tutorials/tutorial.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Tutorial\n",
    "The basic steps of ikarus are similar to the [sklearn workflow](https://scikit-learn.org/stable/getting_started.html#fitting-and-predicting-estimator-basics):\n",
    "\n",
    "0. load data sets\n",
    "1. init model\n",
    "2. fit model\n",
    "3. predict\n",
    "\n",
    "**Further dependencies:**\n",
    "This tutorial uses the python packages matplotlib and seaborn for visualization. These are not required by ikarus and will need to be installed separately."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import urllib.request\n",
    "import anndata\n",
    "import pandas as pd\n",
    "from pathlib import Path\n",
    "from ikarus import classifier, utils, data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## load data sets\n",
    "For this tutorial first, we will download single cell gene expression information for three different cancer types. The download takes about 1.6 GB of disk space.  \n",
    "\n",
    "Laughney lung cancer and Lee colorectal cancer are used for training. Tirosh head & neck cancer is used for testing. Here we rely on annotated data matrices (see [AnnData](https://anndata.readthedocs.io/en/latest/anndata.AnnData.html) for further information.)  \n",
    "\n",
    "The provided AnnData objects are already preprocessed, else it would be recommended to do `adata = data.preprocess_adata(adata)` first."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "Path(\"data\").mkdir(exist_ok=True)\n",
    "\n",
    "if not Path(\"data/laughney20_lung/adata.h5ad\").is_file():\n",
    "    Path(\"data/laughney20_lung\").mkdir(exist_ok=True)\n",
    "    urllib.request.urlretrieve(\n",
    "        url=\"https://bimsbstatic.mdc-berlin.de/akalin/Ikarus/part_1/data/laughney20_lung/adata.h5ad\",\n",
    "        filename=Path(\"data/laughney20_lung/adata.h5ad\")\n",
    "    )\n",
    "    \n",
    "if not Path(\"data/lee20_crc/adata.h5ad\").is_file():\n",
    "    Path(\"data/lee20_crc\").mkdir(exist_ok=True)\n",
    "    urllib.request.urlretrieve(\n",
    "        url=\"https://bimsbstatic.mdc-berlin.de/akalin/Ikarus/part_1/data/lee20_crc/adata.h5ad\",\n",
    "        filename=Path(\"data/lee20_crc/adata.h5ad\")\n",
    "    )\n",
    "\n",
    "if not Path(\"data/tirosh17_headneck/adata.h5ad\").is_file():\n",
    "    Path(\"data/tirosh17_headneck\").mkdir(exist_ok=True)\n",
    "    urllib.request.urlretrieve(\n",
    "        url=\"https://bimsbstatic.mdc-berlin.de/akalin/Ikarus/part_1/data/tirosh17_headneck/adata.h5ad\",\n",
    "        filename=Path(\"data/tirosh17_headneck/adata.h5ad\")\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "paths = [\n",
    "    Path(\"data/laughney20_lung/\"),\n",
    "    Path(\"data/lee20_crc/\"),\n",
    "    Path(\"data/tirosh17_headneck/\")\n",
    "]\n",
    "names = [\n",
    "    \"laughney\",\n",
    "    \"lee\",\n",
    "    \"tirosh\"\n",
    "]\n",
    "adatas = {}\n",
    "for path, name in zip(paths, names):\n",
    "    adatas[name] = anndata.read_h5ad(path / \"adata.h5ad\")\n",
    "    # Uncomment to perform preprocessing. Here, the loaded anndata objects are already preprocessed. \n",
    "    # adatas[name] = data.preprocess_adata(adatas[name])"
   ]
  },
  {
   "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>raw</th>\n",
       "      <th>major</th>\n",
       "      <th>major_hallmark_corrected</th>\n",
       "      <th>tier_0</th>\n",
       "      <th>tier_1</th>\n",
       "      <th>tier_2</th>\n",
       "      <th>tier_3</th>\n",
       "      <th>tier_4</th>\n",
       "      <th>tier_5</th>\n",
       "      <th>tier_0_hallmark_corrected</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CMS2</td>\n",
       "      <td>Tumor</td>\n",
       "      <td>Tumor</td>\n",
       "      <td>Tumor</td>\n",
       "      <td>Epithelial</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "      <td>Tumor</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CMS2</td>\n",
       "      <td>Tumor</td>\n",
       "      <td>Tumor</td>\n",
       "      <td>Tumor</td>\n",
       "      <td>Epithelial</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "      <td>nan</td>\n",
       "      <td>Tumor</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    raw  major major_hallmark_corrected tier_0      tier_1 tier_2 tier_3  \\\n",
       "0  CMS2  Tumor                    Tumor  Tumor  Epithelial    nan    nan   \n",
