Hail_GWAS_tutorial.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "from pyspark import SparkConf, SparkContext\n",
    "from hail import *\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.patches as mpatches\n",
    "import numpy as np\n",
    "from collections import Counter\n",
    "from math import log, isnan\n",
    "from pprint import pprint"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:sparkhpc.sparkjob:Submitted batch job 7302\n",
      "\n",
      "INFO:sparkhpc.sparkjob:Submitted cluster 0\n"
     ]
    }
   ],
   "source": [
    "!sparkcluster start \\\n",
    "--walltime ##:##  \\\n",
    "--queue NAME \\\n",
    "--memory-per-core #### \\\n",
    "--cores-per-executor ## \\\n",
    "##"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "os.environ['SPARK_MASTER_URL'] = 'PATH'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "sc = SparkContext(conf=SparkConf().setAll([\n",
    "   (u'spark.sql.files.maxPartitionBytes', u'#########'),\n",
    "   (u'spark.sql.files.openCostInBytes', u'#########'),\n",
    "   (u'spark.app.name', u'HAIL'),\n",
    "   (u'spark.master', os.environ['SPARK_MASTER_URL']),\n",
    "   (u'spark.submit.deployMode', '####'),\n",
    "   (u'spark.driver.maxResultSize', '##'),\n",
    "   (u'spark.executor.instances', '##')\n",
    "\n",
    "    ]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Running on Apache Spark version 2.1.1\n",
      "SparkUI available at http://172.26.8.132:4041\n",
      "Welcome to\n",
      "     __  __     <>__\n",
      "    / /_/ /__  __/ /\n",
      "   / __  / _ `/ / /\n",
      "  /_/ /_/\\_,_/_/_/   version 0.1-5306854\n"
     ]
    }
   ],
   "source": [
    "#start the Hail context\n",
    "hc = HailContext()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "#load data from disk\n",
    "vds = hc.read('/PATH/TO/gnomAD.vds')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "vtable = vds.variants_table()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Struct{\n",
      "     v: Variant,\n",
      "     va: Struct{\n",
      "         rsid: String,\n",
      "         qual: Double,\n",
      "         filters: Set[String],\n",
      "         info: Struct{\n",
      "             AC: Array[Int],\n",
      "             AF: Array[Double],\n",
      "             AN: Int,\n",
      "             BaseQRankSum: Double,\n",
      "             ClippingRankSum: Double,\n",
      "             DB: Boolean,\n",
      "             DP: Int,\n",
      "             FS: Double,\n",
      "             InbreedingCoeff: Double,\n",
      "             MQ: Double,\n",
      "             MQRankSum: Double,\n",
      "             QD: Double,\n",
      "             ReadPosRankSum: Double,\n",
      "             SOR: Double,\n",
      "             VQSLOD: Double,\n",
      "             VQSR_culprit: String,\n",
      "             VQSR_NEGATIVE_TRAIN_SITE: Boolean,\n",
      "             VQSR_POSITIVE_TRAIN_SITE: Boolean,\n",
      "             GQ_HIST_ALT: Array[String],\n",
      "             DP_HIST_ALT: Array[String],\n",
      "             AB_HIST_ALT: Array[String],\n",
      "             GQ_HIST_ALL: String,\n",
      "             DP_HIST_ALL: String,\n",
      "             AB_HIST_ALL: String,\n",
      "             AC_AFR: Array[Int],\n",
      "             AC_AMR: Array[Int],\n",
      "             AC_ASJ: Array[Int],\n",
      "             AC_EAS: Array[Int],\n",
      "             AC_FIN: Array[Int],\n",
      "             AC_NFE: Array[Int],\n",
      "             AC_OTH: Array[Int],\n",
      "             AC_SAS: Array[Int],\n",
      "             AC_Male: Array[Int],\n",
      "             AC_Female: Array[Int],\n",
