       "6           16.0  \n",
       "20          15.0  \n",
       "141          4.0  \n",
       "184         19.5  \n",
       "194         16.0  \n",
       "...          ...  \n",
       "2548         4.0  \n",
       "2621        10.0  \n",
       "2622        10.0  \n",
       "2667        15.0  \n",
       "2677        18.0  \n",
       "\n",
       "[89 rows x 5 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "species_df[species_df['growth_tmp'] < 20]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Nearby metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_mae(y_true, y_pred):\n",
    "    return np.mean(np.abs(y_true - y_pred))\n",
    "\n",
    "\n",
    "def predict_numerical_field_from_neighbours(graph, species_taxid, trait, control_for_phylum=True):\n",
    "    current_node = graph.nodes[species_taxid]\n",
    "    \n",
    "    if pd.isnull(current_node[trait]):\n",
    "        return None, None, None\n",
    "    \n",
    "    current_phylum = current_node['phylum']\n",
    "    values = []\n",
    "    for key in graph.neighbors(species_taxid):\n",
    "        node = graph.nodes[key]\n",
    "        \n",
    "        if pd.isnull(node[trait]):\n",
    "            continue\n",
    "        \n",
    "        if control_for_phylum and node['phylum'] != current_phylum:\n",
    "            v = node[trait]\n",
    "            values.append(v)\n",
    "        elif not control_for_phylum:\n",
    "            v = node[trait]\n",
    "            values.append(v)\n",
    "            \n",
    "    if len(values) == 0:\n",
    "        prediction = None\n",
    "    else:\n",
    "        prediction = np.mean(values)\n",
    "\n",
    "    return current_node[trait], prediction, values\n",
    "\n",
    "def compute_mae_from_neighbour_predictions(graph, species_taxids, trait, control_for_phylum=True):\n",
    "    actuals = []\n",
    "    predictions = []\n",
    "    for i, species_taxid in enumerate(species_taxids):\n",
    "        actual, prediction, _ = predict_numerical_field_from_neighbours(\n",
    "            graph, \n",
    "            species_taxid, \n",
    "            trait, \n",
    "            control_for_phylum,\n",
    "        )\n",
    "        \n",
    "        if actual is None or prediction is None:\n",
    "            continue\n",
    "        \n",
    "        actuals.append(actual)\n",
    "        predictions.append(prediction)\n",
    "        \n",
    "    return compute_mae(np.array(actuals), np.array(predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_accuracy(y_true, y_pred):\n",
    "    correct = [\n",
    "        1 if y_true[i] == y_pred[i] else 0\n",
    "        for i in range(len(y_pred))\n",
    "    ]\n",
    "    return 100 * np.sum(correct) / len(y_pred)\n",
    "\n",
    "\n",
    "def predict_categorical_field_from_neighbours(graph, species_taxid, trait, control_for_phylum=True):\n",
    "    current_node = graph.nodes[species_taxid]\n",
    "    \n",
    "    if pd.isnull(current_node[trait]):\n",
    "        return None, None, None\n",
    "    \n",
    "    current_phylum = current_node['phylum']\n",
    "    values = collections.defaultdict(int)\n",
    "    for key in graph.neighbors(species_taxid):\n",
    "        node = graph.nodes[key]\n",
    "        \n",
    "        if pd.isnull(node[trait]):\n",
    "            continue\n",
    "        \n",
    "        if control_for_phylum and node['phylum'] != current_phylum:\n",
    "            v = node[trait]\n",
    "            values[v] += 1\n",
    "        elif not control_for_phylum:\n",
    "            v = node[trait]\n",
    "            values[v] += 1\n",
    "            \n",
    "    if len(values) == 0:\n",
    "        prediction = None\n",
    "        sorted_values = []\n",
    "    else:\n",
    "        sorted_values = sorted(values.items(), key=lambda t: t[1], reverse=True)\n",
    "        prediction = sorted_values[0][0]\n",
    "\n",
    "    return current_node[trait], prediction, sorted_values\n",
    "\n",
    "def compute_accuracy_from_neighbour_predictions(graph, species_taxids, trait, control_for_phylum=True):\n",
    "    actuals = []\n",
    "    predictions = []\n",
    "    for i, species_taxid in enumerate(species_taxids):\n",
    "        actual, prediction, _ = predict_categorical_field_from_neighbours(\n",
    "            graph, \n",
    "            species_taxid, \n",
    "            trait, \n",
    "            control_for_phylum,\n",
    "        )\n",
    "        \n",
    "        if actual is None or prediction is None:\n",
    "            continue\n",
    "        \n",
    "        actuals.append(actual)\n",
    "        predictions.append(prediction)\n",
    "        \n",
    "    return compute_accuracy(actuals, predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8.644639843926784"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "compute_mae_from_neighbour_predictions(graph, species_taxids, 'growth_tmp')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6.363267744129107"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "compute_mae_from_neighbour_predictions(graph, species_taxids, 'growth_tmp', control_for_phylum=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8045548369656437"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "compute_mae_from_neighbour_predictions(graph, species_taxids, 'optimum_ph')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.5698871501477607"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "compute_mae_from_neighbour_predictions(graph, species_taxids, 'gc_content')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "14.398689358230122"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "compute_mae_from_neighbour_predictions(graph, species_taxids, 'tRNA_genes')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2.167322515728827"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "compute_mae_from_neighbour_predictions(graph, species_taxids, 'rRNA16S_genes')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1183.0628712801658"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "compute_mae_from_neighbour_predictions(graph, species_taxids, 'coding_genes')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "80.0 78.725 1.2749999999999986\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[77.45, 80.0]"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "actual, prediction, values = predict_numerical_field_from_neighbours(graph, 2336, 'growth_tmp')\n",
    "print(actual, prediction, np.std(values))\n",
    "values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "14.698162729658792"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "compute_accuracy_from_neighbour_predictions(graph, species_taxids, 'isolation_source')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2337 Thermotoga neapolitana Bacteria Thermotogae 79.0 sediment_marine_hydrothermal\n",
      "57487 Pseudothermotoga hypogea Bacteria Thermotogae 69.95 petroleum\n",
      "58290 Archaeoglobus veneficus Archaea Euryarchaeota 77.45 sediment_marine_hydrothermal\n",
      "93929 Thermotoga petrophila Bacteria Thermotogae 77.25 wastewater\n",
      "93930 Thermotoga naphthophila Bacteria Thermotogae 77.15 wastewater\n",
      "565033 Geoglobus acetivorans Archaea Euryarchaeota 80.0 sediment_marine_hydrothermal\n",
      "1184387 Mesotoga prima Bacteria Thermotogae 37.0 sediment_marine\n",
      "1236046 Mesotoga infera Bacteria Thermotogae 45.0 water_marine\n"
     ]
    }
   ],
   "source": [
    "for key in graph.neighbors(2336):\n",
    "    node = graph.nodes[key]\n",
    "    print(\n",
    "        node['species_taxid'], \n",
    "        node['species'], \n",
    "        node['superkingdom'], \n",
    "        node['phylum'], \n",
    "        node['growth_tmp'],\n",
    "        node['isolation_source'],\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'species_taxid': 2336,\n",
       " 'species': 'Thermotoga maritima',\n",
       " 'growth_tmp': 80.0,\n",
       " 'genus': 'Thermotoga',\n",
       " 'family': 'Thermotogaceae',\n",
       " 'order': 'Thermotogales',\n",
       " 'class': 'Thermotogae',\n",
       " 'phylum': 'Thermotogae',\n",
       " 'superkingdom': 'Bacteria',\n",
       " 'optimum_ph': nan,\n",
       " 'gram_stain': 'negative',\n",
       " 'metabolism': 'anaerobic',\n",
       " 'sporulation': 'no',\n",
       " 'motility': 'yes',\n",
       " 'range_salinity': None,\n",
       " 'cell_shape': 'bacillus',\n",
       " 'isolation_source': 'sediment_marine_hydrothermal',\n",
       " 'doubling_h': 1.233,\n",
       " 'genome_size': 1865969.5,\n",
       " 'gc_content': 46.214,\n",
       " 'coding_genes': 1871.167,\n",
       " 'tRNA_genes': 46.0,\n",
       " 'rRNA16S_genes': 1.0,\n",
       " 'n_assemblies': 6,\n",
       " 'assembly_accessions': ('GCA_000008545.1',\n",
       "  'GCA_000230655.3',\n",
       "  'GCA_000390265.1',\n",
       "  'GCA_000978535.1',\n",
       "  'GCA_000978555.1',\n",
       "  'GCA_000978575.1')}"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "graph.nodes[2336]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'species_taxid': 565033,\n",
       " 'species': 'Geoglobus acetivorans',\n",
       " 'growth_tmp': 80.0,\n",
       " 'genus': 'Geoglobus',\n",
       " 'family': 'Archaeoglobaceae',\n",
       " 'order': 'Archaeoglobales',\n",
       " 'class': 'Archaeoglobi',\n",
       " 'phylum': 'Euryarchaeota',\n",
       " 'superkingdom': 'Archaea',\n",
       " 'optimum_ph': nan,\n",
       " 'gram_stain': None,\n",
       " 'metabolism': 'anaerobic',\n",
       " 'sporulation': None,\n",
       " 'motility': 'yes',\n",
       " 'range_salinity': None,\n",
       " 'cell_shape': None,\n",
       " 'isolation_source': 'sediment_marine_hydrothermal',\n",
       " 'doubling_h': nan,\n",
       " 'genome_size': 1860816.25,\n",
       " 'gc_content': 46.8,\n",
       " 'coding_genes': 2172.667,\n",
       " 'tRNA_genes': 48.0,\n",
       " 'rRNA16S_genes': 1.0,\n",
       " 'n_assemblies': 1,\n",
       " 'assembly_accessions': ('GCA_000789255.1',)}"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "graph.nodes[565033]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'species_taxid': 58290,\n",
       " 'species': 'Archaeoglobus veneficus',\n",
       " 'growth_tmp': 77.45,\n",
       " 'genus': 'Archaeoglobus',\n",
       " 'family': 'Archaeoglobaceae',\n",
       " 'order': 'Archaeoglobales',\n",
       " 'class': 'Archaeoglobi',\n",
       " 'phylum': 'Euryarchaeota',\n",
       " 'superkingdom': 'Archaea',\n",
       " 'optimum_ph': nan,\n",
       " 'gram_stain': 'negative',\n",
       " 'metabolism': 'anaerobic',\n",
       " 'sporulation': None,\n",
       " 'motility': 'yes',\n",
       " 'range_salinity': None,\n",
       " 'cell_shape': 'coccus',\n",
       " 'isolation_source': 'sediment_marine_hydrothermal',\n",
       " 'doubling_h': 1.02,\n",
       " 'genome_size': 1901942.4,\n",
       " 'gc_content': 47.0,\n",
       " 'coding_genes': 2076.0,\n",
       " 'tRNA_genes': 46.0,\n",
       " 'rRNA16S_genes': 1.0,\n",
       " 'n_assemblies': 1,\n",
       " 'assembly_accessions': ('GCA_000194625.1',)}"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "graph.nodes[58290]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('sediment_marine_hydrothermal',\n",
