notebooks/rogi-dist.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Relative ROGI plots"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pathlib import Path\n",
    "from typing import Iterable, Optional, Union\n",
    "\n",
    "import matplotlib as mpl\n",
    "from matplotlib import patches, pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "\n",
    "rc = mpl.rc_params_from_file(mpl.matplotlib_fname(), use_default_template=False)\n",
    "sns.set_theme(style=\"white\", rc=rc)\n",
    "\n",
    "REPR_LABELS = dict(\n",
    "    fp=\"FP\", vae=\"VAE\", gin=\"GIN\", chemberta=\"ChemBERTa\", chemgpt=\"ChemGPT\", descriptor='descriptors'\n",
    ")\n",
    "JITTER = 0.1625\n",
    "CSV_DIR = Path(\"../results/raw/rogi/tdc+guac\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def disable_gridlines(ax: plt.Axes, axis: str, vals: Union[float, Iterable[float]]):\n",
    "    if isinstance(vals, float):\n",
    "        vals = [vals]\n",
    "    vals = set(vals)\n",
    "\n",
    "    gridlines = ax.get_xgridlines() if axis == \"x\" else ax.get_ygridlines()\n",
    "    for line in gridlines:\n",
    "        v = line.get_xdata()[0] if axis == \"x\" else line.get_ydata()[0]\n",
    "        if v in vals:\n",
    "            line.set(visible=False)\n",
    "    \n",
    "    return ax\n",
    "\n",
    "def fill_in_boxes(boxes, c: str= \"w\", ax: Optional[plt.Axes] = None):\n",
    "    \"\"\"Fill in the boxes in the given boxplot with the given color\"\"\"\n",
    "    ax = ax or plt.gca()\n",
    "\n",
    "    for box in boxes:\n",
    "        patch = patches.PathPatch(box.get_path(), edgecolor=\"None\", facecolor=c, zorder=1)\n",
    "        ax.add_patch(patch)\n",
    "\n",
    "def rogi_string(x: str):\n",
    "    return rf\"$\\mathrm{{ROGI-XD}}_{{\\mathrm{{{x}}}}}$\""
   ]
  },
  {
   "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>representation</th>\n",
       "      <th>task</th>\n",
       "      <th>rogi</th>\n",
       "      <th>n_valid</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>chemberta</td>\n",
       "      <td>ARIPIPRAZOLE_SIMILARITY</td>\n",
       "      <td>0.158179</td>\n",
       "      <td>10000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>chemberta</td>\n",
       "      <td>CACO2_WANG</td>\n",
       "      <td>0.217590</td>\n",
       "      <td>910.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>chemberta</td>\n",
       "      <td>CELECOXIB_REDISCOVERY</td>\n",
       "      <td>0.156523</td>\n",
       "      <td>10000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>chemberta</td>\n",
       "      <td>CLEARANCE_HEPATOCYTE_AZ</td>\n",
       "      <td>0.446076</td>\n",
       "      <td>1213.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>chemberta</td>\n",
       "      <td>CLEARANCE_MICROSOME_AZ</td>\n",
       "      <td>0.369319</td>\n",
       "      <td>1102.0</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>97</th>\n",
       "      <td>vae</td>\n",
       "      <td>PPBR_AZ</td>\n",
       "      <td>0.213338</td>\n",
       "      <td>1614.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98</th>\n",
       "      <td>vae</td>\n",
       "      <td>SCAFFOLD HOP</td>\n",
       "      <td>0.114157</td>\n",
       "      <td>10000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>vae</td>\n",
       "      <td>SOLUBILITY_AQSOLDB</td>\n",
       "      <td>0.143722</td>\n",
       "      <td>9982.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100</th>\n",
       "      <td>vae</td>\n",
       "      <td>VDSS_LOMBARDO</td>\n",
       "      <td>0.049239</td>\n",
       "      <td>1130.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101</th>\n",