      "             AN_AFR: Int,\n",
      "             AN_AMR: Int,\n",
      "             AN_ASJ: Int,\n",
      "             AN_EAS: Int,\n",
      "             AN_FIN: Int,\n",
      "             AN_NFE: Int,\n",
      "             AN_OTH: Int,\n",
      "             AN_SAS: Int,\n",
      "             AN_Male: Int,\n",
      "             AN_Female: Int,\n",
      "             AF_AFR: Array[Double],\n",
      "             AF_AMR: Array[Double],\n",
      "             AF_ASJ: Array[Double],\n",
      "             AF_EAS: Array[Double],\n",
      "             AF_FIN: Array[Double],\n",
      "             AF_NFE: Array[Double],\n",
      "             AF_OTH: Array[Double],\n",
      "             AF_SAS: Array[Double],\n",
      "             AF_Male: Array[Double],\n",
      "             AF_Female: Array[Double],\n",
      "             GC_AFR: Array[Int],\n",
      "             GC_AMR: Array[Int],\n",
      "             GC_ASJ: Array[Int],\n",
      "             GC_EAS: Array[Int],\n",
      "             GC_FIN: Array[Int],\n",
      "             GC_NFE: Array[Int],\n",
      "             GC_OTH: Array[Int],\n",
      "             GC_SAS: Array[Int],\n",
      "             GC_Male: Array[Int],\n",
      "             GC_Female: Array[Int],\n",
      "             AC_raw: Array[Int],\n",
      "             AN_raw: Int,\n",
      "             AF_raw: Array[Double],\n",
      "             GC_raw: Array[Int],\n",
      "             GC: Array[Int],\n",
      "             Hom_AFR: Array[Int],\n",
      "             Hom_AMR: Array[Int],\n",
      "             Hom_ASJ: Array[Int],\n",
      "             Hom_EAS: Array[Int],\n",
      "             Hom_FIN: Array[Int],\n",
      "             Hom_NFE: Array[Int],\n",
      "             Hom_OTH: Array[Int],\n",
      "             Hom_SAS: Array[Int],\n",
      "             Hom_Male: Array[Int],\n",
      "             Hom_Female: Array[Int],\n",
      "             Hom_raw: Array[Int],\n",
      "             Hom: Array[Int],\n",
      "             STAR_AC: Int,\n",
      "             STAR_AC_raw: Int,\n",
      "             STAR_Hom: Int,\n",
      "             POPMAX: Array[String],\n",
      "             AC_POPMAX: Array[Int],\n",
      "             AN_POPMAX: Array[Int],\n",
      "             AF_POPMAX: Array[Double],\n",
      "             DP_MEDIAN: Array[Int],\n",
      "             DREF_MEDIAN: Array[Double],\n",
      "             GQ_MEDIAN: Array[Int],\n",
      "             AB_MEDIAN: Array[Double],\n",
      "             AS_RF: Array[Double],\n",
      "             AS_FilterStatus: Array[Set[String]],\n",
      "             AS_RF_POSITIVE_TRAIN: Array[Int],\n",
      "             AS_RF_NEGATIVE_TRAIN: Array[Int],\n",
      "             CSQ: Array[String],\n",
      "             AN_FIN_Male: Int,\n",
      "             AN_EAS_Female: Int,\n",
      "             AN_NFE_Female: Int,\n",
      "             AC_AFR_Male: Array[Int],\n",
      "             AN_AMR_Female: Int,\n",
      "             AF_AMR_Male: Array[Double],\n",
      "             Hemi_NFE: Array[Int],\n",
      "             Hemi_AFR: Array[Int],\n",
      "             AC_ASJ_Female: Array[Int],\n",
      "             AF_FIN_Female: Array[Double],\n",
      "             AN_ASJ_Male: Int,\n",
      "             AC_OTH_Female: Array[Int],\n",
      "             GC_OTH_Male: Array[Int],\n",
      "             GC_FIN_Male: Array[Int],\n",
      "             AC_NFE_Female: Array[Int],\n",
      "             AC_EAS_Male: Array[Int],\n",
      "             AC_OTH_Male: Array[Int],\n",
      "             GC_SAS_Male: Array[Int],\n",
      "             Hemi_AMR: Array[Int],\n",
