tutorials/advanced/modelling/PCM_modelling.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# PCM Modeling\n",
    "\n",
    "In this section of the tutorial, we will use the knowledge from all previous tutorials to construct a simple PCM classification model for four  adenosine receptors (A1, A2A, A2B, A3).\n",
    "\n",
    "## Data Preparation\n",
    "\n",
    "First, we need to load the data. For this tutorial we will load the dataset containing data for all four adenosine receptors."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Failed to find the pandas get_adjustment() function to patch\n",
      "Failed to patch pandas - PandasTools will have limited functionality\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SMILES</th>\n",
       "      <th>pchembl_value_Mean</th>\n",
       "      <th>accession</th>\n",
       "      <th>QSPRID</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>QSPRID</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00000</th>\n",
       "      <td>Cc1nn(-c2cc(NC(=O)CCN(C)C)nc(-c3ccc(C)o3)n2)c(...</td>\n",
       "      <td>8.68</td>\n",
       "      <td>P29274</td>\n",
       "      <td>PCMTutorialDataset_00000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00001</th>\n",
       "      <td>Cc1nn(-c2cc(NC(=O)CCN(C)C)nc(-c3ccc(C)o3)n2)c(...</td>\n",
       "      <td>6.68</td>\n",
       "      <td>P30542</td>\n",
       "      <td>PCMTutorialDataset_00001</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00002</th>\n",
       "      <td>Nc1c(C(=O)Nc2ccc([N+](=O)[O-])cc2)sc2c1cc1CCCC...</td>\n",
       "      <td>4.82</td>\n",
       "      <td>P29274</td>\n",
       "      <td>PCMTutorialDataset_00002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00003</th>\n",
       "      <td>O=C(Nc1nc2ncccc2n2c(=O)n(-c3ccccc3)nc12)c1ccccc1</td>\n",
       "      <td>7.15</td>\n",
       "      <td>P0DMS8</td>\n",
       "      <td>PCMTutorialDataset_00003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00004</th>\n",
       "      <td>O=C(Nc1nc2ncccc2n2c(=O)n(-c3ccccc3)nc12)c1ccccc1</td>\n",
       "      <td>5.65</td>\n",
       "      <td>P29274</td>\n",
       "      <td>PCMTutorialDataset_00004</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12452</th>\n",
       "      <td>CCCOc1ccc(C=Cc2cc3c(c(=O)n(C)c(=O)n3C)n2C)cc1</td>\n",
       "      <td>6.51</td>\n",
       "      <td>P29274</td>\n",
       "      <td>PCMTutorialDataset_12452</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12453</th>\n",
       "      <td>CCOC(=O)c1cnc(NCC(C)C)n2nc(-c3ccco3)nc12</td>\n",
       "      <td>7.61</td>\n",
       "      <td>P0DMS8</td>\n",
       "      <td>PCMTutorialDataset_12453</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12454</th>\n",
       "      <td>CCOC(=O)c1cnc(NCC(C)C)n2nc(-c3ccco3)nc12</td>\n",
       "      <td>7.35</td>\n",
       "      <td>P29274</td>\n",
       "      <td>PCMTutorialDataset_12454</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12455</th>\n",
       "      <td>CCOC(=O)c1cnc(NCC(C)C)n2nc(-c3ccco3)nc12</td>\n",
       "      <td>5.15</td>\n",
       "      <td>P29275</td>\n",
       "      <td>PCMTutorialDataset_12455</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12456</th>\n",
       "      <td>CCOC(=O)c1cnc(NCC(C)C)n2nc(-c3ccco3)nc12</td>\n",
       "      <td>7.34</td>\n",
       "      <td>P30542</td>\n",
       "      <td>PCMTutorialDataset_12456</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>12457 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                                     SMILES  \\\n",
       "QSPRID                                                                        \n",
       "PCMTutorialDataset_00000  Cc1nn(-c2cc(NC(=O)CCN(C)C)nc(-c3ccc(C)o3)n2)c(...   \n",
       "PCMTutorialDataset_00001  Cc1nn(-c2cc(NC(=O)CCN(C)C)nc(-c3ccc(C)o3)n2)c(...   \n",
       "PCMTutorialDataset_00002  Nc1c(C(=O)Nc2ccc([N+](=O)[O-])cc2)sc2c1cc1CCCC...   \n",
       "PCMTutorialDataset_00003   O=C(Nc1nc2ncccc2n2c(=O)n(-c3ccccc3)nc12)c1ccccc1   \n",
       "PCMTutorialDataset_00004   O=C(Nc1nc2ncccc2n2c(=O)n(-c3ccccc3)nc12)c1ccccc1   \n",
       "...                                                                     ...   \n",
       "PCMTutorialDataset_12452      CCCOc1ccc(C=Cc2cc3c(c(=O)n(C)c(=O)n3C)n2C)cc1   \n",
       "PCMTutorialDataset_12453           CCOC(=O)c1cnc(NCC(C)C)n2nc(-c3ccco3)nc12   \n",
       "PCMTutorialDataset_12454           CCOC(=O)c1cnc(NCC(C)C)n2nc(-c3ccco3)nc12   \n",
       "PCMTutorialDataset_12455           CCOC(=O)c1cnc(NCC(C)C)n2nc(-c3ccco3)nc12   \n",
       "PCMTutorialDataset_12456           CCOC(=O)c1cnc(NCC(C)C)n2nc(-c3ccco3)nc12   \n",
       "\n",
       "                          pchembl_value_Mean accession  \\\n",
       "QSPRID                                                   \n",
       "PCMTutorialDataset_00000                8.68    P29274   \n",
       "PCMTutorialDataset_00001                6.68    P30542   \n",
       "PCMTutorialDataset_00002                4.82    P29274   \n",
       "PCMTutorialDataset_00003                7.15    P0DMS8   \n",
       "PCMTutorialDataset_00004                5.65    P29274   \n",
       "...                                      ...       ...   \n",
       "PCMTutorialDataset_12452                6.51    P29274   \n",
       "PCMTutorialDataset_12453                7.61    P0DMS8   \n",
       "PCMTutorialDataset_12454                7.35    P29274   \n",
       "PCMTutorialDataset_12455                5.15    P29275   \n",
       "PCMTutorialDataset_12456                7.34    P30542   \n",
       "\n",
       "                                            QSPRID  \n",
       "QSPRID                                              \n",
       "PCMTutorialDataset_00000  PCMTutorialDataset_00000  \n",
       "PCMTutorialDataset_00001  PCMTutorialDataset_00001  \n",
       "PCMTutorialDataset_00002  PCMTutorialDataset_00002  \n",
       "PCMTutorialDataset_00003  PCMTutorialDataset_00003  \n",
       "PCMTutorialDataset_00004  PCMTutorialDataset_00004  \n",
       "...                                            ...  \n",
       "PCMTutorialDataset_12452  PCMTutorialDataset_12452  \n",
       "PCMTutorialDataset_12453  PCMTutorialDataset_12453  \n",
       "PCMTutorialDataset_12454  PCMTutorialDataset_12454  \n",
