process_data.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "b559f3dd-4d65-4a16-ae49-23436f6871d4",
   "metadata": {},
   "source": [
    "# Process the deep mutational scanning data\n",
    "\n",
    "The data come from the following papers from Yunlong Cao's group:\n",
    "\n",
    " - [Imprinted SARS-CoV-2 humoral immunity induces convergent Omicron RBD evolution (2022)](https://www.nature.com/articles/s41586-022-05644-7)\n",
    " \n",
    " - [Repeated Omicron infection alleviates SARS-CoV-2 immune imprinting (2024)](https://www.nature.com/articles/s41586-023-06753-7)\n",
    "\n",
    " - [Evolving antibody response to SARS-CoV-2 antigenic shift from XBB to JN.1](https://www.nature.com/articles/s41586-024-08315-x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "2a9757e8-5572-4173-a5c3-025991bccac5",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:22:46.396861Z",
     "iopub.status.busy": "2024-11-18T23:22:46.396352Z",
     "iopub.status.idle": "2024-11-18T23:22:48.443054Z",
     "shell.execute_reply": "2024-11-18T23:22:48.441713Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import math\n",
    "import os\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import numpy\n",
    "\n",
    "import pandas as pd\n",
    "\n",
    "import yaml\n",
    "\n",
    "with open(\"config.yaml\") as f:\n",
    "    config = yaml.safe_load(f)\n",
    "    \n",
    "antibody_reweighting = config[\"antibody_reweighting\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fa48e2dd-7bb9-4ff3-b214-abe1c42aa5ad",
   "metadata": {},
   "source": [
    "## Process data from [Evolving antibody response to SARS-CoV-2 antigenic shift from XBB to JN.1](https://www.nature.com/articles/s41586-024-08315-x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "525cacd9-6096-4383-9fe8-fe88d2a10083",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:22:48.448185Z",
     "iopub.status.busy": "2024-11-18T23:22:48.447266Z",
     "iopub.status.idle": "2024-11-18T23:22:48.624741Z",
     "shell.execute_reply": "2024-11-18T23:22:48.623782Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "evolving_2024_info['antibody'].nunique()=2915\n"
     ]
    }
   ],
   "source": [
    "# read the antibody data\n",
    "evolving_2024_info = (\n",
    "    pd.read_csv(\n",
    "        \"Cao_data/JN1-evolving-antibody-response/data/_Table_S2.csv\",\n",
    "    )\n",
    "    .assign(study=\"evolving_2024\")\n",
    ")\n",
    "\n",
    "assert len(evolving_2024_info) == evolving_2024_info[\"antibody\"].nunique()\n",
    "\n",
    "print(f\"{evolving_2024_info['antibody'].nunique()=}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "af59a516-902a-4a28-bb83-e6353e93c80d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:22:48.627680Z",
     "iopub.status.busy": "2024-11-18T23:22:48.627364Z",
     "iopub.status.idle": "2024-11-18T23:22:48.658985Z",
     "shell.execute_reply": "2024-11-18T23:22:48.658211Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of antibodies from each source:\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></th>\n",
       "      <th>n_antibodies</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>study</th>\n",
       "      <th>source</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"23\" valign=\"top\">evolving_2024</th>\n",
       "      <th>BA.5 BTI + XBB infection</th>\n",
       "      <td>383</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.5 + XBB infection</th>\n",
       "      <td>314</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XBB BTI</th>\n",
       "      <td>303</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XBB infection</th>\n",
       "      <td>301</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.5 BTI + JN.1 infection</th>\n",
       "      <td>300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.5 BTI + HK.3 infection</th>\n",
       "      <td>300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.5 + JN.1 infection</th>\n",
       "      <td>200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.1 BTI + BA.5/BF.7 infection</th>\n",
       "      <td>162</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.2 BTI + BA.5/BF.7 infection</th>\n",
       "      <td>135</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.2 + BA.5/BF.7 infection</th>\n",
       "      <td>109</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.1 + BA.5/BF.7 infection</th>\n",
       "      <td>97</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.5 BTI</th>\n",
       "      <td>78</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.2 BTI</th>\n",
       "      <td>57</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BF.7 BTI</th>\n",
       "      <td>53</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.1 BTI</th>\n",
       "      <td>52</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>long-term BA.5 BTI</th>\n",
       "      <td>26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>long-term BA.1 BTI</th>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WT convalescents</th>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SARS exposure</th>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>long-term BA.5 infection</th>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>N.D.</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SARS</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WT-phage display</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                              n_antibodies\n",
       "study         source                                      \n",
       "evolving_2024 BA.5 BTI + XBB infection                 383\n",
       "              BA.5 + XBB infection                     314\n",
       "              XBB BTI                                  303\n",
       "              XBB infection                            301\n",
       "              BA.5 BTI + JN.1 infection                300\n",
       "              BA.5 BTI + HK.3 infection                300\n",
       "              BA.5 + JN.1 infection                    200\n",
       "              BA.1 BTI + BA.5/BF.7 infection           162\n",
       "              BA.2 BTI + BA.5/BF.7 infection           135\n",
       "              BA.2 + BA.5/BF.7 infection               109\n",
       "              BA.1 + BA.5/BF.7 infection                97\n",
       "              BA.5 BTI                                  78\n",
       "              BA.2 BTI                                  57\n",
       "              BF.7 BTI                                  53\n",
       "              BA.1 BTI                                  52\n",
       "              long-term BA.5 BTI                        26\n",
       "              long-term BA.1 BTI                        16\n",
       "              WT convalescents                          11\n",
       "              SARS exposure                              7\n",
       "              long-term BA.5 infection                   7\n",
       "              N.D.                                       2\n",
       "              SARS                                       1\n",
       "              WT-phage display                           1"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# get the antibody sources\n",
    "evolving_2024_source = (\n",
    "    evolving_2024_info\n",
    "    [[\"antibody\", \"source\", \"study\"]]\n",
    ")\n",
    "\n",
    "print(\"Number of antibodies from each source:\")\n",
    "display(\n",
    "    evolving_2024_source\n",
    "    .groupby([\"study\", \"source\"])\n",
    "    .aggregate(n_antibodies=pd.NamedAgg(\"antibody\", \"nunique\"))\n",
    "    .sort_values(\"n_antibodies\", ascending=False)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "3888ec98-6dcc-4f29-80f0-e1b212509b08",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:22:48.661679Z",
     "iopub.status.busy": "2024-11-18T23:22:48.661353Z",
     "iopub.status.idle": "2024-11-18T23:22:48.838346Z",
     "shell.execute_reply": "2024-11-18T23:22:48.837695Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Number of IC50s per antibody:\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>n_antibodies</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>viruses_w_IC50s</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; BA.2.86; JN.1; JN.1+F456L; KP.2; JN.1+F456L+A475V; KP.3; KP.3+A475V</th>\n",
       "      <td>2066</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; BA.2.86; JN.1; KP.3</th>\n",
       "      <td>366</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; XBB.1.5</th>\n",
       "      <td>197</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; BA.2.86; JN.1; JN.1+F456L; KP.2; JN.1+F456L+A475V; KP.3</th>\n",
       "      <td>132</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; BA.2.86; JN.1; JN.1+F456L; JN.1+F456L+A475V; KP.3</th>\n",
       "      <td>46</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; BA.2.86; JN.1; JN.1+F456L; JN.1+F456L+A475V; KP.3; KP.3+A475V</th>\n",
       "      <td>45</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; BA.2.86; JN.1; KP.2; KP.3</th>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; BA.2.86; JN.1; JN.1+F456L; KP.2; JN.1+F456L+A475V</th>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; BA.2.86</th>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; JN.1; JN.1+F456L; KP.2; JN.1+F456L+A475V; KP.3; KP.3+A475V</th>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; XBB.1.5; JN.1</th>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.5; XBB.1.5; HK.3.1; BA.2.86; JN.1; JN.1+F456L; KP.2; JN.1+F456L+A475V; KP.3; KP.3+A475V</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; XBB.1.5; BA.2.86; JN.1</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; XBB.1.5; BA.2.86; JN.1; JN.1+F456L; KP.2; JN.1+F456L+A475V; KP.3; KP.3+A475V</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; JN.1; KP.3; KP.3+A475V</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; JN.1; JN.1+F456L; JN.1+F456L+A475V; KP.3; KP.3+A475V</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.5; XBB.1.5; HK.3.1; BA.2.86; JN.1; JN.1+F456L; JN.1+F456L+A475V</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.5; XBB.1.5; HK.3.1; BA.2.86; JN.1; JN.1+F456L; KP.2; JN.1+F456L+A475V; KP.3; KP.3+A475V</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                    n_antibodies\n",
       "viruses_w_IC50s                                                 \n",
       "D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; BA.2....          2066\n",
       "D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; BA.2....           366\n",
       "D614G; BA.1; BA.2; BA.5; XBB.1.5                             197\n",
       "D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; BA.2....           132\n",
       "D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; BA.2....            46\n",
       "D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; BA.2....            45\n",
       "D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; BA.2....            30\n",
       "D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; BA.2....             6\n",
       "D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; BA.2.86              6\n",
       "D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; JN.1;...             5\n",
       "D614G; BA.1; BA.2; BA.5; XBB.1.5; JN.1                         4\n",
       "D614G; BA.1; BA.5; XBB.1.5; HK.3.1; BA.2.86; JN...             2\n",
       "D614G; BA.1; BA.2; BA.5                                        2\n",
       "D614G; BA.1; BA.2; BA.5; XBB.1.5; BA.2.86; JN.1                1\n",
       "D614G; BA.1; BA.2; BA.5; XBB.1.5; BA.2.86; JN.1...             1\n",
       "D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; JN.1;...             1\n",
       "D614G; BA.1; BA.2; BA.5; XBB.1.5; HK.3.1; JN.1;...             1\n",
