overview.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>Run</th>\n",
       "      <th>ReleaseDate</th>\n",
       "      <th>LoadDate</th>\n",
       "      <th>spots</th>\n",
       "      <th>bases</th>\n",
       "      <th>spots_with_mates</th>\n",
       "      <th>avgLength</th>\n",
       "      <th>size_MB</th>\n",
       "      <th>AssemblyName</th>\n",
       "      <th>download_path</th>\n",
       "      <th>...</th>\n",
       "      <th>Affection_Status</th>\n",
       "      <th>Analyte_Type</th>\n",
       "      <th>Histological_Type</th>\n",
       "      <th>Body_Site</th>\n",
       "      <th>CenterName</th>\n",
       "      <th>Submission</th>\n",
       "      <th>dbgap_study_accession</th>\n",
       "      <th>Consent</th>\n",
       "      <th>RunHash</th>\n",
       "      <th>ReadHash</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DRR014176</td>\n",
       "      <td>2015-07-27 10:27:35</td>\n",
       "      <td>2015-07-27 10:32:25</td>\n",
       "      <td>30807540</td>\n",
       "      <td>3734314174</td>\n",
       "      <td>0</td>\n",
       "      <td>121</td>\n",
       "      <td>2297</td>\n",
       "      <td>NaN</td>\n",
       "      <td>https://sra-downloadb.st-va.ncbi.nlm.nih.gov/s...</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NIID</td>\n",
       "      <td>DRA001171</td>\n",
       "      <td>NaN</td>\n",
       "      <td>public</td>\n",
       "      <td>4DA05CDDF555DF45B83261F309ADDDE7</td>\n",
       "      <td>CD8372E963A92B67C96160309CBDD52D</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>SRR8391109</td>\n",
       "      <td>2019-01-07 15:57:36</td>\n",
       "      <td>2019-01-07 15:51:59</td>\n",
       "      <td>1</td>\n",
       "      <td>1332</td>\n",
       "      <td>0</td>\n",
       "      <td>1332</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>https://sra-downloadb.st-va.ncbi.nlm.nih.gov/s...</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>CRASSPHAGE CONSORTIUM</td>\n",
       "      <td>SRA830045</td>\n",
       "      <td>NaN</td>\n",
       "      <td>public</td>\n",
       "      <td>0FCC68EB5CA35EE316539F4C0348E54F</td>\n",
       "      <td>39E03A9EF05D22741DCB6E80175743D0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>DRR019507</td>\n",
       "      <td>2016-03-05 08:47:09</td>\n",
       "      <td>2016-03-05 08:53:45</td>\n",
       "      <td>8517086</td>\n",
       "      <td>4246043700</td>\n",
       "      <td>8517086</td>\n",
       "      <td>498</td>\n",
       "      <td>2687</td>\n",
       "      <td>NaN</td>\n",
       "      <td>https://sra-downloadb.st-va.ncbi.nlm.nih.gov/s...</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>MIYAZAKI</td>\n",
       "      <td>DRA001862</td>\n",
       "      <td>NaN</td>\n",
       "      <td>public</td>\n",
       "      <td>4580B926CA1E057CDED2089BF0BB1003</td>\n",
       "      <td>D456B6384D34959FC2706FACB199D719</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>DRR021939</td>\n",
       "      <td>2015-01-15 10:15:18</td>\n",
       "      <td>2015-01-15 10:18:19</td>\n",
       "      <td>18645</td>\n",
       "      <td>10781479</td>\n",
       "      <td>0</td>\n",
       "      <td>578</td>\n",
       "      <td>28</td>\n",
       "      <td>NaN</td>\n",
       "      <td>https://sra-downloadb.be-md.ncbi.nlm.nih.gov/s...</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>TOHOKUGL</td>\n",
       "      <td>DRA002421</td>\n",
       "      <td>NaN</td>\n",
       "      <td>public</td>\n",
       "      <td>B36925AE092F9ED96FA6EE6435522481</td>\n",
       "      <td>B904CCB0FE4FA1A287DDB5177B781F85</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>DRR025071</td>\n",
       "      <td>2016-03-09 18:20:13</td>\n",
       "      <td>2016-03-10 05:18:09</td>\n",
       "      <td>65366252</td>\n",
       "      <td>19609875600</td>\n",
       "      <td>65366252</td>\n",
       "      <td>300</td>\n",
       "      <td>10905</td>\n",
       "      <td>NaN</td>\n",
       "      <td>https://sra-downloadb.st-va.ncbi.nlm.nih.gov/s...</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>KYOTO_PREF</td>\n",
       "      <td>DRA002733</td>\n",
       "      <td>NaN</td>\n",
       "      <td>public</td>\n",
       "      <td>ED1A01B5C747C37B3641528802099C4C</td>\n",
       "      <td>C55C7EABB80C9A01ADA378250F62561A</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 47 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          Run          ReleaseDate             LoadDate     spots  \\\n",