       "      <td>vae</td>\n",
       "      <td>ZALEPLON_MPO</td>\n",
       "      <td>0.255782</td>\n",
       "      <td>10000.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>102 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    representation                     task      rogi  n_valid\n",
       "0        chemberta  ARIPIPRAZOLE_SIMILARITY  0.158179  10000.0\n",
       "1        chemberta               CACO2_WANG  0.217590    910.0\n",
       "2        chemberta    CELECOXIB_REDISCOVERY  0.156523  10000.0\n",
       "3        chemberta  CLEARANCE_HEPATOCYTE_AZ  0.446076   1213.0\n",
       "4        chemberta   CLEARANCE_MICROSOME_AZ  0.369319   1102.0\n",
       "..             ...                      ...       ...      ...\n",
       "97             vae                  PPBR_AZ  0.213338   1614.0\n",
       "98             vae             SCAFFOLD HOP  0.114157  10000.0\n",
       "99             vae       SOLUBILITY_AQSOLDB  0.143722   9982.0\n",
       "100            vae            VDSS_LOMBARDO  0.049239   1130.0\n",
       "101            vae             ZALEPLON_MPO  0.255782  10000.0\n",
       "\n",
       "[102 rows x 4 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.concat([pd.read_csv(p) for p in CSV_DIR.iterdir() if not p.is_dir()], ignore_index=True)\\\n",
    "    .rename(columns={'dataset_and_task': \"task\"})\\\n",
    "    .groupby([\"representation\", \"task\"])\\\n",
    "    .mean()\\\n",
    "    .reset_index()\\\n",
    "    .replace({\"morgan\": \"FP\", \"descriptor\": \"descriptors\"})\n",
    "\n",
    "REPRS = [\"chemberta\", \"chemgpt\", \"gin\", \"vae\", \"descriptors\", \"FP\"]\n",
    "PRETRAINED_REPRS = [\"chemberta\", \"chemgpt\", \"gin\", \"vae\"]\n",
    "\n",
    "sns.set_palette(\"Set2\", len(REPRS))\n",
    "COLORS = sns.color_palette(\"Set2\", len(REPRS))\n",
    "REPR_COLORS = dict(zip(REPRS, COLORS))\n",
    "\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_desc = df.query(f\"representation == 'descriptors'\").set_index(\"task\")\n",
    "df_fp = df.query(f\"representation == 'FP'\").set_index(\"task\")\n",
    "df_pcms = [df.query(f\"representation == '{f}'\").set_index('task') for f in PRETRAINED_REPRS]\n",
    "\n",
    "for df_ in df_pcms:\n",
    "    df_[\"rogi_desc\"] = df_['rogi'] / df_desc['rogi'] - 1\n",
    "    df_[\"rogi_fp\"] = df_['rogi'] / df_fp['rogi'] - 1\n",
    "\n",
    "df_rel = pd.concat(df_pcms).reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.0465630806273556, 0.4401489869004399)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = df.groupby(\"representation\").rogi.apply(np.array)['descriptors']\n",
    "x.min(), x.max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.04160378, 0.09054517, 0.09799598, 0.10680955, 0.11678901,\n",
       "        0.13587353, 0.15495535, 0.1565227 , 0.15817872, 0.16807419,\n",
       "        0.1723378 , 0.21759034, 0.24683524, 0.26217968, 0.29347398,\n",
       "        0.36931863, 0.44607585],\n",
       "       [0.05208586, 0.08959611, 0.10323407, 0.11002378, 0.1248786 ,\n",
       "        0.14928657, 0.15295035, 0.15589592, 0.15877354, 0.16052551,\n",
       "        0.17361178, 0.19299871, 0.22576629, 0.26885199, 0.28226426,\n",
       "        0.41757403, 0.43932857],\n",
       "       [0.05080547, 0.09418671, 0.09990777, 0.10658723, 0.10862576,\n",
       "        0.11440311, 0.12674707, 0.12987986, 0.13481967, 0.14991195,\n",
       "        0.1619799 , 0.18260657, 0.21483148, 0.24735023, 0.26232079,\n",
       "        0.40002335, 0.43536176],\n",
       "       [0.04923855, 0.09725305, 0.09863696, 0.10381027, 0.10748776,\n",
       "        0.1102204 , 0.11414219, 0.11415692, 0.13209473, 0.14372208,\n",