       "1  CMS2  Tumor                    Tumor  Tumor  Epithelial    nan    nan   \n",
       "\n",
       "  tier_4 tier_5 tier_0_hallmark_corrected  \n",
       "0    nan    nan                     Tumor  \n",
       "1    nan    nan                     Tumor  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "adatas[\"lee\"].obs.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>gene_symbol</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>A1BG</th>\n",
       "      <td>A1BG</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>A1BG-AS1</th>\n",
       "      <td>A1BG-AS1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         gene_symbol\n",
       "A1BG            A1BG\n",
       "A1BG-AS1    A1BG-AS1"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "adatas[\"lee\"].var.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>A1BG</th>\n",
       "      <th>A1BG-AS1</th>\n",
       "      <th>A1CF</th>\n",
       "      <th>A2M</th>\n",
       "      <th>A2M-AS1</th>\n",
       "      <th>A2ML1</th>\n",
       "      <th>A2ML1-AS1</th>\n",
       "      <th>A2ML1-AS2</th>\n",
       "      <th>A3GALT2</th>\n",
       "      <th>A4GALT</th>\n",
       "      <th>...</th>\n",
       "      <th>ZXDC</th>\n",
       "      <th>ZYG11A</th>\n",
       "      <th>ZYG11B</th>\n",
       "      <th>ZYX</th>\n",
       "      <th>ZZEF1</th>\n",
       "      <th>ZZZ3</th>\n",
       "      <th>bP-21264C1.2</th>\n",
       "      <th>bP-2171C21.3</th>\n",
       "      <th>bP-2189O9.3</th>\n",
       "      <th>hsa-mir-1253</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.233330</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.465363</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.259511</td>\n",
       "      <td>0.908856</td>\n",
       "      <td>0.259511</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 33694 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   A1BG  A1BG-AS1      A1CF  A2M  A2M-AS1  A2ML1  A2ML1-AS1  A2ML1-AS2  \\\n",
       "0   0.0       0.0  0.000000  0.0      0.0    0.0        0.0        0.0   \n",
       "1   0.0       0.0  0.465363  0.0      0.0    0.0        0.0        0.0   \n",
       "\n",
       "   A3GALT2  A4GALT  ...  ZXDC  ZYG11A    ZYG11B       ZYX     ZZEF1  ZZZ3  \\\n",
       "0      0.0     0.0  ...   0.0     0.0  0.000000  0.233330  0.000000   0.0   \n",
       "1      0.0     0.0  ...   0.0     0.0  0.259511  0.908856  0.259511   0.0   \n",
       "\n",
       "   bP-21264C1.2  bP-2171C21.3  bP-2189O9.3  hsa-mir-1253  \n",
       "0           0.0           0.0          0.0           0.0  \n",
       "1           0.0           0.0          0.0           0.0  \n",
       "\n",
       "[2 rows x 33694 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "adatas[\"lee\"].to_df().head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## signatures\n",
    "Here we just define the path of our gene signatures of choice (find more information on how the gene lists are created [below](#create-gene-lists)).\n",
    "\n",
    "Note: It is important that the name (see first column in the example) of the gene lists within the `signatures.gmt` file correspond to the cell type for which they are meaningful. `signatures.gmt` is tab-separated."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "signatures_path = Path(\"out/signatures.gmt\")"
   ]
  },
  {
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>1305</th>\n",
       "      <th>1306</th>\n",
       "      <th>1307</th>\n",
       "      <th>1308</th>\n",
       "      <th>1309</th>\n",
       "      <th>1310</th>\n",
       "      <th>1311</th>\n",
       "      <th>1312</th>\n",
       "      <th>1313</th>\n",
       "      <th>1314</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Normal</td>\n",
       "      <td>ikarus</td>\n",
       "      <td>RP11-128M1.1</td>\n",
       "      <td>TRAV8-2</td>\n",
       "      <td>PTPRD-AS1</td>\n",
       "      <td>MEOX2</td>\n",
       "      <td>CXCL12</td>\n",
       "      <td>KLRC4-KLRK1</td>\n",
       "      <td>BCAS1</td>\n",
       "      <td>SCNN1A</td>\n",
       "      <td>...</td>\n",
       "      <td>C22ORF15</td>\n",
       "      <td>CYP4F11</td>\n",
       "      <td>AK8</td>\n",
       "      <td>LRRC18</td>\n",
       "      <td>LMO2</td>\n",
       "      <td>COL12A1</td>\n",
       "      <td>ITGA11</td>\n",
       "      <td>EGFL6</td>\n",