       "D614G; BA.5; XBB.1.5; HK.3.1; BA.2.86; JN.1; JN...             1\n",
       "D614G; BA.5; XBB.1.5; HK.3.1; BA.2.86; JN.1; JN...             1"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "IC50s at different quantiles:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.00    2.119000e-14\n",
       "0.01    5.000000e-04\n",
       "0.10    3.027000e-03\n",
       "0.25    1.940000e-02\n",
       "0.50    1.478600e+00\n",
       "0.70    1.000000e+01\n",
       "0.90    1.000000e+01\n",
       "0.99    1.000000e+01\n",
       "1.00    1.000000e+01\n",
       "Name: IC50, dtype: float64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# get the IC50s\n",
    "\n",
    "evolving_2024_viruses = [\n",
    "    \"D614G\",\n",
    "    \"BA.1\",\n",
    "    \"BA.2\",\n",
    "    \"BA.5\", \n",
    "    \"XBB.1.5\",\n",
    "    \"HK.3.1\",\n",
    "    \"BA.2.86\",\n",
    "    \"JN.1\",\n",
    "    \"JN.1+F456L\",\n",
    "    \"KP.2\",\n",
    "    \"JN.1+F456L+A475V\",\n",
    "    \"KP.3\",\n",
    "    \"KP.3+A475V\",\n",
    "]\n",
    "assert set(evolving_2024_viruses).issubset(evolving_2024_info.columns)\n",
    "\n",
    "evolving_2024_ic50s = (\n",
    "    evolving_2024_info\n",
    "    .melt(\n",
    "        id_vars=[\"study\", \"antibody\"],\n",
    "        value_vars=evolving_2024_viruses,\n",
    "        var_name=\"virus\",\n",
    "        value_name=\"IC50\",\n",
    "    )\n",
    "    .query(\"IC50 != 'N.D.'\")  # IC50 not measured\n",
    "    # 10 is max value, taken to mean no neutralization\n",
    "    .assign(\n",
    "        IC50=lambda x: x[\"IC50\"].map(lambda ic50: 10 if ic50 == \">10\" else min(10, float(ic50)))\n",
    "    )\n",
    ")\n",
    "\n",
    "print(\"\\nNumber of IC50s per antibody:\")\n",
    "display(\n",
    "    evolving_2024_ic50s\n",
    "    .groupby(\"antibody\", as_index=False)\n",
    "    .aggregate(viruses_w_IC50s=pd.NamedAgg(\"virus\", lambda s: \"; \".join(s)))\n",
    "    .groupby(\"viruses_w_IC50s\")\n",
    "    .aggregate(n_antibodies=pd.NamedAgg(\"antibody\", \"nunique\"))\n",
    "    .sort_values(\"n_antibodies\", ascending=False)\n",
    ")\n",
    "\n",
    "print(\"IC50s at different quantiles:\")\n",
    "evolving_2024_ic50s[\"IC50\"].quantile([0, 0.01, 0.1, 0.25, 0.5, 0.7, 0.9, 0.99, 1])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b8bae6b6-3f9d-45cb-8174-956a7b832d72",
   "metadata": {},
   "source": [
    "Now read the actual escape data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "18179aae-056e-4a1f-96f7-21ffb6bc924e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:22:48.841402Z",
     "iopub.status.busy": "2024-11-18T23:22:48.841081Z",
     "iopub.status.idle": "2024-11-18T23:23:15.815869Z",
     "shell.execute_reply": "2024-11-18T23:23:15.814877Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "len(evolving_2024_abs)=2581 have antibody info and DMS, n_missing_dms=332 are missing DMS but have antibody info\n"
     ]
    }
   ],
   "source": [
    "evolving_2024_escape = (\n",
    "    pd.read_csv(\"Cao_data/JN1-evolving-antibody-response/data/DMS/antibody/dms_antibodies_XBB15_JN1_agg.csv\")\n",
    "    .groupby([\"antibody\", \"site\"], as_index=False)\n",
    "    .aggregate(escape=pd.NamedAgg(\"mut_escape\", \"sum\"))\n",
    "    .assign(study=\"evolving_2024\")\n",
    ")\n",
    "\n",
    "# The antibodies are named \"BD57-<xxx>\" in the antibody info, and\n",
    "# \"GC<xxx>\" in the escape data; harmonize\n",
    "assert not evolving_2024_ic50s[\"antibody\"].str.startswith(\"GC\").any()\n",
    "evolving_2024_escape = evolving_2024_escape.assign(\n",
    "    antibody=lambda x: x[\"antibody\"].map(\n",
    "        lambda a: \"BD57-\" + a[2:] if a.startswith(\"GC\") else a\n",
    "    )\n",
    ")\n",
    "\n",
    "evolving_2024_abs = set(evolving_2024_ic50s[\"antibody\"]).intersection(evolving_2024_escape[\"antibody\"])\n",
    "n_missing_dms = len(set(evolving_2024_ic50s[\"antibody\"]) - set(evolving_2024_escape[\"antibody\"]))\n",
    "print(f\"{len(evolving_2024_abs)=} have antibody info and DMS, {n_missing_dms=} are missing DMS but have antibody info\")\n",
    "\n",
    "# only keep escape for antibodies with IC50s\n",
    "evolving_2024_escape = evolving_2024_escape.query(\"antibody in @evolving_2024_abs\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e13d8200-2145-4932-9272-a79ae4d935aa",
   "metadata": {},
   "source": [
    "## Process data from [Repeated Omicron infection alleviates SARS-CoV-2 immune imprinting (2024)](https://www.nature.com/articles/s41586-023-06753-7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "8a8ed2a8-b997-4a01-8677-86785d5d9080",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:23:15.819530Z",
     "iopub.status.busy": "2024-11-18T23:23:15.819162Z",
     "iopub.status.idle": "2024-11-18T23:23:15.838684Z",
     "shell.execute_reply": "2024-11-18T23:23:15.837593Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# read the data\n",
    "repeated_2024_info = (\n",
    "    pd.read_csv(\"Cao_data/SARS-CoV-2-reinfection-DMS/antibody_info.csv\")\n",
    "    .rename(columns={\"Unnamed: 0\": \"antibody\"})\n",
    ").assign(study=\"repeated_2024\")\n",
    "\n",
    "assert len(repeated_2024_info) == repeated_2024_info[\"antibody\"].nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "a8f7058b-20ed-49c3-9bdb-10490c589042",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:23:15.842183Z",
     "iopub.status.busy": "2024-11-18T23:23:15.841778Z",
     "iopub.status.idle": "2024-11-18T23:23:15.865820Z",
     "shell.execute_reply": "2024-11-18T23:23:15.865036Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of antibodies from each source:\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></th>\n",
       "      <th>n_antibodies</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>study</th>\n",
       "      <th>source</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"9\" valign=\"top\">repeated_2024</th>\n",
       "      <th>BA.5 convalescents</th>\n",
       "      <td>445</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.1 convalescents reinfection</th>\n",
       "      <td>284</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BF.7 convalescents</th>\n",
       "      <td>243</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.2 convalescents reinfection</th>\n",
       "      <td>232</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.2 convalescents</th>\n",
       "      <td>53</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>long-term BA.1 convalescents</th>\n",
       "      <td>38</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.1 convalescents</th>\n",
       "      <td>36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WT convalescents</th>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SARS convalescents</th>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                              n_antibodies\n",
       "study         source                                      \n",
       "repeated_2024 BA.5 convalescents                       445\n",
       "              BA.1 convalescents reinfection           284\n",
       "              BF.7 convalescents                       243\n",
       "              BA.2 convalescents reinfection           232\n",
       "              BA.2 convalescents                        53\n",
       "              long-term BA.1 convalescents              38\n",
       "              BA.1 convalescents                        36\n",
       "              WT convalescents                          12\n",
       "              SARS convalescents                         5"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# get the antibody sources\n",
    "repeated_2024_source = (\n",
    "    repeated_2024_info\n",
    "    [[\"antibody\", \"source\", \"study\"]]\n",
    "    .query(\"source != 'WT mouse'\")  # do not keep mouse antibodies\n",
    ")\n",
    "\n",
    "print(\"Number of antibodies from each source:\")\n",
    "display(\n",
    "    repeated_2024_source\n",
    "    .groupby([\"study\", \"source\"])\n",
    "    .aggregate(n_antibodies=pd.NamedAgg(\"antibody\", \"nunique\"))\n",
    "    .sort_values(\"n_antibodies\", ascending=False)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "b9f0f887-a246-4fd4-8c2d-a9508daa1d28",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:23:15.869420Z",
     "iopub.status.busy": "2024-11-18T23:23:15.869065Z",
     "iopub.status.idle": "2024-11-18T23:23:15.962188Z",
     "shell.execute_reply": "2024-11-18T23:23:15.961415Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Number of IC50s per antibody:\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>n_antibodies</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>viruses_w_IC50s</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; SARS; BA.2.75; BA.5; BF.7; BQ.1.1; XBB; XBB.1.5; XBB.1.5.10</th>\n",
       "      <td>428</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; SARS; BA.2.75; BA.5; BF.7; BQ.1.1; XBB; XBB.1.5</th>\n",
       "      <td>181</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.3; SARS; BA.1.1; BA.2.12.1; BA.2.75; BA.5; BA.2.13; BQ.1.1; XBB; XBB.1.5; XBB.1.5.10</th>\n",
       "      <td>137</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; SARS; BA.2.75; BA.5; BQ.1.1; XBB; XBB.1.5; XBB.1.5.10</th>\n",
       "      <td>130</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.2.75; BA.5; BF.7; BQ.1.1; XBB; XBB.1.5; XBB.1.5.10</th>\n",
       "      <td>129</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; SARS; BA.2.75; BA.5; BQ.1.1; XBB; XBB.1.5</th>\n",
       "      <td>92</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.2.75; BA.5; BQ.1.1; XBB; XBB.1.5</th>\n",
       "      <td>48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; SARS; BA.5; BQ.1.1; XBB; XBB.1.5; XBB.1.5.10</th>\n",
       "      <td>42</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.2.75; BA.5; BQ.1.1; XBB; XBB.1.5; XBB.1.5.10</th>\n",
       "      <td>34</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.3; SARS; BA.1.1; BA.2.12.1; BA.2.75; BA.5; BA.2.13; BQ.1.1; XBB; XBB.1.5</th>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; SARS; BQ.1.1; XBB; XBB.1.5</th>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; SARS; BA.5; BQ.1.1; XBB; XBB.1.5</th>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; SARS; BA.2.12.1; BA.2.75; BA.5; BQ.1.1; XBB; XBB.1.5</th>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.5; BQ.1.1; XBB; XBB.1.5; XBB.1.5.10</th>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.2.75; BA.5; BF.7; BQ.1.1; XBB; XBB.1.5</th>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; SARS; BA.5; BF.7; BQ.1.1; XBB; XBB.1.5; XBB.1.5.10</th>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.5; BQ.1.1; XBB; XBB.1.5</th>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.3; SARS; BA.2.12.1; BA.2.75; BA.5; BQ.1.1; XBB; XBB.1.5</th>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; SARS; BA.5; BQ.1.1; XBB; XBB.1.5</th>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; SARS; BA.2.12.1; BA.2.75; BA.5; BQ.1.1; XBB; XBB.1.5; XBB.1.5.10</th>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.3; SARS; BA.2.12.1; BA.2.75; BA.5; BQ.1.1; XBB; XBB.1.5; XBB.1.5.10</th>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; BF.7; BQ.1.1; XBB; XBB.1.5</th>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.2.12.1; BA.2.75; BA.5; BQ.1.1; XBB; XBB.1.5</th>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; SARS; BA.5; BQ.1.1; XBB; XBB.1.5; XBB.1.5.10</th>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; SARS; BA.5; BQ.1.1; XBB; XBB.1.5; XBB.1.5.10</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.3; SARS; BA.1.1; BA.2.12.1; BA.2.75; BA.5; BQ.1.1; XBB; XBB.1.5; XBB.1.5.10</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.2.75; BA.5; BQ.1.1; XBB</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; BQ.1.1; XBB; XBB.1.5; XBB.1.5.10</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; BF.7; BQ.1.1; XBB; XBB.1.5; XBB.1.5.10</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.2; BA.2.75; BA.5; BQ.1.1; XBB; XBB.1.5</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.5; XBB; XBB.1.5; XBB.1.5.10</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SARS; XBB; XBB.1.5; XBB.1.5.10</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                    n_antibodies\n",
