Fig4_functional.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import os\n",
    "import glob\n",
    "from collections import OrderedDict\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.ticker as ticker\n",
    "from matplotlib.colors import hsv_to_rgb\n",
    "\n",
    "import qtl.annotation\n",
    "import qtl.plot\n",
    "import qtl.torus\n",
    "import style\n",
    "\n",
    "annot = qtl.annotation.Annotation('data/gencode.v26.GRCh38.genes.gtf', verbose=False)\n",
    "sample_df = pd.read_csv('data/GTEx_Analysis_v8_RNAseq_samples.txt', sep='\\t')\n",
    "sample_s = sample_df[sample_df['has_genotype']].groupby('tissue_id').apply(lambda x: x['sample_id'].tolist())\n",
    "sample_s = sample_s[sample_s.apply(len)>=70]\n",
    "size_s = sample_s.apply(len)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# VEP annotation for top eQTL/sQTL\n",
    "vep_df = pd.read_csv('data/WGS_Feature_overlap_collapsed_VEP_short_4torus.MAF01.txt.gz', sep='\\t', index_col=0)\n",
    "vep_df = vep_df.astype(np.int8)\n",
    "vep_df.columns = [qtl.torus.torus_dict[i[:-2].upper()] for i in vep_df.columns]\n",
    "\n",
    "# variant IDs for each category\n",
    "variant_sets = {}\n",
    "for p in vep_df:\n",
    "    variant_sets[p] = set(vep_df.loc[vep_df[p]==1].index)\n",
    "variant_set_s = pd.Series(variant_sets)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Panel A: *cis*-QTLs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing Whole_Bloodtine_Terminal_Ileumctionia"
     ]
    }
   ],
   "source": [
    "# eGenes and sGenes files\n",
    "egene_files = {os.path.basename(i).split('.')[0]:i for i in glob.glob('data/GTEx_Analysis_v8_eQTL/*.v8.egenes.txt.gz')}\n",
    "sgene_files = {os.path.basename(i).split('.')[0]:i for i in glob.glob('data/GTEx_Analysis_v8_sQTL/*.v8.sgenes.txt.gz')}\n",
    "tissues = sorted(egene_files.keys())\n",
    "\n",
    "# annotate top eQTLs and sQTLs with variant categories\n",
    "egene_variant_cat_df = []\n",
    "sgene_variant_cat_df = []\n",
    "for t in tissues:\n",
    "    print('\\rProcessing {}'.format(t), end='')\n",
    "    egenes_df = pd.read_csv(egene_files[t], sep='\\t', index_col=0)\n",
    "    egenes_df['biotype'] = egenes_df.index.map(lambda x: annot.gene_dict[x].type)\n",
    "    egenes_df = egenes_df[(egenes_df['qval']<=0.05) & egenes_df['biotype'].isin(['protein_coding', 'lincRNA'])]\n",
    "    variant_ids = egenes_df['variant_id'].values\n",
    "    pct_s = variant_set_s.apply(lambda x: len([i for i in variant_ids if i in x]))\n",
    "    pct_s.name = t\n",
    "    egene_variant_cat_df.append(pct_s)\n",
    "\n",
    "    sgenes_df = pd.read_csv(sgene_files[t], sep='\\t', index_col=0)\n",
    "    sgenes_df['biotype'] = sgenes_df['gene_id'].apply(lambda x: annot.gene_dict[x].type)\n",
    "    sgenes_df = sgenes_df[(sgenes_df['qval']<=0.05) & sgenes_df['biotype'].isin(['protein_coding', 'lincRNA'])]\n",
    "    variant_ids = sgenes_df['variant_id'].values\n",
    "    pct_s = variant_set_s.apply(lambda x: len([i for i in variant_ids if i in x]))\n",
    "    pct_s.name = t\n",
    "    sgene_variant_cat_df.append(pct_s)\n",
    "    \n",
    "egene_variant_cat_df = pd.concat(egene_variant_cat_df, axis=1)    \n",
    "sgene_variant_cat_df = pd.concat(sgene_variant_cat_df, axis=1)    \n",
    "egene_variant_cat_pct_df = egene_variant_cat_df / egene_variant_cat_df.sum(0)\n",
    "sgene_variant_cat_pct_df = sgene_variant_cat_df / sgene_variant_cat_df.sum(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# mean and s.d. of variant categories across tissues\n",
    "mean_df = pd.concat([\n",
    "    egene_variant_cat_pct_df.mean(1),\n",
    "    sgene_variant_cat_pct_df.mean(1)\n",
    "], axis=1)\n",
    "mean_df.columns = ['eQTL', 'sQTL']\n",
    "mean_df.rename(index=qtl.torus.torus_short_dict, inplace =True)\n",
    "\n",
    "std_df = pd.concat([\n",
    "    egene_variant_cat_pct_df.std(1),\n",
    "    sgene_variant_cat_pct_df.std(1)\n",
    "], axis=1)\n",
    "std_df.columns = ['eQTL', 'sQTL']\n",
    "std_df.rename(index=qtl.torus.torus_short_dict, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load Torus outputs for eQTLs and sQTLs\n",