       "      <td>RGS11</td>\n",
       "      <td>PCDHB15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Tumor</td>\n",
       "      <td>ikarus</td>\n",
       "      <td>RP11-277P12.10</td>\n",
       "      <td>RP13-895J2.6</td>\n",
       "      <td>BNC1</td>\n",
       "      <td>MAGEA6</td>\n",
       "      <td>ISX</td>\n",
       "      <td>MAGEA3</td>\n",
       "      <td>RP13-614K11.2</td>\n",
       "      <td>CDH7</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 1315 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     0       1               2             3          4       5       6     \\\n",
       "0  Normal  ikarus    RP11-128M1.1       TRAV8-2  PTPRD-AS1   MEOX2  CXCL12   \n",
       "1   Tumor  ikarus  RP11-277P12.10  RP13-895J2.6       BNC1  MAGEA6     ISX   \n",
       "\n",
       "          7              8       9     ...      1305     1306 1307    1308  \\\n",
       "0  KLRC4-KLRK1          BCAS1  SCNN1A  ...  C22ORF15  CYP4F11  AK8  LRRC18   \n",
       "1       MAGEA3  RP13-614K11.2    CDH7  ...       NaN      NaN  NaN     NaN   \n",
       "\n",
       "   1309     1310    1311   1312   1313     1314  \n",
       "0  LMO2  COL12A1  ITGA11  EGFL6  RGS11  PCDHB15  \n",
       "1   NaN      NaN     NaN    NaN    NaN      NaN  \n",
       "\n",
       "[2 rows x 1315 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.read_csv(signatures_path, sep=\"\\t\", header=None)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## train model\n",
    "Unlike the X and y input from the sklearn fitting, we are using here the AnnData object loaded above and the corresponding observation column (annotation). The input AnnData objects, names and target observation columns are each provided as lists. As one may want to use multiple input data sets.\n",
    "\n",
    "(Note: Here this may take a while, as the used input data sets are quite large. Skip this cell in case you just want load the pretrained model.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = classifier.Ikarus(signatures_gmt=signatures_path, out_dir=\"out\")\n",
    "\n",
    "train_adata_list = [adatas[\"laughney\"], adatas[\"lee\"]]\n",
    "train_names_list = [\"laughney\", \"lee\"]\n",
    "obs_columns_list = [\"tier_0_hallmark_corrected\", \"tier_0_hallmark_corrected\"]\n",
    "\n",
    "model.fit(train_adata_list, train_names_list, obs_columns_list, save=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Alternatively just load the pretrained model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_path = Path(\"out/core_model.joblib\")\n",
    "model = classifier.Ikarus(signatures_gmt=signatures_path, out_dir=\"out\")\n",
    "model.load_core_model(model_path)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## make prediction\n",
    "We provide the AnnData of an unknown data set, make predictions and save the results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Trying to set attribute `.uns` of view, copying.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "converged at iteration step: 9 with 0.0004 < 0.001\n"
     ]
    }
   ],
   "source": [
    "_ = model.predict(adatas[\"tirosh\"], \"tirosh\", save=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Your predictions `prediction.csv` are saved to your output directory as specified in the model initialization. At this point you might try to [correct the predictions by making use of CNV information](make-use-of-cnv-information-and-correct-prediction). Else, go on with the plotting part to verify the results."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## plotting\n",
    "Make plots to investigate the prediction results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "code_folding": [
     5
    ]
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import scanpy as sc\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn import metrics\n",
    "\n",
    "def plot_confusion_matrix(\n",
    "    y_true, y_pred, classes, normalize=False, title=None, cmap=plt.cm.Blues, ax=None\n",
    "):\n",
    "    \"\"\"\n",
    "    This function prints and plots the confusion matrix.\n",
    "    Normalization can be applied by setting `normalize=True`.\n",
    "    \"\"\"\n",
    "    plt.rcParams[\"figure.figsize\"] = [6, 4]\n",
    "    # print(classes)\n",
    "    if not title:\n",
    "        if normalize:\n",
    "            title = \"Normalized confusion matrix\"\n",
    "        else:\n",