       "viruses_w_IC50s                                                 \n",
       "D614G; BA.1; BA.2; SARS; BA.2.75; BA.5; BF.7; B...           428\n",
       "D614G; BA.1; BA.2; SARS; BA.2.75; BA.5; BF.7; B...           181\n",
       "D614G; BA.1; BA.2; BA.3; SARS; BA.1.1; BA.2.12....           137\n",
       "D614G; BA.1; BA.2; SARS; BA.2.75; BA.5; BQ.1.1;...           130\n",
       "D614G; BA.1; BA.2; BA.2.75; BA.5; BF.7; BQ.1.1;...           129\n",
       "D614G; BA.1; BA.2; SARS; BA.2.75; BA.5; BQ.1.1;...            92\n",
       "D614G; BA.1; BA.2; BA.2.75; BA.5; BQ.1.1; XBB; ...            48\n",
       "D614G; SARS; BA.5; BQ.1.1; XBB; XBB.1.5; XBB.1....            42\n",
       "D614G; BA.1; BA.2; BA.2.75; BA.5; BQ.1.1; XBB; ...            34\n",
       "D614G; BA.1; BA.2; BA.3; SARS; BA.1.1; BA.2.12....            17\n",
       "D614G; BA.1; SARS; BQ.1.1; XBB; XBB.1.5                       16\n",
       "D614G; SARS; BA.5; BQ.1.1; XBB; XBB.1.5                       14\n",
       "D614G; BA.1; BA.2; SARS; BA.2.12.1; BA.2.75; BA...            14\n",
       "D614G; BA.5; BQ.1.1; XBB; XBB.1.5; XBB.1.5.10                 12\n",
       "D614G; BA.1; BA.2; BA.2.75; BA.5; BF.7; BQ.1.1;...            10\n",
       "D614G; BA.1; BA.2; SARS; BA.5; BF.7; BQ.1.1; XB...             7\n",
       "D614G; BA.5; BQ.1.1; XBB; XBB.1.5                              5\n",
       "D614G; BA.1; BA.2; BA.3; SARS; BA.2.12.1; BA.2....             4\n",
       "D614G; BA.1; SARS; BA.5; BQ.1.1; XBB; XBB.1.5                  4\n",
       "D614G; BA.1; BA.2; SARS; BA.2.12.1; BA.2.75; BA...             4\n",
       "D614G; BA.1; BA.2; BA.3; SARS; BA.2.12.1; BA.2....             3\n",
       "D614G; BA.1; BA.2; BA.5; BF.7; BQ.1.1; XBB; XBB...             3\n",
       "D614G; BA.1; BA.2; BA.2.12.1; BA.2.75; BA.5; BQ...             3\n",
       "D614G; BA.1; SARS; BA.5; BQ.1.1; XBB; XBB.1.5; ...             3\n",
       "D614G; BA.1; BA.2; SARS; BA.5; BQ.1.1; XBB; XBB...             2\n",
       "D614G; BA.1; BA.2; BA.3; SARS; BA.1.1; BA.2.12....             2\n",
       "D614G; BA.1; BA.2; BA.2.75; BA.5; BQ.1.1; XBB                  1\n",
       "D614G; BA.1; BA.2; BA.5; BQ.1.1; XBB; XBB.1.5; ...             1\n",
       "D614G; BA.1; BA.2; BA.5; BF.7; BQ.1.1; XBB; XBB...             1\n",
       "D614G; BA.2; BA.2.75; BA.5; BQ.1.1; XBB; XBB.1.5               1\n",
       "D614G; BA.5; XBB; XBB.1.5; XBB.1.5.10                          1\n",
       "SARS; XBB; XBB.1.5; XBB.1.5.10                                 1"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# get the IC50s\n",
    "repeated_2024_ic50s = (\n",
    "    repeated_2024_info\n",
    "    .melt(\n",
    "        id_vars=[\"study\", \"antibody\"],\n",
    "        value_vars=[c for c in repeated_2024_info if c.endswith(\"_IC50\")],\n",
    "        var_name=\"virus\",\n",
    "        value_name=\"IC50\",\n",
    "    )\n",
    "    .query(\"IC50.notnull()\")  # no data, 10 means non-neutralizing\n",
    "    .assign(\n",
    "        virus=lambda x: x[\"virus\"].str.replace(\"_IC50\", \"\").map({\n",
    "            \"D614G\": \"D614G\",\n",
    "            \"BA1\": \"BA.1\",\n",
    "            \"BA2\": \"BA.2\",\n",
    "            \"BA3\": \"BA.3\",\n",
    "            \"SARS\": \"SARS\",\n",
    "            \"BA1_1\": \"BA.1.1\",\n",
    "            \"BA2_12_1\": \"BA.2.12.1\",\n",
    "            \"BA2_75\": \"BA.2.75\",\n",
    "            \"BA5\": \"BA.5\",\n",
    "            \"BF7\": \"BF.7\",\n",
    "            \"BA2_13\": \"BA.2.13\",\n",
    "            \"BQ1_1\": \"BQ.1.1\",\n",
    "            \"XBB\": \"XBB\",\n",
    "            \"XBB1_5\": \"XBB.1.5\",\n",
    "            \"XBB1_5_10\": \"XBB.1.5.10\",\n",
    "        }),\n",
    "    )\n",
    ")\n",
    "\n",
    "print(\"\\nNumber of IC50s per antibody:\")\n",
    "display(\n",
    "    repeated_2024_ic50s\n",
    "    .groupby(\"antibody\", as_index=False)\n",
    "    .aggregate(viruses_w_IC50s=pd.NamedAgg(\"virus\", lambda s: \"; \".join(s)))\n",
    "    .groupby(\"viruses_w_IC50s\")\n",
    "    .aggregate(n_antibodies=pd.NamedAgg(\"antibody\", \"nunique\"))\n",
    "    .sort_values(\"n_antibodies\", ascending=False)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "b042a20b-fce1-4361-a3d7-066b00344b51",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:23:15.966214Z",
     "iopub.status.busy": "2024-11-18T23:23:15.965848Z",
     "iopub.status.idle": "2024-11-18T23:23:16.043388Z",
     "shell.execute_reply": "2024-11-18T23:23:16.042554Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Binding targets per antibody:\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>n_antibodies</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>binds</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Wuhan-Hu-1; Omicron BA.2/BA.5</th>\n",
       "      <td>768</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Omicron BA.2/BA.5</th>\n",
       "      <td>562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Wuhan-Hu-1</th>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                               n_antibodies\n",
       "binds                                      \n",
       "Wuhan-Hu-1; Omicron BA.2/BA.5           768\n",
       "Omicron BA.2/BA.5                       562\n",
       "Wuhan-Hu-1                               19"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# get the binding reactivities\n",
    "repeated_2024_binding = (\n",
    "    repeated_2024_info\n",
    "    .query(\"paper_reactivity != '--'\")\n",
    "    .assign(\n",
    "        binds=lambda x: x[\"paper_reactivity\"].map(\n",
    "            {\n",
    "                \"WT\": [\"Wuhan-Hu-1\"],\n",
    "                \"cross\": [\"Wuhan-Hu-1\", \"Omicron BA.2/BA.5\"],\n",
    "                \"specific\": [\"Omicron BA.2/BA.5\"],\n",
    "            }\n",
    "        )\n",
    "    )\n",
    "    [[\"study\", \"antibody\", \"binds\"]]\n",
    "    .explode(\"binds\")\n",
    ")\n",
    "\n",
    "print(\"\\nBinding targets per antibody:\")\n",
    "display(\n",
    "    repeated_2024_binding\n",
    "    .groupby(\"antibody\", as_index=False)\n",
    "    .aggregate(binds=pd.NamedAgg(\"binds\", lambda s: \"; \".join(s)))\n",
    "    .groupby(\"binds\")\n",
    "    .aggregate(n_antibodies=pd.NamedAgg(\"antibody\", \"nunique\"))\n",
    "    .sort_values(\"n_antibodies\", ascending=False)\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a7965ea9-2c45-48fb-abe3-aa0f43910899",
   "metadata": {},
   "source": [
    "Now read the actual escape data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "39e7ab8d-13ab-4a04-a35d-1caed836161b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:23:16.046446Z",
     "iopub.status.busy": "2024-11-18T23:23:16.046000Z",
     "iopub.status.idle": "2024-11-18T23:23:16.506775Z",
     "shell.execute_reply": "2024-11-18T23:23:16.505992Z"
    },
    "tags": []
   },
   "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>antibody</th>\n",
       "      <th>site</th>\n",
       "      <th>escape</th>\n",
       "      <th>study</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>45314</th>\n",
       "      <td>GC1852</td>\n",
       "      <td>390</td>\n",
       "      <td>5.628970e-08</td>\n",
       "      <td>repeated_2024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45311</th>\n",
       "      <td>GC1852</td>\n",
       "      <td>371</td>\n",
       "      <td>6.030392e-08</td>\n",
       "      <td>repeated_2024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45312</th>\n",
       "      <td>GC1852</td>\n",
       "      <td>373</td>\n",
       "      <td>6.448414e-08</td>\n",
       "      <td>repeated_2024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45323</th>\n",
       "      <td>GC1852</td>\n",
       "      <td>514</td>\n",
       "      <td>7.860392e-08</td>\n",
       "      <td>repeated_2024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45313</th>\n",
       "      <td>GC1852</td>\n",
       "      <td>377</td>\n",
       "      <td>9.080642e-08</td>\n",
       "      <td>repeated_2024</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>7512</th>\n",
       "      <td>GC0281</td>\n",
       "      <td>444</td>\n",
       "      <td>1.611308e+01</td>\n",
       "      <td>repeated_2024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4258</th>\n",
       "      <td>GC0177</td>\n",
       "      <td>346</td>\n",
       "      <td>1.637373e+01</td>\n",
       "      <td>repeated_2024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19760</th>\n",
       "      <td>GC0832</td>\n",
       "      <td>444</td>\n",
       "      <td>1.644617e+01</td>\n",
       "      <td>repeated_2024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4775</th>\n",
       "      <td>GC0191</td>\n",
       "      <td>444</td>\n",
       "      <td>1.648478e+01</td>\n",
       "      <td>repeated_2024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4183</th>\n",
       "      <td>GC0174</td>\n",
       "      <td>444</td>\n",
       "      <td>1.669166e+01</td>\n",
       "      <td>repeated_2024</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>54344 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      antibody  site        escape          study\n",
       "45314   GC1852   390  5.628970e-08  repeated_2024\n",
       "45311   GC1852   371  6.030392e-08  repeated_2024\n",
       "45312   GC1852   373  6.448414e-08  repeated_2024\n",
       "45323   GC1852   514  7.860392e-08  repeated_2024\n",
       "45313   GC1852   377  9.080642e-08  repeated_2024\n",
       "...        ...   ...           ...            ...\n",
       "7512    GC0281   444  1.611308e+01  repeated_2024\n",
       "4258    GC0177   346  1.637373e+01  repeated_2024\n",
       "19760   GC0832   444  1.644617e+01  repeated_2024\n",
       "4775    GC0191   444  1.648478e+01  repeated_2024\n",
       "4183    GC0174   444  1.669166e+01  repeated_2024\n",
       "\n",
       "[54344 rows x 4 columns]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "repeated_2024_escape = (\n",
    "    pd.read_csv(\"Cao_data/SARS-CoV-2-reinfection-DMS/antibody_dms_merge_clean.csv.gz\")\n",
    "    .groupby([\"antibody\", \"site\"], as_index=False)\n",
    "    .aggregate(escape=pd.NamedAgg(\"mut_escape\", \"sum\"))\n",
    "    .assign(study=\"repeated_2024\")\n",
    ")\n",
    "\n",
    "repeated_2024_escape.sort_values(\"escape\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "709f94b3-33a1-44b8-88a1-fe3dbcc15200",
   "metadata": {},
   "source": [
    "Now get the antibody re-weightings.\n",
    "These are based on the re-weightings described in [Repeated Omicron infection alleviates SARS-CoV-2 immune imprinting (2023)](https://www.biorxiv.org/content/10.1101/2023.05.01.538516v2).\n",
    "We want to re-weight the cross-reactive antibodies by an amount $w$ so that the desired cross reactive fraction $x$ satisfies $x = \\frac{w N_c}{w N_c + N_s}$ where $w$ is the re-weighting assigned to the cross-reactive antibodies, $N_c$ is the number of cross-reactive antibodies, and $N_s$ is the number of specific antibodies.\n",