      "             AC_NFE_Male: Array[Int],\n",
      "             Hemi: Array[Int],\n",
      "             AN_FIN_Female: Int,\n",
      "             GC_EAS_Male: Array[Int],\n",
      "             GC_ASJ_Female: Array[Int],\n",
      "             GC_SAS_Female: Array[Int],\n",
      "             GC_ASJ_Male: Array[Int],\n",
      "             Hemi_SAS: Array[Int],\n",
      "             AN_ASJ_Female: Int,\n",
      "             AF_FIN_Male: Array[Double],\n",
      "             AN_OTH_Male: Int,\n",
      "             AF_AFR_Male: Array[Double],\n",
      "             STAR_Hemi: Int,\n",
      "             AF_SAS_Male: Array[Double],\n",
      "             Hemi_ASJ: Array[Int],\n",
      "             AN_SAS_Female: Int,\n",
      "             AN_AFR_Female: Int,\n",
      "             Hemi_raw: Array[Int],\n",
      "             AF_OTH_Male: Array[Double],\n",
      "             AC_SAS_Female: Array[Int],\n",
      "             AF_NFE_Female: Array[Double],\n",
      "             AF_EAS_Female: Array[Double],\n",
      "             AN_OTH_Female: Int,\n",
      "             AF_EAS_Male: Array[Double],\n",
      "             AF_SAS_Female: Array[Double],\n",
      "             GC_AFR_Female: Array[Int],\n",
      "             AF_AFR_Female: Array[Double],\n",
      "             AC_FIN_Female: Array[Int],\n",
      "             Hemi_OTH: Array[Int],\n",
      "             GC_AMR_Male: Array[Int],\n",
      "             AC_AFR_Female: Array[Int],\n",
      "             GC_NFE_Male: Array[Int],\n",
      "             AF_AMR_Female: Array[Double],\n",
      "             GC_NFE_Female: Array[Int],\n",
      "             AN_AFR_Male: Int,\n",
      "             AN_NFE_Male: Int,\n",
      "             AC_AMR_Male: Array[Int],\n",
      "             GC_AMR_Female: Array[Int],\n",
      "             AC_SAS_Male: Array[Int],\n",
      "             AF_ASJ_Male: Array[Double],\n",
      "             GC_FIN_Female: Array[Int],\n",
      "             AC_EAS_Female: Array[Int],\n",
      "             AC_AMR_Female: Array[Int],\n",
      "             Hemi_FIN: Array[Int],\n",
      "             AC_FIN_Male: Array[Int],\n",
      "             GC_EAS_Female: Array[Int],\n",
      "             AF_ASJ_Female: Array[Double],\n",
      "             AF_OTH_Female: Array[Double],\n",
      "             GC_AFR_Male: Array[Int],\n",
      "             AN_SAS_Male: Int,\n",
      "             AF_NFE_Male: Array[Double],\n",
      "             AN_EAS_Male: Int,\n",
      "             AC_ASJ_Male: Array[Int],\n",
      "             Hemi_EAS: Array[Int],\n",
      "             AN_AMR_Male: Int,\n",
      "             GC_OTH_Female: Array[Int],\n",
      "             segdup: Boolean,\n",
      "             lcr: Boolean\n",
      "         },\n",
      "         vep: Struct{\n",
      "             assembly_name: String,\n",
      "             allele_string: String,\n",
      "             ancestral: String,\n",
      "             colocated_variants: Array[Struct{\n",
      "                 aa_allele: String,\n",
      "                 aa_maf: Double,\n",
      "                 afr_allele: String,\n",
      "                 afr_maf: Double,\n",
      "                 allele_string: String,\n",
      "                 amr_allele: String,\n",
      "                 amr_maf: Double,\n",
      "                 clin_sig: Array[String],\n",
      "                 end: Int,\n",
      "                 eas_allele: String,\n",
      "                 eas_maf: Double,\n",
      "                 ea_allele: String,\n",
      "                 ea_maf: Double,\n",
      "                 eur_allele: String,\n",
