notebooks/18_profile_recovery.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Summary\n",
    "\n",
    "----"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import concurrent.futures\n",
    "import gzip\n",
    "import os\n",
    "import shutil\n",
    "import subprocess\n",
    "from collections import Counter\n",
    "from pathlib import Path\n",
    "\n",
    "import logomaker\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import proteinsolver\n",
    "import pyarrow as pa\n",
    "import pyarrow.parquet as pq\n",
    "from IPython.display import set_matplotlib_formats\n",
    "from kmbio import PDB\n",
    "from tqdm.notebook import tqdm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.set_option(\"max_columns\", 1000)\n",
    "pd.set_option(\"max_rows\", 1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "set_matplotlib_formats(\"png\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "PosixPath('/home/kimlab1/strokach/workspace/proteinsolver/notebooks/12_profile_recovery')"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "NOTEBOOK_DIR = Path(\"12_profile_recovery\").resolve()\n",
    "NOTEBOOK_DIR.mkdir(exist_ok =True)\n",
    "\n",
    "NOTEBOOK_DIR"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "UNIQUE_ID = \"191f05de\"\n",
    "\n",
    "SEQUENCE_GENERATION_METHOD = \"expectimax\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "PosixPath('/home/kimlab1/strokach/workspace/proteinsolver/notebooks/generate_protein_sequences')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "DESIGNS_PATH = NOTEBOOK_DIR.parent.joinpath(\"generate_protein_sequences\")\n",
    "assert DESIGNS_PATH.is_dir()\n",
    "\n",
    "DESIGNS_PATH"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "tags": [
     "parameters"
    ]
   },
   "outputs": [],
   "source": [
    "# 1n5uA03 4z8jA00 4unuA00 4beuA02\n",
    "STRUCTURE_ID = \"4beuA02\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2228368\n"
     ]
    }
   ],
   "source": [
    "from papermill import execute_notebook\n",
    "\n",
    "\n",
    "def worker(params):\n",
    "    input_path = NOTEBOOK_DIR.with_suffix(\".ipynb\").as_posix()\n",
    "    output_path = NOTEBOOK_DIR.parent.joinpath(\n",
    "        NOTEBOOK_DIR.name + f\"-{params['STRUCTURE_ID']}.ipynb\"\n",
    "    ).as_posix()\n",
    "    out = execute_notebook(input_path, output_path, params)\n",
    "    return out\n",
    "\n",
    "\n",
    "print(os.getpid())\n",
    "if os.getpid() == None:\n",
    "    params_list = [\n",
    "        {\"STRUCTURE_ID\": structure_id}\n",
    "        for structure_id in [\"1n5uA03\", \"4z8jA00\", \"4unuA00\", \"4beuA02\"]\n",
    "    ]\n",
    "    with concurrent.futures.ThreadPoolExecutor(4) as pool:\n",
    "        futures = pool.map(worker, params_list)\n",
    "        results = list(futures)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'map_sequence_identity': 0.41013824884792627,\n",
       " 'map_proba': 0.5304911136627197,\n",
       " 'map_logproba': -0.6995142102241516,\n",
       " 'ref_sequence_identity': 1.0,\n",
       " 'ref_proba': 0.3164460062980652,\n",
       " 'ref_logproba': -1.7113533020019531,\n",
       " 'sequence': 'LGQFQSNIEQFKSHMNANTKICAIMKADAYGNGIRGLMPTIIAQGIPCVGVASNAEARAVRESGFKGELIRVRSASLSEMSSALDLNIEELIGTHQQALDLAELAKQSGKTLKVHIALNDGGMGRNGIDMTTEAGKKEAVSIATQPSLSVVGIMTHFPNYNADEVRAKLAQFKESSTWLMQQANLKREEITLHVANSYTALNVPEAQLDMVRPGGVL'}"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "proteinsolver_stat_dict = {\n",