       " 'sediment_marine_hydrothermal',\n",
       " [('sediment_marine_hydrothermal', 2)])"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predict_categorical_field_from_neighbours(graph, 2336, 'isolation_source')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Term frequency: find proteins with codon bias near the mean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "def protein_description_word_count(assembly_accession, below_threshold=True, count_type='document'):\n",
    "    if count_type not in ('document', 'term'):\n",
    "        raise ValueError('count type should be one of document or term')\n",
    "        \n",
    "    cds_codon_bias_folder = os.path.join(os.getcwd(), 'data/cds_codon_bias')\n",
    "    if below_threshold:\n",
    "        cds_codon_bias_folder = os.path.join(cds_codon_bias_folder, 'below_threshold')\n",
    "    else:\n",
    "        cds_codon_bias_folder = os.path.join(cds_codon_bias_folder, 'all')\n",
    "        \n",
    "    cds_codon_bias_path = os.path.join(cds_codon_bias_folder, f'{assembly_accession}_codon_bias.csv')\n",
    "    \n",
    "    black_list = [\n",
    "        '',\n",
    "        'protein',\n",
    "        'enzyme',\n",
    "        'subunit',\n",
    "        'domain',\n",
    "        'domain-containing',\n",
    "        'hypothetical',\n",
    "        'putative',\n",
    "        'probable',\n",
    "        'conserved',\n",
    "        'factor',\n",
    "        'large',\n",
    "        'small',\n",
    "        'family',\n",
    "        'superfamily',\n",
    "        'outer',\n",
    "        'cell',\n",
    "        'i',\n",
    "        'ii',\n",
    "        'iii',\n",
    "        'iv',\n",
    "        'v',\n",
    "        'vi',\n",
    "        'alpha',\n",
    "        'beta',\n",
    "        'gamma',\n",
    "        'type',\n",
    "        'and',\n",
    "        'or',\n",
    "        'of',\n",
    "        'class',\n",
    "        'assembly',\n",
    "        'acid',\n",
    "        'component',\n",
    "        'response',\n",
    "        '2x',\n",
    "        '1-2',\n",
    "    ]\n",
    "    black_list += [c for c in string.ascii_lowercase]\n",
    "    black_list += [c for c in string.digits]\n",
    "    black_list = set(black_list)\n",
    "    \n",
    "    df = pd.read_csv(cds_codon_bias_path)\n",
    "    descriptions = df['description'].values\n",
    "    \n",
    "    counts = collections.defaultdict(int)\n",
    "    for description in descriptions:\n",
    "        desc = str(description).replace(',', '').replace('.', '').replace('(', '').replace(')', '').strip()\n",
    "        parts_1 = desc.split(' ')\n",
    "        parts = []\n",
    "        for p in parts_1:\n",
    "            parts  += [v.lower() for v in p.split('/')]\n",
    "            \n",
    "        for word in parts:\n",
    "            if word in black_list:\n",
    "                continue\n",
    "                \n",
    "            if count_type == 'document':\n",
    "                counts[word] = 1\n",
    "            else:\n",
    "                counts[word] += 1\n",
    "    \n",
    "    return dict(counts)\n",
    "\n",
    "\n",
    "def compute_word_count(assembly_accessions, below_threshold=True, count_type='document', sorted_output=True):\n",
    "    if count_type not in ('document', 'term'):\n",
    "        raise ValueError('count type should be one of document or term')\n",
    "    \n",
    "    total_counts = {}\n",
    "    for i, assembly_accession in enumerate(assembly_accessions):\n",
    "        if (i+1) % 200 == 0:\n",
    "            print(f'{i+1} / {len(assembly_accessions)}')\n",
    "            \n",
    "        counts = protein_description_word_count(assembly_accession, below_threshold, count_type)\n",
    "        \n",
    "        for k, v in counts.items():\n",
    "            total_counts[k] = total_counts.get(k, 0) + counts[k]\n",
    "    \n",
    "    if sorted_output:\n",
    "        word_counts = [\n",
    "            (\n",
    "                k, \n",
    "                np.round(100 * v / len(assembly_accessions), 2)\n",
    "            ) \n",
    "\n",
    "            for k, v in total_counts.items()\n",
    "        ]\n",
    "\n",
    "        return sorted(word_counts, key=lambda t: t[1], reverse=True)\n",
    "    else:\n",
    "        return total_counts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_word_scores(assembly_accessions, min_term_freq=1):\n",
    "    top_term_freq = compute_word_count(\n",
    "        assembly_accessions,\n",
    "        below_threshold=True,\n",
    "        count_type='term',\n",
    "        sorted_output=False,\n",
    "    )\n",
    "    all_term_freqs = compute_word_count(\n",
    "        assembly_accessions,\n",
    "        below_threshold=False,\n",
    "        count_type='term',\n",
    "        sorted_output=False,\n",
    "    )\n",
    "    all_words = sorted(all_term_freqs.keys())\n",
    "    \n",
    "    word_scores = []\n",
    "    for word in all_words:\n",
    "        term_freq = top_term_freq.get(word, 0)\n",
    "        all_term_freq = all_term_freqs[word]\n",
    "        \n",
    "        term_freq_norm = term_freq / len(assembly_accessions)\n",
    "        all_term_freq_norm = all_term_freq / len(assembly_accessions)\n",
    "        \n",
    "        if term_freq_norm < min_term_freq:\n",
    "            continue\n",
    "            \n",
    "        score = term_freq / all_term_freqs[word]\n",
    "        \n",
    "        word_scores.append((\n",
    "            word, \n",
    "            score, \n",
    "            np.round(term_freq_norm, 2), \n",
    "            np.round(all_term_freq_norm, 2),\n",
    "        ))\n",
    "        \n",
    "    return sorted(word_scores, key=lambda t: t[1], reverse=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "def local_word_scores(graph, species_taxid, control_for_phylum=True, include_neighbours=True):\n",
    "    current_node = graph.nodes[species_taxid]\n",
    "\n",
    "    current_word_scores = compute_word_scores(current_node['assembly_accessions'])\n",
    "\n",
    "    score_dicts = [{t[0]: t for t in current_word_scores}]\n",
    "    word_set = set([t[0] for t in current_word_scores])\n",
    "    \n",
    "    neighbours = list(graph.neighbors(species_taxid))\n",
    "    \n",
    "    if include_neighbours:\n",
    "        for key in neighbours:\n",
    "            node = graph.nodes[key]\n",
    "\n",
    "            if control_for_phylum and node['phylum'] == current_node['phylum']:\n",
    "                continue\n",
    "\n",
    "            word_scores = compute_word_scores(node['assembly_accessions'])\n",
    "\n",
    "            word_set &= set([t[0] for t in word_scores])\n",
    "            score_dicts.append({t[0]: t for t in word_scores})\n",
    "\n",
    "    final_word_scores = []\n",
    "    for word in sorted(word_set):\n",
    "        scores, term_freqs, all_term_freqs = [], [], []\n",
    "        for score_dict in score_dicts:\n",
    "            _, score, term_freq, all_term_freq = score_dict[word]\n",
    "            scores.append(score)\n",
    "            term_freqs.append(term_freq)\n",
    "            all_term_freqs.append(all_term_freq)\n",
    "            \n",
    "        final_word_scores.append((\n",
    "            word, \n",
    "            np.round(np.mean(score), 3),\n",
    "            np.round(np.mean(term_freqs), 2), \n",
    "            np.round(np.mean(all_term_freqs), 2),\n",
    "        ))\n",
    "        \n",
    "    return sorted(final_word_scores, key=lambda t: t[1], reverse=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "def retrieve_orginal_proteins(graph, species_taxid, include_neighbours=True):\n",
    "    cds_codon_bias_folder = os.path.join(os.getcwd(), 'data/cds_codon_bias/below_threshold')\n",
    "    \n",
    "    species_taxids = [species_taxid] \n",
    "    if include_neighbours:\n",
    "        species_taxids += [key for key in graph.neighbors(species_taxid)]\n",
    "    \n",
    "    word_scores = local_word_scores(graph, species_taxid, include_neighbours=include_neighbours)\n",
    "    words = [t[0] for t in word_scores]\n",
    "        \n",
    "    protein_dict = collections.defaultdict(list)\n",
    "    for species_taxid in species_taxids:\n",
    "        node = graph.nodes[species_taxid]\n",
    "        assembly_accessions = node['assembly_accessions']\n",
    "        \n",
    "        for assembly_accession in assembly_accessions:\n",
    "            cds_codon_bias_path = os.path.join(cds_codon_bias_folder, f'{assembly_accession}_codon_bias.csv')\n",
    "            df = pd.read_csv(cds_codon_bias_path)\n",
    "            descriptions = df['description'].values\n",
    "            distances = df['distance'].values\n",
    "            \n",
    "            for word in words:\n",
    "                for i, protein_description in enumerate(descriptions):\n",
    "                    protein_desc = str(protein_description)\n",
    "                    protein_desc_key = protein_desc.replace(',', '').replace(': ', ' ').replace(':', ' ').lower()\n",
    "                    if word.lower() in protein_desc_key:\n",
    "                        protein_dict[protein_desc_key].append((protein_desc, distances[i]))\n",
    "                        \n",
    "    output_proteins = [\n",
    "        (protein_dict[protein_desc][0][0], np.mean([t[1] for t in protein_dict[protein_desc]]))\n",
    "        for protein_desc in protein_dict.keys()\n",
    "    ]\n",
    "                        \n",
    "    return sorted(output_proteins, key=lambda t: t[1])\n",
    "\n",
    "\n",
    "def retrieve_orginal_proteins_global(graph, species_taxids):\n",
    "    cds_codon_bias_folder = os.path.join(os.getcwd(), 'data/cds_codon_bias/below_threshold')\n",
    "\n",
    "    protein_dict = collections.defaultdict(list)\n",
    "    seen_species = set()\n",
    "    for s_ix, species_taxid in enumerate(species_taxids):\n",
    "        if (s_ix + 1) % 10 == 0:\n",
    "            print(f'{s_ix + 1:,} / {len(species_taxids):,}')\n",
    "        \n",
    "        if species_taxid in seen_species:\n",
    "            continue\n",
    "        \n",
    "        node = graph.nodes[species_taxid]\n",
    "        assembly_accessions = node['assembly_accessions']\n",
    "        \n",
    "        neighbors = list(graph.neighbors(species_taxid))\n",
    "        \n",
    "        seen_species.add(species_taxid)\n",
    "        seen_species |= set(neighbors)\n",
    "        \n",
    "        word_scores = local_word_scores(graph, species_taxid)\n",
    "        words = [t[0] for t in word_scores]\n",
    "        \n",
    "        for assembly_accession in assembly_accessions:\n",
    "            cds_codon_bias_path = os.path.join(cds_codon_bias_folder, f'{assembly_accession}_codon_bias.csv')\n",
    "            df = pd.read_csv(cds_codon_bias_path)\n",
    "            descriptions = df['description'].values\n",
    "            distances = df['distance'].values\n",
    "            \n",