       "0   DRR014176  2015-07-27 10:27:35  2015-07-27 10:32:25  30807540   \n",
       "1  SRR8391109  2019-01-07 15:57:36  2019-01-07 15:51:59         1   \n",
       "2   DRR019507  2016-03-05 08:47:09  2016-03-05 08:53:45   8517086   \n",
       "3   DRR021939  2015-01-15 10:15:18  2015-01-15 10:18:19     18645   \n",
       "4   DRR025071  2016-03-09 18:20:13  2016-03-10 05:18:09  65366252   \n",
       "\n",
       "         bases  spots_with_mates  avgLength  size_MB AssemblyName  \\\n",
       "0   3734314174                 0        121     2297          NaN   \n",
       "1         1332                 0       1332        0          NaN   \n",
       "2   4246043700           8517086        498     2687          NaN   \n",
       "3     10781479                 0        578       28          NaN   \n",
       "4  19609875600          65366252        300    10905          NaN   \n",
       "\n",
       "                                       download_path  ... Affection_Status  \\\n",
       "0  https://sra-downloadb.st-va.ncbi.nlm.nih.gov/s...  ...              NaN   \n",
       "1  https://sra-downloadb.st-va.ncbi.nlm.nih.gov/s...  ...              NaN   \n",
       "2  https://sra-downloadb.st-va.ncbi.nlm.nih.gov/s...  ...              NaN   \n",
       "3  https://sra-downloadb.be-md.ncbi.nlm.nih.gov/s...  ...              NaN   \n",
       "4  https://sra-downloadb.st-va.ncbi.nlm.nih.gov/s...  ...              NaN   \n",
       "\n",
       "  Analyte_Type Histological_Type Body_Site             CenterName Submission  \\\n",
       "0          NaN               NaN       NaN                   NIID  DRA001171   \n",
       "1          NaN               NaN       NaN  CRASSPHAGE CONSORTIUM  SRA830045   \n",
       "2          NaN               NaN       NaN               MIYAZAKI  DRA001862   \n",
       "3          NaN               NaN       NaN               TOHOKUGL  DRA002421   \n",
       "4          NaN               NaN       NaN             KYOTO_PREF  DRA002733   \n",
       "\n",
       "   dbgap_study_accession  Consent                           RunHash  \\\n",
       "0                    NaN   public  4DA05CDDF555DF45B83261F309ADDDE7   \n",
       "1                    NaN   public  0FCC68EB5CA35EE316539F4C0348E54F   \n",
       "2                    NaN   public  4580B926CA1E057CDED2089BF0BB1003   \n",
       "3                    NaN   public  B36925AE092F9ED96FA6EE6435522481   \n",
       "4                    NaN   public  ED1A01B5C747C37B3641528802099C4C   \n",
       "\n",
       "                           ReadHash  \n",
       "0  CD8372E963A92B67C96160309CBDD52D  \n",
       "1  39E03A9EF05D22741DCB6E80175743D0  \n",
       "2  D456B6384D34959FC2706FACB199D719  \n",
       "3  B904CCB0FE4FA1A287DDB5177B781F85  \n",
       "4  C55C7EABB80C9A01ADA378250F62561A  \n",
       "\n",
       "[5 rows x 47 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# runinfo: all of the run information, by accession.\n",
    "# note in particular `ScientificName`\n",
    "run_info = pd.read_csv('all.runinfo.csv')\n",
    "run_info.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Of 9988 datasets, 2000 have no matches.\n"
     ]
    }
   ],
   "source": [
    "# our outputs from gather\n",
    "summary_df = pd.read_csv('all.summary.csv')\n",
    "\n",
    "total_num = len(summary_df)\n",
    "num_no_matches = len(summary_df[summary_df['n_matches'] == 0])\n",
    "print(f\"Of {total_num} datasets, {num_no_matches} have no matches.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# build a df for the non-empty ones\n",
    "nonempty_df = summary_df[summary_df['n_matches'] != 0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>sample</th>\n",
       "      <th>n_matches</th>\n",
       "      <th>p_weighted_covered</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>DRR014176</td>\n",
       "      <td>373</td>\n",
       "      <td>88.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>DRR019507</td>\n",
       "      <td>97</td>\n",
       "      <td>4.3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>DRR021939</td>\n",
       "      <td>1</td>\n",
       "      <td>68.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>DRR025071</td>\n",
       "      <td>1069</td>\n",
       "      <td>39.1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>DRR033608</td>\n",
       "      <td>26</td>\n",
       "      <td>12.2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      sample  n_matches  p_weighted_covered\n",