    "This can be achieved by assigning $w = \\frac{x N_s}{N_c + x N_c}$; we then adjust the weights again so the mean weight over all antibodies for the source equals the number of antibodies."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "7251b46e-3ec0-40f2-a9f2-ab2fd1d77c80",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:23:16.510692Z",
     "iopub.status.busy": "2024-11-18T23:23:16.510239Z",
     "iopub.status.idle": "2024-11-18T23:23:16.597979Z",
     "shell.execute_reply": "2024-11-18T23:23:16.597199Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using the following reweightings:\n",
      "                           source  desired_cross_reactive_frac\n",
      "0              BA.5 convalescents                         0.89\n",
      "1              BF.7 convalescents                         0.89\n",
      "2  BA.1 convalescents reinfection                         0.51\n",
      "3  BA.2 convalescents reinfection                         0.51\n",
      "\n",
      "Confirm re-weighting worked:\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>source</th>\n",
       "      <th>desired_cross_reactive_frac</th>\n",
       "      <th>cross_reactive_frac</th>\n",
       "      <th>cross_reactive_reweight_frac</th>\n",
       "      <th>n_antibodies</th>\n",
       "      <th>summed_reweights</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>BA.1 convalescents reinfection</td>\n",
       "      <td>0.51</td>\n",
       "      <td>0.345070</td>\n",
       "      <td>0.51</td>\n",
       "      <td>284</td>\n",
       "      <td>284.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>BA.2 convalescents reinfection</td>\n",
       "      <td>0.51</td>\n",
       "      <td>0.439655</td>\n",
       "      <td>0.51</td>\n",
       "      <td>232</td>\n",
       "      <td>232.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>BA.5 convalescents</td>\n",
       "      <td>0.89</td>\n",
       "      <td>0.713964</td>\n",
       "      <td>0.89</td>\n",
       "      <td>444</td>\n",
       "      <td>444.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>BF.7 convalescents</td>\n",
       "      <td>0.89</td>\n",
       "      <td>0.728395</td>\n",
       "      <td>0.89</td>\n",
       "      <td>243</td>\n",
       "      <td>243.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           source  desired_cross_reactive_frac  \\\n",
       "0  BA.1 convalescents reinfection                         0.51   \n",
       "1  BA.2 convalescents reinfection                         0.51   \n",
       "2              BA.5 convalescents                         0.89   \n",
       "3              BF.7 convalescents                         0.89   \n",
       "\n",
       "   cross_reactive_frac  cross_reactive_reweight_frac  n_antibodies  \\\n",
       "0             0.345070                          0.51           284   \n",
       "1             0.439655                          0.51           232   \n",
       "2             0.713964                          0.89           444   \n",
       "3             0.728395                          0.89           243   \n",
       "\n",
       "   summed_reweights  \n",
       "0             284.0  \n",
       "1             232.0  \n",
       "2             444.0  \n",
       "3             243.0  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Here are the re-weightings:\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>study</th>\n",
       "      <th>source</th>\n",
       "      <th>cross_reactive</th>\n",
       "      <th>reweight</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>BA.1 convalescents reinfection</td>\n",
       "      <td>False</td>\n",
       "      <td>0.748172</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>BA.1 convalescents reinfection</td>\n",
       "      <td>True</td>\n",
       "      <td>1.477959</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>BA.2 convalescents reinfection</td>\n",
       "      <td>False</td>\n",
       "      <td>0.874462</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>BA.2 convalescents reinfection</td>\n",
       "      <td>True</td>\n",
       "      <td>1.160000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>BA.5 convalescents</td>\n",
       "      <td>False</td>\n",
       "      <td>0.384567</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>BA.5 convalescents</td>\n",
       "      <td>True</td>\n",
       "      <td>1.246562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>BF.7 convalescents</td>\n",
       "      <td>False</td>\n",
       "      <td>0.405000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>BF.7 convalescents</td>\n",
       "      <td>True</td>\n",
       "      <td>1.221864</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           study                          source  cross_reactive  reweight\n",
       "0  repeated_2024  BA.1 convalescents reinfection           False  0.748172\n",
       "1  repeated_2024  BA.1 convalescents reinfection            True  1.477959\n",
       "2  repeated_2024  BA.2 convalescents reinfection           False  0.874462\n",
       "3  repeated_2024  BA.2 convalescents reinfection            True  1.160000\n",
       "4  repeated_2024              BA.5 convalescents           False  0.384567\n",
       "5  repeated_2024              BA.5 convalescents            True  1.246562\n",
       "6  repeated_2024              BF.7 convalescents           False  0.405000\n",
       "7  repeated_2024              BF.7 convalescents            True  1.221864"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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>antibody</th>\n",
       "      <th>study</th>\n",
       "      <th>reweight</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CP0590</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.246562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>GC0101</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.246562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>GC0102</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.246562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>GC0103</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.246562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>GC0104</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.246562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1198</th>\n",
       "      <td>GC2054</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.477959</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199</th>\n",
       "      <td>GC2056</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.477959</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1200</th>\n",
       "      <td>GC2057</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.477959</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1201</th>\n",
       "      <td>GC2058</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>0.874462</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1202</th>\n",
       "      <td>GC2059</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.477959</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1203 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     antibody          study  reweight\n",
       "0      CP0590  repeated_2024  1.246562\n",
       "1      GC0101  repeated_2024  1.246562\n",
       "2      GC0102  repeated_2024  1.246562\n",
       "3      GC0103  repeated_2024  1.246562\n",
       "4      GC0104  repeated_2024  1.246562\n",
       "...       ...            ...       ...\n",
       "1198   GC2054  repeated_2024  1.477959\n",
       "1199   GC2056  repeated_2024  1.477959\n",
       "1200   GC2057  repeated_2024  1.477959\n",
       "1201   GC2058  repeated_2024  0.874462\n",
       "1202   GC2059  repeated_2024  1.477959\n",
       "\n",
       "[1203 rows x 3 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "repeated_2024_desired_reweighting = (\n",
    "    pd.Series(antibody_reweighting[\"repeated_2024\"][\"cross_reactive_weights\"])\n",
    "    .rename_axis(\"source\")\n",
    "    .rename(\"desired_cross_reactive_frac\")\n",
    "    .reset_index()\n",
    ")\n",
    "assert set(repeated_2024_desired_reweighting[\"source\"]).issubset(repeated_2024_source[\"source\"])\n",
    "print(f\"Using the following reweightings:\\n{repeated_2024_desired_reweighting}\")\n",
    "\n",
    "repeated_2024_reweighting = (\n",
    "    repeated_2024_binding\n",
    "    .groupby(\"antibody\", as_index=False)\n",
    "    .aggregate(n_binds=pd.NamedAgg(\"binds\", \"nunique\"))\n",
    "    .assign(cross_reactive=lambda x: x[\"n_binds\"] == 2)\n",
    "    .merge(repeated_2024_source, on=\"antibody\", validate=\"one_to_one\")\n",
    "    .merge(repeated_2024_desired_reweighting, on=\"source\", validate=\"many_to_one\")\n",
    "    .assign(\n",
    "        N_c=lambda x: x.groupby(\"source\")[\"cross_reactive\"].transform(\"sum\"),\n",
    "        N=lambda x: x.groupby(\"source\")[\"antibody\"].transform(\"count\"),\n",
    "        N_s=lambda x: x[\"N\"] - x[\"N_c\"],\n",
    "        cross_reactive_frac=lambda x: x[\"N_c\"] / x[\"N\"],\n",
    "        reweight=lambda x: numpy.where(\n",
    "            x[\"cross_reactive\"],\n",
    "            (\n",
    "                x[\"desired_cross_reactive_frac\"] * x[\"N_s\"]\n",
    "                / (x[\"N_c\"] * (1 - x[\"desired_cross_reactive_frac\"]))\n",
    "            ),\n",
    "            1,\n",
    "        ),\n",
    "    )\n",
    "    .assign(\n",
    "        reweight=lambda x: (\n",
    "            x[\"reweight\"] / x.groupby(\"source\")[\"reweight\"].transform(\"mean\")\n",
    "        ),\n",
    "    )\n",
    ")\n",
    "\n",
    "print(\"\\nConfirm re-weighting worked:\")\n",
    "display(\n",
    "    repeated_2024_reweighting\n",
    "    .assign(cross_reactive_reweight=lambda x: x[\"cross_reactive\"] * x[\"reweight\"])\n",
    "    .groupby([\"source\", \"desired_cross_reactive_frac\", \"cross_reactive_frac\"], as_index=False)\n",
    "    .aggregate(\n",
    "        reweighted_cross_reactive_sum=pd.NamedAgg(\"cross_reactive_reweight\", \"sum\"),\n",
    "        reweighted_sum=pd.NamedAgg(\"reweight\", \"sum\"),\n",
    "    )\n",
    "    .assign(\n",
    "        cross_reactive_reweight_frac=lambda x: x[\"reweighted_cross_reactive_sum\"] / x[\"reweighted_sum\"],\n",
    "    )\n",
    "    .drop(columns=[\"reweighted_cross_reactive_sum\", \"reweighted_sum\"])\n",
    "    .merge(\n",
    "        repeated_2024_reweighting\n",
    "        .groupby(\"source\", as_index=False)\n",
    "        .aggregate(\n",
    "            n_antibodies=pd.NamedAgg(\"antibody\", \"nunique\"),\n",
    "            summed_reweights=pd.NamedAgg(\"reweight\", \"sum\"),\n",
    "        )\n",
    "    )\n",
    ")\n",
    "\n",
    "print(\"\\nHere are the re-weightings:\")\n",
    "display(\n",
    "    repeated_2024_reweighting\n",
    "    [[\"study\", \"source\", \"cross_reactive\", \"reweight\"]]\n",
    "    .drop_duplicates()\n",
    "    .sort_values([\"source\", \"cross_reactive\"])\n",
    "    .reset_index(drop=True)\n",
    ")\n",
    "\n",
    "# get the reweightings\n",
    "repeated_2024_reweighting = (\n",
    "    repeated_2024_reweighting\n",
    "    [[\"antibody\", \"study\", \"reweight\"]]\n",
    ")\n",
    "\n",
    "repeated_2024_reweighting"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c0d28d49-bd00-44bb-bd42-4dc6855d001c",
   "metadata": {},
   "source": [
    "## Process data from [Imprinted SARS-CoV-2 humoral immunity induces convergent Omicron RBD evolution (2022)](https://www.nature.com/articles/s41586-022-05644-7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "b249ed38-a3d0-451d-9380-2dcfae9c405b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:23:16.601325Z",
     "iopub.status.busy": "2024-11-18T23:23:16.600958Z",
     "iopub.status.idle": "2024-11-18T23:23:18.510207Z",
     "shell.execute_reply": "2024-11-18T23:23:18.509450Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# read the data, which is in multiple sheets\n",