       "PCMTutorialDataset_12455  PCMTutorialDataset_12455  \n",
       "PCMTutorialDataset_12456  PCMTutorialDataset_12456  \n",
       "\n",
       "[12457 rows x 4 columns]"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import os\n",
    "\n",
    "from qsprpred.data import MoleculeTable\n",
    "\n",
    "os.makedirs(\"../../tutorial_output/data\", exist_ok=True)\n",
    "\n",
    "dataset = MoleculeTable.fromTableFile(\n",
    "    name=\"PCMTutorialDataset\",\n",
    "    filename=\"../../tutorial_data/AR_LIGANDS.tsv\",\n",
    "    store_dir=\"../../tutorial_output/data\",\n",
    ")\n",
    "\n",
    "dataset.getDF()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Fetching Protein Data\n",
    "\n",
    "In addition, it is also possible to easily fetch the sequences for our proteins from [Papyrus](https://jcheminf.biomedcentral.com/articles/10.1186/s13321-022-00672-x) data set. You can find more information on how to fetch data from Papyrus within QSPRpred in the [data collection](../../basics/data/data_collection_with_papyrus.ipynb) tutorial. For now, we are using Papyrus ++ in version 05.6."
   ]
  },
  {
   "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>target_id</th>\n",
       "      <th>HGNC_symbol</th>\n",
       "      <th>UniProtID</th>\n",
       "      <th>Status</th>\n",
       "      <th>Organism</th>\n",
       "      <th>Classification</th>\n",
       "      <th>Length</th>\n",
       "      <th>Sequence</th>\n",
       "      <th>TID</th>\n",
       "      <th>accession</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>P29275_WT</td>\n",
       "      <td>ADORA2B</td>\n",
       "      <td>AA2BR_HUMAN</td>\n",
       "      <td>reviewed</td>\n",
       "      <td>Homo sapiens (Human)</td>\n",
       "      <td>Membrane receptor-&gt;Family A G protein-coupled ...</td>\n",
       "      <td>332</td>\n",
       "      <td>MLLETQDALYVALELVIAALSVAGNVLVCAAVGTANTLQTPTNYFL...</td>\n",
       "      <td>ChEMBL:CHEMBL255;ChEMBL:CHEMBL255;ChEMBL:CHEMB...</td>\n",
       "      <td>P29275</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>P30542_WT</td>\n",
       "      <td>ADORA1</td>\n",
       "      <td>AA1R_HUMAN</td>\n",
       "      <td>reviewed</td>\n",
       "      <td>Homo sapiens (Human)</td>\n",
       "      <td>Membrane receptor-&gt;Family A G protein-coupled ...</td>\n",
       "      <td>326</td>\n",
       "      <td>MPPSISAFQAAYIGIEVLIALVSVPGNVLVIWAVKVNQALRDATFC...</td>\n",
       "      <td>ChEMBL:CHEMBL226;ChEMBL:CHEMBL226;ChEMBL:CHEMB...</td>\n",
       "      <td>P30542</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>P29274_WT</td>\n",
       "      <td>ADORA2A</td>\n",
       "      <td>AA2AR_HUMAN</td>\n",
       "      <td>reviewed</td>\n",
       "      <td>Homo sapiens (Human)</td>\n",
       "      <td>Membrane receptor-&gt;Family A G protein-coupled ...</td>\n",
       "      <td>412</td>\n",
       "      <td>MPIMGSSVYITVELAIAVLAILGNVLVCWAVWLNSNLQNVTNYFVV...</td>\n",
       "      <td>ChEMBL:CHEMBL251;ChEMBL:CHEMBL251;ChEMBL:CHEMB...</td>\n",
       "      <td>P29274</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>P0DMS8_WT</td>\n",
       "      <td>ADORA3</td>\n",
       "      <td>AA3R_HUMAN</td>\n",
       "      <td>reviewed</td>\n",
       "      <td>Homo sapiens (Human)</td>\n",
       "      <td>Membrane receptor-&gt;Family A G protein-coupled ...</td>\n",
       "      <td>318</td>\n",
       "      <td>MPNNSTALSLANVTYITMEIFIGLCAIVGNVLVICVVKLNPSLQTT...</td>\n",
       "      <td>ChEMBL:CHEMBL256;ChEMBL:CHEMBL256;ChEMBL:CHEMB...</td>\n",
       "      <td>P0DMS8</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   target_id HGNC_symbol    UniProtID    Status              Organism  \\\n",
       "0  P29275_WT     ADORA2B  AA2BR_HUMAN  reviewed  Homo sapiens (Human)   \n",
       "1  P30542_WT      ADORA1   AA1R_HUMAN  reviewed  Homo sapiens (Human)   \n",
       "2  P29274_WT     ADORA2A  AA2AR_HUMAN  reviewed  Homo sapiens (Human)   \n",
       "3  P0DMS8_WT      ADORA3   AA3R_HUMAN  reviewed  Homo sapiens (Human)   \n",
       "\n",
       "                                      Classification  Length  \\\n",
       "0  Membrane receptor->Family A G protein-coupled ...     332   \n",
       "1  Membrane receptor->Family A G protein-coupled ...     326   \n",
       "2  Membrane receptor->Family A G protein-coupled ...     412   \n",
       "3  Membrane receptor->Family A G protein-coupled ...     318   \n",
       "\n",
       "                                            Sequence  \\\n",
       "0  MLLETQDALYVALELVIAALSVAGNVLVCAAVGTANTLQTPTNYFL...   \n",
       "1  MPPSISAFQAAYIGIEVLIALVSVPGNVLVIWAVKVNQALRDATFC...   \n",
       "2  MPIMGSSVYITVELAIAVLAILGNVLVCWAVWLNSNLQNVTNYFVV...   \n",
       "3  MPNNSTALSLANVTYITMEIFIGLCAIVGNVLVICVVKLNPSLQTT...   \n",
       "\n",
       "                                                 TID accession  \n",
       "0  ChEMBL:CHEMBL255;ChEMBL:CHEMBL255;ChEMBL:CHEMB...    P29275  \n",
       "1  ChEMBL:CHEMBL226;ChEMBL:CHEMBL226;ChEMBL:CHEMB...    P30542  \n",
       "2  ChEMBL:CHEMBL251;ChEMBL:CHEMBL251;ChEMBL:CHEMB...    P29274  \n",
       "3  ChEMBL:CHEMBL256;ChEMBL:CHEMBL256;ChEMBL:CHEMB...    P0DMS8  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from qsprpred.data.sources.papyrus import Papyrus\n",
    "\n",
    "# Papyrus database version 05.6\n",
    "papyrus = Papyrus(\n",
    "    data_dir=\"../../tutorial_data/\",\n",
    "    version=\"05.6\",  # Papyrus database version\n",
    "    stereo=False,\n",
    "    plus_only=True,\n",
    ")\n",
    "\n",
    "# Get the protein sequences for the four adenosine receptors\n",
    "acc_keys = [\"P29274\", \"P29275\", \"P30542\", \"P0DMS8\"]  # A2AR, A2BR, A1R, A3R\n",
    "dataset_seq = papyrus.getProteinData(\n",
    "    acc_keys,\n",
    "    name=f\"{dataset.name}_seqs\",\n",
    "    use_existing=True,\n",
    "    output_dir=\"../../tutorial_output/data\",\n",
    ")\n",
    "dataset_seq"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The  keys used to fetch the data are saved in the `accession` colum in the resulting data frame:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>accession</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>QSPRID</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00000</th>\n",