    "eqtl_lor_df = qtl.torus.load_summary('data/GTEx_Analysis_v8_eQTL.torus_enrichment.txt', log2=True)\n",
    "sqtl_lor_df = qtl.torus.load_summary('data/GTEx_Analysis_v8_sQTL.torus_enrichment.txt', log2=True)\n",
    "eqtl_lor_df.drop(['Kidney_Cortex'], axis=1, inplace=True)\n",
    "sqtl_lor_df.drop(['Kidney_Cortex'], axis=1, inplace=True)\n",
    "\n",
    "# average across tissues: cis-eQTLs\n",
    "m = (2**eqtl_lor_df).mean(1)\n",
    "s = (2**eqtl_lor_df).std(1)\n",
    "eqtl_tissue_avg_df = np.log2(pd.concat([m, m-s, m+s], axis=1))\n",
    "eqtl_tissue_avg_df.columns = ['mean', 'std_low', 'std_high']\n",
    "eqtl_tissue_avg_df.rename(index=qtl.torus.torus_short_dict, inplace=True)\n",
    "\n",
    "# average across tissues: cis-sQTLs\n",
    "m = (2**sqtl_lor_df).mean(1)\n",
    "s = (2**sqtl_lor_df).std(1)\n",
    "sqtl_tissue_avg_df = np.log2(pd.concat([m, m-s, m+s], axis=1))\n",
    "sqtl_tissue_avg_df.columns = ['mean', 'std_low', 'std_high']\n",
    "sqtl_tissue_avg_df.rename(index=qtl.torus.torus_short_dict, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_enrichment(df1, df2, varmean_df, varstd_df, args1, args2, legend_title=None):\n",
    "    dl = 1.25\n",
    "    aw1 = 1.5\n",
    "    ds = 0.05\n",
    "    aw2 = 0.4\n",
    "    dr = 0.5\n",
    "    db = 0.75\n",
    "    ah = 3/17*df1.shape[0]\n",
    "    dt = 0.25\n",
    "    aw3 = 0.2\n",
    "    \n",
    "    fw = dl + aw1 + ds + aw2 + aw3 + dr\n",
    "    fh = db + ah + dt\n",
    "    fig = plt.figure(facecolor=(1,1,1), figsize=(fw,fh))\n",
    "    ax1 = fig.add_axes([dl/fw, db/fh, aw1/fw, ah/fh])\n",
    "    ax2 = fig.add_axes([(dl+aw1+ds)/fw, db/fh, aw2/fw, ah/fh], sharey=ax1)\n",
    "    ax3 = fig.add_axes([(dl+aw1+ds+aw2+0.05)/fw, db/fh, aw3/fw, ah/fh], sharey=ax1)\n",
    "\n",
    "    y = np.arange(df1.shape[0])\n",
    "    ax1.set_ylim([y[0]-0.5, y[-1]+0.5])\n",
    "    \n",
    "    qtl.plot.plot_effects([df1, df2], [args1, args2],\n",
    "                 xspace=[1.25,2,0.5], yspace=[0.5,ah,0.25], xlim=[-2,9],\n",
    "                 xlabel='log$\\mathregular{_{2}}$(Fold enrichment)', ylabel=None, ax=ax1)\n",
    "\n",
    "    ax1.set_yticks(y)\n",
    "    ax1.set_yticklabels(df1.index)\n",
    "    ax1.legend(title=legend_title, loc='upper right', ncol=1, labelspacing=0.33, \n",
    "               columnspacing=1, handletextpad=0, handlelength=1.66, borderpad=0.25, framealpha=1)\n",
    "    for ax in [ax2, ax3]:\n",
    "        plt.setp(ax.get_yticklabels(), visible=False);\n",
    "        for line in ax.yaxis.get_ticklines():\n",
    "            line.set_markersize(0) # tick length\n",
    "            line.set_markeredgewidth(0) # tick line width\n",
    "    ax2.spines['right'].set_visible(False)\n",
    "    ax3.spines['left'].set_visible(False)\n",
    "\n",
    "    # plot proportion of variants\n",
    "    varmean_df.plot(kind='barh', xerr=1.96*varstd_df, ax=ax2, legend=False, width=0.66, \n",
    "                 color=[args1['c'], args2['c']])\n",
    "    varmean_df.plot(kind='barh', xerr=1.96*varstd_df, ax=ax3, legend=False, width=0.66, \n",
    "                 color=[args1['c'], args2['c']])\n",
    "\n",
    "    ax2.set_xlim([0,0.12])\n",
    "    ax2.set_xticks([0,0.1]) \n",
    "    ax3.set_xlim([0.5, 0.62])\n",
    "    ax3.set_xticks([0.5,0.6])\n",
    "    ax2.set_xticklabels([0,0.1], ha='center', fontsize=8)\n",
    "    ax3.set_xticklabels([0.5,0.6], ha='left', fontsize=8)\n",
    "    ax2.set_xlabel('Proportion\\n of variants', ha='left')\n",
    "\n",
    "    ax1.xaxis.set_major_locator(ticker.MaxNLocator(integer=True, min_n_ticks=5, nbins=4))\n",
    "    cis_xlim = ax1.get_xlim()\n",
    "    ax1.invert_yaxis()\n",
    "    ax2.invert_yaxis()\n",
    "    ax3.invert_yaxis()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 280.8x288 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "eargs = {'fmt':'.', 'ms':8,   'lw':1, 'clip_on':False, 'c':style.qtl_color_s['eqtl'], 'label':'eQTL'}\n",
    "sargs = {'fmt':'D', 'ms':3.5, 'lw':1, 'clip_on':False, 'c':style.qtl_color_s['sqtl'], 'label':'sQTL'}\n",
    "plot_enrichment(eqtl_tissue_avg_df, sqtl_tissue_avg_df, mean_df, std_df, eargs, sargs)\n",