    "    \"1n5uA03\": {\n",
    "        \"map_sequence_identity\": 0.41304347826086957,\n",
    "        \"map_proba\": 0.42337074875831604,\n",
    "        \"map_logproba\": -1.015546441078186,\n",
    "        \"ref_sequence_identity\": 1.0,\n",
    "        \"ref_proba\": 0.28906363248825073,\n",
    "        \"ref_logproba\": -1.9106740951538086,\n",
    "        \"sequence\": \"KFGERAFKAWAVARLSQRFPKAEFAEVSKLVTDLTKVHTECCHGDLLECADDRADLAKYICENQDSISSKLKECCEKPLLEKSHCIAEVEND\",\n",
    "    },\n",
    "    \"4z8jA00\": {\n",
    "        \"map_sequence_identity\": 0.4166666666666667,\n",
    "        \"map_proba\": 0.48954343795776367,\n",
    "        \"map_logproba\": -0.7965440154075623,\n",
    "        \"ref_sequence_identity\": 1.0,\n",
    "        \"ref_proba\": 0.325325071811676,\n",
    "        \"ref_logproba\": -1.840335726737976,\n",
    "        \"sequence\": \"SPRVVRIVKSESGYGFNVRGQVSEGGQLRSINGELYAPLQHVSAVLPGGAADRAGVRKGDRILEVNGVNVEGATHKQVVDLIRAGEKELILTVLSV\",\n",
    "    },\n",
    "    \"4unuA00\": {\n",
    "        \"map_sequence_identity\": 0.43119266055045874,\n",
    "        \"map_proba\": 0.46164458990097046,\n",
    "        \"map_logproba\": -0.8745880126953125,\n",
    "        \"ref_sequence_identity\": 1.0,\n",
    "        \"ref_proba\": 0.3054293692111969,\n",
    "        \"ref_logproba\": -1.7174205780029297,\n",
    "        \"sequence\": \"SALTQPPSASGSLGQSVTISCTGTSSDVGGYNYVSWYQQHAGKAPKVIIYEVNKRPSGVPDRFSGSKSGNTASLTVSGLQAEDEADYYCSSYEGSDNFVFGTGTKVTVL\",\n",
    "    },\n",
    "    \"4beuA02\": {\n",
    "        \"map_sequence_identity\": 0.41013824884792627,\n",
    "        \"map_proba\": 0.5304911136627197,\n",
    "        \"map_logproba\": -0.6995142102241516,\n",
    "        \"ref_sequence_identity\": 1.0,\n",
    "        \"ref_proba\": 0.3164460062980652,\n",
    "        \"ref_logproba\": -1.7113533020019531,\n",
    "        \"sequence\": \"LGQFQSNIEQFKSHMNANTKICAIMKADAYGNGIRGLMPTIIAQGIPCVGVASNAEARAVRESGFKGELIRVRSASLSEMSSALDLNIEELIGTHQQALDLAELAKQSGKTLKVHIALNDGGMGRNGIDMTTEAGKKEAVSIATQPSLSVVGIMTHFPNYNADEVRAKLAQFKESSTWLMQQANLKREEITLHVANSYTALNVPEAQLDMVRPGGVL\",\n",
    "    },\n",
    "}\n",
    "\n",
    "proteinsolver_stats = proteinsolver_stat_dict[STRUCTURE_ID]\n",
    "\n",
    "proteinsolver_stats"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Common functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "columns = list(\"GVALICMFWPDESTYQNKRH\")\n",
    "assert len(columns) == 20\n",
    "assert len(set(columns)) == 20\n",
    "\n",
    "\n",
    "def get_sequence_logo_score(sequence, aa_proba_adj):\n",
    "    return np.mean(\n",
    "        np.log(aa_proba_adj.values[np.arange(len(sequence)), [columns.index(aa) for aa in sequence]])\n",
    "    )\n",
    "\n",
    "\n",
    "def get_sequence_identity(seq1, seq2):\n",
    "    assert len(seq1) == len(seq2)\n",
    "    return sum((aa1 == aa2) for aa1, aa2 in zip(seq1, seq2)) / len(seq1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load reference LOGOs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "AA_PROBA_ADJ = {}\n",
    "for rl in [100, 95, 60, 35]:\n",
    "    AA_PROBA_ADJ[rl] = pq.read_table(\n",
    "        NOTEBOOK_DIR.parents[2].joinpath(\n",
    "            \"proteinsolver\",\n",
    "            \"notebooks\",\n",
    "            \"11_profile_recovery\",\n",
    "            f\"{STRUCTURE_ID}-rl{rl}-aa-proba-adj.parquet\",\n",