       "      <td>P29274</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00001</th>\n",
       "      <td>P30542</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00002</th>\n",
       "      <td>P29274</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00003</th>\n",
       "      <td>P0DMS8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00004</th>\n",
       "      <td>P29274</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12452</th>\n",
       "      <td>P29274</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12453</th>\n",
       "      <td>P0DMS8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12454</th>\n",
       "      <td>P29274</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12455</th>\n",
       "      <td>P29275</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12456</th>\n",
       "      <td>P30542</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>12457 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                         accession\n",
       "QSPRID                            \n",
       "PCMTutorialDataset_00000    P29274\n",
       "PCMTutorialDataset_00001    P30542\n",
       "PCMTutorialDataset_00002    P29274\n",
       "PCMTutorialDataset_00003    P0DMS8\n",
       "PCMTutorialDataset_00004    P29274\n",
       "...                            ...\n",
       "PCMTutorialDataset_12452    P29274\n",
       "PCMTutorialDataset_12453    P0DMS8\n",
       "PCMTutorialDataset_12454    P29274\n",
       "PCMTutorialDataset_12455    P29275\n",
       "PCMTutorialDataset_12456    P30542\n",
       "\n",
       "[12457 rows x 1 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset.getSubset(\"accession\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Calculating Protein Descriptors\n",
    "\n",
    "In this section, we will show how to connect the information about sequences with our data set and calculate protein descriptors using multiple sequence alignment and the `PCMDataSet` class from the `qsprpred.extra` package. First, let us convert the original data set saved in the `dataset` variable to a `PCMDataSet`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from qsprpred.extra.data.tables.pcm import PCMDataSet\n",
    "\n",
    "\n",
    "def sequence_provider(acc_keys):\n",
    "    \"\"\"\n",
    "    A function that provides a mapping from accession key to a protein sequence.\n",
    "\n",
    "    Args:\n",
    "        acc_keys (list): Accession keys of the protein to get a sequences for.\n",
    "\n",
    "    Returns:\n",
    "        (dict) : Mapping of accession keys to protein sequences.\n",
    "        (dict) : Additional information to pass to the MSA provider (can be empty).\n",
    "    \"\"\"\n",
    "    map = dict()\n",
    "    info = dict()\n",
    "    for i, row in dataset_seq.iterrows():\n",
    "        map[row[\"accession\"]] = row[\"Sequence\"]\n",
    "\n",
    "        # can be omitted\n",
    "        info[row[\"accession\"]] = {\n",
    "            \"Organism\": row[\"Organism\"],\n",
    "            \"UniProtID\": row[\"UniProtID\"],\n",
    "        }\n",
    "\n",
    "    return map, info\n",
    "\n",
    "\n",
    "dataset = PCMDataSet.fromMolTable(\n",
    "    dataset,\n",
    "    name=dataset.name,\n",
    "    protein_col=\"accession\",\n",
    "    protein_seq_provider=sequence_provider,\n",
    "    store_dir=dataset.storeDir,\n",
    "    target_props=[{\"name\": \"pchembl_value_Mean\", \"task\": \"REGRESSION\"}],\n",
    ")\n",
    "\n",
    "dataset.save()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`PCMDataset` knows how to connect accession keys to sequences thanks to `proteinCol` and `proteinSeqProvider`. The `proteinSeqProvider` is a function that maps given accession keys to sequences, because this can be done in many different ways there is no default implementation. The `acc_keys` argument is there to make a more general implementation possible.\n",
    "\n",
    "The `ClustalMSA` object facilitates multiple sequence alignment with the Clustal Omega program (needs to be [installed separately](http://www.clustal.org/omega/)). This is done automatically after `proteinSeqProvider` is invoked before descriptor calculation:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "312d7c703f174b8589baafbf6f96efe3",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from qsprpred.extra.data.utils.msa_calculator import ClustalMSA\n",
    "from qsprpred.extra.data.descriptors.sets import ProDec\n",
    "\n",
    "dataset.addDescriptors([\n",
    "    ProDec(\n",
    "        sets=[\"Zscale Hellberg\"],\n",
    "        msa_provider=ClustalMSA(out_dir=dataset.storeDir)\n",
    "    )\n",
    "])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can check the descriptor matrix:"
   ]
  },
  {
   "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>Zscale_1</th>\n",
       "      <th>Zscale_10</th>\n",
       "      <th>Zscale_100</th>\n",
       "      <th>Zscale_1000</th>\n",
       "      <th>Zscale_1001</th>\n",
       "      <th>Zscale_1002</th>\n",
       "      <th>Zscale_1003</th>\n",
       "      <th>Zscale_1004</th>\n",
       "      <th>Zscale_1005</th>\n",
       "      <th>Zscale_1006</th>\n",
       "      <th>...</th>\n",
       "      <th>Zscale_990</th>\n",
       "      <th>Zscale_991</th>\n",
       "      <th>Zscale_992</th>\n",
       "      <th>Zscale_993</th>\n",
       "      <th>Zscale_994</th>\n",
       "      <th>Zscale_995</th>\n",
       "      <th>Zscale_996</th>\n",
       "      <th>Zscale_997</th>\n",
       "      <th>Zscale_998</th>\n",
       "      <th>Zscale_999</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>QSPRID</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00000</th>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.71</td>\n",
       "      <td>1.96</td>\n",
       "      <td>-1.63</td>\n",
       "      <td>0.57</td>\n",
       "      <td>0.07</td>\n",
       "      <td>-1.73</td>\n",
       "      <td>0.09</td>\n",
       "      <td>2.88</td>\n",
       "      <td>...</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.07</td>\n",
       "      <td>-1.73</td>\n",
       "      <td>0.09</td>\n",
       "      <td>2.23</td>\n",
       "      <td>-5.36</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.92</td>\n",
       "      <td>-2.09</td>\n",