    "            for word in words:\n",
    "                for i, protein_description in enumerate(descriptions):\n",
    "                    protein_desc = str(protein_description)\n",
    "                    protein_desc_key = protein_desc.replace(',', '').replace(': ', ' ').replace(':', ' ').lower()\n",
    "                    if word.lower() in protein_desc_key:\n",
    "                        protein_dict[protein_desc_key].append((protein_desc, distances[i]))\n",
    "                        \n",
    "    output_proteins = [\n",
    "        (protein_dict[protein_desc][0][0], np.mean([t[1] for t in protein_dict[protein_desc]]))\n",
    "        for protein_desc in protein_dict.keys()\n",
    "    ]\n",
    "                        \n",
    "    return sorted(output_proteins, key=lambda t: t[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "134 proteins\n",
      "permease\n",
      "DNA-directed RNA polymerase subunit beta\n",
      "alpha-mannosidase-related protein\n",
      "alpha-mannosidase\n",
      "alpha-mannosidase, putative\n",
      "DNA-directed RNA polymerase, beta subunit\n",
      "sensor histidine kinase\n",
      "histidine kinase\n",
      "DNA polymerase I\n",
      "DNA-directed DNA polymerase I\n",
      "DNA polymerase\n",
      "Xylose-regulated ABC transporter, substrate-binding component\n",
      "oligopeptide ABC transporter, periplasmic oligopeptide-binding protein\n",
      "iron transporter FeoB\n",
      "DNA gyrase, subunit A\n",
      "pyrimidine-nucleoside phosphorylase\n",
      "thymidine phosphorylase\n",
      "UDP-N-acetylglucosamine 2-epimerase\n",
      "Flagellar hook protein FlgE\n",
      "flagellar hook-basal body protein\n",
      "helicase\n",
      "single stranded DNA-specific exonuclease, putative\n",
      "single-stranded DNA exonuclease\n",
      "Single-stranded-DNA-specific exonuclease RecJ\n",
      "Superfamily II DNA/RNA helicase, SNF2 family\n",
      "CRISPR-associated protein\n",
      "DNA mismatch repair protein\n",
      "DNA mismatch repair protein MutS\n",
      "DNA ligase\n",
      "NAD-dependent DNA ligase LigA\n",
      "CRISPR-associated helicase Cas3\n",
      "double-stranded DNA repair protein Rad50\n",
      "Exonuclease SbcC\n",
      "DNA repair protein Rad50\n",
      "cation-transporting ATPase, P-type\n",
      "ATPase P\n",
      "Lead, cadmium, zinc and mercury transporting ATPase ; Copper-translocating P-type ATPase\n",
      "diguanylate cyclase\n",
      "Phosphoglycerol transferase\n",
      "(R)-2-hydroxyglutaryl-CoA dehydratase activator-related protein\n",
      "2-hydroxyglutaryl-CoA dehydratase\n",
      "RND transporter\n",
      "membrane protein, putative\n",
      "ATPase AAA\n",
      "MoxR-like ATPase\n",
      "isoleucyl-tRNA synthetase\n",
      "isoleucine--tRNA ligase\n",
      "diguanylate cyclase (GGDEF domain) with PAS/PAC sensor\n",
      "(p)ppGpp synthetase\n",
      "GTP pyrophosphokinase , (p)ppGpp synthetase I\n",
      "6-phosphogluconate dehydrogenase, decarboxylating\n",
      "6-phosphogluconate dehydrogenase\n",
      "peptide ABC transporter substrate-binding protein\n",
      "Fe-S oxidoreductase\n",
      "tRNA nucleotidyl transferase-related protein\n",
      "tRNA nucleotidyltransferase\n",
      "beta-galactosidase\n",
      "lipopolysaccharide biosynthesis protein\n",
      "ribonucleotide reductase, B12-dependent\n",
      "ribonucleotide reductase\n",
      "Ribonucleotide reductase of class II (coenzyme B12-dependent)\n",
      "cellobiose-phosphorylase\n",
      "glycosyl transferase\n",
      "Cellobiose phosphorylase\n",
      "5-methyltetrahydrofolate S-homocysteine methyltransferase\n",
      "5-methyltetrahydrofolate--homocysteine methyltransferase\n",
      "preprotein translocase subunit SecA\n",
      "preprotein translocase SecA subunit\n",
      "Protein export cytoplasm protein SecA ATPase RNA helicase\n",
      "glycyl-tRNA synthetase, beta subunit\n",
      "glycyl-tRNA synthetase subunit beta\n",
      "Glycyl-tRNA synthetase beta chain\n",
      "lipopolysaccharide biosynthesis protein, putative\n",
      "ABC transporter\n",
      "UDP-N-acetylglucosamine 1-carboxyvinyltransferase\n",
      "CRISPR-associated protein Csm1\n",
      "CRISPR-associated protein, Csm1 family\n",
      "oligopeptide ABC transporter, ATP-binding protein\n",
      "Xylobiose ABC transport system, ATP-binding protein 1\n",
      "GTP-binding protein LepA\n",
      "Translation elongation factor LepA\n",
      "ABC transporter, ATP-binding protein\n",
      "Lipid A export ATP-binding/permease protein MsbA\n",
      "multidrug ABC transporter ATP-binding protein\n",
      "DNA topoisomerase\n",
      "DNA topoisomerase I\n",
      "ATP-dependent protease LA, putative\n",
      "ATP-dependent protease\n",
      "ATP-dependent protease La Type II\n",
      "5-methyltetrahydropteroyltriglutamate--homocysteine methyltransferase\n",
      "alcohol dehydrogenase, zinc-containing\n",
      "L-threonine 3-dehydrogenase\n",
      "alanyl-tRNA synthetase\n",
      "alanyl-tRNA synthase\n",
      "dehydrogenase\n",
      "oxidoreductase\n",
      "Myo-inositol 2-dehydrogenase 1\n",
      "FAD dependent oxidoreductase\n",
      "putative electron transfer flavoprotein-quinone oxidoreductase FixC\n",
      "flagellar M-ring protein FliF\n",
      "flagellar M-ring protein\n",
      "alpha-glucuronidase\n",
      "lepA protein\n",
      "flagellar hook-associated protein 1\n",
      "flagellar hook-associated protein FlgK\n",
      "flagellar hook protein FlgK\n",
      "ATPase, AAA family\n",
      "flagellar biosynthesis protein FlhA\n",
      "alcohol dehydrogenase\n",
      "Acetyl-coenzyme A carboxyl transferase alpha chain / Acetyl-coenzyme A carboxyl transferase beta chain ; Propionyl-CoA carboxylase beta chain\n",
      "methylmalonyl-CoA carboxyltransferase\n",
      "protein-export membrane protein SecD, putative\n",
      "Protein-export membrane protein SecD\n",
      "preprotein translocase subunit SecD\n",
      "maltodextrin glycosyltransferase\n",
      "membrane protein\n",
      "transcription termination factor Rho\n",
      "MFS transporter\n",
      "Permease of the major facilitator superfamily\n",
      "3-phosphoshikimate 1-carboxyvinyltransferase\n",
      "ATP-dependent protease LA\n",
      "ATP-dependent protease La Type I\n",
      "DNA-binding protein\n",
      "DNA-directed RNA polymerase subunit beta'\n",
      "pyruvate kinase\n",
      "Xyloglucan ABC transport system, sugar-binding protein\n",
      "glucosamine--fructose-6-phosphate aminotransferase\n",
      "Glucosamine--fructose-6-phosphate aminotransferase [isomerizing\n",
      "phenylalanyl-tRNA synthase subunit alpha\n",
      "phenylalanine--tRNA ligase\n",
      "glycine dehydrogenase subunit 1\n",
      "glycine dehydrogenase\n",
      "transcription-repair coupling factor\n",
      "aspartate aminotransferase\n"
     ]
    }
   ],
   "source": [
    "original_proteins_2336 = retrieve_orginal_proteins(\n",
    "    graph,\n",
    "    species_taxid=2336,\n",
    "    include_neighbours=False,\n",
    ")\n",
    "print(f'{len(original_proteins_2336):,} proteins')\n",
    "\n",
    "for p in [t[0] for t in original_proteins_2336]:\n",
    "    print(p)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "#for v in [t[0] for t in original_proteins_2336 if t[1] < 0.1][:51]:\n",
    "#    print(v)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "164 proteins\n",
      "aconitate hydratase 1\n",
      "alpha2-macroglobulin\n",
      "exodeoxyribonuclease V subunit RecB\n",
      "bacteriophage N4 receptor, outer membrane protein\n",
      "putative outer membrane protein, N-terminal fragment\n",
      "ribonucleoside-diphosphate reductase 2 subunit alpha\n",
      "apo-serine activating enzyme\n",
      "alpha-mannosidase\n",
      "peptidase M16 family protein\n",
      "inner membrane protein - inhibits the Rcs signaling pathway\n",
      "biotin sulfoxide reductase\n",
      "putative selenate reductase YnfE\n",
      "multidrug efflux pump RND permease AcrF\n",
      "DUF5107 domain-containing protein YphG\n",
      "3-hydroxyacyl-CoA dehydrogenase FadJ\n",
      "peptidoglycan DD-transpeptidase FtsI\n",
      "transcription-repair coupling factor\n",
      "protein-tyrosine kinase Etk\n",
      "putative selenate reductase YnfF\n",
      "putative TonB-dependent outer membrane receptor\n",
      "exodeoxyribonuclease V subunit RecC\n",
      "putative xanthine dehydrogenase molybdenum-binding subunit XdhA\n",
      "fused DNA-binding transcriptional repressor/proline dehydrogenase/1-pyrroline-5-carboxylate dehydrogenase PutA\n",
      "DNA mismatch repair protein MutS\n",
      "ATP dependent, structure specific DNA nuclease - SbcC subunit\n",
      "putative transporter YhfK\n",
      "polyphosphate kinase\n",
      "bacteriophage N4 receptor, inner membrane subunit\n",
      "fumarase C\n",
      "soluble pyridine nucleotide transhydrogenase\n",
      "putative ATP-dependent protease YcbZ\n",
      "flagellar biosynthesis protein FlhA\n",
      "regulator of CsrB and CsrC decay\n",
      "oxepin-CoA hydrolase/3-oxo-5,6-dehydrosuberyl-CoA semialdehyde dehydrogenase\n",
      "DUF4132 domain-containing protein YehI\n",
      "multidrug efflux pump RND permease MdtF\n",
      "Type II secretion system protein GspD\n",
      "putative pyruvate-flavodoxin oxidoreductase\n",
      "sensory histidine kinase NarQ\n",
      "intermembrane transport protein YebT\n",
      "N-acetylglutamate synthase\n",
      "AsmA famly protein YicH\n",
      "cellulose phosphoethanolamine transferase\n",
      "carbamoyltransferase HypF\n",
      "beta-galactosidase\n",
      "alpha-D-xyloside xylohydrolase\n",
      "multidrug efflux pump RND permease AcrD\n",
      "beta-glucoside specific PTS enzyme II/BglG kinase/BglG phosphatase\n",
      "1,4-alpha-glucan branching enzyme\n",
      "phosphoglycerol transferase I\n",
      "negative regulator of MalT activity/cystathionine beta-lyase\n",
      "acyltransferase\n",
      "adenylate cyclase\n",
      "ornithine decarboxylase, degradative\n",
      "kojibiose phosphorylase\n",
      "DNA replication terminus site-binding protein\n",
      "DNA-binding transcriptional activator HyfR\n",
      "osmoregulated periplasmic glucans (OPGs) biosynthesis protein H\n",
      "cell division DNA translocase FtsK\n",
      "putative aminohydrolase\n",
      "DUF3029 domain-containing protein YjjI\n",
      "DNA-binding transcriptional activator FhlA\n",
      "multidrug efflux pump RND permease subunit MdtC\n",
      "sensory histidine kinase BarA\n",
      "RcsD phosphotransferase\n",
      "putative autotransporter adhesin YfaL\n",
      "phosphoethanolamine transferase\n",
      "putative oxidoreductase\n",
      "Zn(2(+))/Cd(2(+))/Pb(2(+)) exporting P-type ATPase\n",
      "Cu(+) exporting P-type ATPase\n",
      "1-deoxyxylulose-5-phosphate synthase YajO\n",
      "putative K(+):H(+) antiporter\n",
      "sensory histidine kinase ArcB\n",
      "sensory histidine kinase TorS\n",
      "DNA helicase IV\n",
      "fused indole-3-glycerol phosphate synthase/phosphoribosylanthranilate isomerase\n",
      "putative lipoprotein GfcD\n",
