code/notebooks/major_clade_dimension.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Dimensions of major clades"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dependencies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from statistics import median"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from skbio.tree import TreeNode"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "sns.set_palette('muted')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Helpers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def calc_length_metrics(tree):\n",
    "    for node in tree.postorder(include_self=True):\n",
    "        if node.length is None:\n",
    "            node.length = 0.0\n",
    "        if node.is_tip():\n",
    "            node.depths = [0.0]\n",
    "        else:\n",
    "            node.depths = [y + x.length for x in node.children for y in\n",
    "                           x.depths]\n",
    "    for node in tree.preorder(include_self=True):\n",
    "        if node.is_root():\n",
    "            node.height = 0.0\n",
    "            node.red = 0.0\n",
    "        else:\n",
    "            node.height = node.parent.height + node.length\n",
    "            if node.is_tip():\n",
    "                node.red = 1.0\n",
    "            else:\n",
    "                node.red = node.parent.red + node.length \\\n",
    "                    / (node.length + sum(node.depths) / len(node.depths)) \\\n",
    "                    * (1 - node.parent.red)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Input files"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "File paths"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "meta_fp = 'metadata.tsv'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "taxa_fp = '../../complex/taxa.txt'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "clade_fp = '../../../taxonomy/major/maps/lv1.in.txt'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "dir_map = {'cons': 'conserv', 'rand': 'random', 'rpls': 'r-proteins'}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "tree_fps = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "tree_fps.append(('all.1k', '../../concat/subsample/all_sites_1k_taxa/all.1k.nid.nwk'))\n",
    "tree_fps.append(('all.ft', '../../concat/all/fasttree/concat.fast.nid.nwk'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "for x in ('cons', 'rand', 'rpls'):\n",
    "    tree_fps.append(('%s.g' % x, '../../complex/%s.1k.LG+G.nid.nwk' % x))\n",
    "    tree_fps.append(('%s.r' % x, '../../complex/%s.1k.LG+R.nid.nwk' % x))\n",
    "    tree_fps.append(('%s.pmsf' % x, '../../complex/%s.1k.LG+C60_PMSF.nid.nwk' % x))\n",
    "    tree_fps.append(('%s.full' % x, '../../concat/%s/concat.%s.nid.nwk' % (dir_map[x], x)))\n",
    "    tree_fps.append(('%s.astral' % x, '../../astral/newick/astral.%s.nwk' % x))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Read files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(taxa_fp, 'r') as f:\n",
    "    taxa = set(f.read().splitlines())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(clade_fp, 'r') as f:\n",
    "    clade_map = dict(x.split('\\t') for x in f.read().splitlines())\n",
    "clade_map = {k: v for k, v in clade_map.items() if k in taxa}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "clades = {x: {k for k, v in clade_map.items() if v == x} for x in set(clade_map.values())}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Archaea', 'CPR', 'Eubacteria']"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clade_list = sorted(clades.keys())\n",
    "clade_list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['all.1k', 'all.ft', 'cons.g', 'cons.r', 'cons.pmsf', 'cons.full', 'cons.astral', 'rand.g', 'rand.r', 'rand.pmsf', 'rand.full', 'rand.astral', 'rpls.g', 'rpls.r', 'rpls.pmsf', 'rpls.full', 'rpls.astral']\n"
     ]
    }
   ],
   "source": [
    "ids = [x[0] for x in tree_fps]\n",
    "print(ids)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "trees = {x[0]: TreeNode.read(x[1]) for x in tree_fps}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "for x in ('cons', 'rand', 'rpls'):\n",
    "    for y in ('full', 'astral'):\n",
    "        id_ = '%s.%s' % (x, y)\n",
    "        taxa_ = trees[id_].subset().intersection(taxa)\n",
    "        trees[id_] = trees[id_].shear(taxa_)\n",
    "trees['all.ft'] = trees['all.ft'].shear(trees['all.ft'].subset().intersection(taxa))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Fill empty branch lengths"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "for k, v in trees.items():\n",
    "    for tree in v:\n",
    "        for node in tree.traverse():\n",
    "            if node.length is None:\n",