       "      <td>-1.40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00001</th>\n",
       "      <td>0.00</td>\n",
       "      <td>-2.49</td>\n",
       "      <td>-4.44</td>\n",
       "      <td>-1.22</td>\n",
       "      <td>0.88</td>\n",
       "      <td>2.23</td>\n",
       "      <td>3.08</td>\n",
       "      <td>0.39</td>\n",
       "      <td>-0.07</td>\n",
       "      <td>3.08</td>\n",
       "      <td>...</td>\n",
       "      <td>2.36</td>\n",
       "      <td>3.08</td>\n",
       "      <td>0.39</td>\n",
       "      <td>-0.07</td>\n",
       "      <td>3.64</td>\n",
       "      <td>1.13</td>\n",
       "      <td>2.36</td>\n",
       "      <td>-4.19</td>\n",
       "      <td>-1.03</td>\n",
       "      <td>-0.98</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00002</th>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.71</td>\n",
       "      <td>1.96</td>\n",
       "      <td>-1.63</td>\n",
       "      <td>0.57</td>\n",
       "      <td>0.07</td>\n",
       "      <td>-1.73</td>\n",
       "      <td>0.09</td>\n",
       "      <td>2.88</td>\n",
       "      <td>...</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.07</td>\n",
       "      <td>-1.73</td>\n",
       "      <td>0.09</td>\n",
       "      <td>2.23</td>\n",
       "      <td>-5.36</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.92</td>\n",
       "      <td>-2.09</td>\n",
       "      <td>-1.40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00003</th>\n",
       "      <td>-2.49</td>\n",
       "      <td>3.22</td>\n",
       "      <td>-4.44</td>\n",
       "      <td>3.08</td>\n",
       "      <td>0.39</td>\n",
       "      <td>-0.07</td>\n",
       "      <td>2.84</td>\n",
       "      <td>1.41</td>\n",
       "      <td>-3.14</td>\n",
       "      <td>3.22</td>\n",
       "      <td>...</td>\n",
       "      <td>2.36</td>\n",
       "      <td>0.92</td>\n",
       "      <td>-2.09</td>\n",
       "      <td>-1.40</td>\n",
       "      <td>1.96</td>\n",
       "      <td>-1.63</td>\n",
       "      <td>0.57</td>\n",
       "      <td>-4.44</td>\n",
       "      <td>-1.68</td>\n",
       "      <td>-1.03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00004</th>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.71</td>\n",
       "      <td>1.96</td>\n",
       "      <td>-1.63</td>\n",
       "      <td>0.57</td>\n",
       "      <td>0.07</td>\n",
       "      <td>-1.73</td>\n",
       "      <td>0.09</td>\n",
       "      <td>2.88</td>\n",
       "      <td>...</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.07</td>\n",
       "      <td>-1.73</td>\n",
       "      <td>0.09</td>\n",
       "      <td>2.23</td>\n",
       "      <td>-5.36</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.92</td>\n",
       "      <td>-2.09</td>\n",
       "      <td>-1.40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12452</th>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.71</td>\n",
       "      <td>1.96</td>\n",
       "      <td>-1.63</td>\n",
       "      <td>0.57</td>\n",
       "      <td>0.07</td>\n",
       "      <td>-1.73</td>\n",
       "      <td>0.09</td>\n",
       "      <td>2.88</td>\n",
       "      <td>...</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.07</td>\n",
       "      <td>-1.73</td>\n",
       "      <td>0.09</td>\n",
       "      <td>2.23</td>\n",
       "      <td>-5.36</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.92</td>\n",
       "      <td>-2.09</td>\n",
       "      <td>-1.40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12453</th>\n",
       "      <td>-2.49</td>\n",
       "      <td>3.22</td>\n",
       "      <td>-4.44</td>\n",
       "      <td>3.08</td>\n",
       "      <td>0.39</td>\n",
       "      <td>-0.07</td>\n",
       "      <td>2.84</td>\n",
       "      <td>1.41</td>\n",
       "      <td>-3.14</td>\n",
       "      <td>3.22</td>\n",
       "      <td>...</td>\n",
       "      <td>2.36</td>\n",
       "      <td>0.92</td>\n",
       "      <td>-2.09</td>\n",
       "      <td>-1.40</td>\n",
       "      <td>1.96</td>\n",
       "      <td>-1.63</td>\n",
       "      <td>0.57</td>\n",
       "      <td>-4.44</td>\n",
       "      <td>-1.68</td>\n",
       "      <td>-1.03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12454</th>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.71</td>\n",
       "      <td>1.96</td>\n",
       "      <td>-1.63</td>\n",
       "      <td>0.57</td>\n",
       "      <td>0.07</td>\n",
       "      <td>-1.73</td>\n",
       "      <td>0.09</td>\n",
       "      <td>2.88</td>\n",
       "      <td>...</td>\n",
       "      <td>0.09</td>\n",
       "      <td>0.07</td>\n",
       "      <td>-1.73</td>\n",
       "      <td>0.09</td>\n",
       "      <td>2.23</td>\n",
       "      <td>-5.36</td>\n",
       "      <td>0.30</td>\n",
       "      <td>0.92</td>\n",
       "      <td>-2.09</td>\n",
       "      <td>-1.40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12455</th>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.71</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>...</td>\n",
       "      <td>0.30</td>\n",
       "      <td>3.22</td>\n",
       "      <td>1.45</td>\n",
       "      <td>0.84</td>\n",
       "      <td>2.23</td>\n",
       "      <td>-5.36</td>\n",
       "      <td>0.30</td>\n",
       "      <td>2.18</td>\n",
       "      <td>0.53</td>\n",
       "      <td>-1.14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12456</th>\n",
       "      <td>0.00</td>\n",
       "      <td>-2.49</td>\n",
       "      <td>-4.44</td>\n",
       "      <td>-1.22</td>\n",
       "      <td>0.88</td>\n",
       "      <td>2.23</td>\n",
       "      <td>3.08</td>\n",
       "      <td>0.39</td>\n",
       "      <td>-0.07</td>\n",
       "      <td>3.08</td>\n",
       "      <td>...</td>\n",
       "      <td>2.36</td>\n",
       "      <td>3.08</td>\n",
       "      <td>0.39</td>\n",
       "      <td>-0.07</td>\n",
       "      <td>3.64</td>\n",
       "      <td>1.13</td>\n",
       "      <td>2.36</td>\n",
       "      <td>-4.19</td>\n",
       "      <td>-1.03</td>\n",
       "      <td>-0.98</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>12457 rows × 1278 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                          Zscale_1  Zscale_10  Zscale_100  Zscale_1000  \\\n",