      "fused putative oxidoreductase, Fe-S subunit and nucleotide-binding subunit\n",
      "putative transporter YfcC\n",
      "putative NAD(P)-dependent oxidoreductase YbjT\n",
      "putative transporter YccS\n",
      "selenocysteyl-tRNA-specific translation elongation factor\n",
      "ATP-dependent RNA helicase HrpA\n",
      "carnitine-CoA ligase\n",
      "DUF418 domain-containing protein YeiB\n",
      "membrane-bound lytic murein transglycosylase F\n",
      "N-succinylarginine dihydrolase\n",
      "aromatic carboxylic acid efflux pump subunit AaeB\n",
      "putative diguanylate cyclase DgcE\n",
      "toxin YjjJ\n",
      "3-ketoacyl-CoA thiolase FadI\n",
      "carbamoyl phosphate synthetase subunit alpha\n",
      "DNA-binding transcriptional dual regulator ArgP\n",
      "putative DNA repair helicase RadD\n",
      "dimethyl sulfoxide reductase subunit A\n",
      "multidrug efflux pump RND permease subunit MdtB\n",
      "putative hydratase YbhJ\n",
      "trimethylamine N-oxide reductase 2\n",
      "DUF1479 domain-containing protein YbiU\n",
      "putative uncharacterized protein YjiV\n",
      "putative diguanylate cyclase DgcQ\n",
      "DNA-binding transcriptional activator/ornithine decarboxylase inhibitor AtoC\n",
      "metal phosphate:H(+) symporter PitB\n",
      "chaperone protein HscC\n",
      "ClpB chaperone\n",
      "Fe-S cluster scaffold complex subunit SufB\n",
      "L-tartrate:succinate antiporter\n",
      "inverse autotransporter adhesin\n",
      "sensory histidine kinase KdpD\n",
      "putative 5' to 3' DNA helicase implicated in DNA repair\n",
      "putative TonB-dependent receptor\n",
      "putative formate acetyltransferase 2\n",
      "putative oxidoreductase, Fe-S subunit\n",
      "putative PTS enzyme IIC component SgcC\n",
      "glycerol-3-phosphate dehydrogenase\n",
      "putative oxidoreductase YdeP\n",
      "hydrogenase 4 component B\n",
      "putative FAD-linked oxidoreductase\n",
      "fatty acyl-CoA synthetase\n",
      "exopolyphosphatase\n",
      "peptidoglycan DD-transpeptidase MrdA\n",
      "protein ViaA\n",
      "putative c-di-GMP phosphodiesterase PdeK\n",
      "primosome factor N'\n",
      "RNase D\n",
      "poly-N-acetyl-D-glucosamine synthase subunit PgaC\n",
      "DNA-binding transcriptional dual regulator GalR\n",
      "copper-containing amine oxidase\n",
      "putative D,D-dipeptide ABC transporter periplasmic binding protein\n",
      "putative fimbrial usher protein YbgQ\n",
      "putative multidrug efflux pump outer membrane channel\n",
      "phosphatidylserine synthase\n",
      "fermentation-respiration switch protein\n",
      "tail-specific protease\n",
      "exodeoxyribonuclease V subunit RecD\n",
      "DNA-binding transcriptional activator MalT\n",
      "NADP(+)-dependent aldehyde reductase\n",
      "dGTP triphosphohydrolase\n",
      "acetohydroxy acid synthase I subunit IlvB\n",
      "malate synthase A\n",
      "ubiquinone biosynthesis protein UbiB\n",
      "23S rRNA 5-hydroxycytidine C2501 synthase\n",
      "fused putative xanthine/hypoxanthine oxidase: molybdopterin-binding subunit and Fe-S binding subunit\n",
      "putative major facilitator superfamily transport protein YajR\n",
      "putative xyloside transporter YicJ\n",
      "AsmA2 domain-containing protein\n",
      "aspartate-semialdehyde dehydrogenase\n",
      "peptidoglycan DD-endopeptidase MepM\n",
      "putative iron siderophore outer membrane transporter\n",
      "radical SAM superfamily protein YgiQ\n",
      "fused chorismate mutase/prephenate dehydratase\n",
      "peptidyl-lysine acetyltransferase\n",
      "putative hydrolase YieL\n",
      "fused aspartate kinase/homoserine dehydrogenase 1\n",
      "DNA topoisomerase III\n",
      "putative c-di-GMP phosphodiesterase PdeD\n",
      "oligopeptide ABC transporter periplasmic binding protein\n",
      "D-amino acid dehydrogenase\n",
      "transaldolase A\n",
      "DNA polymerase III subunit tau\n",
      "phosphoenolpyruvate-protein phosphotransferase PtsP\n",
      "putative transporter YdhK\n",
      "outer membrane channel TolC\n",
      "inorganic triphosphatase\n"
     ]
    }
   ],
   "source": [
    "original_proteins_562 = retrieve_orginal_proteins(\n",
    "    graph,\n",
    "    species_taxid=562,\n",
    "    include_neighbours=False,\n",
    ")\n",
    "print(f'{len(original_proteins_562):,} proteins')\n",
    "for p in [t[0] for t in original_proteins_562]:\n",
    "    print(p)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "79 proteins\n",
      "FAD linked oxidase-like protein\n",
      "Large subunit of NADH-dependent glutamate synthase\n",
      "HAE1 efflux family protein\n",
      "oxidoreductase alpha (molybdopterin) subunit\n",
      "ATP-dependent DNA helicase\n",
      "DNA polymerase I\n",
      "glycine dehydrogenase\n",
      "probable glycosyl hydrolase\n",
      "probable periplasmic tail-specific proteinase\n",
      "bifunctional GMP synthase/glutamine amidotransferase protein\n",
      "putative DNA polymerase III alpha subunit\n",
      "ATP-dependent Clp protease, ATP-binding subunit ClpC\n",
      "phosphoribosylformylglycinamidine synthase\n",
      "DNA mismatch repair protein\n",
      "isoleucyl-tRNA synthetase\n",
      "TonB-dependent siderophore receptor\n",
      "putative NADP-dependent malic enzyme\n",
      "probable long chain fatty-acid CoA ligase\n",
      "glutamine synthetase\n",
      "AcrB/AcrD family multidrug resistance protein\n",
      "putative chloride channel protein\n",
      "DNA topoisomerase IV subunit A\n",
      "methionyl-tRNA synthetase\n",
      "putative protease\n",
      "ATP-dependent helicase\n",
      "sensory transduction histidine kinase\n",
      "acetoin dehydrogenase (TPP-dependent) beta chain\n",
      "aminoacyl-histidine dipeptidase\n",
      "FixL-related histidine kinase\n",
      "ferric enterobactin receptor\n",
      "penicillin-binding protein 1A\n",
      "GTP-binding protein LepA\n",
      "Secreted protein containing N-terminal zinc-dependent carboxypeptidase related domain\n",
      "primosomal protein N'\n",
      "aspartate kinase III\n",
      "hypothetical protein\n",
      "aminopeptidase M1 family protein\n",
      "Penicillin amidase\n",
      "putative outer membrane protein, probably involved in nutrient binding\n",
      "DNA topoisomerase III\n",
      "phosphoglucomutase phosphomannomutase\n",
      "molybdopterin oxidoreductase, iron-sulfur binding subunit\n",
      "putative outer membrane receptor protein\n",
      "two-component regulatory system, sensor kinase protein\n",
      "ribonuclease R\n",
      "putative FtsK/SpoIIIE-like protein\n",
      "2-succinyl-6-hydroxy-2,4-cyclohexadiene-1-carboxylic acid synthase/2-oxoglutarate decarboxylase\n",
      "alanyl-tRNA synthetase\n",
      "hypothetical carboxypeptidase\n",
      "putative peptidase/protease family protein\n",
      "phosphate acetyltransferase\n",
      "putative ATP-dependent DNA helicase\n",
      "Alpha-glucosidase, family 31 of glycosyl hydrolase\n",
      "GTP pyrophosphokinase\n",
      "putative DNA ligase\n",
      "pyruvate carboxylase\n",
      "methylmalonyl-CoA mutase\n",
      "aspartyl-tRNA synthetase\n",
      "glutaminyl-tRNA synthetase\n",
      "glutaryl-7-ACA acylase precursor\n",
      "Nitrilase/cyanide hydratase and apolipoprotein N-acyltransferase\n",
      "isocitrate dehydrogenase, NADP\n",
      "histidine kinase sensor protein\n",
      "putative outer membrane protein\n",
      "molecular chaperone DnaK\n",
      "beta-glucanase precursor\n",
      "isoquinoline 1-oxidoreductase, beta subunit\n",
      "glycosyl hydrolase\n",
      "FKBP-type peptidyl-prolyl cis-transisomerase\n",
      "Amidohydrolase family, putative\n",
      "aminopeptidase\n",
      "TonB-dependent receptor, putative\n",
      "phosphoenolpyruvate carboxylase\n",
      "excinuclease ABC subunit A\n",
      "putative capsular polysaccharide biosynthesis protein\n",
      "subtilisin-like serine protease\n",
      "putative auxin-regulated protein\n",
      "2-isopropylmalate synthase\n",
      "acyl-CoA dehydrogenase\n"
     ]
    }
   ],
   "source": [
    "original_proteins_531 = retrieve_orginal_proteins(\n",
    "    graph,\n",
    "    species_taxid=531,\n",
    "    include_neighbours=False,\n",
    ")\n",
    "print(f'{len(original_proteins_531):,} proteins')\n",
    "for p in [t[0] for t in original_proteins_531]:\n",
    "    print(p)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## tRNA expression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "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>in_test_set</th>\n",
       "      <th>AAA_ratio</th>\n",
       "      <th>AAG_ratio</th>\n",
       "      <th>AAT_ratio</th>\n",
       "      <th>AAC_ratio</th>\n",
       "      <th>ACT_ratio</th>\n",
       "      <th>ACC_ratio</th>\n",
       "      <th>ACA_ratio</th>\n",
       "      <th>ACG_ratio</th>\n",
       "      <th>CGT_ratio</th>\n",
       "      <th>...</th>\n",
       "      <th>motility</th>\n",
       "      <th>range_salinity</th>\n",
       "      <th>cell_shape</th>\n",
       "      <th>isolation_source</th>\n",
       "      <th>doubling_h</th>\n",
       "      <th>genome_size</th>\n",
       "      <th>gc_content</th>\n",
       "      <th>coding_genes</th>\n",
       "      <th>tRNA_genes</th>\n",
       "      <th>rRNA16S_genes</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>species_taxid</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>True</td>\n",
       "      <td>0.082550</td>\n",
       "      <td>0.917450</td>\n",
       "      <td>0.327724</td>\n",
       "      <td>0.672276</td>\n",
       "      <td>0.024511</td>\n",
       "      <td>0.575353</td>\n",
       "      <td>0.036262</td>\n",
       "      <td>0.363875</td>\n",
       "      <td>0.078732</td>\n",
       "      <td>...</td>\n",
       "      <td>yes</td>\n",
       "      <td>None</td>\n",
       "      <td>None</td>\n",
       "      <td>host_plant</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5369771.500</td>\n",
       "      <td>67.300</td>\n",
       "      <td>4713.667</td>\n",
       "      <td>53.000</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>False</td>\n",
       "      <td>0.919627</td>\n",
       "      <td>0.080373</td>\n",
       "      <td>0.858168</td>\n",
       "      <td>0.141832</td>\n",
       "      <td>0.454569</td>\n",
       "      <td>0.050026</td>\n",
       "      <td>0.446018</td>\n",
       "      <td>0.049387</td>\n",
       "      <td>0.325065</td>\n",
       "      <td>...</td>\n",
       "      <td>yes</td>\n",
       "      <td>None</td>\n",
       "      <td>coccobacillus</td>\n",
       "      <td>host_animal_ectotherm</td>\n",
       "      <td>35.40</td>\n",
       "      <td>601699.243</td>\n",
       "      <td>25.469</td>\n",
       "      <td>517.549</td>\n",
       "      <td>30.485</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>True</td>\n",
       "      <td>0.007302</td>\n",
       "      <td>0.992698</td>\n",
       "      <td>0.010730</td>\n",
       "      <td>0.989270</td>\n",
       "      <td>0.008586</td>\n",
       "      <td>0.425527</td>\n",
       "      <td>0.015789</td>\n",