    "plt.savefig('pdfs/Fig4A_torus_enrichment_summary.pdf')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Panel B: *trans*-eQTLs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 280.8x199.059 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "trans_df = pd.read_csv('data/GTEx_Analysis_v8_trans_eGenes_fdr05.txt', sep='\\t')\n",
    "torus_trans_df = qtl.torus.load('data/GTEx_Analysis_v8_eQTL.trans.torus_enrichment.txt', log2=True, short_labels=True)\n",
    "\n",
    "# drop categories with insufficient overlap\n",
    "drop_ix = [\n",
    "    'TF binding site',\n",
    "    'Frameshift',\n",
    "    'Splice acceptor',\n",
    "    'Splice donor',\n",
    "    'Splice region',\n",
    "    'Stop gained',\n",
    "    'Synonymous',\n",
    "]\n",
    "# proportion of variants\n",
    "counts_s = variant_set_s.apply(lambda x: len([i for i in trans_df['variant_id'] if i in x]))\n",
    "counts_s.rename(index=qtl.torus.torus_short_dict, inplace=True)\n",
    "counts_s.drop(drop_ix, inplace=True)\n",
    "pct_s = counts_s/counts_s.sum()\n",
    "pct_s.name = 'trans'\n",
    "\n",
    "cargs = {'fmt':'.', 'ms':8,   'lw':1, 'clip_on':True, 'c':style.qtl_color_s['eqtl'], 'label':'cis'}\n",
    "targs = {'fmt':'D', 'ms':3.5, 'lw':1, 'clip_on':True, 'c':style.qtl_color_s['eqtl_trans'], 'label':'trans'}\n",
    "plot_enrichment(eqtl_tissue_avg_df.drop(drop_ix), torus_trans_df.drop(drop_ix),\n",
    "                mean_df[['eQTL']].join(pct_s.to_frame(), how='right'),\n",
    "                std_df[['eQTL']].join((pct_s*0).to_frame(), how='right'), cargs, targs, legend_title='eQTLs')\n",
    "plt.savefig('pdfs/Fig4B_torus_enrichment_trans.pdf')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Panel C: *cis* enrichment of *trans*-eVariants"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 252x149.76 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "or_df = pd.DataFrame([[5.8, 1.6], [2.4, 1.4]], \n",
    "                     columns=['cis-eQTLs', 'cis-sQTLs'],\n",
    "                     index=['trans-\\neQTLs', 'trans-\\nsQTLs'])\n",
    "ax = qtl.plot.setup_figure(2, 1.33, yspace=[0.5,0.25], xspace=[1.25,0.25])\n",
    "w = 0.66\n",
    "or_df.plot(kind='barh', ax=ax, width=w, color=[style.qtl_color_s['eqtl'], style.qtl_color_s['sqtl']])\n",
    "ax.legend(loc='lower right', title='Enrichment', handlelength=0.66, labelspacing=0.33, \n",
    "          handletextpad=0.33, borderaxespad=0, bbox_to_anchor=(1,0.1))\n",
    "ax.set_xlabel('Odds ratio', fontsize=12)\n",
    "qtl.plot.format_plot(ax, fontsize=12)\n",
    "ax.spines['left'].set_smart_bounds(True)\n",
    "ax.spines['bottom'].set_position(('outward', -4))\n",
    "ax.invert_yaxis()\n",
    "plt.savefig('pdfs/Fig4C_trans_cis_enrichment.pdf')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Panel D: proportion of *trans*-eVariants that are mediated by *cis*-eQTLs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_summary(summary_df, phenotype):\n",
    "    test_color_s = pd.Series(OrderedDict([\n",
    "        ('top_evariant', hsv_to_rgb([0, 0, 0.4      ])),\n",
    "        ('caveman',        hsv_to_rgb([0.5, 0.8, 0.7  ])),\n",
    "        ('caviar',      hsv_to_rgb([0.025, 0.8, 0.8])),\n",
    "        ('dap-g' ,      hsv_to_rgb([0.1, 0.8, 0.9  ])),\n",
    "        ('consensus',    hsv_to_rgb([0.3, 0.99, 0.7 ])),\n",
    "    ]))\n",
    "\n",
    "    dl = 0.75\n",
    "    aw = 0.3*summary_df.shape[0]\n",
    "    dr = 1.5\n",
    "    db = 1\n",
    "    ah = 2\n",
    "    dt = 0.5\n",
    "    fw = dl+aw+dr\n",
    "    fh = db+ah+dt\n",
    "\n",
    "    fig = plt.figure(facecolor=(1,1,1), figsize=(fw,fh))\n",
    "    ax = fig.add_axes([dl/fw, db/fh, aw/fw, ah/fh])\n",
    "    ax.margins(0.03,0)\n",
    "\n",
    "    colors = [style.qtl_color_s['eqtl'], style.qtl_color_s['eqtl_dark'], [0.75]*3, \n",
    "                         [0.5]*3,  test_color_s['consensus']]\n",
    "\n",
    "    x_pos = np.arange(summary_df.shape[0]).astype(np.float32)\n",
    "    x_pos[-1] += 0.5\n",
    "    cols = summary_df.columns[:-2]\n",
    "    k = 0\n",
    "    ax.bar(x_pos, summary_df[cols[k]], color=colors[k], label=cols[k], width=0.8)\n",
    "    for k in range(1,4):\n",