      "                 eur_maf: Double,\n",
      "                 exac_adj_allele: String,\n",
      "                 exac_adj_maf: Double,\n",
      "                 exac_allele: String,\n",
      "                 exac_afr_allele: String,\n",
      "                 exac_afr_maf: Double,\n",
      "                 exac_amr_allele: String,\n",
      "                 exac_amr_maf: Double,\n",
      "                 exac_eas_allele: String,\n",
      "                 exac_eas_maf: Double,\n",
      "                 exac_fin_allele: String,\n",
      "                 exac_fin_maf: Double,\n",
      "                 exac_maf: Double,\n",
      "                 exac_nfe_allele: String,\n",
      "                 exac_nfe_maf: Double,\n",
      "                 exac_oth_allele: String,\n",
      "                 exac_oth_maf: Double,\n",
      "                 exac_sas_allele: String,\n",
      "                 exac_sas_maf: Double,\n",
      "                 id: String,\n",
      "                 minor_allele: String,\n",
      "                 minor_allele_freq: Double,\n",
      "                 phenotype_or_disease: Int,\n",
      "                 pubmed: Array[Int],\n",
      "                 sas_allele: String,\n",
      "                 sas_maf: Double,\n",
      "                 somatic: Int,\n",
      "                 start: Int,\n",
      "                 strand: Int\n",
      "             }],\n",
      "             context: String,\n",
      "             end: Int,\n",
      "             id: String,\n",
      "             input: String,\n",
      "             intergenic_consequences: Array[Struct{\n",
      "                 allele_num: Int,\n",
      "                 consequence_terms: Array[String],\n",
      "                 impact: String,\n",
      "                 minimised: Int,\n",
      "                 variant_allele: String\n",
      "             }],\n",
      "             most_severe_consequence: String,\n",
      "             motif_feature_consequences: Array[Struct{\n",
      "                 allele_num: Int,\n",
      "                 consequence_terms: Array[String],\n",
      "                 high_inf_pos: String,\n",
      "                 impact: String,\n",
      "                 minimised: Int,\n",
      "                 motif_feature_id: String,\n",
      "                 motif_name: String,\n",
      "                 motif_pos: Int,\n",
      "                 motif_score_change: Double,\n",
      "                 strand: Int,\n",
      "                 variant_allele: String\n",
      "             }],\n",
      "             regulatory_feature_consequences: Array[Struct{\n",
      "                 allele_num: Int,\n",
      "                 biotype: String,\n",
      "                 consequence_terms: Array[String],\n",
      "                 impact: String,\n",
      "                 minimised: Int,\n",
      "                 regulatory_feature_id: String,\n",
      "                 variant_allele: String\n",
      "             }],\n",
      "             seq_region_name: String,\n",
      "             start: Int,\n",
      "             strand: Int,\n",
      "             transcript_consequences: Array[Struct{\n",
      "                 allele_num: Int,\n",
      "                 amino_acids: String,\n",
      "                 biotype: String,\n",
      "                 canonical: Int,\n",
      "                 ccds: String,\n",
      "                 cdna_start: Int,\n",
      "                 cdna_end: Int,\n",
      "                 cds_end: Int,\n",
      "                 cds_start: Int,\n",
      "                 codons: String,\n",