    "            title = \"Confusion matrix, without normalization\"\n",
    "\n",
    "    # Compute confusion matrix\n",
    "    cm = metrics.confusion_matrix(y_true, y_pred, labels=classes)\n",
    "    # Only use the labels that appear in the data\n",
    "    # classes = classes[unique_labels(y_true, y_pred)]\n",
    "    if normalize:\n",
    "        cm = cm.astype(\"float\") / cm.sum(axis=1)[:, np.newaxis]\n",
    "\n",
    "    if ax is None:\n",
    "        (fig, ax) = plt.subplots()\n",
    "\n",
    "    im = ax.imshow(cm, interpolation=\"nearest\", cmap=cmap)\n",
    "    ax.figure.colorbar(im, ax=ax)\n",
    "    # We want to show all ticks...\n",
    "    ax.set(\n",
    "        xticks=np.arange(cm.shape[1]),\n",
    "        yticks=np.arange(cm.shape[0]),\n",
    "        # ... and label them with the respective list entries\n",
    "        xticklabels=classes,\n",
    "        yticklabels=classes,\n",
    "        title=title,\n",
    "        ylabel=\"True label\",\n",
    "        xlabel=\"Predicted label\",\n",
    "    )\n",
    "    for item in (\n",
    "        [ax.title, ax.xaxis.label, ax.yaxis.label]\n",
    "        + ax.get_xticklabels()\n",
    "        + ax.get_yticklabels()\n",
    "    ):\n",
    "        item.set_fontsize(12)\n",
    "\n",
    "    # Rotate the tick labels and set their alignment.\n",
    "    plt.setp(ax.get_xticklabels(), rotation=45, ha=\"right\", rotation_mode=\"anchor\")\n",
    "\n",
    "    # Loop over data dimensions and create text annotations.\n",
    "    fmt = \".2f\" if normalize else \"d\"\n",
    "    thresh = cm.max() / 2.0\n",
    "    for i in range(cm.shape[0]):\n",
    "        for j in range(cm.shape[1]):\n",
    "            ax.text(\n",
    "                j,\n",
    "                i,\n",
    "                format(cm[i, j], fmt),\n",
    "                ha=\"center\",\n",
    "                va=\"center\",\n",
    "                color=\"white\" if cm[i, j] > thresh else \"black\",\n",
    "            )\n",
    "\n",
    "    return fig, ax"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Compute UMAPs and store the adata object `adata_umap.h5ad`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/jdohmen/.miniconda3/envs/ikarus-test/lib/python3.8/site-packages/anndata/_core/anndata.py:1220: FutureWarning: The `inplace` parameter in pandas.Categorical.reorder_categories is deprecated and will be removed in a future version. Removing unused categories will always return a new Categorical object.\n",
      "  c.reorder_categories(natsorted(c.categories), inplace=True)\n",
      "... storing 'core_pred' as categorical\n",
      "/home/jdohmen/.miniconda3/envs/ikarus-test/lib/python3.8/site-packages/anndata/_core/anndata.py:1220: FutureWarning: The `inplace` parameter in pandas.Categorical.reorder_categories is deprecated and will be removed in a future version. Removing unused categories will always return a new Categorical object.\n",
      "  c.reorder_categories(natsorted(c.categories), inplace=True)\n",
      "... storing 'final_pred' as categorical\n"
     ]
    }
   ],
   "source": [
    "_ = model.get_umap(adatas[\"tirosh\"], \"tirosh\", save=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Load prediction results and the AnnData object including the just computed UMAP coordinates."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "path = Path(\"out/tirosh\")\n",
    "results = pd.read_csv(path / \"prediction.csv\", index_col=0)\n",
    "adata = anndata.read_h5ad(path / \"adata_umap.h5ad\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Print classification metrics and show confusion matrix and UMAPs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "      Normal       1.00      1.00      1.00      3363\n",
      "       Tumor       1.00      1.00      1.00      2215\n",
      "\n",
      "    accuracy                           1.00      5578\n",
      "   macro avg       1.00      1.00      1.00      5578\n",
      "weighted avg       1.00      1.00      1.00      5578\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y = adata.obs.loc[:, \"tier_0\"]\n",
    "y_pred_lr = results[\"final_pred\"]\n",
    "acc = metrics.balanced_accuracy_score(y, y_pred_lr)\n",
    "print(metrics.classification_report(y, y_pred_lr, labels=[\"Normal\", \"Tumor\"]))\n",
    "fig, ax = plot_confusion_matrix(\n",
    "    y,\n",
    "    y_pred_lr,\n",
    "    classes=[\"Normal\", \"Tumor\"],\n",