    "        ax.bar(x_pos, summary_df[cols[k]], bottom=summary_df[cols[:k]].sum(1), color=colors[k], label=cols[k], width=0.8)\n",
    "    ax.set_xticks(x_pos)\n",
    "    ax.set_xticklabels(summary_df.index, rotation=-45, ha='left')\n",
    "\n",
    "    x = np.arange(summary_df.shape[0]-2)\n",
    "    ylim = ax.get_ylim()\n",
    "    d = ylim[1]-ylim[0]\n",
    "    s = 1.08\n",
    "    b = ylim[1] - s*d\n",
    "    ax.set_ylim(ylim)\n",
    "    \n",
    "    colors_df = style.colors_df.append(pd.DataFrame({'color_hex':'#FFFFFF', 'tissue_abbrv':'All tissues (mean)'},\n",
    "                                                    index=['All tissues (mean)']), sort=False)\n",
    "    \n",
    "    ax.scatter(x, b*np.ones(len(x)), s=summary_df['sample_size'].iloc[:-2]/5, \n",
    "               c=colors_df.set_index('tissue_abbrv').loc[summary_df.index[:-2], 'color_hex'], clip_on=False)\n",
    "    for i,n in enumerate(summary_df['trans_genes'].iloc[:-2].astype(int)):\n",
    "        ax.text(i,1.02,n, ha='center', fontsize=12)\n",
    "    ax.set_title(phenotype, pad=18, fontsize=14)\n",
    "\n",
    "    qtl.plot.format_plot(ax, fontsize=12)\n",
    "    ax.set_xticklabels(summary_df.index, rotation=-36, ha='left')\n",
    "    ax.tick_params(axis='x', which='major', pad=15, labelsize=10)\n",
    "\n",
    "    handles, labels = ax.get_legend_handles_labels()\n",
    "    leg = ax.legend(handles, labels, loc='center left', ncol=1, labelspacing=0.25, \n",
    "               columnspacing=1, handletextpad=0.5, handlelength=.66, framealpha=1, bbox_to_anchor=(0.98,0.5))\n",
    "\n",
    "    ax.set_xlabel(None)\n",
    "    ax.set_ylabel('Fraction of '+phenotype, fontsize=14)\n",
    "    ax.spines['bottom'].set_smart_bounds(True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing Whole_Bloodposed_Lower_legpubiction"
     ]
    }
   ],
   "source": [
    "# enrichment in cis-eQTLs\n",
    "\n",
    "# restrict to protein coding or lincRNA genes\n",
    "gene_set = set([g.id for g in annot.genes if g.type in ['protein_coding', 'lincRNA']])\n",
    "\n",
    "# load trans-eQTL results (0.05 FDR)\n",
    "trans_df = pd.read_csv('data/GTEx_Analysis_v8_trans_eGenes_fdr05.txt', sep='\\t')\n",
    "for c in ['tested_cis_eqtl', 'tested_cis_sqtl', 'cis_eQTL', 'cis_sQTL']:\n",
    "    trans_df[c] = False\n",
    "\n",
    "for tissue_id, gdf in trans_df.groupby('tissue_id'):\n",
    "    print('\\rProcessing {}'.format(tissue_id), end='')\n",
    "    bed_df = pd.read_csv('data/GTEx_Analysis_v8_eQTL_expression_matrices/{}.v8.normalized_expression.bed.gz'.format(tissue_id),\n",
    "                    sep='\\t', usecols=['#chr', 'end', 'gene_id'], index_col=2).rename(columns={'end':'tss'})\n",
    "    bed_df['biotype'] = bed_df.index.map(lambda x: annot.gene_dict[x].type)\n",
    "    bed_df = bed_df[bed_df['biotype'].isin(['protein_coding', 'lincRNA'])]\n",
    "\n",
    "    group_s = pd.read_csv('data/GTEx_Analysis_v8_sQTL_groups/{}.leafcutter.phenotype_groups.txt'.format(tissue_id),\n",
    "                          sep='\\t', index_col=0, header=None, squeeze=True)\n",
    "    sqtl_gene_ids = group_s.unique()\n",
    "    \n",
    "    signif_eqtl_df = pd.read_csv('data/GTEx_Analysis_v8_eQTL/{}.v8.signif_variant_gene_pairs.txt.gz'.format(tissue_id),\n",
    "                                 sep='\\t', usecols=['variant_id', 'gene_id'])\n",
    "    signif_eqtl_df = signif_eqtl_df[signif_eqtl_df['gene_id'].apply(lambda x: x in gene_set)]\n",
    "\n",
    "    signif_sqtl_df = pd.read_csv('data/GTEx_Analysis_v8_sQTL/{}.v8.sqtl_signifpairs.txt.gz'.format(tissue_id),\n",
    "                                 sep='\\t', usecols=['variant_id', 'phenotype_id'])\n",
    "    signif_sqtl_df['gene_id'] = signif_sqtl_df['phenotype_id'].apply(lambda x: x.split(':')[-1])\n",
    "    signif_sqtl_df = signif_sqtl_df[signif_sqtl_df['gene_id'].apply(lambda x: x in gene_set)]\n",
    "    \n",
    "    for i,r in gdf.iterrows():\n",
    "        chrom, pos = r['variant_id'].split('_')[:2]\n",
    "        pos = int(pos)\n",
    "        tss_s = bed_df.loc[bed_df['#chr']==chrom, 'tss']\n",
    "        if (tss_s-pos).abs().min() <= 1000000:\n",
    "            trans_df.loc[i, 'tested_cis_eqtl'] = True\n",