       "      <td>0.550098</td>\n",
       "      <td>0.092597</td>\n",
       "      <td>...</td>\n",
       "      <td>yes</td>\n",
       "      <td>None</td>\n",
       "      <td>bacillus</td>\n",
       "      <td>host_animal_endotherm</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3526440.800</td>\n",
       "      <td>73.805</td>\n",
       "      <td>3139.333</td>\n",
       "      <td>45.000</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>True</td>\n",
       "      <td>0.664866</td>\n",
       "      <td>0.335134</td>\n",
       "      <td>0.775571</td>\n",
       "      <td>0.224429</td>\n",
       "      <td>0.458203</td>\n",
       "      <td>0.188052</td>\n",
       "      <td>0.304183</td>\n",
       "      <td>0.049563</td>\n",
       "      <td>0.035122</td>\n",
       "      <td>...</td>\n",
       "      <td>no</td>\n",
       "      <td>None</td>\n",
       "      <td>bacillus</td>\n",
       "      <td>water_hotspring</td>\n",
       "      <td>2.47</td>\n",
       "      <td>1959987.600</td>\n",
       "      <td>33.700</td>\n",
       "      <td>1876.333</td>\n",
       "      <td>46.000</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>True</td>\n",
       "      <td>0.551810</td>\n",
       "      <td>0.448190</td>\n",
       "      <td>0.440559</td>\n",
       "      <td>0.559441</td>\n",
       "      <td>0.099134</td>\n",
       "      <td>0.591478</td>\n",
       "      <td>0.097796</td>\n",
       "      <td>0.211592</td>\n",
       "      <td>0.204394</td>\n",
       "      <td>...</td>\n",
       "      <td>yes</td>\n",
       "      <td>None</td>\n",
       "      <td>bacillus</td>\n",
       "      <td>petroleum</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3722544.667</td>\n",
       "      <td>55.100</td>\n",
       "      <td>3222.667</td>\n",
       "      <td>54.000</td>\n",
       "      <td>2.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 85 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "               in_test_set  AAA_ratio  AAG_ratio  AAT_ratio  AAC_ratio  \\\n",
       "species_taxid                                                            \n",
       "7                     True   0.082550   0.917450   0.327724   0.672276   \n",
       "9                    False   0.919627   0.080373   0.858168   0.141832   \n",
       "11                    True   0.007302   0.992698   0.010730   0.989270   \n",
       "14                    True   0.664866   0.335134   0.775571   0.224429   \n",
       "19                    True   0.551810   0.448190   0.440559   0.559441   \n",
       "\n",
       "               ACT_ratio  ACC_ratio  ACA_ratio  ACG_ratio  CGT_ratio  ...  \\\n",
       "species_taxid                                                         ...   \n",
       "7               0.024511   0.575353   0.036262   0.363875   0.078732  ...   \n",
       "9               0.454569   0.050026   0.446018   0.049387   0.325065  ...   \n",
       "11              0.008586   0.425527   0.015789   0.550098   0.092597  ...   \n",
       "14              0.458203   0.188052   0.304183   0.049563   0.035122  ...   \n",
       "19              0.099134   0.591478   0.097796   0.211592   0.204394  ...   \n",
       "\n",
       "               motility  range_salinity     cell_shape       isolation_source  \\\n",
       "species_taxid                                                                   \n",
       "7                   yes            None           None             host_plant   \n",
       "9                   yes            None  coccobacillus  host_animal_ectotherm   \n",
       "11                  yes            None       bacillus  host_animal_endotherm   \n",
       "14                   no            None       bacillus        water_hotspring   \n",
       "19                  yes            None       bacillus              petroleum   \n",
       "\n",
       "               doubling_h  genome_size  gc_content  coding_genes  tRNA_genes  \\\n",
       "species_taxid                                                                  \n",
       "7                     NaN  5369771.500      67.300      4713.667      53.000   \n",
       "9                   35.40   601699.243      25.469       517.549      30.485   \n",
       "11                    NaN  3526440.800      73.805      3139.333      45.000   \n",
       "14                   2.47  1959987.600      33.700      1876.333      46.000   \n",
       "19                    NaN  3722544.667      55.100      3222.667      54.000   \n",
       "\n",
       "               rRNA16S_genes  \n",
       "species_taxid                 \n",
       "7                        3.0  \n",
       "9                        1.0  \n",
       "11                       2.0  \n",
       "14                       2.0  \n",
       "19                       2.0  \n",
       "\n",
       "[5 rows x 85 columns]"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "species_codon_ratios_path = os.path.join(os.getcwd(), 'data/species_codon_ratios.csv')\n",
    "species_codon_df = load_codon_bias(engine, species_codon_ratios_path).set_index('species_taxid')\n",
    "species_codon_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "ratio_columns = [c for c in species_codon_df.columns if c.endswith('_ratio')]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_trnas(engine, assembly_accession):\n",
    "    trna_query = \"\"\"\n",
    "    select * from sequences where sequence_type = 'tRNA' and assembly_accession = ?\n",
    "    \"\"\"\n",
    "    trna_df = pd.read_sql(trna_query, engine, params=(assembly_accession,))\n",
    "    metadata = [json.loads(v) if v is not None else {} for v in trna_df['metadata_json'].values]\n",
    "    trna_df['codon'] = [v.get('codon') for v in metadata]\n",
    "    trna_df['anticodon'] = [v.get('anticodon') for v in metadata]\n",
    "    trna_df['amino_acid'] = [v.get('amino_acid') for v in metadata]\n",
    "    return trna_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "86 tRNA records\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>assembly_accession</th>\n",
       "      <th>species_taxid</th>\n",
       "      <th>sequence_type</th>\n",
       "      <th>chromosome_id</th>\n",
       "      <th>location_json</th>\n",
       "      <th>strand</th>\n",
       "      <th>length</th>\n",
       "      <th>description</th>\n",
       "      <th>metadata_json</th>\n",
       "      <th>sequence</th>\n",
       "      <th>codon</th>\n",
       "      <th>anticodon</th>\n",
       "      <th>amino_acid</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>GCA_000005845.2</td>\n",
       "      <td>562</td>\n",
       "      <td>tRNA</td>\n",
       "      <td>U00096.3</td>\n",
       "      <td>[[225381, 225457]]</td>\n",
       "      <td>+</td>\n",
       "      <td>77</td>\n",
       "      <td>lcl|U00096.3_trna_5 [gene=ileV] [locus_tag=b02...</td>\n",
       "      <td>{\"amino_acid\": \"Ile\", \"anticodon\": \"GAT\", \"cod...</td>\n",
       "      <td>AGGCTTGTAGCTCAGGTGGTTAGAGCGCACCCCTGATAAGGGTGAG...</td>\n",
       "      <td>ATC</td>\n",
       "      <td>GAT</td>\n",
       "      <td>Ile</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>GCA_000005845.2</td>\n",
       "      <td>562</td>\n",
       "      <td>tRNA</td>\n",
       "      <td>U00096.3</td>\n",
       "      <td>[[225500, 225575]]</td>\n",
       "      <td>+</td>\n",
       "      <td>76</td>\n",
       "      <td>lcl|U00096.3_trna_6 [gene=alaV] [locus_tag=b02...</td>\n",
       "      <td>{\"amino_acid\": \"Ala\", \"anticodon\": \"TGC\", \"cod...</td>\n",
       "      <td>GGGGCTATAGCTCAGCTGGGAGAGCGCCTGCTTTGCACGCAGGAGG...</td>\n",
       "      <td>GCA</td>\n",
       "      <td>TGC</td>\n",
       "      <td>Ala</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>GCA_000005845.2</td>\n",
       "      <td>562</td>\n",
       "      <td>tRNA</td>\n",
       "      <td>U00096.3</td>\n",
       "      <td>[[228928, 229004]]</td>\n",
       "      <td>+</td>\n",
       "      <td>77</td>\n",
       "      <td>lcl|U00096.3_trna_9 [gene=aspU] [locus_tag=b02...</td>\n",
       "      <td>{\"amino_acid\": \"Asp\", \"anticodon\": \"GTC\", \"cod...</td>\n",
       "      <td>GGAGCGGTAGTTCAGTCGGTTAGAATACCTGCCTGTCACGCAGGGG...</td>\n",
       "      <td>GAC</td>\n",
       "      <td>GTC</td>\n",
       "      <td>Asp</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>GCA_000005845.2</td>\n",
       "      <td>562</td>\n",
       "      <td>tRNA</td>\n",
       "      <td>U00096.3</td>\n",
       "      <td>[[236931, 237007]]</td>\n",
       "      <td>+</td>\n",
       "      <td>77</td>\n",
       "      <td>lcl|U00096.3_trna_10 [gene=aspV] [locus_tag=b0...</td>\n",
       "      <td>{\"amino_acid\": \"Asp\", \"anticodon\": \"GTC\", \"cod...</td>\n",
       "      <td>GGAGCGGTAGTTCAGTCGGTTAGAATACCTGCCTGTCACGCAGGGG...</td>\n",
       "      <td>GAC</td>\n",
       "      <td>GTC</td>\n",
       "      <td>Asp</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>GCA_000005845.2</td>\n",
       "      <td>562</td>\n",
       "      <td>tRNA</td>\n",
       "      <td>U00096.3</td>\n",
       "      <td>[[262871, 262946]]</td>\n",
       "      <td>+</td>\n",
       "      <td>76</td>\n",
       "      <td>lcl|U00096.3_trna_11 [gene=thrW] [locus_tag=b0...</td>\n",
       "      <td>{\"amino_acid\": \"Thr\", \"anticodon\": \"CGT\", \"cod...</td>\n",
       "      <td>GCCGATATAGCTCAGTTGGTAGAGCAGCGCATTCGTAATGCGAAGG...</td>\n",
       "      <td>ACG</td>\n",
       "      <td>CGT</td>\n",
       "      <td>Thr</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  assembly_accession  species_taxid sequence_type chromosome_id  \\\n",
       "0    GCA_000005845.2            562          tRNA      U00096.3   \n",
       "1    GCA_000005845.2            562          tRNA      U00096.3   \n",
       "2    GCA_000005845.2            562          tRNA      U00096.3   \n",
       "3    GCA_000005845.2            562          tRNA      U00096.3   \n",
       "4    GCA_000005845.2            562          tRNA      U00096.3   \n",
       "\n",
       "        location_json strand  length  \\\n",
       "0  [[225381, 225457]]      +      77   \n",
       "1  [[225500, 225575]]      +      76   \n",
       "2  [[228928, 229004]]      +      77   \n",
       "3  [[236931, 237007]]      +      77   \n",
       "4  [[262871, 262946]]      +      76   \n",
       "\n",
       "                                         description  \\\n",
       "0  lcl|U00096.3_trna_5 [gene=ileV] [locus_tag=b02...   \n",
       "1  lcl|U00096.3_trna_6 [gene=alaV] [locus_tag=b02...   \n",
       "2  lcl|U00096.3_trna_9 [gene=aspU] [locus_tag=b02...   \n",
       "3  lcl|U00096.3_trna_10 [gene=aspV] [locus_tag=b0...   \n",
       "4  lcl|U00096.3_trna_11 [gene=thrW] [locus_tag=b0...   \n",
       "\n",
       "                                       metadata_json  \\\n",
       "0  {\"amino_acid\": \"Ile\", \"anticodon\": \"GAT\", \"cod...   \n",
       "1  {\"amino_acid\": \"Ala\", \"anticodon\": \"TGC\", \"cod...   \n",
       "2  {\"amino_acid\": \"Asp\", \"anticodon\": \"GTC\", \"cod...   \n",
       "3  {\"amino_acid\": \"Asp\", \"anticodon\": \"GTC\", \"cod...   \n",