    "                node.length = 0.0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Branch statistics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Calculate branch length-related metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "for _, tree in trees.items():\n",
    "    calc_length_metrics(tree)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Radius of each tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "radia = {k: median(v.depths) for k, v in trees.items()}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Per-clade statistics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "data = []\n",
    "for name, tree in trees.items():\n",
    "    data.append([name, radia[name]])\n",
    "    for clade in clade_list:\n",
    "        taxa_ = tree.subset().intersection(clades[clade])\n",
    "        node = tree.lca(taxa_)\n",
    "        if node.subset() != taxa_:\n",
    "            data[-1].extend([None, None, None])\n",
    "            continue\n",
    "        height = node.length\n",
    "        if clade == 'Archaea':\n",
    "            height *= 2\n",
    "        depth = median(node.depths)\n",
    "        ratio = height / depth\n",
    "        data[-1].extend([height, depth, ratio])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>radius</th>\n",
       "      <th>Archaea_height</th>\n",
       "      <th>Archaea_depth</th>\n",
       "      <th>Archaea_ratio</th>\n",
       "      <th>CPR_height</th>\n",
       "      <th>CPR_depth</th>\n",
       "      <th>CPR_ratio</th>\n",
       "      <th>Eubacteria_height</th>\n",
       "      <th>Eubacteria_depth</th>\n",
       "      <th>Eubacteria_ratio</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tree</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>all.1k</th>\n",
       "      <td>1.332852</td>\n",
       "      <td>0.169972</td>\n",
       "      <td>1.046816</td>\n",
       "      <td>0.162370</td>\n",
       "      <td>0.153735</td>\n",
       "      <td>1.113456</td>\n",
       "      <td>0.138071</td>\n",
       "      <td>0.069774</td>\n",
       "      <td>1.188847</td>\n",
       "      <td>0.058690</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>all.ft</th>\n",
       "      <td>1.749837</td>\n",
       "      <td>0.210401</td>\n",
       "      <td>1.329346</td>\n",
       "      <td>0.158274</td>\n",
       "      <td>0.238584</td>\n",
       "      <td>1.446355</td>\n",
       "      <td>0.164955</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cons.g</th>\n",
       "      <td>0.706015</td>\n",
       "      <td>0.109772</td>\n",
       "      <td>0.669294</td>\n",
       "      <td>0.164012</td>\n",
       "      <td>0.081533</td>\n",
       "      <td>0.731724</td>\n",
       "      <td>0.111426</td>\n",
       "      <td>0.038142</td>\n",
       "      <td>0.588887</td>\n",
       "      <td>0.064769</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cons.r</th>\n",
       "      <td>0.614697</td>\n",
       "      <td>0.093849</td>\n",
       "      <td>0.589953</td>\n",
       "      <td>0.159079</td>\n",
       "      <td>0.071903</td>\n",
       "      <td>0.625998</td>\n",
       "      <td>0.114861</td>\n",
       "      <td>0.018443</td>\n",
       "      <td>0.525590</td>\n",
       "      <td>0.035090</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cons.pmsf</th>\n",
       "      <td>0.981989</td>\n",
       "      <td>0.167806</td>\n",
       "      <td>1.049143</td>\n",
       "      <td>0.159946</td>\n",
       "      <td>0.120048</td>\n",
       "      <td>1.128478</td>\n",
       "      <td>0.106380</td>\n",
       "      <td>0.041744</td>\n",
       "      <td>0.807889</td>\n",
       "      <td>0.051671</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             radius  Archaea_height  Archaea_depth  Archaea_ratio  CPR_height  \\\n",
       "tree                                                                            \n",
       "all.1k     1.332852        0.169972       1.046816       0.162370    0.153735   \n",
       "all.ft     1.749837        0.210401       1.329346       0.158274    0.238584   \n",
       "cons.g     0.706015        0.109772       0.669294       0.164012    0.081533   \n",
       "cons.r     0.614697        0.093849       0.589953       0.159079    0.071903   \n",
       "cons.pmsf  0.981989        0.167806       1.049143       0.159946    0.120048   \n",
       "\n",
       "           CPR_depth  CPR_ratio  Eubacteria_height  Eubacteria_depth  \\\n",
       "tree                                                                   \n",
       "all.1k      1.113456   0.138071           0.069774          1.188847   \n",