       "QSPRID                                                                   \n",
       "PCMTutorialDataset_00000      0.00       0.00        0.71         1.96   \n",
       "PCMTutorialDataset_00001      0.00      -2.49       -4.44        -1.22   \n",
       "PCMTutorialDataset_00002      0.00       0.00        0.71         1.96   \n",
       "PCMTutorialDataset_00003     -2.49       3.22       -4.44         3.08   \n",
       "PCMTutorialDataset_00004      0.00       0.00        0.71         1.96   \n",
       "...                            ...        ...         ...          ...   \n",
       "PCMTutorialDataset_12452      0.00       0.00        0.71         1.96   \n",
       "PCMTutorialDataset_12453     -2.49       3.22       -4.44         3.08   \n",
       "PCMTutorialDataset_12454      0.00       0.00        0.71         1.96   \n",
       "PCMTutorialDataset_12455      0.00       0.00        0.71         0.00   \n",
       "PCMTutorialDataset_12456      0.00      -2.49       -4.44        -1.22   \n",
       "\n",
       "                          Zscale_1001  Zscale_1002  Zscale_1003  Zscale_1004  \\\n",
       "QSPRID                                                                         \n",
       "PCMTutorialDataset_00000        -1.63         0.57         0.07        -1.73   \n",
       "PCMTutorialDataset_00001         0.88         2.23         3.08         0.39   \n",
       "PCMTutorialDataset_00002        -1.63         0.57         0.07        -1.73   \n",
       "PCMTutorialDataset_00003         0.39        -0.07         2.84         1.41   \n",
       "PCMTutorialDataset_00004        -1.63         0.57         0.07        -1.73   \n",
       "...                               ...          ...          ...          ...   \n",
       "PCMTutorialDataset_12452        -1.63         0.57         0.07        -1.73   \n",
       "PCMTutorialDataset_12453         0.39        -0.07         2.84         1.41   \n",
       "PCMTutorialDataset_12454        -1.63         0.57         0.07        -1.73   \n",
       "PCMTutorialDataset_12455         0.00         0.00         0.00         0.00   \n",
       "PCMTutorialDataset_12456         0.88         2.23         3.08         0.39   \n",
       "\n",
       "                          Zscale_1005  Zscale_1006  ...  Zscale_990  \\\n",
       "QSPRID                                              ...               \n",
       "PCMTutorialDataset_00000         0.09         2.88  ...        0.09   \n",
       "PCMTutorialDataset_00001        -0.07         3.08  ...        2.36   \n",
       "PCMTutorialDataset_00002         0.09         2.88  ...        0.09   \n",
       "PCMTutorialDataset_00003        -3.14         3.22  ...        2.36   \n",
       "PCMTutorialDataset_00004         0.09         2.88  ...        0.09   \n",
       "...                               ...          ...  ...         ...   \n",
       "PCMTutorialDataset_12452         0.09         2.88  ...        0.09   \n",
       "PCMTutorialDataset_12453        -3.14         3.22  ...        2.36   \n",
       "PCMTutorialDataset_12454         0.09         2.88  ...        0.09   \n",
       "PCMTutorialDataset_12455         0.00         0.00  ...        0.30   \n",
       "PCMTutorialDataset_12456        -0.07         3.08  ...        2.36   \n",
       "\n",
       "                          Zscale_991  Zscale_992  Zscale_993  Zscale_994  \\\n",
       "QSPRID                                                                     \n",
       "PCMTutorialDataset_00000        0.07       -1.73        0.09        2.23   \n",
       "PCMTutorialDataset_00001        3.08        0.39       -0.07        3.64   \n",
       "PCMTutorialDataset_00002        0.07       -1.73        0.09        2.23   \n",
       "PCMTutorialDataset_00003        0.92       -2.09       -1.40        1.96   \n",
       "PCMTutorialDataset_00004        0.07       -1.73        0.09        2.23   \n",
       "...                              ...         ...         ...         ...   \n",
       "PCMTutorialDataset_12452        0.07       -1.73        0.09        2.23   \n",
       "PCMTutorialDataset_12453        0.92       -2.09       -1.40        1.96   \n",
       "PCMTutorialDataset_12454        0.07       -1.73        0.09        2.23   \n",
       "PCMTutorialDataset_12455        3.22        1.45        0.84        2.23   \n",
       "PCMTutorialDataset_12456        3.08        0.39       -0.07        3.64   \n",
       "\n",
       "                          Zscale_995  Zscale_996  Zscale_997  Zscale_998  \\\n",
       "QSPRID                                                                     \n",
       "PCMTutorialDataset_00000       -5.36        0.30        0.92       -2.09   \n",
       "PCMTutorialDataset_00001        1.13        2.36       -4.19       -1.03   \n",
       "PCMTutorialDataset_00002       -5.36        0.30        0.92       -2.09   \n",
       "PCMTutorialDataset_00003       -1.63        0.57       -4.44       -1.68   \n",
       "PCMTutorialDataset_00004       -5.36        0.30        0.92       -2.09   \n",
       "...                              ...         ...         ...         ...   \n",
       "PCMTutorialDataset_12452       -5.36        0.30        0.92       -2.09   \n",
       "PCMTutorialDataset_12453       -1.63        0.57       -4.44       -1.68   \n",
       "PCMTutorialDataset_12454       -5.36        0.30        0.92       -2.09   \n",
       "PCMTutorialDataset_12455       -5.36        0.30        2.18        0.53   \n",
       "PCMTutorialDataset_12456        1.13        2.36       -4.19       -1.03   \n",
       "\n",
       "                          Zscale_999  \n",
       "QSPRID                                \n",
       "PCMTutorialDataset_00000       -1.40  \n",
       "PCMTutorialDataset_00001       -0.98  \n",
       "PCMTutorialDataset_00002       -1.40  \n",
       "PCMTutorialDataset_00003       -1.03  \n",
       "PCMTutorialDataset_00004       -1.40  \n",
       "...                              ...  \n",
       "PCMTutorialDataset_12452       -1.40  \n",
       "PCMTutorialDataset_12453       -1.03  \n",
       "PCMTutorialDataset_12454       -1.40  \n",
       "PCMTutorialDataset_12455       -1.14  \n",