    "        tss_s = pd.Series([annot.gene_dict[x].tss for x in sqtl_gene_ids], index=sqtl_gene_ids)\n",
    "        if (tss_s.loc[sqtl_gene_ids]-pos).abs().min() <= 1000000:\n",
    "            trans_df.loc[i, 'tested_cis_sqtl'] = True\n",
    "        if r['variant_id'] in signif_eqtl_df['variant_id'].values:\n",
    "            trans_df.loc[i, 'cis_eQTL'] = True\n",
    "        if r['variant_id'] in signif_sqtl_df['variant_id'].values:\n",
    "            trans_df.loc[i, 'cis_sQTL'] = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "scrolled": true
   },
   "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>tested_cis_eqtl</th>\n",
       "      <th>tested_cis_sqtl</th>\n",
       "      <th>cis_eQTL</th>\n",
       "      <th>cis_sQTL</th>\n",
       "      <th>trans_egenes</th>\n",
       "      <th>sample_size</th>\n",
       "      <th>cis_eqtl_mediator</th>\n",
       "      <th>cis_sqtl_mediator</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tissue_abbrv</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>TESTIS</td>\n",
       "      <td>49.000000</td>\n",
       "      <td>49.0</td>\n",
       "      <td>20.000000</td>\n",
       "      <td>13.000000</td>\n",
       "      <td>49.0</td>\n",
       "      <td>322.000000</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>13.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>FIBRBLS</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>9.0</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>9.0</td>\n",
       "      <td>483.000000</td>\n",
       "      <td>6.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>SKINNS</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>7.0</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>7.0</td>\n",
       "      <td>517.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>NERVET</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>8.0</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>8.0</td>\n",
       "      <td>532.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>THYROID</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>16.0</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>16.0</td>\n",
       "      <td>574.000000</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>2.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>SKINS</td>\n",
       "      <td>16.000000</td>\n",
       "      <td>16.0</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>16.0</td>\n",
       "      <td>605.000000</td>\n",
       "      <td>7.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>WHLBLD</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>13.0</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>13.0</td>\n",
       "      <td>670.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>MSCLSK</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>9.0</td>\n",
       "      <td>9.000000</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>9.0</td>\n",
       "      <td>706.000000</td>\n",
       "      <td>8.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>All tissues (mean)</td>\n",
       "      <td>5.851852</td>\n",
       "      <td>6.0</td>\n",
       "      <td>2.851852</td>\n",
       "      <td>0.814815</td>\n",
       "      <td>6.0</td>\n",
       "      <td>395.814815</td>\n",
       "      <td>2.074074</td>\n",
       "      <td>0.740741</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    tested_cis_eqtl  tested_cis_sqtl   cis_eQTL   cis_sQTL  \\\n",
       "tissue_abbrv                                                                 \n",
       "TESTIS                    49.000000             49.0  20.000000  13.000000   \n",
       "FIBRBLS                    9.000000              9.0   6.000000   2.000000   \n",
       "SKINNS                     7.000000              7.0   3.000000   0.000000   \n",
       "NERVET                     8.000000              8.0   5.000000   1.000000   \n",
       "THYROID                   16.000000             16.0  11.000000   2.000000   \n",
       "SKINS                     16.000000             16.0   9.000000   1.000000   \n",
       "WHLBLD                    11.000000             13.0   2.000000   0.000000   \n",
