train_SOM_general.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import r2_score\n",
    "import pickle\n",
    "import gentrl\n",
    "import numpy as np\n",
    "\n",
    "import susi\n",
    "from susi.SOMPlots import plot_estimation_map, plot_som_histogram, plot_umatrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\mored\\Miniconda3\\envs\\leishnn\\lib\\site-packages\\IPython\\core\\interactiveshell.py:3146: DtypeWarning: Columns (1,4,5,14,17,18,19,38) have mixed types.Specify dtype option on import or set low_memory=False.\n",
      "  has_raised = await self.run_ast_nodes(code_ast.body, cell_name,\n"
     ]
    }
   ],
   "source": [
    "df = pd.read_csv(\"./train_ligase_2_filtered.csv\")\n",
    "#df = df[(df[\"Standard Units\"] == \"%\") | (df[\"Standard Units\"] == \"nM\")]\n",
    "#df = df[df[\"Standard Units\"] == \"%\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "smiles = []\n",
    "for i in df[\"Smiles\"]:\n",
    "    smiles.append(gentrl.tokenizer.encode([i])[0][0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "smiles = np.array([i.numpy() for i in smiles])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "sort = df[df[\"Standard Units\"] == \"nM\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "values = list(df[\"Standard Value\"]/max(list(df[\"Standard Value\"])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "biggest = max(list(sort[\"Standard Value\"]))\n",
    "\n",
    "for i, row in enumerate(df[\"Standard Value\"]):\n",
    "    if row < 0:\n",
    "        values[i] = -1\n",
    "    elif df[\"Standard Units\"][i] != \"nM\":\n",
    "        values[i] = -1\n",
    "    else:\n",
    "        try:\n",
    "            values[i]= -math.log(df[\"Standard Value\"][i]/biggest)\n",
    "        except:\n",
    "             values[i] = -1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "import math\n",
    "ic50 = []\n",
    "biggest = max(list(sort[\"Standard Value\"]))\n",
    "for i, row in enumerate(df[\"Standard Units\"]):\n",
    "    if row == \"%\":\n",
    "        ic50.append(0)\n",
    "    elif row == \"nM\":\n",
    "        try:\n",
    "            ic50.append(\n",
    "                -math.log(\n",
    "                    df[\"Standard Value\"][i]/biggest\n",
    "                )\n",
    "            )\n",
    "        except:\n",
    "             ic50.append(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(45644, 50) (5072, 50) (45644,) (5072,)\n"
     ]
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "    np.array(smiles), np.array(values),\n",
    "    test_size=0.1, random_state=42, shuffle=True)\n",
    "print(X_train.shape, X_test.shape, y_train.shape, y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 448,
   "metadata": {},
   "outputs": [],
   "source": [
    "som = susi.SOMRegressor(\n",
    "    n_rows=60,\n",
    "    n_columns=60,\n",
    "    n_iter_unsupervised=1000000,\n",
    "    n_iter_supervised=1000000,\n",
    "    missing_label_placeholder=-1,\n",
    "    #neighborhood_mode_unsupervised=\"root\",\n",
    "    #neighborhood_mode_supervised=\"root\",\n",
    "    learn_mode_unsupervised=\"min\",\n",
    "    learn_mode_supervised=\"min\",\n",
    "    learning_rate_start=0.5,\n",
    "    learning_rate_end=0.05,\n",
    "    random_state=None,\n",
    "    n_jobs=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 449,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SOMRegressor(missing_label_placeholder=-1, n_columns=60,\n",
       "             n_iter_supervised=1000000, n_iter_unsupervised=1000000, n_jobs=1,\n",
       "             n_rows=60)"
      ]
     },
     "execution_count": 449,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "som.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_Test = []\n",
    "y_Test = []\n",
    "for i,v in enumerate(y_test):\n",
    "    if v > 0:\n",
    "        y_Test.append(v)\n",
    "        x_Test.append(X_test[i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = []\n",
    "y = []\n",
    "for i,v in enumerate(values):\n",
    "    if v > 0:\n",
    "        y.append(v)\n",
    "        x.append(smiles[i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train: 55.0 %\n"
     ]
    }
   ],
   "source": [
    "#y_pred = som.predict(x_Test)\n",
    "y_pred_train = som.predict(x)\n",
    "\n",
    "print(\"Train: {0:.1f} %\".format(r2_score(y, y_pred_train)*100))\n",
    "#print(\"Test: {0:.1f} %\".format(r2_score(y_Test, y_pred)*100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 348,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.056895430151244104, 0.09681841016882288)"
      ]
     },
     "execution_count": 348,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train[31],y_pred_train[31]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 456,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "OK\n"
     ]
    }
   ],
   "source": [
    "import requests\n",
    "\n",
    "url = \"https://notificationer.glitch.me/notify\"\n",
    "\n",
    "payload = \"{ \\\"id\\\":\\\"k1r9fznrir0wwxgf5j7nx\\\", \\\"text\\\":\\\"Project Ready!\\\" }\"\n",
    "headers = {\n",
    "    'Content-Type': \"application/json\",\n",
    "    'Authorization': \"Basic Og==\"\n",
    "    }\n",
    "\n",
    "response = requests.request(\"POST\", url, data=payload, headers=headers)\n",
    "\n",
    "print(response.text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "estimation_map = np.squeeze(som.get_estimation_map())\n",
    "plot_estimation_map(\n",
    "    estimation_map, cbar_label=\"Log IC50\", cmap=\"viridis\")\n",
    "plt.title(\"Figure 3: General SOM Map\",  fontsize=18)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 457,
   "metadata": {},
   "outputs": [],
   "source": [
    "pickle.dump(som, open(\"general_leish_som_1M.p\", \"wb\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "som = pickle.load(open(\"general_leish_som_500K.p\", \"rb\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "11.569821233577905"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "som.predict([list(x[0])])[0]"
   ]
  }
 ],
 "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": "3.8.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}