    "    title=f\"confusion matrix \\n train on laughney+lee, test on tirosh \\n balanced acc: {acc:.3f}\",\n",
    ")\n",
    "fig.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/jdohmen/.miniconda3/envs/ikarus-test/lib/python3.8/site-packages/anndata/_core/anndata.py:1220: FutureWarning: The `inplace` parameter in pandas.Categorical.reorder_categories is deprecated and will be removed in a future version. Removing unused categories will always return a new Categorical object.\n",
      "  c.reorder_categories(natsorted(c.categories), inplace=True)\n",
      "... storing 'tier_0_pred_correctness' as categorical\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "adata.obs.loc[:, \"tier_0_pred_correctness\"] = \"wrongly assigned\"\n",
    "adata.obs.loc[\n",
    "    adata.obs[\"tier_0\"] == adata.obs[\"final_pred\"],\n",
    "    \"tier_0_pred_correctness\"\n",
    "] = \"correctly assigned\"\n",
    "adata.obs.loc[:, \"tier_0_pred_wrong\"] = pd.Categorical(\n",
    "    adata.obs[\"tier_0\"].copy(),\n",
    "    categories=np.array([\"Normal\", \"Tumor\", \"wrongly assigned\"]),\n",
    "    ordered=True\n",
    ")\n",
    "adata.obs.loc[\n",
    "    adata.obs[\"tier_0_pred_correctness\"] == \"wrongly assigned\",\n",
    "    \"tier_0_pred_wrong\"\n",
    "] = \"wrongly assigned\"\n",
    "\n",
    "plt.rcParams[\"figure.figsize\"] = [9, 6]\n",
    "\n",
    "colors = [\n",
    "    [\"major\"],\n",
    "    [\"tier_0_pred_wrong\"]\n",
    "    ]\n",
    "titles = [\n",
    "    [\"major types\"],\n",
    "    [\"prediction\"]\n",
    "    ]\n",
    "palettes = [\n",
    "    [\"#7f7f7f\", \"#98df8a\", \"#aec7e8\", \"#9467bd\", \"#ff0000\"],\n",
    "    [\"#aec7e8\", \"#ff0000\", \"#0b559f\"], \n",
    "]\n",
    "for color, title, palette in zip(colors, titles, palettes):\n",
    "    ax = sc.pl.umap(\n",
    "        adata, ncols=1, size=20, \n",
    "        color=color,\n",
    "        title=title,\n",
    "        wspace=0.25,\n",
    "        vmax=\"p99\",\n",
    "        legend_fontsize=12,\n",
    "        palette=palette,\n",
    "        show=False\n",
    "    )\n",
    "    for item in ([ax.title, ax.xaxis.label, ax.yaxis.label] + \n",
    "                 ax.get_xticklabels() + ax.get_yticklabels()):\n",
    "        item.set_fontsize(12)\n",
    "    ax.legend(loc=\"best\")\n",
    "    plt.tight_layout()\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## extra"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### create gene lists"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "import upsetplot\n",
    "import anndata\n",
    "import pandas as pd\n",
    "from pathlib import Path\n",
    "from ikarus import gene_list, utils"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### tumor"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, we define the input datasets used to create the gene lists, the upregulated and downregulated cell types. Here, we consider six different pair-wise comparisons:\n",
    "- Tumor vs. Other\n",
    "- Tumor vs. Epithelial\n",
    "- Tumor vs. Immune\n",
    "- Tumor vs. Fibroblast\n",
    "- Tumor vs. Endothelial\n",
    "- Tumor vs. Pericyte"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "paths = [\n",
    "    Path(\"data/laughney20_lung/\"),\n",
    "    Path(\"data/lee20_crc/\")\n",
    "]\n",
    "names = [\n",
    "    \"laughney\",\n",
    "    \"lee\"\n",
    "]\n",
    "obs_names = [\n",
    "    \"major_hallmark_corrected\",\n",
    "    \"major_hallmark_corrected\"\n",
    "]\n",
    "label_upregs = [\n",
    "    \"Tumor\",\n",
    "    \"Tumor\",\n",
    "    \"Tumor\",\n",
    "    \"Tumor\",\n",
    "    \"Tumor\",\n",
    "    \"Tumor\"\n",
    "]\n",
    "label_downregs = [\n",
    "    \"Other\",\n",
    "    \"Epithelial\",\n",
    "    \"Immune\",\n",
    "    \"Fibroblast\",\n",
    "    \"Endothelial\",\n",
    "    \"Pericyte\"\n",
    "]\n",
    "adatas = {}\n",
    "for path, name in zip(paths, names):\n",
    "    adatas[name] = anndata.read_h5ad(path / \"adata.h5ad\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We loop over all pair-wise comparisons and over both input datasets, and perform t-tests with overestimated variance so that we obtain for each gene two (one for each dataset) estimates of log2fold changes. For integrating the outcome of both datasets, first of all we just consider genes available in both datasets (intersection). Then we compute for each gene the weighted average of log2fold changes. For each pair-wise comparison we take the top 300 of upregulated (highest averaged log2fold changes) genes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Either Tumor or Fibroblast is not available in given obs_name. None is returned.\n",