      "                 consequence_terms: Array[String],\n",
      "                 distance: Int,\n",
      "                 domains: Array[Struct{\n",
      "                     db: String,\n",
      "                     name: String\n",
      "                 }],\n",
      "                 exon: String,\n",
      "                 gene_id: String,\n",
      "                 gene_pheno: Int,\n",
      "                 gene_symbol: String,\n",
      "                 gene_symbol_source: String,\n",
      "                 hgnc_id: Int,\n",
      "                 hgvsc: String,\n",
      "                 hgvsp: String,\n",
      "                 hgvs_offset: Int,\n",
      "                 impact: String,\n",
      "                 intron: String,\n",
      "                 lof: String,\n",
      "                 lof_flags: String,\n",
      "                 lof_filter: String,\n",
      "                 lof_info: String,\n",
      "                 minimised: Int,\n",
      "                 polyphen_prediction: String,\n",
      "                 polyphen_score: Double,\n",
      "                 protein_end: Int,\n",
      "                 protein_start: Int,\n",
      "                 protein_id: String,\n",
      "                 sift_prediction: String,\n",
      "                 sift_score: Double,\n",
      "                 strand: Int,\n",
      "                 swissprot: String,\n",
      "                 transcript_id: String,\n",
      "                 trembl: String,\n",
      "                 uniparc: String,\n",
      "                 variant_allele: String\n",
      "             }],\n",
      "             variant_class: String\n",
      "         }\n",
      "     }\n",
      " }\n"
     ]
    }
   ],
   "source": [
    "#look at the structure of the table using \"schema\"\n",
    "pprint(vtable.schema)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "#split multi-allelic variants\n",
    "vds = vds.split_multi()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#map the variants to their alternate allele, filter to SNPs, and count by unique ref/alt pair\n",
    "snp_counts = vds.query_variants('variants.map(v => v.altAllele()).filter(aa => aa.isSNP()).counter()')\n",
    "pprint(Counter(snp_counts).most_common())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZMAAAD8CAYAAACyyUlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4xLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvAOZPmwAAFGhJREFUeJzt3X+sX/V93/HnayaklC4xPyxEbZi9xlpFUNcQC6hSdRFsYJJqphJBsLW4GYs3BbZ0TGpM94e7ZEhk60KDlFB5sRdTpXEskhZrceJYhCitNAgmofxsyh2BYguwiw00yxJG8t4f34+zL7ffey++H9tff6+fD+mre877fM75nKOD74vPOed7bqoKSZJ6/J1x74AkafIZJpKkboaJJKmbYSJJ6maYSJK6GSaSpG6GiSSpm2EiSepmmEiSup007h04Vs4888xavnz5uHdDkibKgw8++NdVtWSudidMmCxfvpzdu3ePezckaaIkeeaNtPMylySpm2EiSepmmEiSuhkmkqRuc4ZJks1J9iV5dKj2X5L8RZKHk/xxksVDy25OMpXkO0kuH6qvbrWpJOuH6iuS3N/qn09ycqu/uc1PteXL5+pDkjQeb2Rk8hlg9bTaLuD8qvoF4C+BmwGSnAdcA7y9rfOpJIuSLAI+CVwBnAdc29oCfAy4rareBhwErm/164GDrX5bazdjH4d53JKkI2jOMKmqbwAHptW+WlWvtdn7gGVteg2wtap+WFXfBaaAC9tnqqqeqqpXga3AmiQBLgHuautvAa4c2taWNn0XcGlrP1MfkqQxORL3TP4F8OU2vRR4dmjZnlabqX4G8NJQMB2qv25bbfnLrf1M25IkjUlXmCT5D8BrwGePzO4cWUnWJdmdZPf+/fvHvTuStGDN+xvwSX4T+FXg0qqqVt4LnDPUbFmrMUP9RWBxkpPa6GO4/aFt7UlyEvDW1n62Pl6nqjYCGwFWrVpVo9osZMvXf2nW5U/f+t5jtCeSFrp5hUmS1cBvA/+oqr4/tGg78EdJPg78LLAS+CYQYGWSFQx+8V8D/LOqqiT3AlcxuI+yFrh7aFtrgf/Zln+ttZ+pjxPCXAEBhoSkY2/OMEnyOeDdwJlJ9gAbGDy99WZg1+CeOPdV1b+uqseSbAMeZ3D564aq+lHbzo3ATmARsLmqHmtdfBjYmuQ/Ad8GNrX6JuAPk0wxeADgGoDZ+pAkjcecYVJV144obxpRO9T+FuCWEfUdwI4R9acY8TRWVf0AeN/h9CFJGg+/AS9J6maYSJK6GSaSpG6GiSSpm2EiSepmmEiSuhkmkqRuhokkqZthIknqNu8XPWrh8H1fkno5MpEkdTNMJEndDBNJUjfDRJLUzTCRJHUzTCRJ3QwTSVI3w0SS1M0wkSR1M0wkSd0ME0lSN9/NdRzw3ViSJp0