    "imprinted_2022_raw_info = pd.read_excel(\n",
    "    \"Cao_data/convergent_RBD_evolution/antibody_info.xlsx\",\n",
    "    header=1,\n",
    "    sheet_name=None,\n",
    "    usecols=lambda c: not c.startswith(\"Unnamed\"),\n",
    ")\n",
    "\n",
    "common_cols = None\n",
    "\n",
    "for sheetname, sheet in imprinted_2022_raw_info.items():\n",
    "    columns = sheet.columns.tolist()\n",
    "    if common_cols is None:\n",
    "        common_cols = columns\n",
    "    else:\n",
    "        common_cols = [\n",
    "            c for c in columns if c in set(common_cols).intersection(columns)\n",
    "        ]\n",
    "\n",
    "imprinted_2022_info = pd.concat(\n",
    "    [\n",
    "        sheet[common_cols].assign(sheet_name=sheet_name)\n",
    "        for sheet_name, sheet in imprinted_2022_raw_info.items()\n",
    "    ],\n",
    "    ignore_index=True,\n",
    ").rename(columns={\"Antibody  Name\": \"antibody\"}).assign(study=\"imprinted_2022\")\n",
    "\n",
    "assert len(imprinted_2022_info) == imprinted_2022_info[\"antibody\"].nunique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "706122b4-74af-4cba-9d7a-d256db440fea",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:23:18.514366Z",
     "iopub.status.busy": "2024-11-18T23:23:18.514020Z",
     "iopub.status.idle": "2024-11-18T23:23:18.536101Z",
     "shell.execute_reply": "2024-11-18T23:23:18.535330Z"
    },
    "tags": []
   },
   "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></th>\n",
       "      <th>n_antibodies</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>study</th>\n",
       "      <th>source</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"6\" valign=\"top\">imprinted_2022</th>\n",
       "      <th>BA.1 convalescents</th>\n",
       "      <td>752</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.2 convalescents</th>\n",
       "      <td>751</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.5 convalescents</th>\n",
       "      <td>269</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SARS convalescents</th>\n",
       "      <td>881</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WT convalescents</th>\n",
       "      <td>438</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WT vaccinees</th>\n",
       "      <td>240</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                   n_antibodies\n",
       "study          source                          \n",
       "imprinted_2022 BA.1 convalescents           752\n",
       "               BA.2 convalescents           751\n",
       "               BA.5 convalescents           269\n",
       "               SARS convalescents           881\n",
       "               WT convalescents             438\n",
       "               WT vaccinees                 240"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# get the antibody sources\n",
    "imprinted_2022_source = (\n",
    "    imprinted_2022_info\n",
    "    [[\"antibody\", \"source\", \"study\"]]\n",
    "    .query(\"source != 'WT-engineered'\")\n",
    ")\n",
    "\n",
    "display(\n",
    "    imprinted_2022_source\n",
    "    .groupby([\"study\", \"source\"])\n",
    "    .aggregate(n_antibodies=pd.NamedAgg(\"antibody\", \"nunique\"))\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "6e7dc881-b3ea-4af6-9e09-95e99b898440",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:23:18.540098Z",
     "iopub.status.busy": "2024-11-18T23:23:18.539687Z",
     "iopub.status.idle": "2024-11-18T23:23:18.672872Z",
     "shell.execute_reply": "2024-11-18T23:23:18.672140Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Number of IC50s per antibody:\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>n_antibodies</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>viruses_w_IC50s</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; BA.2.75; BQ.1.1; XBB</th>\n",
       "      <td>3309</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; BA.2.75</th>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; BA.2.75; BQ.1.1</th>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BA.1; BA.2; BA.5; BA.2.75; XBB</th>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BQ.1.1; XBB</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.1; BQ.1.1; XBB</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.1; BA.2; BA.5; BA.2.75; BQ.1.1; XBB</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>D614G; BQ.1.1; XBB</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                               n_antibodies\n",
       "viruses_w_IC50s                                            \n",
       "D614G; BA.1; BA.2; BA.5; BA.2.75; BQ.1.1; XBB          3309\n",
       "D614G; BA.1; BA.2; BA.5; BA.2.75                         13\n",
       "D614G; BA.1; BA.2; BA.5; BA.2.75; BQ.1.1                  3\n",
       "D614G; BA.1; BA.2; BA.5; BA.2.75; XBB                     3\n",
       "BQ.1.1; XBB                                               2\n",
       "BA.1; BQ.1.1; XBB                                         1\n",
       "BA.1; BA.2; BA.5; BA.2.75; BQ.1.1; XBB                    1\n",
       "D614G; BQ.1.1; XBB                                        1"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# get the IC50s\n",
    "imprinted_2022_ic50s = (\n",
    "    imprinted_2022_info\n",
    "    .melt(\n",
    "        id_vars=[\"study\", \"antibody\"],\n",
    "        value_vars=[\"D614G\", \"BA.1\", \"BA.2\", \"BA.5\", \"BA.2.75\", \"BQ.1.1\", \"XBB\"],\n",
    "        var_name=\"virus\",\n",
    "        value_name=\"IC50\",\n",
    "    )\n",
    "    # values of >10 to 10\n",
    "    .query(\"IC50 != '--'\")\n",
    "    .assign(IC50=lambda x: x[\"IC50\"].map(lambda ic50: 10 if ic50 == \">10\" else float(ic50)))\n",
    ")\n",
    "\n",
    "print(\"\\nNumber of IC50s per antibody:\")\n",
    "display(\n",
    "    imprinted_2022_ic50s\n",
    "    .groupby(\"antibody\", as_index=False)\n",
    "    .aggregate(viruses_w_IC50s=pd.NamedAgg(\"virus\", lambda s: \"; \".join(s)))\n",
    "    .groupby(\"viruses_w_IC50s\")\n",
    "    .aggregate(n_antibodies=pd.NamedAgg(\"antibody\", \"nunique\"))\n",
    "    .sort_values(\"n_antibodies\", ascending=False)\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e1600244-34ad-4013-ba98-cede879266e2",
   "metadata": {},
   "source": [
    "Now read the actual escape data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "8ddc06a9-bf6a-43ff-8c70-b3ffc38d3364",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:23:18.677673Z",
     "iopub.status.busy": "2024-11-18T23:23:18.677264Z",
     "iopub.status.idle": "2024-11-18T23:23:19.084571Z",
     "shell.execute_reply": "2024-11-18T23:23:19.083703Z"
    },
    "tags": []
   },
   "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>antibody</th>\n",
       "      <th>site</th>\n",
       "      <th>escape</th>\n",
       "      <th>study</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1-57</td>\n",
       "      <td>338</td>\n",
       "      <td>0.521154</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1-57</td>\n",
       "      <td>359</td>\n",
       "      <td>0.140217</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1-57</td>\n",
       "      <td>370</td>\n",
       "      <td>0.285110</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1-57</td>\n",
       "      <td>394</td>\n",
       "      <td>0.112732</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1-57</td>\n",
       "      <td>396</td>\n",
       "      <td>0.194379</td>\n",
       "      <td>imprinted_2022</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>63387</th>\n",
       "      <td>XGv-422</td>\n",
       "      <td>465</td>\n",
       "      <td>0.769640</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63388</th>\n",
       "      <td>XGv-422</td>\n",
       "      <td>466</td>\n",
       "      <td>0.886558</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63389</th>\n",
       "      <td>XGv-422</td>\n",
       "      <td>468</td>\n",
       "      <td>0.313522</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63390</th>\n",
       "      <td>XGv-422</td>\n",
       "      <td>484</td>\n",
       "      <td>0.165163</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63391</th>\n",
       "      <td>XGv-422</td>\n",
       "      <td>514</td>\n",
       "      <td>0.224545</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>63392 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      antibody  site    escape           study\n",
       "0         1-57   338  0.521154  imprinted_2022\n",
       "1         1-57   359  0.140217  imprinted_2022\n",
       "2         1-57   370  0.285110  imprinted_2022\n",
       "3         1-57   394  0.112732  imprinted_2022\n",
       "4         1-57   396  0.194379  imprinted_2022\n",
       "...        ...   ...       ...             ...\n",
       "63387  XGv-422   465  0.769640  imprinted_2022\n",
       "63388  XGv-422   466  0.886558  imprinted_2022\n",
       "63389  XGv-422   468  0.313522  imprinted_2022\n",
       "63390  XGv-422   484  0.165163  imprinted_2022\n",
       "63391  XGv-422   514  0.224545  imprinted_2022\n",
       "\n",
       "[63392 rows x 4 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "imprinted_2022_escape = (\n",
    "    pd.read_csv(\"Cao_data/convergent_RBD_evolution/use_res_clean.csv\")\n",
    "    .groupby([\"antibody\", \"site\"], as_index=False)\n",
    "    .aggregate(escape=pd.NamedAgg(\"mut_escape\", \"sum\"))\n",
    "    .assign(study=\"imprinted_2022\")\n",
    ")\n",
    "\n",
    "imprinted_2022_escape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ef6347f8-ad27-41a3-a884-6991f91cc450",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Aggregate the antibody information from the different studies\n",
    "\n",
    "First, read the information on duplicated antibodies to get duplicate names:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "08c9c6a6-2de8-4882-92c2-2738edffe090",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:23:19.089111Z",
     "iopub.status.busy": "2024-11-18T23:23:19.088641Z",
     "iopub.status.idle": "2024-11-18T23:23:19.432421Z",
     "shell.execute_reply": "2024-11-18T23:23:19.431650Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "antibody_renaming = (\n",
    "    pd.read_excel(\"Cao_data/all-repo-duplicates.xlsx\")\n",
    "    .set_index(\"Antibody  Name\")\n",
    "    [\"BD equivalent name\"]\n",
    "    .to_dict()\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "40434c66-7697-48ca-942a-6ba0fd4cca9a",
   "metadata": {},
   "source": [
    "Now aggregate the antibody sources:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "4895396a-b117-4463-8fed-18e512d9e180",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:23:19.437842Z",
     "iopub.status.busy": "2024-11-18T23:23:19.437021Z",
     "iopub.status.idle": "2024-11-18T23:23:19.513902Z",
     "shell.execute_reply": "2024-11-18T23:23:19.512589Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dropping the following antibodies duplicated across studies:\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>antibody</th>\n",
       "      <th>source</th>\n",
       "      <th>study</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>BD45-105</td>\n",
       "      <td>WT convalescents</td>\n",
       "      <td>repeated_2024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>BD55-1205</td>\n",
       "      <td>WT convalescents</td>\n",
       "      <td>repeated_2024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>BD55-3372</td>\n",
       "      <td>SARS convalescents</td>\n",
       "      <td>repeated_2024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>BD55-4637</td>\n",
       "      <td>SARS convalescents</td>\n",