       "4  {\"amino_acid\": \"Thr\", \"anticodon\": \"CGT\", \"cod...   \n",
       "\n",
       "                                            sequence codon anticodon  \\\n",
       "0  AGGCTTGTAGCTCAGGTGGTTAGAGCGCACCCCTGATAAGGGTGAG...   ATC       GAT   \n",
       "1  GGGGCTATAGCTCAGCTGGGAGAGCGCCTGCTTTGCACGCAGGAGG...   GCA       TGC   \n",
       "2  GGAGCGGTAGTTCAGTCGGTTAGAATACCTGCCTGTCACGCAGGGG...   GAC       GTC   \n",
       "3  GGAGCGGTAGTTCAGTCGGTTAGAATACCTGCCTGTCACGCAGGGG...   GAC       GTC   \n",
       "4  GCCGATATAGCTCAGTTGGTAGAGCAGCGCATTCGTAATGCGAAGG...   ACG       CGT   \n",
       "\n",
       "  amino_acid  \n",
       "0        Ile  \n",
       "1        Ala  \n",
       "2        Asp  \n",
       "3        Asp  \n",
       "4        Thr  "
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trna_df = load_trnas(engine, 'GCA_000005845.2')\n",
    "print(len(trna_df), 'tRNA records')\n",
    "trna_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## tRNA adaptation index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "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>assembly_accession</th>\n",
       "      <th>species_taxid</th>\n",
       "      <th>row_id</th>\n",
       "      <th>protein_id</th>\n",
       "      <th>protein</th>\n",
       "      <th>gene</th>\n",
       "      <th>adaptation_index</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>GCA_000005845.2</td>\n",
       "      <td>562</td>\n",
       "      <td>255469</td>\n",
       "      <td>AAC74347.1</td>\n",
       "      <td>trp operon leader peptide</td>\n",
       "      <td>trpL</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>GCA_000005845.2</td>\n",
       "      <td>562</td>\n",
       "      <td>258414</td>\n",
       "      <td>ABP93450.1</td>\n",
       "      <td>uncharacterized protein YpfM</td>\n",
       "      <td>ypfM</td>\n",
       "      <td>0.858984</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>GCA_000005845.2</td>\n",
       "      <td>562</td>\n",
       "      <td>257893</td>\n",
       "      <td>AAC77015.1</td>\n",
       "      <td>putative stress response protein</td>\n",
       "      <td>yjbJ</td>\n",
       "      <td>0.848080</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>GCA_000005845.2</td>\n",
       "      <td>562</td>\n",
       "      <td>254740</td>\n",
       "      <td>AAC73573.1</td>\n",
       "      <td>putative nucleoid-associated protein YbaB</td>\n",
       "      <td>ybaB</td>\n",
       "      <td>0.812893</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>GCA_000005845.2</td>\n",
       "      <td>562</td>\n",
       "      <td>257793</td>\n",
       "      <td>AAC76910.1</td>\n",
       "      <td>cell division factor ZapB</td>\n",
       "      <td>zapB</td>\n",
       "      <td>0.775375</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  assembly_accession  species_taxid  row_id  protein_id  \\\n",
       "0    GCA_000005845.2            562  255469  AAC74347.1   \n",
       "1    GCA_000005845.2            562  258414  ABP93450.1   \n",
       "2    GCA_000005845.2            562  257893  AAC77015.1   \n",
       "3    GCA_000005845.2            562  254740  AAC73573.1   \n",
       "4    GCA_000005845.2            562  257793  AAC76910.1   \n",
       "\n",
       "                                     protein  gene  adaptation_index  \n",
       "0                  trp operon leader peptide  trpL          1.000000  \n",
       "1               uncharacterized protein YpfM  ypfM          0.858984  \n",
       "2           putative stress response protein  yjbJ          0.848080  \n",
       "3  putative nucleoid-associated protein YbaB  ybaB          0.812893  \n",
       "4                  cell division factor ZapB  zapB          0.775375  "
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trna_ai_path = os.path.join(os.getcwd(), f'data/trn_adaptation_index/GCA_000005845.2_tai.csv')\n",
    "trna_ai = pd.read_csv(trna_ai_path)\n",
    "trna_ai.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "adaptation_ix = trna_ai['adaptation_index'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "_, ax = plt.subplots(1, 1, figsize=(12, 6))\n",
    "ax.hist(adaptation_ix, bins=100);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "matching_df = trna_ai[\n",
    "    trna_ai['protein'].isin(np.unique([t[0] for t in original_proteins_562]))\n",
    "]\n",
    "adaptation_ix_top_codon = matching_df['adaptation_index'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "_, ax2 = plt.subplots(1, 1, figsize=(12, 6))\n",
    "ax2.hist(adaptation_ix_top_codon, bins=20);\n",
    "ax2.set_xlim(ax.get_xlim());"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "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>protein</th>\n",
       "      <th>adaptation_index</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>677</th>\n",
       "      <td>ClpB chaperone</td>\n",
       "      <td>0.616203</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>720</th>\n",
       "      <td>malate synthase A</td>\n",
       "      <td>0.613474</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>817</th>\n",
       "      <td>peptidoglycan DD-transpeptidase MrdA</td>\n",
       "      <td>0.607896</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>819</th>\n",
       "      <td>tail-specific protease</td>\n",
       "      <td>0.607810</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>828</th>\n",
       "      <td>copper-containing amine oxidase</td>\n",
       "      <td>0.607257</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3616</th>\n",
       "      <td>putative transporter YdhK</td>\n",
       "      <td>0.493055</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3723</th>\n",
       "      <td>fused putative oxidoreductase, Fe-S subunit an...</td>\n",
       "      <td>0.486201</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3856</th>\n",
       "      <td>exodeoxyribonuclease V subunit RecD</td>\n",
       "      <td>0.476493</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3878</th>\n",
       "      <td>DNA-binding transcriptional activator HyfR</td>\n",
       "      <td>0.474149</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4144</th>\n",
       "      <td>putative oxidoreductase</td>\n",
       "      <td>0.440617</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>172 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                protein  adaptation_index\n",
       "677                                      ClpB chaperone          0.616203\n",
       "720                                   malate synthase A          0.613474\n",
       "817                peptidoglycan DD-transpeptidase MrdA          0.607896\n",
       "819                              tail-specific protease          0.607810\n",
       "828                     copper-containing amine oxidase          0.607257\n",
       "...                                                 ...               ...\n",
       "3616                          putative transporter YdhK          0.493055\n",
       "3723  fused putative oxidoreductase, Fe-S subunit an...          0.486201\n",
       "3856                exodeoxyribonuclease V subunit RecD          0.476493\n",
       "3878         DNA-binding transcriptional activator HyfR          0.474149\n",
       "4144                            putative oxidoreductase          0.440617\n",
       "\n",
       "[172 rows x 2 columns]"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "matching_df[['protein', 'adaptation_index']]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Compute mean diff "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_adaptation_mean_diff(assembly_accessions):\n",
    "    cached_proteins = {}\n",
    "    \n",
    "    diff = []\n",
    "    for i, assembly_accession in enumerate(assembly_accessions):\n",
    "        if (i+1) % 100 == 0:\n",
    "            print(f'{i+1:,} / {len(assembly_accessions):,}')\n",
    "        \n",
    "        trna_ai_path = os.path.join(os.getcwd(), f'data/trn_adaptation_index/{assembly_accession}_tai.csv')\n",
    "        trna_ai = pd.read_csv(trna_ai_path)\n",
    "        \n",
    "        species_taxid = trna_ai['species_taxid'].iloc[0]\n",
    "    \n",
    "        if species_taxid in cached_proteins:\n",
    "            proteins = cached_proteins[species_taxid]\n",
    "        else:\n",
    "            proteins = retrieve_orginal_proteins(\n",
    "                graph,\n",
    "                species_taxid=species_taxid,\n",
    "                include_neighbours=False,\n",
    "            )\n",
    "            cached_proteins[species_taxid] = proteins\n",
    "            \n",
    "        mean_1 = trna_ai['adaptation_index'].mean()\n",
    "        mean_2 = trna_ai[\n",
    "            trna_ai['protein'].isin(np.unique([t[0] for t in proteins]))\n",
    "        ]['adaptation_index'].mean()\n",
    "        \n",
    "        if pd.isnull(mean_1) or pd.isnull(mean_2):\n",
    "            print(species_taxid, assembly_accession, mean_1, mean_2)\n",
    "            continue\n",
    "        \n",
    "        diff.append(mean_2 - mean_1)\n",
    "        \n",
    "    return np.mean(diff), np.std(diff), np.min(diff), np.max(diff)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "q = \"\"\"\n",
    "select assembly_accession, species_taxid from assembly_source\n",
    "\"\"\"\n",
    "response_df = pd.read_sql(q, engine)\n",
    "assembly_accessions = response_df['assembly_accession'].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100 / 3,282\n",
      "200 / 3,282\n",
      "680 GCA_000017705.1 0.6611196531617023 nan\n",
      "300 / 3,282\n",
      "400 / 3,282\n",
      "500 / 3,282\n",
      "600 / 3,282\n",
      "700 / 3,282\n",
      "800 / 3,282\n",
      "900 / 3,282\n",
      "1,000 / 3,282\n",
      "1,100 / 3,282\n",
      "1,200 / 3,282\n",
      "1,300 / 3,282\n",
      "1,400 / 3,282\n",
      "1,500 / 3,282\n",
      "1,600 / 3,282\n",
      "1,700 / 3,282\n",
      "1,800 / 3,282\n",
      "1,900 / 3,282\n",
      "2,000 / 3,282\n",
      "2,100 / 3,282\n",
      "2,200 / 3,282\n",
      "2,300 / 3,282\n",
      "2,400 / 3,282\n",
      "2,500 / 3,282\n",
      "2,600 / 3,282\n",
      "2,700 / 3,282\n",
      "2,800 / 3,282\n",
      "2,900 / 3,282\n",
      "3,000 / 3,282\n",
      "3,100 / 3,282\n",
      "3,200 / 3,282\n",
      "CPU times: user 5min 37s, sys: 22.4 s, total: 5min 59s\n",
      "Wall time: 6min 8s\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(-0.004329882242986181,\n",
       " 0.008983127220666189,\n",
       " -0.058914431101381504,\n",
       " 0.030915092541605182)"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%%time\n",
    "mean, std, min_, max_ = compute_adaptation_mean_diff(assembly_accessions)\n",
    "mean, std, min_, max_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_mean_adaptation_per_species(assembly_accessions):\n",
    "    cached_proteins = {}\n",
    "    \n",
    "    res = collections.defaultdict(list)\n",
    "    for i, assembly_accession in enumerate(assembly_accessions):\n",
    "        if (i+1) % 100 == 0:\n",
    "            print(f'{i+1:,} / {len(assembly_accessions):,}')\n",