    "        )\n",
    "    ).to_pandas()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load reference structure"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "PosixPath('/home/kimlab1/strokach/workspace/proteinsolver/proteinsolver/data/inputs/4beuA02.pdb')"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "STRUCTURE_FILE = Path(proteinsolver.__path__[0]).joinpath(\"data\", \"inputs\", f\"{STRUCTURE_ID}.pdb\")\n",
    "assert STRUCTURE_FILE.is_file()\n",
    "\n",
    "STRUCTURE_FILE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "structure = PDB.load(STRUCTURE_FILE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "structure_df = structure.to_dataframe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Ingraham"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "coarse\tdist  full  hbonds\n"
     ]
    }
   ],
   "source": [
    "ingraham_data_dir = NOTEBOOK_DIR.parents[2].joinpath(\n",
    "    \"neurips19-graph-protein-design\", \"notebooks\", \"03_generate_designs\", STRUCTURE_ID\n",
    ")\n",
    "\n",
    "!ls {ingraham_data_dir}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "def parse_fasta_file(fin):\n",
    "    results = []\n",
    "    row = None\n",
    "    for line in fin:\n",
    "        if line.startswith(\">\"):\n",
    "            if row is not None:\n",
    "                results.append(row)\n",
    "                row = None\n",
    "            header = [l.strip(\"> \") for l in line.strip().split(\",\")]\n",
    "            if header[0] == \"Native\":\n",
    "                row = {\n",
    "                    \"T\": None,\n",
    "                    \"sample\": -1,\n",
    "                    \"score\": float(header[1].replace(\"score=\", \"\")),\n",
    "                    \"sequence\": \"\",\n",
    "                }\n",
    "            else:\n",
    "                row = {\n",
    "                    \"T\": header[0].replace(\"T=\", \"\"),\n",
    "                    \"sample\": header[1].replace(\"sample=\", \"\"),\n",
    "                    \"score\": float(header[2].replace(\"score=\", \"\")),\n",
    "                    \"sequence\": \"\",\n",
    "                }\n",
    "        else:\n",
    "            row[\"sequence\"] += line.strip()\n",
    "    if row is not None:\n",
    "        results.append(row)\n",
    "    return results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "result_dfs = {}\n",
    "for model_features in [\"full\", \"dist\", \"coarse\", \"hbonds\"]:\n",
    "    with gzip.open(\n",
    "        ingraham_data_dir.joinpath(model_features, \"alignments\", f\"{STRUCTURE_ID}.fa.gz\"), \"rt\"\n",
    "    ) as fin:\n",
    "        results = parse_fasta_file(fin)\n",
    "        results_df = pd.DataFrame(results)\n",
    "        result_dfs[model_features] = results_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>T</th>\n",
       "      <th>sample</th>\n",
       "      <th>score</th>\n",
       "      <th>sequence</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>None</td>\n",
       "      <td>-1</td>\n",
       "      <td>1.409763</td>\n",
       "      <td>LGQFQSNIEQFKSHMNANTKICAIMKADAYGNGIRGLMPTIIAQGI...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>1.825835</td>\n",
       "      <td>ASLLKIRLQNMSSELRMDTRYLIMYKADAKGKGVESLLPSIKCLKW...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>1.762215</td>\n",