       "      <td>repeated_2024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>BD55-5483</td>\n",
       "      <td>SARS convalescents</td>\n",
       "      <td>repeated_2024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>389</th>\n",
       "      <td>BD57-0300</td>\n",
       "      <td>BA.5 convalescents</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>390</th>\n",
       "      <td>BD57-0308</td>\n",
       "      <td>BA.5 convalescents</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>391</th>\n",
       "      <td>BD57-0318</td>\n",
       "      <td>BA.5 convalescents</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>392</th>\n",
       "      <td>BD57-0332</td>\n",
       "      <td>BA.5 convalescents</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>393</th>\n",
       "      <td>BD57-0398</td>\n",
       "      <td>BA.5 convalescents</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>394 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      antibody              source           study\n",
       "0     BD45-105    WT convalescents   repeated_2024\n",
       "1    BD55-1205    WT convalescents   repeated_2024\n",
       "2    BD55-3372  SARS convalescents   repeated_2024\n",
       "3    BD55-4637  SARS convalescents   repeated_2024\n",
       "4    BD55-5483  SARS convalescents   repeated_2024\n",
       "..         ...                 ...             ...\n",
       "389  BD57-0300  BA.5 convalescents  imprinted_2022\n",
       "390  BD57-0308  BA.5 convalescents  imprinted_2022\n",
       "391  BD57-0318  BA.5 convalescents  imprinted_2022\n",
       "392  BD57-0332  BA.5 convalescents  imprinted_2022\n",
       "393  BD57-0398  BA.5 convalescents  imprinted_2022\n",
       "\n",
       "[394 rows x 3 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Retained sources:\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></th>\n",
       "      <th>n_antibodies</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>source</th>\n",
       "      <th>study</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>BA.1 + BA.5/BF.7 infection</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>97</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.1 BTI</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>52</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.1 BTI + BA.5/BF.7 infection</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>162</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">BA.1 convalescents</th>\n",
       "      <th>imprinted_2022</th>\n",
       "      <td>735</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>repeated_2024</th>\n",
       "      <td>36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.1 convalescents reinfection</th>\n",
       "      <th>repeated_2024</th>\n",
       "      <td>284</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.2 + BA.5/BF.7 infection</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>109</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.2 BTI</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>57</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.2 BTI + BA.5/BF.7 infection</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>135</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">BA.2 convalescents</th>\n",
       "      <th>imprinted_2022</th>\n",
       "      <td>685</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>repeated_2024</th>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.2 convalescents reinfection</th>\n",
       "      <th>repeated_2024</th>\n",
       "      <td>232</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.5 + JN.1 infection</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.5 + XBB infection</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>314</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.5 BTI</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>78</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.5 BTI + HK.3 infection</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.5 BTI + JN.1 infection</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BA.5 BTI + XBB infection</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>383</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">BA.5 convalescents</th>\n",
       "      <th>imprinted_2022</th>\n",
       "      <td>42</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>repeated_2024</th>\n",
       "      <td>428</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BF.7 BTI</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>53</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BF.7 convalescents</th>\n",
       "      <th>repeated_2024</th>\n",
       "      <td>243</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>N.D.</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SARS</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">SARS convalescents</th>\n",
       "      <th>imprinted_2022</th>\n",
       "      <td>872</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>repeated_2024</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SARS exposure</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">WT convalescents</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>imprinted_2022</th>\n",
       "      <td>418</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>repeated_2024</th>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WT vaccinees</th>\n",
       "      <th>imprinted_2022</th>\n",
       "      <td>240</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WT-phage display</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XBB BTI</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>303</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>XBB infection</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>301</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>long-term BA.1 BTI</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>long-term BA.1 convalescents</th>\n",
       "      <th>repeated_2024</th>\n",
       "      <td>38</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>long-term BA.5 BTI</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>long-term BA.5 infection</th>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                               n_antibodies\n",
       "source                         study                       \n",
       "BA.1 + BA.5/BF.7 infection     evolving_2024             97\n",
       "BA.1 BTI                       evolving_2024             52\n",
       "BA.1 BTI + BA.5/BF.7 infection evolving_2024            162\n",
       "BA.1 convalescents             imprinted_2022           735\n",
       "                               repeated_2024             36\n",
       "BA.1 convalescents reinfection repeated_2024            284\n",
       "BA.2 + BA.5/BF.7 infection     evolving_2024            109\n",
       "BA.2 BTI                       evolving_2024             57\n",
       "BA.2 BTI + BA.5/BF.7 infection evolving_2024            135\n",
       "BA.2 convalescents             imprinted_2022           685\n",
       "                               repeated_2024             21\n",
       "BA.2 convalescents reinfection repeated_2024            232\n",
       "BA.5 + JN.1 infection          evolving_2024            200\n",
       "BA.5 + XBB infection           evolving_2024            314\n",
       "BA.5 BTI                       evolving_2024             78\n",
       "BA.5 BTI + HK.3 infection      evolving_2024            300\n",
       "BA.5 BTI + JN.1 infection      evolving_2024            300\n",
       "BA.5 BTI + XBB infection       evolving_2024            383\n",
       "BA.5 convalescents             imprinted_2022            42\n",
       "                               repeated_2024            428\n",
       "BF.7 BTI                       evolving_2024             53\n",
       "BF.7 convalescents             repeated_2024            243\n",
       "N.D.                           evolving_2024              2\n",
       "SARS                           evolving_2024              1\n",
       "SARS convalescents             imprinted_2022           872\n",
       "                               repeated_2024              1\n",
       "SARS exposure                  evolving_2024              7\n",
       "WT convalescents               evolving_2024             11\n",
       "                               imprinted_2022           418\n",
       "                               repeated_2024             10\n",
       "WT vaccinees                   imprinted_2022           240\n",
       "WT-phage display               evolving_2024              1\n",
       "XBB BTI                        evolving_2024            303\n",
       "XBB infection                  evolving_2024            301\n",
       "long-term BA.1 BTI             evolving_2024             16\n",
       "long-term BA.1 convalescents   repeated_2024             38\n",
       "long-term BA.5 BTI             evolving_2024             26\n",
       "long-term BA.5 infection       evolving_2024              7"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "antibody_source = (\n",
    "    pd.concat(\n",
    "        [evolving_2024_source, repeated_2024_source, imprinted_2022_source],\n",
    "        ignore_index=True,\n",
    "    )\n",
    "    .assign(\n",
    "        antibody=lambda x: x[\"antibody\"].map(\n",
    "            lambda a: antibody_renaming[a] if a in antibody_renaming else a\n",
    "        ),\n",
    "    )\n",
    ")\n",
    "\n",
    "print(\"Dropping the following antibodies duplicated across studies:\")\n",
    "display(antibody_source.groupby(\"antibody\").tail(-1).reset_index(drop=True))\n",
    "# keep just one of antibodies repeated across studies\n",
    "antibody_source = antibody_source.groupby(\"antibody\").first().reset_index()\n",
    "\n",
    "assert (\n",
    "    antibody_source[\"antibody\"].nunique()\n",
    "    == len(antibody_source[[\"antibody\", \"source\"]].drop_duplicates())\n",
    ")\n",
    "\n",
    "print(\"\\nRetained sources:\")\n",
    "display(\n",
    "    antibody_source\n",
    "    .groupby([\"source\", \"study\"])\n",
    "    .aggregate(n_antibodies=pd.NamedAgg(\"antibody\", \"count\"))\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bcfea729-282d-47e1-a20f-dfba1aa4444f",
   "metadata": {},
   "source": [
    "Now the IC50s, clipping to be between 1e-5 and 10:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "d8f51f4d-8fd6-49a9-acd8-832dbcf83334",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:23:19.517871Z",
     "iopub.status.busy": "2024-11-18T23:23:19.517321Z",
     "iopub.status.idle": "2024-11-18T23:23:19.752662Z",
     "shell.execute_reply": "2024-11-18T23:23:19.751781Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dropping the following IC50s duplicated across studies:\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>study</th>\n",
       "      <th>antibody</th>\n",
       "      <th>virus</th>\n",
       "      <th>IC50</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>BD45-105</td>\n",
       "      <td>D614G</td>\n",
       "      <td>0.57400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>BD55-1205</td>\n",
       "      <td>D614G</td>\n",
       "      <td>0.00331</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>BD55-3372</td>\n",
       "      <td>D614G</td>\n",
       "      <td>0.00680</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>BD55-4637</td>\n",
       "      <td>D614G</td>\n",
       "      <td>0.02010</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>BD55-5483</td>\n",
       "      <td>D614G</td>\n",
       "      <td>0.01380</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>2528</th>\n",
       "      <td>imprinted_2022</td>\n",
       "      <td>BD57-0300</td>\n",
       "      <td>XBB</td>\n",