       "PCMTutorialDataset_12456       -0.98  \n",
       "\n",
       "[12457 rows x 1278 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset.getDescriptors()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can of course combine this with molecular descriptors as well:"
   ]
  },
  {
   "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>Zscale_1</th>\n",
       "      <th>Zscale_10</th>\n",
       "      <th>Zscale_100</th>\n",
       "      <th>Zscale_1000</th>\n",
       "      <th>Zscale_1001</th>\n",
       "      <th>Zscale_1002</th>\n",
       "      <th>Zscale_1003</th>\n",
       "      <th>Zscale_1004</th>\n",
       "      <th>Zscale_1005</th>\n",
       "      <th>Zscale_1006</th>\n",
       "      <th>...</th>\n",
       "      <th>MorganFP_1014</th>\n",
       "      <th>MorganFP_1015</th>\n",
       "      <th>MorganFP_1016</th>\n",
       "      <th>MorganFP_1017</th>\n",
       "      <th>MorganFP_1018</th>\n",
       "      <th>MorganFP_1019</th>\n",
       "      <th>MorganFP_1020</th>\n",
       "      <th>MorganFP_1021</th>\n",
       "      <th>MorganFP_1022</th>\n",
       "      <th>MorganFP_1023</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>QSPRID</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00000</th>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.71</td>\n",
       "      <td>1.96</td>\n",
       "      <td>-1.63</td>\n",
       "      <td>0.57</td>\n",
       "      <td>0.07</td>\n",
       "      <td>-1.73</td>\n",
       "      <td>0.09</td>\n",
       "      <td>2.88</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00001</th>\n",
       "      <td>0.00</td>\n",
       "      <td>-2.49</td>\n",
       "      <td>-4.44</td>\n",
       "      <td>-1.22</td>\n",
       "      <td>0.88</td>\n",
       "      <td>2.23</td>\n",
       "      <td>3.08</td>\n",
       "      <td>0.39</td>\n",
       "      <td>-0.07</td>\n",
       "      <td>3.08</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00002</th>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.71</td>\n",
       "      <td>1.96</td>\n",
       "      <td>-1.63</td>\n",
       "      <td>0.57</td>\n",
       "      <td>0.07</td>\n",
       "      <td>-1.73</td>\n",
       "      <td>0.09</td>\n",
       "      <td>2.88</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00003</th>\n",
       "      <td>-2.49</td>\n",
       "      <td>3.22</td>\n",
       "      <td>-4.44</td>\n",
       "      <td>3.08</td>\n",
       "      <td>0.39</td>\n",
       "      <td>-0.07</td>\n",
       "      <td>2.84</td>\n",
       "      <td>1.41</td>\n",
       "      <td>-3.14</td>\n",
       "      <td>3.22</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_00004</th>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.71</td>\n",
       "      <td>1.96</td>\n",
       "      <td>-1.63</td>\n",
       "      <td>0.57</td>\n",
       "      <td>0.07</td>\n",
       "      <td>-1.73</td>\n",
       "      <td>0.09</td>\n",
       "      <td>2.88</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12452</th>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.71</td>\n",
       "      <td>1.96</td>\n",
       "      <td>-1.63</td>\n",
       "      <td>0.57</td>\n",
       "      <td>0.07</td>\n",
       "      <td>-1.73</td>\n",
       "      <td>0.09</td>\n",
       "      <td>2.88</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12453</th>\n",
       "      <td>-2.49</td>\n",
       "      <td>3.22</td>\n",
       "      <td>-4.44</td>\n",
       "      <td>3.08</td>\n",
       "      <td>0.39</td>\n",
       "      <td>-0.07</td>\n",
       "      <td>2.84</td>\n",
       "      <td>1.41</td>\n",
       "      <td>-3.14</td>\n",
       "      <td>3.22</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12454</th>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.71</td>\n",
       "      <td>1.96</td>\n",
       "      <td>-1.63</td>\n",
       "      <td>0.57</td>\n",
       "      <td>0.07</td>\n",
       "      <td>-1.73</td>\n",
       "      <td>0.09</td>\n",
       "      <td>2.88</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12455</th>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.71</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PCMTutorialDataset_12456</th>\n",
       "      <td>0.00</td>\n",
       "      <td>-2.49</td>\n",
       "      <td>-4.44</td>\n",
       "      <td>-1.22</td>\n",
       "      <td>0.88</td>\n",
       "      <td>2.23</td>\n",
       "      <td>3.08</td>\n",
       "      <td>0.39</td>\n",
       "      <td>-0.07</td>\n",
       "      <td>3.08</td>\n",
       "      <td>...</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>12457 rows × 2302 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                          Zscale_1  Zscale_10  Zscale_100  Zscale_1000  \\\n",
       "QSPRID                                                                   \n",
       "PCMTutorialDataset_00000      0.00       0.00        0.71         1.96   \n",
       "PCMTutorialDataset_00001      0.00      -2.49       -4.44        -1.22   \n",
       "PCMTutorialDataset_00002      0.00       0.00        0.71         1.96   \n",
       "PCMTutorialDataset_00003     -2.49       3.22       -4.44         3.08   \n",
       "PCMTutorialDataset_00004      0.00       0.00        0.71         1.96   \n",
       "...                            ...        ...         ...          ...   \n",
       "PCMTutorialDataset_12452      0.00       0.00        0.71         1.96   \n",
       "PCMTutorialDataset_12453     -2.49       3.22       -4.44         3.08   \n",
       "PCMTutorialDataset_12454      0.00       0.00        0.71         1.96   \n",
       "PCMTutorialDataset_12455      0.00       0.00        0.71         0.00   \n",
       "PCMTutorialDataset_12456      0.00      -2.49       -4.44        -1.22   \n",
       "\n",
       "                          Zscale_1001  Zscale_1002  Zscale_1003  Zscale_1004  \\\n",
       "QSPRID                                                                         \n",
       "PCMTutorialDataset_00000        -1.63         0.57         0.07        -1.73   \n",
       "PCMTutorialDataset_00001         0.88         2.23         3.08         0.39   \n",
       "PCMTutorialDataset_00002        -1.63         0.57         0.07        -1.73   \n",