jE0lSN8NEktTNMJEkdTNMJEnd5gyTJJuT7Evy6FDt9CS7kjzZfp7W6klye5KpJA8nuWBonbWt/ZNJ1g7V35nkkbbO7Uky3z4kSePxRkYmnwFWT6utB+6pqpXAPW0e4ApgZfusA+6AQTAAG4CLgAuBDYfCobX5wNB6q+fThyRpfOYMk6r6BnBgWnkNsKVNbwGuHKrfWQP3AYuTnA1cDuyqqgNVdRDYBaxuy95SVfdVVQF3TtvW4fQhSRqT+d4zOauqnmvTzwNntemlwLND7fa02mz1PSPq8+lDkjQm3Tfg24iijsC+HPE+kqxLsjvJ7v379x+FPZMkwfy/Af9CkrOr6rl2iWlfq+8Fzhlqt6zV9gLvnlb/eqsvG9F+Pn38LVW1EdgIsGrVqqMaeCcKv60vaZT5jky2A4eeyFoL3D1Uv649cXUx8HK7VLUTuCzJae3G+2XAzrbslSQXt6e4rpu2rcPpQ5I0JnOOTJJ8jsGo4swkexg8lXUrsC3J9cAzwNWt+Q7gPcAU8H3g/QBVdSDJR4EHWruPVNWhm/ofZPDE2CnAl9uHw+1DkjQ+c4ZJVV07w6JLR7Qt4IYZtrMZ2Dyivhs4f0T9xcPtQ5I0Hn4DXpLUzTCRJHUzTCRJ3QwTSVI3w0SS1M0wkSR1M0wkSd0ME0lSN8NEktTNMJEkdTNMJEndDBNJUjfDRJLUzTCRJHUzTCRJ3QwTSVI3w0SS1M0wkSR1M0wkSd0ME0lSN8NEktTNMJEkdTNMJEndThr3DmjhWr7+S3O2efrW9x6DPZF0tDkykSR1M0wkSd26wiTJv0vyWJJHk3wuyU8lWZHk/iRTST6f5OTW9s1tfqotXz60nZtb/TtJLh+qr261qSTrh+oj+5Akjce8wyTJUuDfAquq6nxgEXAN8DHgtqp6G3AQuL6tcj1wsNVva+1Icl5b7+3AauBTSRYlWQR8ErgCOA+4trVllj4kSWPQe5nrJOCUJCcBPw08B1wC3NWWbwGubNNr2jxt+aVJ0upbq+qHVfVdYAq4sH2mquqpqnoV2AqsaevM1IckaQzm/TRXVe1N8nvAXwH/B/gq8CDwUlW91prtAZa26aXAs23d15K8DJzR6vcNbXp4nWen1S9q68zUx+skWQesAzj33HPnd6Dz5JNMkk4kPZe5TmMwqlgB/CxwKoPLVMeNqtpYVauqatWSJUvGvTuStGD1XOb6x8B3q2p/Vf1f4IvAu4DF7bIXwDJgb5veC5wD0Ja/FXhxuD5tnZnqL87ShyRpDHrC5K+Ai5P8dLuPcSnwOHAvcFVrsxa4u01vb/O05V+rqmr1a9rTXiuAlcA3gQeAle3JrZMZ3KTf3taZqQ9J0hjMO0yq6n4GN8G/BTzStrUR+DBwU5IpBvc3NrVVNgFntPpNwPq2nceAbQyC6CvADVX1o3ZP5EZgJ/AEsK21ZZY+JElj0PU6laraAGyYVn6KwZNY09v+AHjfDNu5BbhlRH0HsGNEfWQfkqTx8BvwkqRuhokkqZthIknqZphIkroZJpKkboaJJKmbYSJJ6maYSJK6GSaSpG6GiSSpm2EiSepmmEiSuhkmkqRuXW8Nlo4U/8yxNNkcmUiSuhkmkqRuhokkqZthIknqZphIkroZJpKkboaJJKmbYSJJ6maYSJK6GSaSpG6GiSSpW1eYJFmc5K4kf5HkiSS/lOT0JLuSPNl+ntbaJsntSaaSPJzkgqHtrG3tn0yydqj+ziSPtHVuT5JWH9mHJGk8ekcmnwC+UlU/D/xD4AlgPXBPVa0E7mnzAFcAK9tnHXAHDIIB2ABcBFwIbBgKhzuADwytt7rVZ+pDkjQG8w6TJG8FfgXYBFBVr1bVS8AaYEtrtgW4sk2vAe6sgfuAxUnOBi4HdlXVgao6COwCVrdlb6mq+6qqgDunbWtUH5KkMegZmawA9gP/Pcm3k3w6yanAWVX1XGvzPHBWm14KPDu0/p5Wm62+Z0SdWfqQJI1BT5icBFwA3FFV7wD+N9MuN7URRXX0MafZ+kiyLsnuJLv3799/NHdDkk5oPWGyB9hTVfe3+bsYhMsL7RIV7ee+tnwvcM7Q+stabbb6shF1ZunjdapqY1WtqqpVS5YsmddBSpLmNu8wqarngWeT/INWuhR4HNgOHHoiay1wd5veDlzXnuq6GHi5XaraCVyW5LR24/0yYGdb9kqSi9tTXNdN29aoPiRJY9D7Z3v/DfDZJCcDTwHvZxBQ25JcDzwDXN3a7gDeA0wB329tqaoDST4KPNDafaSqDrTpDwKfAU4Bvtw+ALfO0IckaQy6wqSqHgJWjVh06Yi2Bdwww3Y2A5tH1HcD54+ovziqD0nSePgNeElSN8NEktTNMJEkdeu9AS8dc8vXf2nW5U/f+t5jtCeSDnFkIknqZphIkroZJpKkboaJJKmbYSJJ6maYSJK6GSaSpG6GiSSpm2EiSepmmEiSuhkmkqRuhokkqZsvejwMc71gEHzJoKQTkyMTSVI3w0SS1M0wkSR1M0wkSd0ME0lSN8NEktTNR4O1YPkot3TsODKRJHUzTCRJ3brDJMmiJN9O8j/a/Iok9yeZSvL5JCe3+pvb/FRbvnxoGze3+neSXD5UX91qU0nWD9VH9iFJGo8jMTL5EPDE0PzHgNuq6m3AQeD6Vr8eONjqt7V2JDkPuAZ4O7Aa+FQLqEXAJ4ErgPOAa1vb2fqQJI1BV5gkWQa8F/h0mw9wCXBXa7IFuLJNr2nztOWXtvZrgK1V9cOq+i4wBVzYPlNV9VRVvQpsBdbM0YckaQx6Rya/D/w28OM2fwbwUlW91ub3AEvb9FLgWYC2/OXW/if1aevMVJ+tD0nSGMw7TJL8KrCvqh48gvtzRCVZl2R3kt379+8f9+5I0oLVMzJ5F/BPkzzN4BLUJcAngMVJDn1/ZRmwt03vBc4BaMvfCrw4XJ+2zkz1F2fp43WqamNVraqqVUuWLJn/kUqSZjXvMKmqm6tqWVUtZ3AD/WtV9c+Be4GrWrO1wN1tenubpy3/WlVVq1/TnvZaAawEvgk8AKxsT26d3PrY3taZqQ9J0hgcje+ZfBi4KckUg/sbm1p9E3BGq98ErAeoqseAbcDjwFeAG6rqR+2eyI3ATgZPi21rbWfrQ5I0BkfkdSpV9XXg6236KQZPYk1v8wPgfTOsfwtwy4j6DmDHiPrIPiRJ4+E34CVJ3QwTSVI3w0SS1M0wkSR1M0wkSd3841gS/iEtqZcjE0lSN8NEktTNMJEkdTNMJEndDBNJUjfDRJLUzTCRJHUzTCRJ3QwTSVI3w0SS1M0wkSR1M0wkSd180aN0mHwppPS3OTKRJHUzTCRJ3QwTSVI3w0SS1M0wkSR1M0wkSd0ME0lSt3mHSZJzktyb5PEkjyX5UKufnmRXkifbz9NaPUluTzKV5OEkFwxta21r/2SStUP1dyZ5pK1ze5LM1ockaTx6RiavAf++qs4DLgZuSHIesB64p6pWAve0eYArgJXtsw64AwbBAGwALgIuBDYMhcMdwAeG1lvd6jP1IUkag3mHSVU9V1XfatN/AzwBLAXWAFtasy3AlW16DXBnDdwHLE5yNnA5sKuqDlTVQWAXsLote0tV3VdVBdw5bVuj+pAkjcEReZ1KkuXAO4D7gbOq6rm26HngrDa9FHh2aLU9rTZbfc+IOrP0MX2/1jEYBXHuuece5lFJ/Xz1ik4U3Tfgk/wM8AXgt6rqleFlbURRvX3MZrY+qmpjVa2qqlVLliw5mrshSSe0rjBJ8iYGQfLZqvpiK7/QLlHRfu5r9b3AOUOrL2u12erLRtRn60OSNAY9T3MF2AQ8UVUfH1q0HTj0RNZa4O6h+nXtqa6LgZfbpaqdwGVJTms33i8DdrZlryS5uPV13bRtjepDkjQGPfdM3gX8BvBIkoda7XeAW4FtSa4HngGubst2AO8BpoDvA+8HqKoDST4KPNDafaSqDrTpDwKfAU4Bvtw+zNKHJGkM5h0mVfVnQGZYfOmI9gXcMMO2NgObR9R3A+ePqL84qg9J0nj4DXhJUjfDRJLUzT/bKx0n5vpOit9H0fHMkYkkqZthIknqZphIkroZJpKkboaJJKmbYSJJ6uajwdKE8bX2Oh45MpEkdTNMJEndDBNJUjfDRJLUzRvw0gLmzXodK45MJEndDBNJUjfDRJLUzXsmkgDvr6iPIxNJUjfDRJLUzctckg6bl8Q0nSMTSVI3RyaSjqq5RjGOYBYGw0TSccFLZ5NtosMkyWrgE8Ai4NNVdeuYd0nSMWDwHH8mNkySLAI+CfwTYA/wQJLtVfX4ePdM0vHE4Dk2JjZMgAuBqap6CiDJVmANYJhImpfDCR7vBb3eJIfJUuDZofk9wEVj2hdJGulIBtSxaDtfqaquDYxLkquA1VX1L9v8bwAXVdWNQ23WAeva7PnAo8d8R4++M4G/HvdOHAUe12TxuCbL4RzX36uqJXM1muSRyV7gnKH5Za32E1W1EdgIkGR3Va06drt3bHhck8Xjmiwe1xs3yV9afABYmWRFkpOBa4DtY94nSTohTezIpKpeS3IjsJPBo8Gbq+qxMe+WJJ2QJjZMAKpqB7DjDTbfeDT3ZYw8rsnicU0Wj+sNmtgb8JKk48ck3zORJB0nTogwSbI6yXeSTCVZP+79OVKSPJ3kkSQPJdk97v2ZrySbk+xL8uhQ7fQku5I82X6eNs59nI8Zjut3k+xt5+yhJO8Z5z7OR5Jzktyb5PEkjyX5UKtP9Dmb5bgm+pwl+akk30zy5+24/mOrr0hyf/u9+Pn2INP8+1nol7naa1f+kqHXrgDXLoTXriR5GlhVVRP9HHySXwG+B9xZVee32n8GDlTVre1/AE6rqg+Pcz8P1wzH9bvA96rq98a5bz2SnA2cXVXfSvJ3gQeBK4HfZILP2SzHdTUTfM6SBDi1qr6X5E3AnwEfAm4CvlhVW5P8AfDnVXXHfPs5EUYmP3ntSlW9Chx67YqOE1X1DeDAtPIaYEub3sLgH/VEmeG4Jl5VPVdV32rTfwM8weCNFBN9zmY5rolWA99rs29qnwIuAe5q9e7zdSKEyajXrkz8fyBNAV9N8mD7tv9CclZVPdemnwfOGufOHGE3Jnm4XQabqEtB0yVZDrwDuJ8FdM6mHRdM+DlLsijJQ8A+YBfwv4CXquq11qT79+KJECYL2S9X1QXAFcAN7bLKglODa7EL5XrsHcDPAb8IPAf81/Huzvwl+RngC8BvVdUrw8sm+ZyNOK6JP2dV9aOq+kUGbwq5EPj5I93HiRAmc752ZVJV1d72cx/wxwz+I1koXmjXsA9dy9435v05IqrqhfYP+8fAf2NCz1m79v4F4LNV9cVWnvhzNuq4Fso5A6iql4B7gV8CFic59F3D7t+LJ0KYLMjXriQ5td0kJMmpwGUsrBdZbgfWtum1wN1j3Jcj5tAv2+bXmMBz1m7obgKeqKqPDy2a6HM203FN+jlLsiTJ4jZ9CoOHkZ5gECpXtWbd52vBP80F0B7l+33+/2tXbhnzLnVL8vcZjEZg8CaDP5rU40ryOeDdDN5k+gKwAfgTYBtwLvAMcHVVTdTN7BmO690MLpcU8DTwr4buM0yEJL8M/CnwCPDjVv4dBvcXJvaczXJc1zLB5yzJLzC4wb6IwQBiW1V9pP0O2QqcDnwb+PWq+uG8+zkRwkSSdHSdCJe5JElHmWEiSepmmEiSuhkmkqRuhokkqZthIknqZphIkroZJpKkbv8P1AhD3ZUOtjsAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2aaafb09add0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Here we will use the hist aggregator to produce and plot a histogram of DP values for genotypes in our thousand genomes dataset.\n",
    "dp_hist = vds.query_genotypes('gs.map(g => g.dp).hist(0, 30, 30)')\n",
    "plt.xlim(0, 31)\n",
    "plt.bar(dp_hist.binEdges[1:], dp_hist.binFrequencies)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Struct{\n",
      "     Population: String,\n",
      "     SuperPopulation: String,\n",
      "     isFemale: Boolean,\n",
      "     PurpleHair: Boolean,\n",
      "     CaffeineConsumption: Int\n",
      " }\n"
     ]
    }
   ],
   "source": [
    "#QC is entirely based on the ability to understand the properties of a dataset. \n",
    "#Hail attempts to make this easier by providing the sample_qc method, which produces a set of \n",
    "#useful metrics as sample annotations.\n",
    "pprint(vds.sample_schema)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "vds = vds.sample_qc()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Struct{\n",
      "     Population: String,\n",
      "     SuperPopulation: String,\n",
      "     isFemale: Boolean,\n",
      "     PurpleHair: Boolean,\n",
      "     CaffeineConsumption: Int,\n",
      "     qc: Struct{\n",
      "         callRate: Double,\n",
      "         nCalled: Int,\n",
      "         nNotCalled: Int,\n",
      "         nHomRef: Int,\n",
      "         nHet: Int,\n",
      "         nHomVar: Int,\n",
      "         nSNP: Int,\n",
      "         nInsertion: Int,\n",
      "         nDeletion: Int,\n",
      "         nSingleton: Int,\n",
