figureS3.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Why prefilter paired chain data?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib\n",
    "\n",
    "import pyrepseq as prs\n",
    "import pyrepseq.plotting as pp\n",
    "\n",
    "from lib import *\n",
    "\n",
    "plt.style.use('tcrspace.mplstyle')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "ratios = []\n",
    "pc_raw = []\n",
    "pc_corrected = []\n",
    "for name in tannofiles[:9]:\n",
    "    df = pd.read_csv(f'{data_directory}/tanno/{name}.txt', sep='\\t')\n",
    "    df = df.sort_values('Clustered', ascending=False)\n",
    "    valid_light = df['CDRL3_AA'].apply(prs.isvalidcdr3)\n",
    "    valid_heavy = df['CDRH3_AA'].apply(prs.isvalidcdr3)\n",
    "    df = df[valid_light & valid_heavy]\n",
    "    ratio = prs.pc(df['CDRL3_NT'])/prs.pc(df['CDRL3_AA'])\n",
    "    pc_raw.append(prs.pc(df['CDRL3_AA']))\n",
    "    df.drop_duplicates('CDRL3_NT', inplace=True)\n",
    "    pc_corrected.append(prs.pc(df['CDRL3_AA']))\n",
    "    ratios.append(ratio)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(f'{data_directory}/post_sample_alpha.csv',\n",
    "                 names=['amino_acid', 'v_gene', 'j_gene', 'nucleotide'])\n",
    "ratio_expected = prs.pc(df['nucleotide'])/prs.pc(df['amino_acid'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "chain='alpha'\n",
    "pc_data = []\n",
    "for subject in ['W', 'M']:\n",
    "    df_back = pd.read_csv(f'{data_directory}/minervina/{chain}/{subject}_F1_2018_{chain}.txt.gz',\n",
    "                          sep='\\t')\n",
    "    df_back = df_back[df_back['aaSeqCDR3'].apply(prs.isvalidcdr3)]\n",
    "    pc_data.append(prs.pc(df_back['aaSeqCDR3']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "colors = matplotlib.rcParams['axes.prop_cycle'].by_key()['color']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 246.24x144 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axes = plt.subplots(figsize=(3.42, 2.0), ncols=2)\n",
    "ax = axes[0]\n",
    "ax.plot(['Model'], [ratio_expected], 'o')\n",
    "ax.scatter(['Data']*len(ratios), ratios, color=colors[:len(ratios)])\n",
    "ax.set_ylabel('NT $p_C$ / AA $p_C$')\n",
    "ax.set_ylim(0, 1.0)\n",
    "ax.set_xlim(-0.3, 1.3)\n",
    "ax=axes[1]\n",
    "pcs = np.array(list(zip(pc_raw, pc_corrected)))\n",
    "for raw, corrected in pcs:\n",
    "    ax.plot(['Raw', 'Filtered'], [raw, corrected], 'o-')\n",
    "ax.plot(['Bulk']*len(pc_data), pc_data, 'o')\n",
    "ax.set_ylabel('$p_C$')\n",
    "ax.set_yscale('log')\n",
    "ax.set_xlim(-0.3, 2.3)\n",
    "ax.set_ylim(1e-6, 1.7e-4)\n",
    "fig.tight_layout()\n",
    "pp.label_axes(fig, xy=(-0.4, 1.0))\n",
    "fig.savefig('figures/figureS3.svg')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Rationale for collapsing alpha chains paired with multiple beta chains in the paired chain data from Tanno. The ratio between nucleotide and amino acid coincidences probabilities for alpha chains is substantially higher than expected from models of VDJ recombination suggesting potential cross-contamination. Collapsing redundant alpha chains reduces variation in amino acid coincidence probabilities across samples and makes them more comparable to those found in bulk single chain datasets (subjects M and W at baseline from Minervina 2021)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}