       "      <td>MDGIESALGAFHSKLPDNTQILAVLRATANGRGVASIMPIFAKMNV...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.603048</td>\n",
       "      <td>LGNIRQQLQDFQSQLEEETSISLALKANAHGYGVEKLMPVFIEQNV...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>3</td>\n",
       "      <td>2.078782</td>\n",
       "      <td>SEKTRERLETLSRHLADNAFICPQVTPDAFGTGVTGCLPLLQEEKI...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      T sample     score                                           sequence\n",
       "0  None     -1  1.409763  LGQFQSNIEQFKSHMNANTKICAIMKADAYGNGIRGLMPTIIAQGI...\n",
       "1   1.0      0  1.825835  ASLLKIRLQNMSSELRMDTRYLIMYKADAKGKGVESLLPSIKCLKW...\n",
       "2   1.0      1  1.762215  MDGIESALGAFHSKLPDNTQILAVLRATANGRGVASIMPIFAKMNV...\n",
       "3   1.0      2  1.603048  LGNIRQQLQDFQSQLEEETSISLALKANAHGYGVEKLMPVFIEQNV...\n",
       "4   1.0      3  2.078782  SEKTRERLETLSRHLADNAFICPQVTPDAFGTGVTGCLPLLQEEKI..."
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result_dfs[\"full\"].head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Calculate scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# for feature_set in result_dfs:\n",
    "#     result_dfs[feature_set][\"seq_identity\"] = list(\n",
    "#         tqdm(\n",
    "#             (\n",
    "#                 get_sequence_identity(sequence, proteinsolver_stats[\"sequence\"])\n",
    "#                 for sequence in result_dfs[feature_set][\"sequence\"].values\n",
    "#             ),\n",
    "#             total=len(result_dfs[feature_set]),\n",
    "#         )\n",
    "#     )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# for feature_set in result_dfs:\n",
    "#     for rl in AA_PROBA_ADJ:\n",
    "#         result_dfs[feature_set][f\"avg_logprob_{rl}\"] = list(\n",
    "#             tqdm(\n",
    "#                 (\n",
    "#                     get_sequence_logo_score(sequence, AA_PROBA_ADJ[rl])\n",
    "#                     for sequence in result_dfs[feature_set][\"sequence\"].values\n",
    "#                 ),\n",
    "#                 total=len(result_dfs[feature_set]),\n",
    "#             )\n",
    "#         )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Accuracy plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "summary_dfs = {}\n",
    "for model_features in [\"full\", \"dist\", \"coarse\", \"hbonds\"]:\n",
    "    df = pd.read_csv(ingraham_data_dir.joinpath(model_features, \"results.csv\"))\n",
    "    df[\"model_features\"] = model_features\n",
    "    summary_dfs[model_features] = df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fg, ax = plt.subplots()\n",
    "for model_features, df in summary_dfs.items():\n",
    "    x = np.arange(len(df))\n",
    "    ax.plot(x, df[\"similarity\"], \"o-\", label=model_features)\n",
    "    ax.set_xticks(x)\n",
    "    ax.set_xticklabels(df[\"T\"])\n",
    "    ax.grid(True, linestyle=\"--\")\n",
    "    ax.set_ylim(0.1, 0.65)\n",
    "ax.hlines(proteinsolver_stats[\"map_sequence_identity\"], *ax.set_xlim(), linestyle='--')\n",
    "ax.legend()\n",
    "ax.set_title(STRUCTURE_ID)\n",
    "fg.savefig(NOTEBOOK_DIR.joinpath(f\"{STRUCTURE_ID}-ingraham-seq-identity.svg\"))\n",
    "fg.savefig(NOTEBOOK_DIR.joinpath(f\"{STRUCTURE_ID}-ingraham-seq-identity.pdf\"))\n",
    "fg.savefig(NOTEBOOK_DIR.joinpath(f\"{STRUCTURE_ID}-ingraham-seq-identity.png\"), dpi=300)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Proteinsolver"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[PosixPath('/home/kimlab1/strokach/workspace/proteinsolver/notebooks/generate_protein_sequences/designs-191f05de-expectimax-1n5uA03-2044-c02.parquet'),\n",