       "PCMTutorialDataset_00003         0.39        -0.07         2.84         1.41   \n",
       "PCMTutorialDataset_00004        -1.63         0.57         0.07        -1.73   \n",
       "...                               ...          ...          ...          ...   \n",
       "PCMTutorialDataset_12452        -1.63         0.57         0.07        -1.73   \n",
       "PCMTutorialDataset_12453         0.39        -0.07         2.84         1.41   \n",
       "PCMTutorialDataset_12454        -1.63         0.57         0.07        -1.73   \n",
       "PCMTutorialDataset_12455         0.00         0.00         0.00         0.00   \n",
       "PCMTutorialDataset_12456         0.88         2.23         3.08         0.39   \n",
       "\n",
       "                          Zscale_1005  Zscale_1006  ...  MorganFP_1014  \\\n",
       "QSPRID                                              ...                  \n",
       "PCMTutorialDataset_00000         0.09         2.88  ...          False   \n",
       "PCMTutorialDataset_00001        -0.07         3.08  ...          False   \n",
       "PCMTutorialDataset_00002         0.09         2.88  ...          False   \n",
       "PCMTutorialDataset_00003        -3.14         3.22  ...          False   \n",
       "PCMTutorialDataset_00004         0.09         2.88  ...          False   \n",
       "...                               ...          ...  ...            ...   \n",
       "PCMTutorialDataset_12452         0.09         2.88  ...          False   \n",
       "PCMTutorialDataset_12453        -3.14         3.22  ...          False   \n",
       "PCMTutorialDataset_12454         0.09         2.88  ...          False   \n",
       "PCMTutorialDataset_12455         0.00         0.00  ...          False   \n",
       "PCMTutorialDataset_12456        -0.07         3.08  ...          False   \n",
       "\n",
       "                          MorganFP_1015  MorganFP_1016  MorganFP_1017  \\\n",
       "QSPRID                                                                  \n",
       "PCMTutorialDataset_00000          False          False          False   \n",
       "PCMTutorialDataset_00001          False          False          False   \n",
       "PCMTutorialDataset_00002          False          False          False   \n",
       "PCMTutorialDataset_00003          False          False          False   \n",
       "PCMTutorialDataset_00004          False          False          False   \n",
       "...                                 ...            ...            ...   \n",
       "PCMTutorialDataset_12452          False          False          False   \n",
       "PCMTutorialDataset_12453          False          False          False   \n",
       "PCMTutorialDataset_12454          False          False          False   \n",
       "PCMTutorialDataset_12455          False          False          False   \n",
       "PCMTutorialDataset_12456          False          False          False   \n",
       "\n",
       "                          MorganFP_1018  MorganFP_1019  MorganFP_1020  \\\n",
       "QSPRID                                                                  \n",
       "PCMTutorialDataset_00000          False          False          False   \n",
       "PCMTutorialDataset_00001          False          False          False   \n",
       "PCMTutorialDataset_00002          False          False           True   \n",
       "PCMTutorialDataset_00003          False          False          False   \n",
       "PCMTutorialDataset_00004          False          False          False   \n",
       "...                                 ...            ...            ...   \n",
       "PCMTutorialDataset_12452          False          False          False   \n",
       "PCMTutorialDataset_12453          False          False          False   \n",
       "PCMTutorialDataset_12454          False          False          False   \n",
       "PCMTutorialDataset_12455          False          False          False   \n",
       "PCMTutorialDataset_12456          False          False          False   \n",
       "\n",
       "                          MorganFP_1021  MorganFP_1022  MorganFP_1023  \n",
       "QSPRID                                                                 \n",
       "PCMTutorialDataset_00000          False          False          False  \n",
       "PCMTutorialDataset_00001          False          False          False  \n",
       "PCMTutorialDataset_00002          False          False          False  \n",
       "PCMTutorialDataset_00003          False          False          False  \n",
       "PCMTutorialDataset_00004          False          False          False  \n",
       "...                                 ...            ...            ...  \n",
       "PCMTutorialDataset_12452          False          False          False  \n",
       "PCMTutorialDataset_12453          False          False          False  \n",
       "PCMTutorialDataset_12454          False          False          False  \n",
       "PCMTutorialDataset_12455          False          False          False  \n",
       "PCMTutorialDataset_12456          False          False          False  \n",
       "\n",
       "[12457 rows x 2302 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from qsprpred.data.descriptors.fingerprints import MorganFP\n",
    "\n",
    "dataset.nJobs = 12  # allowing parallelization to speed things up a little\n",
    "dataset.addDescriptors([MorganFP(radius=2, nBits=1024)])\n",
    "dataset.getDescriptors()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To split the dataset we use the `PCMsplit`, which is a wrapper around the `DataSplit` class. It splits a dataset into train and test set such that the subsets are balanced with respect to each of the protein targets. This is done with https://github.com/sohviluukkonen/gbmt-splits, linear programming of initial clusters (random-, scaffold- or cluster-based) to get a balanced split.\n",
    "It therefore expects either a `RandomSplit`, `ScaffoldSplit` or `ClusterSplit` object as input."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "from qsprpred.extra.data.sampling.splits import PCMSplit\n",
    "from qsprpred.data import RandomSplit\n",
    "\n",
    "dataset.split(PCMSplit(splitter=RandomSplit(test_fraction=0.2),\n",