       "      <td>10.00000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2529</th>\n",
       "      <td>imprinted_2022</td>\n",
       "      <td>BD57-0308</td>\n",
       "      <td>XBB</td>\n",
       "      <td>10.00000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2530</th>\n",
       "      <td>imprinted_2022</td>\n",
       "      <td>BD57-0318</td>\n",
       "      <td>XBB</td>\n",
       "      <td>10.00000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2531</th>\n",
       "      <td>imprinted_2022</td>\n",
       "      <td>BD57-0332</td>\n",
       "      <td>XBB</td>\n",
       "      <td>10.00000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2532</th>\n",
       "      <td>imprinted_2022</td>\n",
       "      <td>BD57-0398</td>\n",
       "      <td>XBB</td>\n",
       "      <td>10.00000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2533 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "               study   antibody  virus      IC50\n",
       "0      repeated_2024   BD45-105  D614G   0.57400\n",
       "1      repeated_2024  BD55-1205  D614G   0.00331\n",
       "2      repeated_2024  BD55-3372  D614G   0.00680\n",
       "3      repeated_2024  BD55-4637  D614G   0.02010\n",
       "4      repeated_2024  BD55-5483  D614G   0.01380\n",
       "...              ...        ...    ...       ...\n",
       "2528  imprinted_2022  BD57-0300    XBB  10.00000\n",
       "2529  imprinted_2022  BD57-0308    XBB  10.00000\n",
       "2530  imprinted_2022  BD57-0318    XBB  10.00000\n",
       "2531  imprinted_2022  BD57-0332    XBB  10.00000\n",
       "2532  imprinted_2022  BD57-0398    XBB  10.00000\n",
       "\n",
       "[2533 rows x 4 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Only keep virus with IC50s against at least 1000 viruses\n",
      "Keeping 17 of 23 viruses\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Retained IC50s:\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>antibody</th>\n",
       "      <th>virus</th>\n",
       "      <th>study</th>\n",
       "      <th>IC50</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1-57</td>\n",
       "      <td>BA.1</td>\n",
       "      <td>imprinted_2022</td>\n",
       "      <td>10.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1-57</td>\n",
       "      <td>BA.2</td>\n",
       "      <td>imprinted_2022</td>\n",
       "      <td>10.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1-57</td>\n",
       "      <td>BA.2.75</td>\n",
       "      <td>imprinted_2022</td>\n",
       "      <td>10.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1-57</td>\n",
       "      <td>BA.5</td>\n",
       "      <td>imprinted_2022</td>\n",
       "      <td>10.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1-57</td>\n",
       "      <td>BQ.1.1</td>\n",
       "      <td>imprinted_2022</td>\n",
       "      <td>10.000</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>69038</th>\n",
       "      <td>XGv-422</td>\n",
       "      <td>BA.2.75</td>\n",
       "      <td>imprinted_2022</td>\n",
       "      <td>0.495</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>69039</th>\n",
       "      <td>XGv-422</td>\n",
       "      <td>BA.5</td>\n",
       "      <td>imprinted_2022</td>\n",
       "      <td>0.681</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>69040</th>\n",
       "      <td>XGv-422</td>\n",
       "      <td>BQ.1.1</td>\n",
       "      <td>imprinted_2022</td>\n",
       "      <td>1.060</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>69041</th>\n",
       "      <td>XGv-422</td>\n",
       "      <td>D614G</td>\n",
       "      <td>imprinted_2022</td>\n",
       "      <td>0.394</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>69042</th>\n",
       "      <td>XGv-422</td>\n",
       "      <td>XBB</td>\n",
       "      <td>imprinted_2022</td>\n",
       "      <td>0.789</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>66690 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      antibody    virus           study    IC50\n",
       "0         1-57     BA.1  imprinted_2022  10.000\n",
       "1         1-57     BA.2  imprinted_2022  10.000\n",
       "2         1-57  BA.2.75  imprinted_2022  10.000\n",
       "3         1-57     BA.5  imprinted_2022  10.000\n",
       "4         1-57   BQ.1.1  imprinted_2022  10.000\n",
       "...        ...      ...             ...     ...\n",
       "69038  XGv-422  BA.2.75  imprinted_2022   0.495\n",
       "69039  XGv-422     BA.5  imprinted_2022   0.681\n",
       "69040  XGv-422   BQ.1.1  imprinted_2022   1.060\n",
       "69041  XGv-422    D614G  imprinted_2022   0.394\n",
       "69042  XGv-422      XBB  imprinted_2022   0.789\n",
       "\n",
       "[66690 rows x 4 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Number of antibodies with no neutralization on any retained virus:\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>n_antibodies</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>non_neut</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>False</th>\n",
       "      <td>5456</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>True</th>\n",
       "      <td>1746</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          n_antibodies\n",
       "non_neut              \n",
       "False             5456\n",
       "True              1746"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "antibody_ic50s = (\n",
    "    pd.concat(\n",
    "        [evolving_2024_ic50s, repeated_2024_ic50s, imprinted_2022_ic50s],\n",
    "        ignore_index=True,\n",
    "    )\n",
    "    .assign(\n",
    "        antibody=lambda x: x[\"antibody\"].map(\n",
    "            lambda a: antibody_renaming[a] if a in antibody_renaming else a\n",
    "        ),\n",
    "        IC50=lambda x: x[\"IC50\"].clip(lower=1e-5, upper=10)\n",
    "    )\n",
    ")\n",
    "\n",
    "print(\"Dropping the following IC50s duplicated across studies:\")\n",
    "display(antibody_ic50s.groupby([\"antibody\", \"virus\"]).tail(-1).reset_index(drop=True))\n",
    "\n",
    "# keep just one of IC50s repeated across studies\n",
    "antibody_ic50s = antibody_ic50s.groupby([\"antibody\", \"virus\"]).first().reset_index()\n",
    "\n",
    "# only keep IC50s for viruses w enough antibodies\n",
    "virus_min_antibodies = config[\"virus_min_antibodies_w_ic50\"]\n",
    "print(f\"\\nOnly keep virus with IC50s against at least {virus_min_antibodies} viruses\")\n",
    "n_antibodies_per_virus = (\n",
    "    antibody_ic50s\n",
    "    .groupby(\"virus\")\n",
    "    .aggregate(n_antibodies=pd.NamedAgg(\"antibody\", \"nunique\"))\n",
    "    .sort_values(\"n_antibodies\", ascending=False)\n",
    "    .reset_index()\n",
    ")\n",
    "viruses_w_enough_antibodies = (\n",
    "    n_antibodies_per_virus\n",
    "    .query(\"n_antibodies >= @virus_min_antibodies\")\n",
    "    [\"virus\"]\n",
    "    .tolist()\n",
    ")\n",
    "print(\n",
    "    f\"Keeping {len(viruses_w_enough_antibodies)} of \"\n",
    "    f\"{len(n_antibodies_per_virus)} viruses\"\n",
    ")\n",
    "antibody_ic50s = antibody_ic50s.query(\"virus in @viruses_w_enough_antibodies\")\n",
    "\n",
    "print(\"\\nRetained IC50s:\")\n",
    "display(antibody_ic50s)\n",
    "\n",
    "print(\"\\nNumber of antibodies with no neutralization on any retained virus:\")\n",
    "assert antibody_ic50s[\"IC50\"].dtype == float\n",
    "assert (antibody_ic50s[\"IC50\"] <= 10).all()\n",
    "display(\n",
    "    antibody_ic50s\n",
    "    .assign(non_neut=lambda x: x[\"IC50\"] == 10)\n",
    "    .groupby(\"antibody\", as_index=False)\n",
    "    .aggregate({\"non_neut\": \"all\"})\n",
    "    .groupby(\"non_neut\")\n",
    "    .aggregate(n_antibodies=pd.NamedAgg(\"antibody\", \"nunique\"))\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9465462b-17ea-4ddc-b9a5-1ebd911a394c",
   "metadata": {},
   "source": [
    "Now combine the escape data, keeping only the escape data in the antibody source data frame for that study.\n",
    "We normalize so that the max site escape for each antibody is 1.\n",
    "We also determine how many sites have significant escape, and drop antibodies with too many such sites:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "2e953763-edf0-45e4-81a5-4dc87f214b3a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:23:19.756196Z",
     "iopub.status.busy": "2024-11-18T23:23:19.755796Z",
     "iopub.status.idle": "2024-11-18T23:23:36.805490Z",
     "shell.execute_reply": "2024-11-18T23:23:36.803977Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x600 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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>antibody</th>\n",
       "      <th>site</th>\n",
       "      <th>escape</th>\n",
       "      <th>study</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>115</th>\n",
       "      <td>BD55-1041</td>\n",
       "      <td>453</td>\n",
       "      <td>0.081796</td>\n",
       "      <td>evolving_2024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>119</th>\n",
       "      <td>BD55-1041</td>\n",
       "      <td>457</td>\n",
       "      <td>0.047790</td>\n",
       "      <td>evolving_2024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>124</th>\n",
       "      <td>BD55-1041</td>\n",
       "      <td>462</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>evolving_2024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>128</th>\n",
       "      <td>BD55-1041</td>\n",
       "      <td>466</td>\n",
       "      <td>0.332554</td>\n",
       "      <td>evolving_2024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>193</th>\n",
       "      <td>BD55-1205</td>\n",
       "      <td>331</td>\n",
       "      <td>0.055369</td>\n",
       "      <td>evolving_2024</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>500689</th>\n",
       "      <td>XGv-422</td>\n",
       "      <td>465</td>\n",
       "      <td>0.229740</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>500690</th>\n",
       "      <td>XGv-422</td>\n",
       "      <td>466</td>\n",
       "      <td>0.264641</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>500691</th>\n",
       "      <td>XGv-422</td>\n",
       "      <td>468</td>\n",
       "      <td>0.093587</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>500692</th>\n",
       "      <td>XGv-422</td>\n",
       "      <td>484</td>\n",
       "      <td>0.049302</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>500693</th>\n",
       "      <td>XGv-422</td>\n",
       "      <td>514</td>\n",
       "      <td>0.067027</td>\n",
       "      <td>imprinted_2022</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>220959 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         antibody  site    escape           study\n",
       "115     BD55-1041   453  0.081796   evolving_2024\n",
       "119     BD55-1041   457  0.047790   evolving_2024\n",
       "124     BD55-1041   462  1.000000   evolving_2024\n",
       "128     BD55-1041   466  0.332554   evolving_2024\n",
       "193     BD55-1205   331  0.055369   evolving_2024\n",
       "...           ...   ...       ...             ...\n",
       "500689    XGv-422   465  0.229740  imprinted_2022\n",
       "500690    XGv-422   466  0.264641  imprinted_2022\n",
       "500691    XGv-422   468  0.093587  imprinted_2022\n",
       "500692    XGv-422   484  0.049302  imprinted_2022\n",
       "500693    XGv-422   514  0.067027  imprinted_2022\n",
       "\n",
       "[220959 rows x 4 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Escape measurements for each study:\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>n</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>study</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>evolving_2024</th>\n",