       "        0.16521254, 0.19640767, 0.21333778, 0.25450324, 0.25578156,\n",
       "        0.39015835, 0.44436971],\n",
       "       [0.04656308, 0.05760821, 0.07911512, 0.08490194, 0.08530468,\n",
       "        0.09289114, 0.10685921, 0.10738952, 0.10762028, 0.13206975,\n",
       "        0.14221757, 0.16798704, 0.20347591, 0.21761633, 0.22715576,\n",
       "        0.39393872, 0.44014899],\n",
       "       [0.04281969, 0.08177538, 0.09203702, 0.09375421, 0.10637054,\n",
       "        0.11263546, 0.13268132, 0.13476453, 0.14877897, 0.15643553,\n",
       "        0.15830168, 0.1771371 , 0.22057402, 0.23308387, 0.26269833,\n",
       "        0.38436712, 0.4117816 ]])"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "s = df.groupby(\"representation\").rogi.apply(np.sort)\n",
    "X = np.array([s[r] for r in REPRS])\n",
    "X"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Relative ROGI distribution boxplot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "def add_boxplot(df: pd.DataFrame, x_key: str = \"descriptor\", ax = None):\n",
    "    ax = ax or plt.gca()\n",
    "    \n",
    "    s_rogis = df.groupby(\"representation\").rogi.apply(np.array)\n",
    "    x = s_rogis[x_key]\n",
    "    ys = [s_rogis[f] for f in PRETRAINED_REPRS]\n",
    "\n",
    "    Z = np.array(ys) / x - 1\n",
    "    Z *= 100\n",
    "\n",
    "    bp = ax.boxplot(\n",
    "        Z.T,\n",
    "        vert=False,\n",
    "        labels=[REPR_LABELS.get(f, f) for f in PRETRAINED_REPRS],\n",
    "        widths=4 * JITTER,\n",
    "        # showmeans=True,\n",
    "        patch_artist=True,\n",
    "        # boxprops=dict(fc=\"None\"),\n",
    "        medianprops=dict(zorder=2),\n",
    "        meanprops=dict(ms=8),\n",
    "        showfliers=False,\n",
    "        zorder=3\n",
    "    )\n",
    "    fill_in_boxes(bp['boxes'], 'w', ax)\n",
    "    for l, r in zip(bp['medians'], PRETRAINED_REPRS):\n",
    "        l.set_color(REPR_COLORS[r])\n",
    "\n",
    "\n",
    "    for j, r in zip(range(len(Z)), PRETRAINED_REPRS):\n",
    "        z = Z[j]\n",
    "        y = (j + 1) + np.random.uniform(-JITTER, JITTER, z.shape)\n",
    "\n",
    "        ax.plot(z, y, 'o', color=REPR_COLORS[r], zorder=4)\n",
    "    \n",
    "    ax.grid(False, axis='y')\n",
    "    ax.axvline(0, c='darkgrey', ls='-', lw=1.25, zorder=0.9)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1100x350 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set_palette(\"Set2\")\n",
    "\n",
    "fig, axs = plt.subplots(1, 2, figsize=(1 + 2 * 5, 3.5), sharey=True)\n",
    "\n",
    "add_boxplot(df, \"descriptors\", axs[0])\n",
    "add_boxplot(df, \"FP\", axs[1])\n",
    "\n",
    "fig.supxlabel(\"Relative difference between ROGI-XD values (%)\", fontsize=\"medium\", y=0.05)\n",
    "fig.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "fig.savefig(\"../results/figures/boxplot-both-v2.pdf\", bbox_inches=\"tight\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## ROGI distribution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "JITTER = 0.1625\n",
    "COLORS = sns.color_palette(None, len(REPRS))\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(6, 4))\n",
    "\n",
    "s = df.groupby(\"representation\").rogi.apply(np.sort)\n",
    "X = np.array([s[r] for r in REPRS])\n",
    "\n",
    "bp = ax.boxplot(\n",
    "    X.T,\n",
    "    vert=False,\n",
    "    labels=[REPR_LABELS.get(r, r) for r in REPRS],\n",
    "    widths=4 * JITTER,\n",
    "    showmeans=True,\n",
    "    patch_artist=True,\n",
    "    boxprops=dict(fc=\"None\"),\n",
    "    medianprops=dict(c=None),\n",
    "    meanprops=dict(ms=8),\n",
    "    showfliers=False,\n",
    ")\n",
    "fill_in_boxes(bp['boxes'], 'w', ax)\n",