       "all.ft      1.446355   0.164955                NaN               NaN   \n",
       "cons.g      0.731724   0.111426           0.038142          0.588887   \n",
       "cons.r      0.625998   0.114861           0.018443          0.525590   \n",
       "cons.pmsf   1.128478   0.106380           0.041744          0.807889   \n",
       "\n",
       "           Eubacteria_ratio  \n",
       "tree                         \n",
       "all.1k             0.058690  \n",
       "all.ft                  NaN  \n",
       "cons.g             0.064769  \n",
       "cons.r             0.035090  \n",
       "cons.pmsf          0.051671  "
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.DataFrame(data)\n",
    "df.columns = ['tree', 'radius'] + ['%s_%s' % (x, y) for x in clade_list\n",
    "              for y in ['height', 'depth', 'ratio']]\n",
    "df.set_index('tree', inplace=True)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## A-B distances"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "clades['Bacteria'] = clades['Eubacteria'].union(clades['CPR'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Tip-to-tip distances"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "dms = {name: tree.tip_tip_distances() for name, tree in trees.items()}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "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>Archaea_n</th>\n",
       "      <th>Bacteria_n</th>\n",
       "      <th>Archaea_sum</th>\n",
       "      <th>Bacteria_sum</th>\n",
       "      <th>A-B_sum</th>\n",
       "      <th>Archaea_mean</th>\n",
       "      <th>Bacteria_mean</th>\n",
       "      <th>A-B_mean</th>\n",
       "      <th>A-B_ratio</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tree</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>all.1k</th>\n",
       "      <td>74</td>\n",
       "      <td>926</td>\n",
       "      <td>5014.241641</td>\n",
       "      <td>9.527379e+05</td>\n",
       "      <td>172545.805273</td>\n",
       "      <td>1.856439</td>\n",
       "      <td>2.224594</td>\n",
       "      <td>2.518035</td>\n",
       "      <td>1.535296</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>all.ft</th>\n",
       "      <td>74</td>\n",
       "      <td>926</td>\n",
       "      <td>6668.136751</td>\n",
       "      <td>1.271256e+06</td>\n",
       "      <td>224488.904535</td>\n",
       "      <td>2.468766</td>\n",
       "      <td>2.968318</td>\n",
       "      <td>3.276062</td>\n",
       "      <td>1.464583</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cons.g</th>\n",
       "      <td>74</td>\n",
       "      <td>926</td>\n",
       "      <td>3162.296967</td>\n",
       "      <td>4.977385e+05</td>\n",
       "      <td>101044.460915</td>\n",
       "      <td>1.170787</td>\n",
       "      <td>1.162194</td>\n",
       "      <td>1.474585</td>\n",
       "      <td>1.598023</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cons.r</th>\n",
       "      <td>74</td>\n",
       "      <td>926</td>\n",
       "      <td>2753.995592</td>\n",
       "      <td>4.332583e+05</td>\n",
       "      <td>88411.081814</td>\n",
       "      <td>1.019621</td>\n",
       "      <td>1.011636</td>\n",
       "      <td>1.290221</td>\n",
       "      <td>1.613857</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cons.pmsf</th>\n",
       "      <td>74</td>\n",
       "      <td>926</td>\n",
       "      <td>4952.094619</td>\n",
       "      <td>7.109533e+05</td>\n",
       "      <td>150551.202280</td>\n",
       "      <td>1.833430</td>\n",
       "      <td>1.660039</td>\n",
       "      <td>2.197058</td>\n",
       "      <td>1.585990</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Archaea_n  Bacteria_n  Archaea_sum  Bacteria_sum        A-B_sum  \\\n",
       "tree                                                                         \n",
       "all.1k            74         926  5014.241641  9.527379e+05  172545.805273   \n",
       "all.ft            74         926  6668.136751  1.271256e+06  224488.904535   \n",
       "cons.g            74         926  3162.296967  4.977385e+05  101044.460915   \n",
       "cons.r            74         926  2753.995592  4.332583e+05   88411.081814   \n",
       "cons.pmsf         74         926  4952.094619  7.109533e+05  150551.202280   \n",
       "\n",
       "           Archaea_mean  Bacteria_mean  A-B_mean  A-B_ratio  \n",
       "tree                                                         \n",
       "all.1k         1.856439       2.224594  2.518035   1.535296  \n",
       "all.ft         2.468766       2.968318  3.276062   1.464583  \n",
       "cons.g         1.170787       1.162194  1.474585   1.598023  \n",