    "                       dataset=dataset))  # 20% test set per target"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model Construction\n",
    "\n",
    "Now that we have defined our preprocessing steps, we can construct our model. We will use a simple K-nearest neighbors model in this case, but notice that the training code is almost the same as in the simple [quick start](../../quick_start.ipynb). The only difference is that we are using the `SklearnPCMModel` class instead of `SklearnModel`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.neighbors import KNeighborsRegressor\n",
    "from qsprpred.extra.models.pcm import SklearnPCMModel\n",
    "from qsprpred.models import CrossValAssessor, TestSetAssessor\n",
    "\n",
    "os.makedirs('../../tutorial_output/data', exist_ok=True)\n",
    "\n",
    "model = SklearnPCMModel(\n",
    "    base_dir='../../tutorial_output/models',\n",
    "    alg=KNeighborsRegressor,\n",
    "    name='PCMTutorialModel',\n",
    ")\n",
    "\n",
    "CrossValAssessor(scoring='r2')(model, dataset)\n",
    "TestSetAssessor(scoring='r2')(model, dataset)\n",
    "_ = model.fitDataset(dataset)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "from qsprpred.extra.models.pcm import SklearnPCMModel\n",
    "\n",
    "model = SklearnPCMModel.fromFile(\n",
    "    \"../../tutorial_output/models/PCMTutorialModel/PCMTutorialModel_meta.json\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The `SklearnPCMModel` class is a subclass of `SklearnModel` and thus inherits all of its functionality. The only difference is in how it handles predictions, which we will [show later](#model-prediction). Model evaluation and plotting is also the same as in the [quick start](../../quick_start.ipynb) tutorial:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "PCMTutorialModel\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 569.5x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "(<seaborn.axisgrid.FacetGrid at 0x7f0923d00730>,\n",
       "               Model              Fold            Property        R2      RMSE  \\\n",
       " 0  PCMTutorialModel               0.0  pchembl_value_Mean  0.593952  0.709219   \n",
       " 1  PCMTutorialModel               1.0  pchembl_value_Mean  0.623370  0.711725   \n",
       " 2  PCMTutorialModel               2.0  pchembl_value_Mean  0.606751  0.702374   \n",
       " 3  PCMTutorialModel               3.0  pchembl_value_Mean  0.601332  0.712577   \n",
       " 4  PCMTutorialModel               4.0  pchembl_value_Mean  0.634756  0.687840   \n",
       " 5  PCMTutorialModel  Independent Test  pchembl_value_Mean  0.657932  0.640084   \n",
       " \n",
       "                 Set  \n",
       " 0  Cross Validation  \n",
       " 1  Cross Validation  \n",
       " 2  Cross Validation  \n",
       " 3  Cross Validation  \n",
       " 4  Cross Validation  \n",
       " 5  Independent Test  )"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 640x480 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from qsprpred.plotting.regression import CorrelationPlot\n",
    "\n",
    "plot = CorrelationPlot([model])\n",
    "plot.make(save=False, show=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model Prediction\n",
    "\n",
    "Prediction with the PCM model works similarly to regular models. We can load the saved model using its name (`PCMTutorialModel`) as usual:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<qsprpred.extra.models.pcm.SklearnPCMModel at 0x7f09069d7010>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_from_files = SklearnPCMModel(name='PCMTutorialModel',\n",
    "                                   base_dir='../../tutorial_output/models')\n",
    "model_from_files"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "But naturally when doing the prediction, we need to provide the protein ID to fetch protein descriptors for as well. When we created the model all necessary information to use the generated MSA and calculate protein descriptors was saved:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "sample_mols = [\n",
    "    'CN1C=NC2=C1C(=O)N(C(=O)N2C)C',  # caffeine\n",
    "    'c1ccccc1CCN1N=CC2=C1N=C(N)N3C2=NC(C4=CC=CO4)=N3'\n",
    "    # SCH-58261 (50x more selective to A2A than caffeine -> more potent)\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['P29274', 'P30542', 'P0DMS8', 'P29275'], dtype=object)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset.getSubset(dataset.proteinCol)[dataset.proteinCol].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2e3a379cf1624c719dcb0018731e01a2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "array([[6.35933333],\n",
       "       [8.69461538]])"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_from_files.predictMols(\n",
    "    sample_mols,\n",
    "    protein_id='P29274',  # human adenosine A2A receptor\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that makes sense. We can see that for the selective compound (`SCH-58261`), the model predicts it as very active. If we use the same compounds with the other isoforms, we can see that although the model predicts it as active for all isoforms, it is not as active for the A2A isoform:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "98a22894da704babb716a7c14207069b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "array([[5.328],\n",
       "       [6.969]])"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_from_files.predictMols(\n",
    "    sample_mols,\n",
    "    protein_id='P30542',  # human adenosine A1 receptor\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e29b0b1a599442589c7fa9094fe40e42",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "array([[6.03  ],\n",
       "       [6.1895]])"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_from_files.predictMols(\n",
    "    sample_mols,\n",
    "    protein_id='P0DMS8',  # human adenosine A3 receptor\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d51019238c6442e2a59f7ef4a0292196",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "array([[5.266],\n",
       "       [7.209]])"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_from_files.predictMols(\n",
    "    sample_mols,\n",
    "    protein_id='P29275',  # human adenosine A2B receptor\n",
    ")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}