    "        \n",
    "        trna_ai_path = os.path.join(os.getcwd(), f'data/trn_adaptation_index/{assembly_accession}_tai.csv')\n",
    "        trna_ai = pd.read_csv(trna_ai_path)\n",
    "        \n",
    "        species_taxid = trna_ai['species_taxid'].iloc[0]\n",
    "    \n",
    "        if species_taxid in cached_proteins:\n",
    "            proteins = cached_proteins[species_taxid]\n",
    "        else:\n",
    "            proteins = retrieve_orginal_proteins(\n",
    "                graph,\n",
    "                species_taxid=species_taxid,\n",
    "                include_neighbours=False,\n",
    "            )\n",
    "            cached_proteins[species_taxid] = proteins\n",
    "            \n",
    "        mean = trna_ai['adaptation_index'].mean()\n",
    "        res[species_taxid].append(mean)\n",
    "        \n",
    "    output = {}\n",
    "    for key in sorted(res.keys()):\n",
    "        output[key] = np.mean([m for m in res[key] if not pd.isnull(m)])\n",
    "        \n",
    "    return output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100 / 3,282\n",
      "200 / 3,282\n",
      "300 / 3,282\n",
      "400 / 3,282\n",
      "500 / 3,282\n",
      "600 / 3,282\n",
      "700 / 3,282\n",
      "800 / 3,282\n",
      "900 / 3,282\n",
      "1,000 / 3,282\n",
      "1,100 / 3,282\n",
      "1,200 / 3,282\n",
      "1,300 / 3,282\n",
      "1,400 / 3,282\n",
      "1,500 / 3,282\n",
      "1,600 / 3,282\n",
      "1,700 / 3,282\n",
      "1,800 / 3,282\n",
      "1,900 / 3,282\n",
      "2,000 / 3,282\n",
      "2,100 / 3,282\n",
      "2,200 / 3,282\n",
      "2,300 / 3,282\n",
      "2,400 / 3,282\n",
      "2,500 / 3,282\n",
      "2,600 / 3,282\n",
      "2,700 / 3,282\n",
      "2,800 / 3,282\n",
      "2,900 / 3,282\n",
      "3,000 / 3,282\n",
      "3,100 / 3,282\n",
      "3,200 / 3,282\n",
      "CPU times: user 5min 53s, sys: 24 s, total: 6min 17s\n",
      "Wall time: 6min 26s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "mean_adaptation_per_species = compute_mean_adaptation_per_species(assembly_accessions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7526456316462177\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "_, ax3 = plt.subplots(1, 1, figsize=(12, 6))\n",
    "ax3.hist(list(mean_adaptation_per_species.values()), bins=100);\n",
    "print(np.mean(list(mean_adaptation_per_species.values())))\n",
    "ax3.set_title('Histogram of mean tRNA adaptation index across all species\\n');\n",
    "ax3.set_xlabel('Mean tRNA adaptation index');\n",
    "ax3.set_ylabel('Species count');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Does length have an effect on genes close to the mean?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_distributions_of_length(assembly_accessions):\n",
    "    cds_codon_bias_folder = os.path.join(os.getcwd(), 'data/cds_codon_bias/')\n",
    "    \n",
    "    lengths_all = []\n",
    "    lengths_below = []\n",
    "    for i, assembly_accession in enumerate(assembly_accessions):\n",
    "        if (i+1) % 200 == 0:\n",
    "            print(f'{i+1} / {len(assembly_accessions)}')\n",
    "        \n",
    "        path_all = os.path.join(cds_codon_bias_folder, f'all/{assembly_accession}_codon_bias.csv')\n",
    "        path_below = os.path.join(cds_codon_bias_folder, f'below_threshold/{assembly_accession}_codon_bias.csv')\n",
    "        \n",
    "        df_all = pd.read_csv(path_all)\n",
    "        df_below = pd.read_csv(path_below)\n",
    "        \n",
    "        lengths_all.extend(df_all['length'].values.tolist())\n",
    "        lengths_below.extend(df_below['length'].values.tolist())\n",
    "        \n",
    "    return np.array(lengths_all), np.array(lengths_below)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "200 / 3282\n",
      "400 / 3282\n",
      "600 / 3282\n",
      "800 / 3282\n",
      "1000 / 3282\n",
      "1200 / 3282\n",
      "1400 / 3282\n",
      "1600 / 3282\n",
      "1800 / 3282\n",
      "2000 / 3282\n",
      "2200 / 3282\n",
      "2400 / 3282\n",
      "2600 / 3282\n",
      "2800 / 3282\n",
      "3000 / 3282\n",
      "3200 / 3282\n",
      "CPU times: user 2min 8s, sys: 17.3 s, total: 2min 25s\n",
      "Wall time: 2min 30s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "lengths_all, lengths_below = compute_distributions_of_length(assembly_accessions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_distributions(lengths_all, lengths_below):\n",
    "    palette = sns.color_palette()\n",
    "    _, axes = plt.subplots(1, 2, figsize=(20, 6))\n",
    "    \n",
    "    ax1 = axes[0]\n",
    "    ax2 = axes[1]\n",
    "    \n",
    "    ax1.hist(lengths_all, bins=100, log=True)\n",
    "    ax1.set_title('Length distribution (all)')\n",
    "    xlim = ax1.get_xlim()\n",
    "    \n",
    "    ax2.hist(lengths_below, bins=100, color=palette[1], log=True)\n",
    "    ax2.set_title('Length distribution (below threshold)')\n",
    "    ax2.set_xlim(xlim)\n",
    "    \n",
    "    return ax1, ax2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x432 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_distributions(lengths_all, lengths_below);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Distance to perfectly balanced"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "distance_matrix_path = os.path.join(os.getcwd(), 'data/distance_matrix.npy')\n",
    "distance_matrix = np.load(distance_matrix_path, allow_pickle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "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>AAA_ratio</th>\n",
       "      <th>AAG_ratio</th>\n",
       "      <th>AAT_ratio</th>\n",
       "      <th>AAC_ratio</th>\n",
       "      <th>ACT_ratio</th>\n",
       "      <th>ACC_ratio</th>\n",
       "      <th>ACA_ratio</th>\n",
       "      <th>ACG_ratio</th>\n",
       "      <th>CGT_ratio</th>\n",
       "      <th>CGC_ratio</th>\n",
       "      <th>...</th>\n",
       "      <th>GTG_ratio</th>\n",
       "      <th>TAA_ratio</th>\n",
       "      <th>TAG_ratio</th>\n",
       "      <th>TGA_ratio</th>\n",
       "      <th>TAC_ratio</th>\n",
       "      <th>TAT_ratio</th>\n",
       "      <th>TGC_ratio</th>\n",
       "      <th>TGT_ratio</th>\n",
       "      <th>TTC_ratio</th>\n",
       "      <th>TTT_ratio</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>species_taxid</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.082550</td>\n",
       "      <td>0.917450</td>\n",
       "      <td>0.327724</td>\n",
       "      <td>0.672276</td>\n",
       "      <td>0.024511</td>\n",
       "      <td>0.575353</td>\n",
       "      <td>0.036262</td>\n",
       "      <td>0.363875</td>\n",
       "      <td>0.078732</td>\n",
       "      <td>0.615278</td>\n",
       "      <td>...</td>\n",
       "      <td>0.603462</td>\n",
       "      <td>0.059996</td>\n",
       "      <td>0.093068</td>\n",
       "      <td>0.846937</td>\n",
       "      <td>0.387338</td>\n",
       "      <td>0.612662</td>\n",
       "      <td>0.903673</td>\n",
       "      <td>0.096327</td>\n",
       "      <td>0.897180</td>\n",
       "      <td>0.102820</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.919627</td>\n",
       "      <td>0.080373</td>\n",
       "      <td>0.858168</td>\n",
       "      <td>0.141832</td>\n",
       "      <td>0.454569</td>\n",
       "      <td>0.050026</td>\n",
       "      <td>0.446018</td>\n",
       "      <td>0.049387</td>\n",
       "      <td>0.325065</td>\n",
       "      <td>0.043907</td>\n",
       "      <td>...</td>\n",
       "      <td>0.064608</td>\n",
       "      <td>0.790826</td>\n",
       "      <td>0.100917</td>\n",
       "      <td>0.108257</td>\n",
       "      <td>0.130608</td>\n",
       "      <td>0.869392</td>\n",
       "      <td>0.176688</td>\n",
       "      <td>0.823312</td>\n",
       "      <td>0.080224</td>\n",
       "      <td>0.919776</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.007302</td>\n",
       "      <td>0.992698</td>\n",
       "      <td>0.010730</td>\n",
       "      <td>0.989270</td>\n",
       "      <td>0.008586</td>\n",
       "      <td>0.425527</td>\n",
       "      <td>0.015789</td>\n",
       "      <td>0.550098</td>\n",
       "      <td>0.092597</td>\n",
       "      <td>0.559134</td>\n",
       "      <td>...</td>\n",
       "      <td>0.480311</td>\n",
       "      <td>0.009482</td>\n",
       "      <td>0.085967</td>\n",
       "      <td>0.904551</td>\n",
       "      <td>0.981660</td>\n",
       "      <td>0.018340</td>\n",
       "      <td>0.966107</td>\n",
       "      <td>0.033893</td>\n",
       "      <td>0.994594</td>\n",
       "      <td>0.005406</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.664866</td>\n",
       "      <td>0.335134</td>\n",
       "      <td>0.775571</td>\n",
       "      <td>0.224429</td>\n",
       "      <td>0.458203</td>\n",
       "      <td>0.188052</td>\n",
       "      <td>0.304183</td>\n",
       "      <td>0.049563</td>\n",
       "      <td>0.035122</td>\n",
       "      <td>0.008361</td>\n",
       "      <td>...</td>\n",
       "      <td>0.153708</td>\n",
       "      <td>0.563059</td>\n",
       "      <td>0.206130</td>\n",
       "      <td>0.230811</td>\n",
       "      <td>0.208150</td>\n",
       "      <td>0.791850</td>\n",
       "      <td>0.204657</td>\n",
       "      <td>0.795343</td>\n",
       "      <td>0.189441</td>\n",
       "      <td>0.810559</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>0.551810</td>\n",
       "      <td>0.448190</td>\n",
       "      <td>0.440559</td>\n",
       "      <td>0.559441</td>\n",
       "      <td>0.099134</td>\n",
       "      <td>0.591478</td>\n",
       "      <td>0.097796</td>\n",
       "      <td>0.211592</td>\n",
       "      <td>0.204394</td>\n",
       "      <td>0.366337</td>\n",
       "      <td>...</td>\n",
       "      <td>0.334484</td>\n",
       "      <td>0.340503</td>\n",
       "      <td>0.164193</td>\n",
       "      <td>0.495304</td>\n",
       "      <td>0.486740</td>\n",
       "      <td>0.513260</td>\n",
       "      <td>0.636140</td>\n",
       "      <td>0.363860</td>\n",
       "      <td>0.502549</td>\n",
       "      <td>0.497451</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1813871</th>\n",
       "      <td>0.772490</td>\n",
       "      <td>0.227510</td>\n",
       "      <td>0.681741</td>\n",
       "      <td>0.318259</td>\n",
       "      <td>0.240711</td>\n",
       "      <td>0.192149</td>\n",
       "      <td>0.446839</td>\n",
       "      <td>0.120302</td>\n",
       "      <td>0.243327</td>\n",
       "      <td>0.187369</td>\n",
       "      <td>...</td>\n",
       "      <td>0.205557</td>\n",
       "      <td>0.680140</td>\n",
       "      <td>0.164518</td>\n",
       "      <td>0.155342</td>\n",
       "      <td>0.300637</td>\n",
       "      <td>0.699363</td>\n",
       "      <td>0.476391</td>\n",
       "      <td>0.523609</td>\n",
       "      <td>0.250815</td>\n",
       "      <td>0.749185</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1833852</th>\n",
       "      <td>0.589533</td>\n",
       "      <td>0.410467</td>\n",