       "      <td>134429</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>imprinted_2022</th>\n",
       "      <td>46301</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>repeated_2024</th>\n",
       "      <td>40229</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     n\n",
       "study                 \n",
       "evolving_2024   134429\n",
       "imprinted_2022   46301\n",
       "repeated_2024    40229"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "escape = (\n",
    "    pd.concat(\n",
    "        [\n",
    "            # get just antibodies from that study\n",
    "            (\n",
    "                escape_df\n",
    "                .assign(\n",
    "                    antibody=lambda x: x[\"antibody\"].map(\n",
    "                        lambda a: antibody_renaming[a] if a in antibody_renaming else a\n",
    "                    ),\n",
    "                )\n",
    "                .merge(antibody_source.query(\"study == @study\")[[\"antibody\"]])\n",
    "            )\n",
    "            for study, escape_df in [\n",
    "                (\"evolving_2024\", evolving_2024_escape),\n",
    "                (\"repeated_2024\", repeated_2024_escape),\n",
    "                (\"imprinted_2022\", imprinted_2022_escape),\n",
    "            ]\n",
    "        ],\n",
    "        ignore_index=True,\n",
    "    )\n",
    "    .assign(\n",
    "        escape=lambda x: (\n",
    "            x[\"escape\"] / x.groupby([\"antibody\", \"study\"])[\"escape\"].transform(\"max\")\n",
    "        )\n",
    "    )\n",
    ")\n",
    "\n",
    "assert len(escape) == len(escape.groupby([\"antibody\", \"site\"]))\n",
    "\n",
    "# get and plot n sites with significant escape, then eliminate antibodies with\n",
    "# too many sites of substantial escape\n",
    "sig_escape_threshold = config[\"filter_antibodies\"][\"sig_escape_threshold\"]\n",
    "max_sig_sites = config[\"filter_antibodies\"][\"max_sig_sites\"]\n",
    "n_sig_escape = (\n",
    "    escape\n",
    "    .query(\"escape >= @sig_escape_threshold\")\n",
    "    .groupby([\"antibody\", \"study\"], as_index=False)\n",
    "    .aggregate(n_sites=pd.NamedAgg(\"site\", \"nunique\"))\n",
    ")\n",
    "\n",
    "ncols = 2\n",
    "nrows = int(math.ceil(n_sig_escape[\"study\"].nunique() / ncols))\n",
    "axes = n_sig_escape.hist(\n",
    "    column=\"n_sites\",\n",
    "    by=\"study\",\n",
    "    bins=range(1, n_sig_escape[\"n_sites\"].max() + 1, 1),\n",
    "    grid=False,\n",
    "    layout=(nrows, ncols),\n",
    "    figsize=(4 * ncols, 3 * nrows),\n",
    ")\n",
    "for ax in axes.ravel():\n",
    "    ax.axvline(max_sig_sites, color=\"red\")\n",
    "fig = plt.gcf()\n",
    "fig.suptitle(f\"number of sites with escape >={sig_escape_threshold}\")\n",
    "fig.tight_layout()\n",
    "plt.show(plt.gcf())\n",
    "\n",
    "# drop antibodies with too many sites of escape\n",
    "escape = (\n",
    "    escape\n",
    "    .merge(\n",
    "        n_sig_escape.query(\"n_sites <= @max_sig_sites\")[[\"antibody\", \"study\"]],\n",
    "        how=\"inner\",\n",
    "        validate=\"many_to_one\",\n",
    "    )\n",
    ")\n",
    "\n",
    "# drop low escape values\n",
    "drop_escape_below = config[\"drop_escape_below\"]\n",
    "escape = escape.query(\"escape >= @drop_escape_below\")\n",
    "    \n",
    "assert escape[\"antibody\"].nunique() == len(escape.groupby([\"antibody\", \"study\"]))\n",
    "\n",
    "display(escape)\n",
    "\n",
    "print(\"Escape measurements for each study:\")\n",
    "display(escape.groupby(\"study\").aggregate(n=pd.NamedAgg(\"site\", \"count\")))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b96b0fc0-3123-4c19-80be-5d2a46923aeb",
   "metadata": {},
   "source": [
    "Only keep the antibodies that have all types of measurements when writing files in next cell:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "9d526e0c-4e59-4ee0-91bf-d2f163ba069f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:23:36.809823Z",
     "iopub.status.busy": "2024-11-18T23:23:36.809190Z",
     "iopub.status.idle": "2024-11-18T23:23:36.996136Z",
     "shell.execute_reply": "2024-11-18T23:23:36.995181Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "This many antibodies have all measurements:\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_22101/248179622.py:17: FutureWarning: Downcasting object dtype arrays on .fillna, .ffill, .bfill is deprecated and will change in a future version. Call result.infer_objects(copy=False) instead. To opt-in to the future behavior, set `pd.set_option('future.no_silent_downcasting', True)`\n",
      "  .fillna(False)\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></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>n_antibodies</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>has_source</th>\n",
       "      <th>has_ic50</th>\n",
       "      <th>has_escape</th>\n",
       "      <th>has_all</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>False</th>\n",
       "      <th>True</th>\n",
       "      <th>False</th>\n",
       "      <th>False</th>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"3\" valign=\"top\">True</th>\n",
       "      <th>False</th>\n",
       "      <th>False</th>\n",
       "      <th>False</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">True</th>\n",
       "      <th>False</th>\n",
       "      <th>False</th>\n",
       "      <td>1130</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>True</th>\n",
       "      <th>True</th>\n",
       "      <td>6068</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                        n_antibodies\n",
       "has_source has_ic50 has_escape has_all              \n",
       "False      True     False      False               4\n",
       "True       False    False      False               2\n",
       "           True     False      False            1130\n",
       "                    True       True             6068"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "antibodies = (\n",
    "    antibody_source\n",
    "    [[\"antibody\"]]\n",
    "    .assign(has_source=True)\n",
    "    .drop_duplicates()\n",
    "    .merge(\n",
    "        antibody_ic50s[[\"antibody\"]].assign(has_ic50=True).drop_duplicates(),\n",
    "        how=\"outer\",\n",
    "        validate=\"one_to_one\",\n",
    "    )\n",
    "    .merge(\n",
    "        escape[[\"antibody\"]].assign(has_escape=True).drop_duplicates(),\n",
    "        how=\"outer\",\n",
    "        validate=\"one_to_one\",\n",
    "    )\n",
    "    .set_index(\"antibody\")\n",
    "    .fillna(False)\n",
    "    .assign(has_all=lambda x: x.all(axis=1))\n",
    "    .reset_index()\n",
    ")\n",
    "\n",
    "print(\"This many antibodies have all measurements:\")\n",
    "display(\n",
    "    antibodies\n",
    "    .groupby([c for c in antibodies.columns if c != \"antibody\"])\n",
    "    .aggregate(n_antibodies=pd.NamedAgg(\"antibody\", \"count\"))\n",
    ")\n",
    "\n",
    "antibodies_to_keep = set(antibodies.query(\"has_all\")[\"antibody\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a1d58139-be75-4b1c-a94d-04199cdde12e",
   "metadata": {},
   "source": [
    "Get the antibody reweightings.\n",
    "Unlike for other parameters, we only expect reweightings for some antibodies:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "ba27a03c-aa27-42f4-b2bf-347f4799b931",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:23:36.999502Z",
     "iopub.status.busy": "2024-11-18T23:23:36.999053Z",
     "iopub.status.idle": "2024-11-18T23:23:37.012171Z",
     "shell.execute_reply": "2024-11-18T23:23:37.011385Z"
    },
    "tags": []
   },
   "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>antibody</th>\n",
       "      <th>study</th>\n",
       "      <th>reweight</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CP0590</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.246562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>GC0101</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.246562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>GC0102</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.246562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>GC0103</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.246562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>GC0104</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.246562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1198</th>\n",
       "      <td>GC2054</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.477959</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199</th>\n",
       "      <td>GC2056</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.477959</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1200</th>\n",
       "      <td>GC2057</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.477959</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1201</th>\n",
       "      <td>GC2058</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>0.874462</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1202</th>\n",
       "      <td>GC2059</td>\n",
       "      <td>repeated_2024</td>\n",
       "      <td>1.477959</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1203 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     antibody          study  reweight\n",
       "0      CP0590  repeated_2024  1.246562\n",
       "1      GC0101  repeated_2024  1.246562\n",
       "2      GC0102  repeated_2024  1.246562\n",
       "3      GC0103  repeated_2024  1.246562\n",
       "4      GC0104  repeated_2024  1.246562\n",
       "...       ...            ...       ...\n",
       "1198   GC2054  repeated_2024  1.477959\n",
       "1199   GC2056  repeated_2024  1.477959\n",
       "1200   GC2057  repeated_2024  1.477959\n",
       "1201   GC2058  repeated_2024  0.874462\n",
       "1202   GC2059  repeated_2024  1.477959\n",
       "\n",
       "[1203 rows x 3 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "antibody_reweighting = repeated_2024_reweighting\n",
    "\n",
    "antibody_reweighting"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "05dcf7f7-db55-4113-a3ac-3ab855de56c3",
   "metadata": {},
   "source": [
    "Write the output files, only with antibodies to keep:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "d0db434b-838d-49f4-9d54-cb54e28b651b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-11-18T23:23:37.015063Z",
     "iopub.status.busy": "2024-11-18T23:23:37.014701Z",
     "iopub.status.idle": "2024-11-18T23:23:38.222318Z",
     "shell.execute_reply": "2024-11-18T23:23:38.221465Z"
    }
   },
   "outputs": [],
   "source": [
    "os.makedirs(\"results\", exist_ok=True)\n",
    "\n",
    "(\n",
    "    antibody_source\n",
    "    .query(\"antibody in @antibodies_to_keep\")\n",
    "    .to_csv(\"results/antibody_sources.csv\", index=False)\n",
    ")\n",
    "\n",
    "(\n",
    "    antibody_ic50s\n",
    "    .query(\"antibody in @antibodies_to_keep\")\n",
    "    .drop(columns=\"study\")\n",
    "    .to_csv(\"results/antibody_IC50s.csv\", index=False)\n",
    ")\n",
    "\n",
    "(\n",
    "    escape\n",
    "    .query(\"antibody in @antibodies_to_keep\")\n",
    "    .drop(columns=\"study\")\n",
    "    .to_csv(\"results/escape.csv\", float_format=\"%.4g\", index=False)\n",
    ")\n",
    "\n",
    "(\n",
    "    antibody_reweighting\n",
    "    .query(\"antibody in @antibodies_to_keep\")\n",
    "    .drop(columns=\"study\")\n",
    "    .to_csv(\"results/antibody_reweighting.csv\", float_format=\"%.4g\", index=False)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fa5d193d-0598-414f-8641-eba126a2fa6f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "09bfe902-f5d9-49ae-9ec7-3a64f73f7adb",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.11.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}