       "MSCLSK                     9.000000              9.0   9.000000   2.000000   \n",
       "All tissues (mean)         5.851852              6.0   2.851852   0.814815   \n",
       "\n",
       "                    trans_egenes  sample_size  cis_eqtl_mediator  \\\n",
       "tissue_abbrv                                                       \n",
       "TESTIS                      49.0   322.000000          10.000000   \n",
       "FIBRBLS                      9.0   483.000000           6.000000   \n",
       "SKINNS                       7.0   517.000000           2.000000   \n",
       "NERVET                       8.0   532.000000           4.000000   \n",
       "THYROID                     16.0   574.000000           9.000000   \n",
       "SKINS                       16.0   605.000000           7.000000   \n",
       "WHLBLD                      13.0   670.000000           2.000000   \n",
       "MSCLSK                       9.0   706.000000           8.000000   \n",
       "All tissues (mean)           6.0   395.814815           2.074074   \n",
       "\n",
       "                    cis_sqtl_mediator  \n",
       "tissue_abbrv                           \n",
       "TESTIS                      13.000000  \n",
       "FIBRBLS                      1.000000  \n",
       "SKINNS                       0.000000  \n",
       "NERVET                       1.000000  \n",
       "THYROID                      2.000000  \n",
       "SKINS                        1.000000  \n",
       "WHLBLD                       0.000000  \n",
       "MSCLSK                       1.000000  \n",
       "All tissues (mean)           0.740741  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# summarize by tissue\n",
    "summary_df = []\n",
    "for tissue_id, gdf in trans_df.groupby('tissue_id'):\n",
    "    sum_s = gdf[['tested_cis_eqtl', 'tested_cis_sqtl', 'cis_eQTL', 'cis_sQTL']].sum()\n",
    "    sum_s['trans_egenes'] = gdf.shape[0]\n",
    "    sum_s.name = tissue_id\n",
    "    summary_df.append(sum_s)\n",
    "summary_df = pd.concat(summary_df, axis=1).T.astype(int).sort_values('tested_cis_eqtl', ascending=False)\n",
    "summary_df['sample_size'] = size_s.loc[summary_df.index]\n",
    "summary_df.index = style.colors_df.loc[summary_df.index, 'tissue_abbrv']\n",
    "\n",
    "# add mediation results: cis-eQTLs\n",
    "mediation_df = pd.read_excel('/xchip/gtex/francois/manuscripts/gtex_v8/data/GTEx_v8_mediation_filtered.xlsx',\n",
    "                             sheet_name='GTEx_v8_mediation_consortium_fi')\n",
    "mediation_df = mediation_df[mediation_df['BH_FDR']<0.05]\n",
    "med_s = mediation_df[['trans_gene', 'tissue']].drop_duplicates()['tissue'].value_counts()\n",
    "med_s.index = style.colors_df.loc[med_s.index, 'tissue_abbrv']\n",
    "med_s.name = 'cis_eqtl_mediator'\n",
    "summary_df = summary_df.join(med_s, how='left').fillna(0).astype(int)\n",
    "\n",
    "# add mediation results: cis-sQTLs\n",
    "mediation_df = pd.read_excel('/xchip/gtex/francois/manuscripts/gtex_v8/data/GTEx_v8_mediation_filtered.xlsx',\n",
    "                             sheet_name='GTEx_v8_sQTL_mediation_by_trio_')\n",
    "mediation_df = mediation_df[mediation_df['BH_FDR']<0.05]\n",
    "med_s = mediation_df[['trans_gene', 'tissue']].drop_duplicates()['tissue'].value_counts()\n",
    "med_s.index = style.colors_df.loc[med_s.index, 'tissue_abbrv']\n",
    "med_s.name = 'cis_sqtl_mediator'\n",
    "summary_df = summary_df.join(med_s, how='left').fillna(0).astype(int)\n",
    "\n",
    "summary_df = summary_df.sort_values('sample_size')\n",
    "\n",
    "# add mean across tissues\n",
    "mean_s = summary_df.mean()\n",
    "mean_s.name = 'All tissues (mean)'\n",
    "summary_df = summary_df.append(mean_s)\n",
    "\n",
    "# summary_df = summary_df[['trans_egenes', 'tested_cis', 'cis_eQTL', 'cis_eqtl_mediator', 'cis_sQTL', 'cis_sqtl_mediator', 'sample_size']]\n",
    "subset_df = summary_df[summary_df['trans_egenes']>5]\n",
    "subset_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 378x252 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "trans_eqtl_frac_df = (pd.concat([\n",
    "    subset_df['cis_eqtl_mediator'],\n",