    "for l, r in zip(bp['medians'], REPRS):\n",
    "    l.set_color(REPR_COLORS[r])\n",
    "    \n",
    "\n",
    "for j, r in zip(range(len(X)), REPRS):\n",
    "    z = X[j]\n",
    "    y = (j + 1) + np.random.uniform(-JITTER, JITTER, z.shape)\n",
    "\n",
    "    ax.plot(z, y, 'o', color=REPR_COLORS[r])\n",
    "\n",
    "ax.set_xlim(-0.01)\n",
    "# ax.grid(which=\"both\", axis=\"x\", ls=\"--\", dashes=(4, 4))\n",
    "ax.grid(False, axis='y')\n",
    "ax.axvline(0, c='darkgrey', ls='-', lw=1.25, zorder=0.9)\n",
    "ax.set_xlabel(\"ROGI-XD\")\n",
    "\n",
    "fig.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "fig.savefig(\"../results/figures/rogi-dist.pdf\", bbox_inches=\"tight\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 352,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(representation\n",
       " FP             0.042820\n",
       " chemberta      0.041604\n",
       " chemgpt        0.052086\n",
       " descriptors    0.046563\n",
       " gin            0.050805\n",
       " vae            0.049239\n",
       " Name: rogi, dtype: float64,\n",
       " representation\n",
       " FP             0.411782\n",
       " chemberta      0.446076\n",
       " chemgpt        0.439329\n",
       " descriptors    0.440149\n",
       " gin            0.435362\n",
       " vae            0.444370\n",
       " Name: rogi, dtype: float64)"
      ]
     },
     "execution_count": 352,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "s.apply(min), s.apply(max)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "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>representation</th>\n",
       "      <th>dataset</th>\n",
       "      <th>rogi</th>\n",
       "      <th>n_valid</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>chemberta</td>\n",
       "      <td>SCAFFOLD HOP</td>\n",
       "      <td>0.172338</td>\n",
       "      <td>10000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>chemgpt</td>\n",
       "      <td>SCAFFOLD HOP</td>\n",
       "      <td>0.173612</td>\n",
       "      <td>10000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>descriptor</td>\n",
       "      <td>SCAFFOLD HOP</td>\n",
       "      <td>0.084902</td>\n",
       "      <td>10000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68</th>\n",
       "      <td>gin</td>\n",
       "      <td>SCAFFOLD HOP</td>\n",
       "      <td>0.149912</td>\n",
       "      <td>10000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86</th>\n",
       "      <td>morgan</td>\n",
       "      <td>SCAFFOLD HOP</td>\n",
       "      <td>0.156436</td>\n",
       "      <td>10000.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>104</th>\n",
       "      <td>vae</td>\n",
       "      <td>SCAFFOLD HOP</td>\n",
       "      <td>0.117476</td>\n",
       "      <td>10000.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    representation       dataset      rogi  n_valid\n",
       "14       chemberta  SCAFFOLD HOP  0.172338  10000.0\n",
       "32         chemgpt  SCAFFOLD HOP  0.173612  10000.0\n",
       "50      descriptor  SCAFFOLD HOP  0.084902  10000.0\n",
       "68             gin  SCAFFOLD HOP  0.149912  10000.0\n",
       "86          morgan  SCAFFOLD HOP  0.156436  10000.0\n",
       "104            vae  SCAFFOLD HOP  0.117476  10000.0"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.query('task == \"SCAFFOLD HOP\"')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "pcmr",
   "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": "3.9.16"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "dceaf6c72880e525c78341673ab20302021b9b62719726c1d3662256ded9b8d0"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}