       "0  DRR014176        373                88.1\n",
       "1  DRR019507         97                 4.3\n",
       "2  DRR021939          1                68.0\n",
       "3  DRR025071       1069                39.1\n",
       "4  DRR033608         26                12.2"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nonempty_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>n_matches</th>\n",
       "      <th>p_weighted_covered</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>7988.000000</td>\n",
       "      <td>7988.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>273.413495</td>\n",
       "      <td>49.273748</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>455.414336</td>\n",
       "      <td>33.120600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>3.000000</td>\n",
       "      <td>16.100000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>64.000000</td>\n",
       "      <td>50.600000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>376.000000</td>\n",
       "      <td>81.600000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>6215.000000</td>\n",
       "      <td>100.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         n_matches  p_weighted_covered\n",
       "count  7988.000000         7988.000000\n",
       "mean    273.413495           49.273748\n",
       "std     455.414336           33.120600\n",
       "min       1.000000            0.000000\n",
       "25%       3.000000           16.100000\n",
       "50%      64.000000           50.600000\n",
       "75%     376.000000           81.600000\n",
       "max    6215.000000          100.000000"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nonempty_df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>n_matches</th>\n",
       "      <th>p_weighted_covered</th>\n",
       "      <th>ScientificName</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>DRR014176</th>\n",
       "      <td>373</td>\n",
       "      <td>88.1</td>\n",
       "      <td>human metagenome</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DRR019507</th>\n",
       "      <td>97</td>\n",
       "      <td>4.3</td>\n",
       "      <td>Bos taurus</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DRR021939</th>\n",
       "      <td>1</td>\n",
       "      <td>68.0</td>\n",
       "      <td>uncultured bacterium</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DRR025071</th>\n",
       "      <td>1069</td>\n",
       "      <td>39.1</td>\n",
       "      <td>pig gut metagenome</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>DRR033608</th>\n",
       "      <td>26</td>\n",
       "      <td>12.2</td>\n",
       "      <td>human oral metagenome</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           n_matches  p_weighted_covered         ScientificName\n",
       "DRR014176        373                88.1       human metagenome\n",
       "DRR019507         97                 4.3             Bos taurus\n",
       "DRR021939          1                68.0   uncultured bacterium\n",
       "DRR025071       1069                39.1     pig gut metagenome\n",
       "DRR033608         26                12.2  human oral metagenome"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# CTB note: this join may duplicate columns. FIXME.\n",
    "nonempty_df2 = nonempty_df.set_index('sample').join(run_info.set_index(\"Run\")['ScientificName'])\n",
    "nonempty_df2[~nonempty_df2.ScientificName.isnull()].head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8055"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(nonempty_df2.ScientificName.isnull())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>Organism Name</th>\n",
       "      <th>sample_type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>viral metagenome</td>\n",
       "      <td>viral metagenome</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>human viral metagenome</td>\n",
       "      <td>viral metagenome</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>marine metagenome</td>\n",
       "      <td>marine metagenome</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>human oral metagenome</td>\n",
       "      <td>human metagenome</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Equus caballus</td>\n",
       "      <td>animal metagenome</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            Organism Name        sample_type\n",
       "0        viral metagenome   viral metagenome\n",