    "    subset_df['cis_eQTL'] - subset_df['cis_eqtl_mediator'] ,\n",
    "    subset_df['tested_cis_eqtl'] - subset_df['cis_eQTL'],\n",
    "    subset_df['trans_egenes'] - subset_df['tested_cis_eqtl'],\n",
    "], axis=1).T / subset_df['trans_egenes']).T\n",
    "trans_eqtl_frac_df.columns = ['cis-eQTL\\nmediator', 'cis-eQTL', 'Not a\\ncis-eQTL', 'Not tested\\nfor cis']\n",
    "trans_eqtl_frac_df['sample_size'] = summary_df.loc[trans_eqtl_frac_df.index, 'sample_size']\n",
    "trans_eqtl_frac_df['trans_genes'] = summary_df['trans_egenes']\n",
    "# add background\n",
    "background_df = pd.DataFrame(np.array([[216, 1527-216, 1620-1527]])/1620, \n",
    "                             columns=['cis-eQTL', 'Not a\\ncis-eQTL', 'Not tested\\nfor cis'], \n",
    "                             index=['Background'])\n",
    "trans_eqtl_frac_df = trans_eqtl_frac_df.append(background_df, sort=False).fillna(0)\n",
    "\n",
    "plot_summary(trans_eqtl_frac_df, phenotype='trans-eGenes')\n",
    "plt.savefig('pdfs/Fig4D_trans_eqtl_mediation_cis_eqtl.pdf')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Supplementary figures associated with Fig. 4D"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 378x252 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "trans_eqtl_sqtl_frac_df = (pd.concat([\n",
    "    subset_df['cis_sqtl_mediator'],\n",
    "    subset_df['cis_sQTL'] - subset_df['cis_sqtl_mediator'] ,\n",
    "    subset_df['tested_cis_sqtl'] - subset_df['cis_sQTL'],\n",
    "    subset_df['trans_egenes'] - subset_df['tested_cis_sqtl'],\n",
    "], axis=1).T / subset_df['trans_egenes']).T\n",
    "\n",
    "trans_eqtl_sqtl_frac_df.columns = ['cis-sQTL mediator', 'cis-sQTL', 'Not a cis-sQTL', 'Not tested for cis']\n",
    "trans_eqtl_sqtl_frac_df['sample_size'] = summary_df.loc[trans_eqtl_sqtl_frac_df.index, 'sample_size']\n",
    "trans_eqtl_sqtl_frac_df['trans_genes'] = summary_df['trans_egenes']\n",
    "# add background\n",
    "background_df = pd.DataFrame(np.array([[155, 1416-155, 1620-1416]])/1620, \n",
    "                             columns=['cis-sQTL', 'Not a cis-sQTL', 'Not tested for cis'], \n",
    "                             index=['Background'])\n",
    "trans_eqtl_sqtl_frac_df = trans_eqtl_sqtl_frac_df.append(background_df, sort=False).fillna(0)\n",
    "plot_summary(trans_eqtl_sqtl_frac_df, phenotype='trans-eGenes')\n",
    "plt.savefig('pdfs/Fig4Dsup_trans_eqtl_mediation_cis_sqtl.pdf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing Whole_Bloodtine_Terminal_Ileum"
     ]
    }
   ],
   "source": [
    "# load trans-sQTL results (0.05 FDR)\n",
    "trans_sqtl_df = pd.read_csv('data/GTEx_Analysis_v8_trans_sGenes_fdr05.txt', sep='\\t')\n",
    "for c in ['tested_cis_eqtl', 'tested_cis_sqtl', 'cis_eQTL', 'cis_sQTL']:\n",
    "    trans_sqtl_df[c] = False\n",
    "\n",
    "for tissue_id, gdf in trans_sqtl_df.groupby('tissue_id'):\n",
    "    print('\\rProcessing {}'.format(tissue_id), end='')\n",
    "    bed_df = pd.read_csv('data/GTEx_Analysis_v8_eQTL_expression_matrices/{}.v8.normalized_expression.bed.gz'.format(tissue_id),\n",
    "                    sep='\\t', usecols=['#chr', 'end', 'gene_id'], index_col=2).rename(columns={'end':'tss'})\n",
    "    bed_df['biotype'] = bed_df.index.map(lambda x: annot.gene_dict[x].type)\n",
    "    bed_df = bed_df[bed_df['biotype'].isin(['protein_coding', 'lincRNA'])]\n",
    "\n",
    "    group_s = pd.read_csv('data/GTEx_Analysis_v8_sQTL_groups/Whole_Blood.leafcutter.phenotype_groups.txt'.format(tissue_id),\n",
    "                          sep='\\t', index_col=0, header=None, squeeze=True)\n",
    "    sqtl_gene_ids = group_s.unique()\n",
    "    \n",
    "    signif_eqtl_df = pd.read_csv('data/GTEx_Analysis_v8_eQTL/{}.v8.signif_variant_gene_pairs.txt.gz'.format(tissue_id),\n",
    "                                 sep='\\t', usecols=['variant_id', 'gene_id'])\n",
    "    signif_eqtl_df = signif_eqtl_df[signif_eqtl_df['gene_id'].apply(lambda x: x in gene_set)]\n",
    "\n",
    "    signif_sqtl_df = pd.read_csv('data/GTEx_Analysis_v8_sQTL/{}.v8.sqtl_signifpairs.txt.gz'.format(tissue_id),\n",
    "                                 sep='\\t', usecols=['variant_id', 'phenotype_id'])\n",