       " PosixPath('/home/kimlab1/strokach/workspace/proteinsolver/notebooks/generate_protein_sequences/designs-191f05de-expectimax-1n5uA03-2012-c02.parquet'),\n",
       " PosixPath('/home/kimlab1/strokach/workspace/proteinsolver/notebooks/generate_protein_sequences/designs-191f05de-expectimax-1n5uA03-3028-c02.parquet'),\n",
       " PosixPath('/home/kimlab1/strokach/workspace/proteinsolver/notebooks/generate_protein_sequences/designs-191f05de-expectimax-1n5uA03-3099-c02.parquet'),\n",
       " PosixPath('/home/kimlab1/strokach/workspace/proteinsolver/notebooks/generate_protein_sequences/designs-191f05de-expectimax-1n5uA03-1037-c02.parquet')]"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "design_files = list(DESIGNS_PATH.glob(f\"designs-{UNIQUE_ID}-{SEQUENCE_GENERATION_METHOD}-{STRUCTURE_FILE.stem}-*.parquet\"))\n",
    "design_files[-5:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2e945b19ed444047af45fabe79225a67",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=376.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "dfs = []\n",
    "for design_file in tqdm(design_files):\n",
    "    df = pq.read_table(design_file).to_pandas(integer_object_nulls=True)\n",
    "    dfs.append(df)\n",
    "proteinsolver_df = pd.concat(dfs, sort=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5472\n",
      "5472\n"
     ]
    }
   ],
   "source": [
    "print(len(df))\n",
    "proteinsolver_df = proteinsolver_df.drop_duplicates(subset=[\"sequence\"])\n",
    "print(len(df))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Calculate scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "82e763f296a54c65b3ae4190060c313f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=2057472.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "proteinsolver_df[\"seq_identity\"] = list(\n",
    "    tqdm(\n",
    "        (\n",
    "            get_sequence_identity(sequence, proteinsolver_stats[\"sequence\"])\n",
    "            for sequence in proteinsolver_df[\"sequence\"].values\n",
    "        ),\n",
    "        total=len(proteinsolver_df),\n",
    "    )\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e09a966b76804b9585ed567406bb76ca",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=2057472.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "034ddb80925e463fba8adb154270e845",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=2057472.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4483c33521b0429792f369be4272baf7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=2057472.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6e0e6c521e944f9b809942465059c796",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=2057472.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "for rl in AA_PROBA_ADJ:\n",
    "    proteinsolver_df[f\"avg_logprob_{rl}\"] = list(\n",
    "        tqdm(\n",
    "            (\n",
    "                get_sequence_logo_score(sequence, AA_PROBA_ADJ[rl])\n",
    "                for sequence in proteinsolver_df[\"sequence\"].values\n",
    "            ),\n",
    "            total=len(proteinsolver_df),\n",
    "        )\n",