       "      <td>0.567693</td>\n",
       "      <td>0.432307</td>\n",
       "      <td>0.202326</td>\n",
       "      <td>0.538744</td>\n",
       "      <td>0.183109</td>\n",
       "      <td>0.075821</td>\n",
       "      <td>0.195981</td>\n",
       "      <td>0.173208</td>\n",
       "      <td>...</td>\n",
       "      <td>0.343317</td>\n",
       "      <td>0.582863</td>\n",
       "      <td>0.286616</td>\n",
       "      <td>0.130521</td>\n",
       "      <td>0.346503</td>\n",
       "      <td>0.653497</td>\n",
       "      <td>0.459139</td>\n",
       "      <td>0.540861</td>\n",
       "      <td>0.237507</td>\n",
       "      <td>0.762493</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1835254</th>\n",
       "      <td>0.555758</td>\n",
       "      <td>0.444242</td>\n",
       "      <td>0.685138</td>\n",
       "      <td>0.314862</td>\n",
       "      <td>0.374892</td>\n",
       "      <td>0.272930</td>\n",
       "      <td>0.209890</td>\n",
       "      <td>0.142287</td>\n",
       "      <td>0.341983</td>\n",
       "      <td>0.323291</td>\n",
       "      <td>...</td>\n",
       "      <td>0.254571</td>\n",
       "      <td>0.660029</td>\n",
       "      <td>0.156387</td>\n",
       "      <td>0.183584</td>\n",
       "      <td>0.391660</td>\n",
       "      <td>0.608340</td>\n",
       "      <td>0.547508</td>\n",
       "      <td>0.452492</td>\n",
       "      <td>0.294507</td>\n",
       "      <td>0.705493</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1945662</th>\n",
       "      <td>0.131137</td>\n",
       "      <td>0.868863</td>\n",
       "      <td>0.229159</td>\n",
       "      <td>0.770841</td>\n",
       "      <td>0.020880</td>\n",
       "      <td>0.561770</td>\n",
       "      <td>0.044037</td>\n",
       "      <td>0.373313</td>\n",
       "      <td>0.050484</td>\n",
       "      <td>0.587127</td>\n",
       "      <td>...</td>\n",
       "      <td>0.458268</td>\n",
       "      <td>0.073481</td>\n",
       "      <td>0.117497</td>\n",
       "      <td>0.809021</td>\n",
       "      <td>0.314323</td>\n",
       "      <td>0.685677</td>\n",
       "      <td>0.909613</td>\n",
       "      <td>0.090387</td>\n",
       "      <td>0.738885</td>\n",
       "      <td>0.261115</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1985254</th>\n",
       "      <td>0.163778</td>\n",
       "      <td>0.836222</td>\n",
       "      <td>0.293950</td>\n",
       "      <td>0.706050</td>\n",
       "      <td>0.048542</td>\n",
       "      <td>0.627622</td>\n",
       "      <td>0.044654</td>\n",
       "      <td>0.279182</td>\n",
       "      <td>0.150186</td>\n",
       "      <td>0.649024</td>\n",
       "      <td>...</td>\n",
       "      <td>0.619580</td>\n",
       "      <td>0.197077</td>\n",
       "      <td>0.142955</td>\n",
       "      <td>0.659968</td>\n",
       "      <td>0.667897</td>\n",
       "      <td>0.332103</td>\n",
       "      <td>0.870500</td>\n",
       "      <td>0.129500</td>\n",
       "      <td>0.803614</td>\n",
       "      <td>0.196386</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2680 rows × 62 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "               AAA_ratio  AAG_ratio  AAT_ratio  AAC_ratio  ACT_ratio  \\\n",
       "species_taxid                                                          \n",
       "7               0.082550   0.917450   0.327724   0.672276   0.024511   \n",
       "9               0.919627   0.080373   0.858168   0.141832   0.454569   \n",
       "11              0.007302   0.992698   0.010730   0.989270   0.008586   \n",
       "14              0.664866   0.335134   0.775571   0.224429   0.458203   \n",
       "19              0.551810   0.448190   0.440559   0.559441   0.099134   \n",
       "...                  ...        ...        ...        ...        ...   \n",
       "1813871         0.772490   0.227510   0.681741   0.318259   0.240711   \n",
       "1833852         0.589533   0.410467   0.567693   0.432307   0.202326   \n",
       "1835254         0.555758   0.444242   0.685138   0.314862   0.374892   \n",
       "1945662         0.131137   0.868863   0.229159   0.770841   0.020880   \n",
       "1985254         0.163778   0.836222   0.293950   0.706050   0.048542   \n",
       "\n",
       "               ACC_ratio  ACA_ratio  ACG_ratio  CGT_ratio  CGC_ratio  ...  \\\n",
       "species_taxid                                                         ...   \n",
       "7               0.575353   0.036262   0.363875   0.078732   0.615278  ...   \n",
       "9               0.050026   0.446018   0.049387   0.325065   0.043907  ...   \n",
       "11              0.425527   0.015789   0.550098   0.092597   0.559134  ...   \n",
       "14              0.188052   0.304183   0.049563   0.035122   0.008361  ...   \n",
       "19              0.591478   0.097796   0.211592   0.204394   0.366337  ...   \n",
       "...                  ...        ...        ...        ...        ...  ...   \n",
       "1813871         0.192149   0.446839   0.120302   0.243327   0.187369  ...   \n",
       "1833852         0.538744   0.183109   0.075821   0.195981   0.173208  ...   \n",
       "1835254         0.272930   0.209890   0.142287   0.341983   0.323291  ...   \n",
       "1945662         0.561770   0.044037   0.373313   0.050484   0.587127  ...   \n",
       "1985254         0.627622   0.044654   0.279182   0.150186   0.649024  ...   \n",
       "\n",
       "               GTG_ratio  TAA_ratio  TAG_ratio  TGA_ratio  TAC_ratio  \\\n",
       "species_taxid                                                          \n",
       "7               0.603462   0.059996   0.093068   0.846937   0.387338   \n",
       "9               0.064608   0.790826   0.100917   0.108257   0.130608   \n",
       "11              0.480311   0.009482   0.085967   0.904551   0.981660   \n",
       "14              0.153708   0.563059   0.206130   0.230811   0.208150   \n",
       "19              0.334484   0.340503   0.164193   0.495304   0.486740   \n",
       "...                  ...        ...        ...        ...        ...   \n",
       "1813871         0.205557   0.680140   0.164518   0.155342   0.300637   \n",
       "1833852         0.343317   0.582863   0.286616   0.130521   0.346503   \n",
       "1835254         0.254571   0.660029   0.156387   0.183584   0.391660   \n",
       "1945662         0.458268   0.073481   0.117497   0.809021   0.314323   \n",
       "1985254         0.619580   0.197077   0.142955   0.659968   0.667897   \n",
       "\n",
       "               TAT_ratio  TGC_ratio  TGT_ratio  TTC_ratio  TTT_ratio  \n",
       "species_taxid                                                         \n",
       "7               0.612662   0.903673   0.096327   0.897180   0.102820  \n",
       "9               0.869392   0.176688   0.823312   0.080224   0.919776  \n",
       "11              0.018340   0.966107   0.033893   0.994594   0.005406  \n",
       "14              0.791850   0.204657   0.795343   0.189441   0.810559  \n",
       "19              0.513260   0.636140   0.363860   0.502549   0.497451  \n",
       "...                  ...        ...        ...        ...        ...  \n",
       "1813871         0.699363   0.476391   0.523609   0.250815   0.749185  \n",
       "1833852         0.653497   0.459139   0.540861   0.237507   0.762493  \n",
       "1835254         0.608340   0.547508   0.452492   0.294507   0.705493  \n",
       "1945662         0.685677   0.909613   0.090387   0.738885   0.261115  \n",
       "1985254         0.332103   0.870500   0.129500   0.803614   0.196386  \n",
       "\n",
       "[2680 rows x 62 columns]"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "species_codon_df[ratio_columns]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_uniform_bias(codon_to_aa_table, aa_to_codons, ratio_columns):\n",
    "    uniform_bias = np.zeros((len(ratio_columns),))\n",
    "    for i, col in enumerate(ratio_columns):\n",
    "        codon = col.replace('_ratio', '')\n",
    "        aa = codon_to_aa_table[codon]\n",
    "        synonyms = aa_to_codons[aa]\n",
    "        \n",
    "        uniform_bias[i] = 1 / len(synonyms)\n",
    "        \n",
    "    return uniform_bias"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.5       , 0.5       , 0.5       , 0.5       , 0.25      ,\n",
       "       0.25      , 0.25      , 0.25      , 0.16666667, 0.16666667,\n",
       "       0.16666667, 0.16666667, 0.16666667, 0.16666667, 0.16666667,\n",
       "       0.16666667, 0.16666667, 0.16666667, 0.16666667, 0.16666667,\n",
       "       0.33333333, 0.33333333, 0.33333333, 0.5       , 0.5       ,\n",
       "       0.5       , 0.5       , 0.25      , 0.25      , 0.25      ,\n",
       "       0.25      , 0.16666667, 0.16666667, 0.16666667, 0.16666667,\n",
       "       0.16666667, 0.16666667, 0.5       , 0.5       , 0.5       ,\n",
       "       0.5       , 0.25      , 0.25      , 0.25      , 0.25      ,\n",
       "       0.25      , 0.25      , 0.25      , 0.25      , 0.25      ,\n",
       "       0.25      , 0.25      , 0.25      , 0.33333333, 0.33333333,\n",
       "       0.33333333, 0.5       , 0.5       , 0.5       , 0.5       ,\n",
       "       0.5       , 0.5       ])"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "uniform_bias = make_uniform_bias(codon_to_aa_table, aa_to_codons, ratio_columns)\n",
    "uniform_bias"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_distance_to_uniform_bias(species_codon_df, uniform_bias, ratio_columns):\n",
    "    X = species_codon_df[ratio_columns].values\n",
    "    return compute_codon_bias_distance(X, uniform_bias)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "distance_to_uniform = compute_distance_to_uniform_bias(species_codon_df, uniform_bias, ratio_columns)\n",
    "species_codon_df['uniform_dist'] = distance_to_uniform"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2020-12-04 12:34:41,529 (WARNING) *c* argument looks like a single numeric RGB or RGBA sequence, which should be avoided as value-mapping will have precedence in case its length matches with *x* & *y*.  Please use the *color* keyword-argument or provide a 2-D array with a single row if you intend to specify the same RGB or RGBA value for all points.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "species_codon_df.plot.scatter(\n",
    "    x='growth_tmp',\n",
    "    y='uniform_dist',\n",
    "    figsize=(12, 6),\n",
    ");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "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>uniform_dist</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>species_taxid</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>562</th>\n",
       "      <td>0.145268</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2336</th>\n",
       "      <td>0.133543</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               uniform_dist\n",
       "species_taxid              \n",
       "562                0.145268\n",
       "2336               0.133543"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "species_codon_df.loc[[562, 2336]][['uniform_dist']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}