    "    signif_sqtl_df['gene_id'] = signif_sqtl_df['phenotype_id'].apply(lambda x: x.split(':')[-1])\n",
    "    signif_sqtl_df = signif_sqtl_df[signif_sqtl_df['gene_id'].apply(lambda x: x in gene_set)]\n",
    "\n",
    "    for i,r in gdf.iterrows():\n",
    "        chrom, pos = r['variant_id'].split('_')[:2]\n",
    "        pos = int(pos)\n",
    "        tss_s = bed_df.loc[bed_df['#chr']==chrom, 'tss']\n",
    "        if (tss_s-pos).abs().min() <= 1000000:\n",
    "            trans_sqtl_df.loc[i, 'tested_cis_eqtl'] = True\n",
    "        tss_s = pd.Series([annot.gene_dict[x].tss for x in sqtl_gene_ids], index=sqtl_gene_ids)\n",
    "        if (tss_s.loc[sqtl_gene_ids]-pos).abs().min() <= 1000000:\n",
    "            trans_sqtl_df.loc[i, 'tested_cis_sqtl'] = True\n",
    "        if r['variant_id'] in signif_eqtl_df['variant_id'].values:\n",
    "            trans_sqtl_df.loc[i, 'cis_eQTL'] = True\n",
    "        if r['variant_id'] in signif_sqtl_df['variant_id'].values:\n",
    "            trans_sqtl_df.loc[i, 'cis_sQTL'] = True    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# summarize by tissue\n",
    "trans_sqtl_summary_df = []\n",
    "for tissue_id, gdf in trans_sqtl_df.groupby('tissue_id'):\n",
    "    sum_s = gdf[['tested_cis_eqtl', 'tested_cis_sqtl', 'cis_eQTL', 'cis_sQTL']].sum()\n",
    "    sum_s['trans_sgenes'] = gdf.shape[0]\n",
    "    sum_s.name = tissue_id\n",
    "    trans_sqtl_summary_df.append(sum_s)\n",
    "trans_sqtl_summary_df = pd.concat(trans_sqtl_summary_df, axis=1).T.astype(int).sort_values('tested_cis_eqtl', ascending=False)\n",
    "trans_sqtl_summary_df['sample_size'] = size_s.loc[trans_sqtl_summary_df.index]\n",
    "trans_sqtl_summary_df.index = style.colors_df.loc[trans_sqtl_summary_df.index, 'tissue_abbrv']\n",
    "\n",
    "# add mediation results: cis-eQTLs\n",
    "mediation_df = pd.read_excel('/xchip/gtex/francois/manuscripts/gtex_v8/data/GTEx_v8_mediation_filtered.xlsx',\n",
    "                             sheet_name='GTEx_v8_trans_sQTL_mediation_by')\n",
    "mediation_df = mediation_df[mediation_df['BH_FDR']<0.05]\n",
    "med_s = mediation_df[['trans_gene', 'tissue']].drop_duplicates()['tissue'].value_counts()\n",
    "med_s.index = style.colors_df.loc[med_s.index, 'tissue_abbrv']\n",
    "med_s.name = 'cis_eqtl_mediator'\n",
    "trans_sqtl_summary_df = trans_sqtl_summary_df.join(med_s, how='left').fillna(0).astype(int)\n",
    "\n",
    "trans_sqtl_summary_df = trans_sqtl_summary_df.sort_values('sample_size')\n",
    "\n",
    "# add mean across tissues\n",
    "mean_s = trans_sqtl_summary_df.mean()\n",
    "mean_s.name = 'All tissues (mean)'\n",
    "trans_sqtl_summary_df = trans_sqtl_summary_df.append(mean_s)\n",
    "\n",
    "subset_df = trans_sqtl_summary_df[trans_sqtl_summary_df['trans_sgenes']>0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 529.2x252 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "trans_sqtl_frac_df = (pd.concat([\n",
    "    subset_df['cis_eqtl_mediator'],\n",
    "    subset_df['cis_eQTL'] - subset_df['cis_eqtl_mediator'] ,\n",
    "    subset_df['tested_cis_eqtl'] - subset_df['cis_eQTL'],\n",
    "    subset_df['trans_sgenes'] - subset_df['tested_cis_eqtl'],\n",
    "], axis=1).T / subset_df['trans_sgenes']).T\n",
    "trans_sqtl_frac_df.columns = ['cis-eQTL mediator', 'cis-eQTL', 'Not a cis-eQTL', 'Not tested for cis']\n",
    "trans_sqtl_frac_df['sample_size'] = trans_sqtl_summary_df.loc[trans_sqtl_frac_df.index, 'sample_size']\n",
    "trans_sqtl_frac_df['trans_genes'] = trans_sqtl_summary_df['trans_sgenes']\n",
    "# add background\n",
    "background_df = pd.DataFrame(np.array([[171, 1252-171, 1300-1252]])/1300, \n",
    "                             columns=['cis-eQTL', 'Not a cis-eQTL', 'Not tested for cis'], \n",
    "                             index=['Background'])\n",
    "trans_sqtl_frac_df = trans_sqtl_frac_df.append(background_df, sort=False).fillna(0)\n",
    "plot_summary(trans_sqtl_frac_df, phenotype='trans-sGenes')\n",
    "plt.savefig('pdfs/Fig4Dsup_trans_sqtl_mediation_cis_eqtl.pdf')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "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": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}