    "    )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Make plots"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "result_dfs[\"proteinsolver\"] = proteinsolver_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x288 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fg, axs = plt.subplots(1, 5, figsize=(9, 4))\n",
    "\n",
    "ylim = [None, None]\n",
    "ps_median = None\n",
    "for i, (featureset, df) in enumerate(result_dfs.items()):\n",
    "    data = {}\n",
    "    if featureset in [\"proteinsolver\"]:\n",
    "        data[\"all\\ndesigns\"] = df[\"seq_identity\"]\n",
    "    else:\n",
    "        for temp, gp in df.groupby(\"T\"):\n",
    "            data[temp] = gp[\"seq_identity\"].values\n",
    "\n",
    "    ax = axs[i]\n",
    "    box_values = list(data.values())\n",
    "    positions = np.linspace(-0.5, 0.5, len(data) + 2)\n",
    "    width = 0.1\n",
    "    _ = ax.boxplot(box_values, positions=positions[1:-1], sym=\".\", widths=width)\n",
    "    if featureset in [\"proteinsolver\"]:\n",
    "        ax.set_xticklabels([f\" T=1.00\" for key in data.keys()], rotation=\"vertical\")\n",
    "        ax.set_xlabel(f\"ProteinSolver\", fontsize=\"large\")\n",
    "    else:\n",
    "        ax.set_xticklabels([f\" T={float(key):.2f}\" for key in data.keys()], rotation=\"vertical\")\n",
    "        ax.set_xlabel(f\"ST {featureset}\", fontsize=\"large\")\n",
    "\n",
    "    ax.set_xlim(positions[0], positions[-1])\n",
    "\n",
    "    y_min_ = min([v.min() for v in box_values])\n",
    "    if ylim[0] is None or ylim[0] > y_min_:\n",
    "        ylim[0] = y_min_\n",
    "    y_max_ = max([v.max() for v in box_values])\n",
    "    if ylim[1] is None or ylim[1] < y_max_:\n",
    "        ylim[1] = y_max_\n",
    "    if featureset in [\"proteinsolver\"]:\n",
    "        ps_median = np.median(box_values[0])\n",
    "\n",
    "    if i == 0:\n",
    "        ax.set_ylabel(\"Average sequence identity\")\n",
    "\n",
    "    if i > 0:\n",
    "        ax.set_yticklabels(\"\")\n",
    "\n",
    "    for tic in ax.yaxis.get_major_ticks():\n",
    "        tic.tick1line.set_visible(False)\n",
    "        tic.tick2line.set_visible(False)\n",
    "\n",
    "    ax.spines[\"top\"].set_visible(False)\n",
    "    ax.spines[\"right\"].set_visible(False)\n",
    "    ax.spines[\"left\"].set_visible(False)\n",
    "    ax.grid(axis=\"y\", linestyle=\":\")\n",
    "\n",
    "ylim = [ylim[0] * 10 // 1 / 10, np.ceil(ylim[1] * 10) / 10]\n",
    "for ax in axs:\n",
    "    ax.set_ylim(ylim)\n",
    "    ax.hlines(\n",
    "        proteinsolver_stats[\"map_sequence_identity\"],\n",
    "        *ax.set_xlim(),\n",
    "        linestyle=\"--\",\n",
    "        linewidth=1,\n",
    "        color=plt.get_cmap(\"Set2\")(2),\n",
    "    )\n",
    "    ax.hlines(\n",
    "        ps_median, *ax.set_xlim(), linestyle=\"--\", linewidth=1, color=plt.get_cmap(\"Set2\")(3),\n",
    "    )\n",
    "\n",
    "fg.subplots_adjust(left=0.08, bottom=0.23, top=0.97, right=0.99)\n",
    "fg.savefig(NOTEBOOK_DIR.joinpath(f\"{STRUCTURE_ID}-seq-identity.svg\"))\n",
    "fg.savefig(NOTEBOOK_DIR.joinpath(f\"{STRUCTURE_ID}-seq-identity.pdf\"))\n",
    "fg.savefig(NOTEBOOK_DIR.joinpath(f\"{STRUCTURE_ID}-seq-identity.png\"), dpi=300)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x288 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x288 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x288 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x288 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for rl in AA_PROBA_ADJ:\n",