      "Either Tumor or Endothelial is not available in given obs_name. None is returned.\n",
      "Either Tumor or Pericyte is not available in given obs_name. None is returned.\n"
     ]
    }
   ],
   "source": [
    "signatures = gene_list.create_all(\n",
    "    label_upregs_list=label_upregs,\n",
    "    label_downregs_list=label_downregs,\n",
    "    adatas_dict=adatas,\n",
    "    names_list=names,\n",
    "    obs_names_list=obs_names,\n",
    "    integration_fun=utils.intersection_fun,\n",
    "    top_x=300\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here we use UpSetPlots to visualize the amount of genes overlapping across multiple pair-wise comparisons."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 320x384 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "contents = upsetplot.from_contents(signatures)\n",
    "_ = upsetplot.plot(contents, show_counts=\"%d\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Get the intersection of all the different gene lists. This leads to our final tumor gene list of 162 genes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_30579/1056897521.py:1: PerformanceWarning: indexing past lexsort depth may impact performance.\n",
      "  tumor_genes_intersection = contents.loc[\n"
     ]
    }
   ],
   "source": [
    "tumor_genes_intersection = contents.loc[\n",
    "    (True, True, True, True, True, True)\n",
    "].values.ravel().tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "tumor_genes = tumor_genes_intersection"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Later, when creating the normal gene list, we will remove those genes, which were also upregulated in the pair-wise comparisons from above (Tumor vs. X). For that purpose, we create here the union of the tumor genes coming from all pair-wise comparisons. This is not used as our final Tumor gene list. It's just used to refine the Normal gene list."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "tumor_genes_union = []\n",
    "for i in signatures.values():\n",
    "    tumor_genes_union += i\n",
    "tumor_genes_union = list(set(tumor_genes_union)) # unique genes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### normal\n",
    "Basically repeat the above procedure for the normal gene list. Though, here we take the union of genes upregulated in non-tumor cells.\n",
    "\n",
    "First, we define the input datasets used to create the gene lists, the upregulated and downregulated cell types. Here, we consider six different pair-wise comparisons:\n",
    "- Other vs. Tumor\n",
    "- Epithelial vs. Tumor\n",
    "- Immune vs. Tumor\n",
    "- Fibroblast vs. Tumor\n",
    "- Endothelial vs. Tumor\n",
    "- Pericyte vs. Tumor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "paths = [\n",
    "    Path(\"data/laughney20_lung/\"),\n",
    "    Path(\"data/lee20_crc/\")\n",
    "]\n",
    "names = [\n",
    "    \"laughney\",\n",
    "    \"lee\"\n",
    "]\n",
    "obs_names = [\n",
    "    \"major_hallmark_corrected\",\n",
    "    \"major_hallmark_corrected\"\n",
    "]\n",
    "label_downregs = [\n",
    "    \"Tumor\",\n",
    "    \"Tumor\",\n",
    "    \"Tumor\",\n",
    "    \"Tumor\",\n",
    "    \"Tumor\",\n",
    "    \"Tumor\"\n",
    "]\n",
    "label_upregs = [\n",
    "    \"Other\",\n",
    "    \"Epithelial\",\n",
    "    \"Immune\",\n",
    "    \"Fibroblast\",\n",
    "    \"Endothelial\",\n",
    "    \"Pericyte\"\n",
    "]\n",
    "adatas = {}\n",
    "for path, name in zip(paths, names):\n",
    "    adatas[name] = anndata.read_h5ad(path / \"adata.h5ad\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We loop over all pair-wise comparisons and over both input datasets, and perform t-tests with overestimated variance so that we obtain for each gene two (one for each dataset) estimates of log2fold changes. For integrating the outcome of both datasets, first of all we just consider genes available in both datasets (intersection). Then we compute for each gene the weighted average of log2fold changes. For each pair-wise comparison we take the top 300 of upregulated (highest averaged log2fold changes) genes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Either Fibroblast or Tumor is not available in given obs_name. None is returned.\n",