       "cons.r         1.019621       1.011636  1.290221   1.613857  \n",
       "cons.pmsf      1.833430       1.660039  2.197058   1.585990  "
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = []\n",
    "for name, tree in trees.items():\n",
    "    dm = dms[name]\n",
    "    \n",
    "    all_n = dm.shape[0]\n",
    "    all_sum = dm.condensed_form().sum()\n",
    "\n",
    "    arc_dm = dm.filter(clades['Archaea'], strict=False)\n",
    "    arc_n = len(arc_dm.ids)\n",
    "    arc_m = arc_n * (arc_n - 1) / 2\n",
    "    arc_sum = arc_dm.condensed_form().sum()\n",
    "    arc_mean = arc_sum / arc_m\n",
    "\n",
    "    bac_dm = dm.filter(clades['Bacteria'], strict=False)\n",
    "    bac_n = len(bac_dm.ids)\n",
    "    bac_m = bac_n * (bac_n - 1) / 2\n",
    "    bac_sum = bac_dm.condensed_form().sum()\n",
    "    bac_mean = bac_sum / bac_m\n",
    "\n",
    "    ab_sum = all_sum - arc_sum - bac_sum\n",
    "    ab_mean = ab_sum / (all_n ** 2 - arc_n ** 2 - bac_n ** 2) * 2\n",
    "\n",
    "    ratio = ab_mean ** 2 / arc_mean / bac_mean\n",
    "\n",
    "    data.append([name, arc_n, bac_n, arc_sum, bac_sum, ab_sum,\n",
    "                 arc_mean, bac_mean, ab_mean, ratio])\n",
    "\n",
    "df_ = pd.DataFrame(data)\n",
    "df_.columns = ['tree', 'Archaea_n', 'Bacteria_n', 'Archaea_sum', 'Bacteria_sum', 'A-B_sum',\n",
    "               'Archaea_mean', 'Bacteria_mean', 'A-B_mean', 'A-B_ratio']\n",
    "df_.set_index('tree', inplace=True)\n",
    "df_.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.concat([df, df_], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.to_csv('output.tsv', sep='\\t')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plotting"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A-B branch length"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "labels = list(df.index)\n",
    "xticks = list(range(len(labels)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(12, 3))\n",
    "data = df['Archaea_height'] / df['radius']\n",
    "plt.bar(xticks, data)\n",
    "plt.yscale('log')\n",
    "yticks = [0.1, 0.2, 0.4, 0.6, 0.8, 1.0]\n",
    "plt.yticks(yticks, yticks)\n",
    "plt.ylabel('Norm. A-B branch length')\n",
    "plt.xticks(xticks, labels, rotation=30, ha='right');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "fig.tight_layout()\n",
    "fig.savefig('ab.brlen.pdf', bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A-B distance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(12, 3))\n",
    "data = df['A-B_ratio']\n",
    "plt.bar(xticks, data)\n",
    "plt.ylabel('Relative A-B distance')\n",
    "plt.xticks(xticks, labels, rotation=30, ha='right');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "fig.tight_layout()\n",
    "fig.savefig('ab.dist.pdf', bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Depth and height"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_ = df[df['Eubacteria_height'].notnull()].copy()\n",
    "df_ = df_[~df_.index.to_series().str.startswith('rpls')]\n",
    "df_ = df_[~df_.index.to_series().str.endswith('.r')]\n",
    "df_ = df_[~df_.index.to_series().str.endswith('.astral')]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "for x in ('depth', 'height'):\n",
    "    df_['%s_sum' % x] = sum(df_['%s_%s' % (y, x)] for y in ('Archaea', 'CPR', 'Eubacteria'))\n",
    "    for y in ('Archaea', 'CPR', 'Eubacteria'):\n",
    "        df_['%s_%s_norm' % (y, x)] = df_['%s_%s' % (y, x)] / df_['%s_sum' % x]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "labels = list(df_.index)\n",
    "xticks = range(len(labels))\n",
    "fig = plt.figure(figsize=(6, 3))\n",
    "bottom = [0] * df_.shape[0]\n",
    "for x in ('Archaea', 'CPR', 'Eubacteria'):\n",
    "    plt.bar(xticks, df_['%s_depth_norm' % x], bottom=bottom)\n",
    "    for i in range(len(bottom)):\n",
    "        bottom[i] += df_['%s_depth_norm' % x][i]\n",
    "plt.xticks(xticks, labels, rotation=30, ha='right')\n",
    "plt.ylabel('Median depth')\n",
    "plt.legend(labels=['Archaea', 'CPR', 'Eubacteria']);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "fig.tight_layout()\n",
    "fig.savefig('depth.pdf', bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x216 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "labels = list(df_.index)\n",
    "xticks = range(len(labels))\n",
    "fig = plt.figure(figsize=(6, 3))\n",
    "bottom = [0] * df_.shape[0]\n",
    "for x in ('Archaea', 'CPR', 'Eubacteria'):\n",
    "    plt.bar(xticks, df_['%s_height_norm' % x], bottom=bottom)\n",
    "    for i in range(len(bottom)):\n",
    "        bottom[i] += df_['%s_height_norm' % x][i]\n",
    "plt.ylabel('Height')\n",