    "    fg, axs = plt.subplots(1, 5, figsize=(9, 4), constrained_layout=False)\n",
    "\n",
    "    ylim = [None, None]\n",
    "    ps_median = None\n",
    "\n",
    "    for i, (featureset, df) in enumerate(result_dfs.items()):\n",
    "        data = {}\n",
    "        if featureset in [\"proteinsolver\"]:\n",
    "            data[\"all\\ndesigns\"] = np.exp(df[f\"avg_logprob_{rl}\"].values)\n",
    "        else:\n",
    "            for temp, gp in df.groupby(\"T\"):\n",
    "                data[temp] = np.exp(gp[f\"avg_logprob_{rl}\"].values)\n",
    "\n",
    "        ax = axs[i]\n",
    "        box_values = list(data.values())\n",
    "        positions = np.linspace(-0.5, 0.5, len(data) + 2)\n",
    "        width = 0.1\n",
    "        _ = ax.boxplot(box_values, positions=positions[1:-1], sym=\".\", widths=width)\n",
    "        if featureset in [\"proteinsolver\"]:\n",
    "            ax.set_xticklabels([f\" T=1.00\" for key in data.keys()], rotation=\"vertical\")\n",
    "            ax.set_xlabel(f\"ProteinSolver\", fontsize=\"large\")\n",
    "        else:\n",
    "            ax.set_xticklabels([f\" T={float(key):.2f}\" for key in data.keys()], rotation=\"vertical\")\n",
    "            ax.set_xlabel(f\"ST {featureset}\", fontsize=\"large\")\n",
    "\n",
    "        ax.set_xlim(positions[0], positions[-1])\n",
    "\n",
    "        y_min_ = min([v.min() for v in box_values])\n",
    "        if ylim[0] is None or ylim[0] > y_min_:\n",
    "            ylim[0] = y_min_\n",
    "        y_max_ = max([v.max() for v in box_values])\n",
    "        if ylim[1] is None or ylim[1] < y_max_:\n",
    "            ylim[1] = y_max_\n",
    "        if featureset in [\"proteinsolver\"]:\n",
    "            ps_median = np.median(box_values[0])\n",
    "\n",
    "        if i == 0:\n",
    "            ax.set_ylabel(\"Average residue probability\")\n",
    "\n",
    "        if i > 0:\n",
    "            ax.set_yticklabels(\"\")\n",
    "\n",
    "        for tic in ax.yaxis.get_major_ticks():\n",
    "            tic.tick1line.set_visible(False)\n",
    "            tic.tick2line.set_visible(False)\n",
    "\n",
    "        ax.spines[\"top\"].set_visible(False)\n",
    "        ax.spines[\"right\"].set_visible(False)\n",
    "        ax.spines[\"left\"].set_visible(False)\n",
    "        ax.grid(axis=\"y\", linestyle=\":\")\n",
    "\n",
    "    ylim = [ylim[0] * 100 // 1 / 100, np.ceil(ylim[1] * 100) / 100]\n",
    "    for ax in axs:\n",
    "        ax.set_ylim(ylim)\n",
    "        ax.hlines(\n",
    "            np.exp(get_sequence_logo_score(proteinsolver_stats[\"sequence\"], AA_PROBA_ADJ[rl])),\n",
    "            *ax.set_xlim(),\n",
    "            linestyle=\"--\",\n",
    "            linewidth=1,\n",
    "            color=plt.get_cmap(\"Set2\")(2),\n",
    "        )\n",
    "        ax.hlines(\n",
    "            ps_median, *ax.set_xlim(), linestyle=\"--\", linewidth=1, color=plt.get_cmap(\"Set2\")(3),\n",
    "        )\n",
    "\n",
    "    fg.subplots_adjust(left=0.08, bottom=0.23, top=0.97, right=0.99)\n",
    "    fg.savefig(NOTEBOOK_DIR.joinpath(f\"{STRUCTURE_ID}-avg-logprog-rl{rl}.svg\"))\n",
    "    fg.savefig(NOTEBOOK_DIR.joinpath(f\"{STRUCTURE_ID}-avg-logprog-rl{rl}.pdf\"))\n",
    "    fg.savefig(NOTEBOOK_DIR.joinpath(f\"{STRUCTURE_ID}-avg-logprog-rl{rl}.png\"), dpi=300)"
   ]
  },
  {
   "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
}