      "Either Endothelial or Tumor is not available in given obs_name. None is returned.\n",
      "Either Pericyte or Tumor is not available in given obs_name. None is returned.\n"
     ]
    }
   ],
   "source": [
    "signatures = gene_list.create_all(\n",
    "    label_upregs_list=label_upregs,\n",
    "    label_downregs_list=label_downregs,\n",
    "    adatas_dict=adatas,\n",
    "    names_list=names,\n",
    "    obs_names_list=obs_names,\n",
    "    integration_fun=utils.intersection_fun,\n",
    "    top_x=300\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Again. here we use UpSetPlots to visualize the amount of genes overlapping across multiple pair-wise comparisons."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 672x384 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "contents = upsetplot.from_contents(signatures)\n",
    "_ = upsetplot.plot(contents, show_counts=\"%d\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Get unique genes from the six pairwise comparisons."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "normal_genes_union = []\n",
    "for i in signatures.values():\n",
    "    normal_genes_union += i\n",
    "normal_genes_union = list(set(normal_genes_union)) # unique genes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Remove genes which are overlapping with the union of tumor genes from all pair-wise comparisons with Tumor being the upregulated class (see above)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "overlap = list(set(tumor_genes_union) & set(normal_genes_union))\n",
    "normal_genes = list(set(normal_genes_union) - set(overlap))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Save signatures as `.gmt` file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "gene_list.save_gmt([normal_genes, tumor_genes], [\"Normal\", \"Tumor\"], out_dir=\"out\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### make use of cnv information and correct prediction\n",
    "Altough, for this particular test data set this step might seem unnecessary, incorporating cnv information might help to improve prediction results.\n",
    "\n",
    "[inferCNV, v1.3.3](https://github.com/broadinstitute/inferCNV) was used to compute CNV scores.  \n",
    "\n",
    "The download takes about 1.1 GB of disk space.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "Path(\"data\").mkdir(exist_ok=True)\n",
    "\n",
    "if not Path(\"data/tirosh17_headneck/infercnv.observations.txt\").is_file():\n",
    "    Path(\"data/tirosh17_headneck\").mkdir(exist_ok=True)\n",
    "    urllib.request.urlretrieve(\n",
    "        url=\"https://bimsbstatic.mdc-berlin.de/akalin/Ikarus/part_1/data/tirosh17_headneck/infercnv.observations.txt\",\n",
    "        filename=Path(\"data/tirosh17_headneck/infercnv.observations.txt\")\n",
    "    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "cnv_path = Path(\"data/tirosh17_headneck/infercnv.observations.txt\")\n",
    "cnv = pd.read_csv(cnv_path, sep=\" \")\n",
    "cnv.columns = cnv.columns.astype(\"int\")\n",
    "cnv.sort_index(axis=1, inplace=True)\n",
    "cnv = cnv.T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "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>WASH7P</th>\n",
       "      <th>LINC00115</th>\n",
       "      <th>FAM41C</th>\n",
       "      <th>NOC2L</th>\n",
       "      <th>PLEKHN1</th>\n",
       "      <th>HES4</th>\n",
       "      <th>ISG15</th>\n",
       "      <th>AGRN</th>\n",
       "      <th>C1orf159</th>\n",
       "      <th>TNFRSF18</th>\n",
       "      <th>...</th>\n",
       "      <th>SBF1</th>\n",
       "      <th>LMF2</th>\n",
       "      <th>NCAPH2</th>\n",
       "      <th>SCO2</th>\n",
       "      <th>TYMP</th>\n",
       "      <th>ODF3B</th>\n",
       "      <th>CPT1B</th>\n",
       "      <th>CHKB</th>\n",
       "      <th>ARSA</th>\n",
       "      <th>RABL2B</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>...</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>1.002759</td>\n",
       "      <td>...</td>\n",
       "      <td>0.850724</td>\n",
       "      <td>0.850990</td>\n",
       "      <td>0.851188</td>\n",
       "      <td>0.850058</td>\n",
       "      <td>0.849016</td>\n",
       "      <td>0.848378</td>\n",
       "      <td>0.847733</td>\n",
       "      <td>0.847032</td>\n",
       "      <td>0.846197</td>\n",