    "plt.xticks(xticks, labels, rotation=30, ha='right');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "fig.tight_layout()\n",
    "fig.savefig('height.pdf', bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Depth distribution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "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>tree</th>\n",
       "      <th>site</th>\n",
       "      <th>model</th>\n",
       "      <th>clade</th>\n",
       "      <th>depth</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>cons.g</td>\n",
       "      <td>cons</td>\n",
       "      <td>g</td>\n",
       "      <td>Archaea</td>\n",
       "      <td>0.584646</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>cons.g</td>\n",
       "      <td>cons</td>\n",
       "      <td>g</td>\n",
       "      <td>Archaea</td>\n",
       "      <td>0.545367</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>cons.g</td>\n",
       "      <td>cons</td>\n",
       "      <td>g</td>\n",
       "      <td>Archaea</td>\n",
       "      <td>0.559824</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>cons.g</td>\n",
       "      <td>cons</td>\n",
       "      <td>g</td>\n",
       "      <td>Archaea</td>\n",
       "      <td>0.529409</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>cons.g</td>\n",
       "      <td>cons</td>\n",
       "      <td>g</td>\n",
       "      <td>Archaea</td>\n",
       "      <td>0.588778</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     tree  site model    clade     depth\n",
       "0  cons.g  cons     g  Archaea  0.584646\n",
       "1  cons.g  cons     g  Archaea  0.545367\n",
       "2  cons.g  cons     g  Archaea  0.559824\n",
       "3  cons.g  cons     g  Archaea  0.529409\n",
       "4  cons.g  cons     g  Archaea  0.588778"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = []\n",
    "for site in ('cons', 'rand'):\n",
    "    for model in ('g', 'pmsf'):\n",
    "        for clade in ('Archaea', 'CPR', 'Eubacteria'):\n",
    "            name = '%s.%s' % (site, model)\n",
    "            tree = trees[name]\n",
    "            node = tree.lca(tree.subset().intersection(clades[clade]))\n",
    "            for depth in node.depths:\n",
    "                data.append([name, site, model, clade, depth])\n",
    "dfd = pd.DataFrame(data)\n",
    "dfd.columns = ['tree', 'site', 'model', 'clade', 'depth']\n",
    "dfd.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 216x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 216x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for site in ('cons', 'rand'):\n",
    "    fig, ax = plt.subplots(figsize=(3, 5))\n",
    "    sns.violinplot(x='clade', y='depth', hue='model', linewidth=1,\n",
    "                   data=dfd.query('site == \"%s\"' % site), ax=ax)\n",
    "    ax.set_title(site)\n",
    "    ax.set_ylim(bottom=0)\n",
    "    ax.set_xlabel('')\n",
    "    ax.set_ylabel('')\n",
    "    for i, clade in enumerate(('Archaea', 'CPR', 'Eubacteria')):\n",
    "        h = df['%s_height' % clade]['%s.g' % site]\n",
    "        ax.plot([i - 0.35, i - 0.05], [h, h], color=sns.color_palette()[0])\n",
    "        h = df['%s_height' % clade]['%s.pmsf' % site]\n",
    "        ax.plot([i + 0.05, i + 0.35], [h, h], color=sns.color_palette()[1])\n",
    "    fig.tight_layout()\n",
    "    fig.savefig('%s.violin.pdf' % site, bbox_inches='tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "dfd['radius'] = dfd['tree'].map(radia)\n",
    "dfd['depth_norm'] = dfd['depth'] / dfd['radius']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 216x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 216x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for site in ('cons', 'rand'):\n",
    "    fig, ax = plt.subplots(figsize=(3, 5))\n",
    "    sns.violinplot(x='clade', y='depth_norm', hue='model', split=True, linewidth=1,\n",
    "                   data=dfd.query('site == \"%s\"' % site), ax=ax)\n",
    "    ax.set_title(site)\n",
    "    ax.set_ylim(bottom=0)\n",
    "    ax.set_xlabel('')\n",
    "    ax.set_ylabel('')\n",
    "    for i, clade in enumerate(('Archaea', 'CPR', 'Eubacteria')):\n",
    "        h = df['%s_height' % clade]['%s.g' % site] / radia['%s.g' % site]\n",
    "        ax.plot([i - 0.35, i - 0.05], [h, h], color=sns.color_palette()[0])\n",
    "        h = df['%s_height' % clade]['%s.pmsf' % site] / radia['%s.pmsf' % site]\n",
    "        ax.plot([i + 0.05, i + 0.35], [h, h], color=sns.color_palette()[1])\n",
    "    fig.tight_layout()\n",
    "    fig.savefig('%s.norm.violin.pdf' % site, bbox_inches='tight')"
   ]
  },
  {
   "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": "3.7.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}