       "1  human viral metagenome   viral metagenome\n",
       "2       marine metagenome  marine metagenome\n",
       "3   human oral metagenome   human metagenome\n",
       "4          Equus caballus  animal metagenome"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# our hand-coded remapping information for `ScientificName` to collate.\n",
    "sra_remapping = pd.read_csv('categories/mapping.csv')\n",
    "sra_remapping.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# remap all the `ScientificName` into `sample_type`, paying attention to our hand coding.\n",
    "# if not recoded, just return original value.\n",
    "#\n",
    "# this could be made much faster with pandas magic. FIXME!\n",
    "def do_remap(val):\n",
    "    new_vals = list(sra_remapping[sra_remapping['Organism Name'] == val].sample_type)\n",
    "    if new_vals:\n",
    "        return new_vals[0]\n",
    "    return val\n",
    "    \n",
    "nonempty_df2['sample_type'] = nonempty_df2['ScientificName'].apply(do_remap)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "human metagenome            3907\n",
       "animal metagenome            524\n",
       "unknown                      502\n",
       "soil metagenome              396\n",
       "marine metagenome            260\n",
       "plant metagenome             127\n",
       "freshwater metagenome        113\n",
       "sediment metagenome          107\n",
       "viral metagenome             100\n",
       "bacterium                     68\n",
       "wastewater metagenome         64\n",
       "indoor metagenome             64\n",
       "aquatic metagenome            49\n",
       "food metagenome               42\n",
       "groundwater metagenome        38\n",
       "bioreactor metagenome         34\n",
       "hot springs metagenome        33\n",
       "human vaginal metagenome      32\n",
       "milk metagenome               32\n",
       "biofilm metagenome            31\n",
       "Name: sample_type, dtype: int64"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# count the things.\n",
    "nonempty_df2[\"sample_type\"].value_counts()[:20]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['human metagenome', 'animal metagenome', 'unknown', 'soil metagenome', 'marine metagenome']\n"
     ]
    }
   ],
   "source": [
    "#plot_names = ['human metagenome', 'marine metagenome']\n",
    "# pick the top NUM to plot\n",
    "NUM=5\n",
    "plot_names = list(nonempty_df2[\"sample_type\"].value_counts()[0:NUM].index)\n",
    "print(plot_names)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "fig, axes = plt.subplots(figsize=[20, 5])\n",
    "\n",
    "datasets = []\n",
    "for name in plot_names:\n",
    "    datasets.append(nonempty_df2[nonempty_df2.sample_type == name][\"p_weighted_covered\"].values)\n",
    "\n",
    "axes.violinplot(dataset = datasets)\n",
    "#                           nonempty_df[df.Scenario == 'S2'][\"LMP\"].values,\n",
    "\n",
    "axes.set_title('percent classified (abund weighted)')\n",
    "axes.yaxis.grid(True)\n",
    "#axes.set_xlabel('Scenario')\n",
    "#axes.set_ylabel('LMP ($/MWh)')\n",
    "\n",
    "labels = [ f\"{n} ({len(d)})\" for (n, d) in zip(plot_names, datasets) ]\n",
    "\n",
    "axes.set_xticks([y + 1 for y in range(len(plot_names))],\n",
    "                labels=labels)\n",
    "axes.set_ylabel('percent covered by known genomes')\n",
    "\n",
    "#plt.show()\n",
    "None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "fig, axes = plt.subplots(figsize=[20, 5])\n",
    "\n",
    "datasets = []\n",
    "for name in plot_names:\n",
    "    datasets.append(nonempty_df2[nonempty_df2.sample_type == name][\"n_matches\"].values)\n",
    "\n",
    "axes.violinplot(dataset = datasets)\n",
    "#                           nonempty_df[df.Scenario == 'S2'][\"LMP\"].values,\n",
    "\n",
    "axes.set_title('total number of matches')\n",
    "axes.yaxis.grid(True)\n",
    "#axes.set_xlabel('Scenario')\n",
    "#axes.set_ylabel('LMP ($/MWh)')\n",
    "\n",
    "labels = [ f\"{n} ({len(d)})\" for (n, d) in zip(plot_names, datasets) ]\n",
    "\n",
    "axes.set_xticks([y + 1 for y in range(len(plot_names))],\n",
    "                labels=labels)\n",
    "#axes.set_xlabel('Four separate samples')\n",
    "\n",
    "#plt.show()\n",
    "None"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python (myenv)",
   "language": "python",
   "name": "myenv"
  },
  "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
}