       "      <td>0.845542</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 11991 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     WASH7P  LINC00115    FAM41C     NOC2L   PLEKHN1      HES4     ISG15  \\\n",
       "0  1.002759   1.002759  1.002759  1.002759  1.002759  1.002759  1.002759   \n",
       "1  1.002759   1.002759  1.002759  1.002759  1.002759  1.002759  1.002759   \n",
       "\n",
       "       AGRN  C1orf159  TNFRSF18  ...      SBF1      LMF2    NCAPH2      SCO2  \\\n",
       "0  1.002759  1.002759  1.002759  ...  1.002759  1.002759  1.002759  1.002759   \n",
       "1  1.002759  1.002759  1.002759  ...  0.850724  0.850990  0.851188  0.850058   \n",
       "\n",
       "       TYMP     ODF3B     CPT1B      CHKB      ARSA    RABL2B  \n",
       "0  1.002759  1.002759  1.002759  1.002759  1.002759  1.002759  \n",
       "1  0.849016  0.848378  0.847733  0.847032  0.846197  0.845542  \n",
       "\n",
       "[2 rows x 11991 columns]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cnv.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "_ = model.cnv_correct(cnv, adatas[\"tirosh\"], save=True, name=\"tirosh\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Verify CNV corrected results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "path = Path(\"out/tirosh\")\n",
    "results = pd.read_csv(path / \"prediction.csv\", index_col=0)\n",
    "adata = anndata.read_h5ad(path / \"adata_umap.h5ad\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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>Normal</th>\n",
       "      <th>Tumor</th>\n",
       "      <th>core_pred</th>\n",
       "      <th>core_pred_proba_Normal</th>\n",
       "      <th>core_pred_proba_Tumor</th>\n",
       "      <th>final_pred</th>\n",
       "      <th>final_pred_proba_Normal</th>\n",
       "      <th>final_pred_proba_Tumor</th>\n",
       "      <th>final_pred_cnv_corrected</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.594332</td>\n",
       "      <td>0.022227</td>\n",
       "      <td>Normal</td>\n",
       "      <td>0.860585</td>\n",
       "      <td>0.139415</td>\n",
       "      <td>Normal</td>\n",
       "      <td>0.855562</td>\n",
       "      <td>0.144438</td>\n",
       "      <td>Normal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.497360</td>\n",
       "      <td>0.018565</td>\n",
       "      <td>Normal</td>\n",
       "      <td>0.869966</td>\n",
       "      <td>0.130034</td>\n",
       "      <td>Normal</td>\n",
       "      <td>0.844266</td>\n",
       "      <td>0.155734</td>\n",
       "      <td>Normal</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Normal     Tumor core_pred  core_pred_proba_Normal  \\\n",
       "0  0.594332  0.022227    Normal                0.860585   \n",
       "1  0.497360  0.018565    Normal                0.869966   \n",
       "\n",
       "   core_pred_proba_Tumor final_pred  final_pred_proba_Normal  \\\n",
       "0               0.139415     Normal                 0.855562   \n",
       "1               0.130034     Normal                 0.844266   \n",
       "\n",
       "   final_pred_proba_Tumor final_pred_cnv_corrected  \n",
       "0                0.144438                   Normal  \n",
       "1                0.155734                   Normal  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "      Normal       1.00      1.00      1.00      3363\n",
      "       Tumor       1.00      1.00      1.00      2215\n",
      "\n",
      "    accuracy                           1.00      5578\n",
      "   macro avg       1.00      1.00      1.00      5578\n",
      "weighted avg       1.00      1.00      1.00      5578\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y = adata.obs.loc[:, \"tier_0\"]\n",
    "y_pred_lr = results[\"final_pred_cnv_corrected\"]\n",
    "acc = metrics.balanced_accuracy_score(y, y_pred_lr)\n",
    "print(metrics.classification_report(y, y_pred_lr, labels=[\"Normal\", \"Tumor\"]))\n",
    "fig, ax = plot_confusion_matrix(\n",
    "    y,\n",
    "    y_pred_lr,\n",
    "    classes=[\"Normal\", \"Tumor\"],\n",
    "    title=f\"confusion matrix \\n train on laughney+lee, test on tirosh \\n cnv corrected \\n balanced acc: {acc:.3f}\",\n",
    ")\n",
    "fig.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "ik_install_test",
   "language": "python",
   "name": "ik_install_test"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}