      "         nTransition: Int,\n",
      "         nTransversion: Int,\n",
      "         dpMean: Double,\n",
      "         dpStDev: Double,\n",
      "         gqMean: Double,\n",
      "         gqStDev: Double,\n",
      "         nNonRef: Int,\n",
      "         rTiTv: Double,\n",
      "         rHetHomVar: Double,\n",
      "         rInsertionDeletion: Double\n",
      "     }\n",
      " }\n"
     ]
    }
   ],
   "source": [
    "pprint(vds.sample_schema)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "#We can pull out these new metrics to a Pandas dataframe with one line of code:\n",
    "df = vds.samples_table().to_pandas()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>s</th>\n",
       "      <th>sa.Population</th>\n",
       "      <th>sa.SuperPopulation</th>\n",
       "      <th>sa.isFemale</th>\n",
       "      <th>sa.PurpleHair</th>\n",
       "      <th>sa.CaffeineConsumption</th>\n",
       "      <th>sa.qc.callRate</th>\n",
       "      <th>sa.qc.nCalled</th>\n",
       "      <th>sa.qc.nNotCalled</th>\n",
       "      <th>sa.qc.nHomRef</th>\n",
       "      <th>...</th>\n",
       "      <th>sa.qc.nTransition</th>\n",
       "      <th>sa.qc.nTransversion</th>\n",
       "      <th>sa.qc.dpMean</th>\n",
       "      <th>sa.qc.dpStDev</th>\n",
       "      <th>sa.qc.gqMean</th>\n",
       "      <th>sa.qc.gqStDev</th>\n",
       "      <th>sa.qc.nNonRef</th>\n",
       "      <th>sa.qc.rTiTv</th>\n",
       "      <th>sa.qc.rHetHomVar</th>\n",
       "      <th>sa.qc.rInsertionDeletion</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>HG00096</td>\n",
       "      <td>GBR</td>\n",
       "      <td>EUR</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>6</td>\n",
       "      <td>0.979017</td>\n",
       "      <td>10731</td>\n",
       "      <td>230</td>\n",
       "      <td>6321</td>\n",
       "      <td>...</td>\n",
       "      <td>5233</td>\n",
       "      <td>1190</td>\n",
       "      <td>4.566828</td>\n",
       "      <td>2.343927</td>\n",
       "      <td>22.506570</td>\n",
       "      <td>22.531111</td>\n",
       "      <td>4410</td>\n",
       "      <td>4.397479</td>\n",
       "      <td>1.190760</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>HG00097</td>\n",
       "      <td>GBR</td>\n",
       "      <td>EUR</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>6</td>\n",
       "      <td>0.993705</td>\n",
       "      <td>10892</td>\n",
       "      <td>69</td>\n",
       "      <td>6152</td>\n",
       "      <td>...</td>\n",
       "      <td>5279</td>\n",
       "      <td>1251</td>\n",
       "      <td>9.440750</td>\n",
       "      <td>3.653101</td>\n",
       "      <td>42.334833</td>\n",
       "      <td>28.494689</td>\n",
       "      <td>4740</td>\n",
       "      <td>4.219824</td>\n",
       "      <td>1.648045</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HG00099</td>\n",
       "      <td>GBR</td>\n",
       "      <td>EUR</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>6</td>\n",
       "      <td>0.988687</td>\n",
       "      <td>10837</td>\n",
       "      <td>124</td>\n",
       "      <td>6113</td>\n",
       "      <td>...</td>\n",
       "      <td>5377</td>\n",
       "      <td>1229</td>\n",
       "      <td>8.240004</td>\n",
       "      <td>3.720313</td>\n",
       "      <td>37.523761</td>\n",
       "      <td>28.130359</td>\n",
       "      <td>4724</td>\n",
       "      <td>4.375102</td>\n",
       "      <td>1.510096</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HG00100</td>\n",
       "      <td>GBR</td>\n",
       "      <td>EUR</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>6</td>\n",
       "      <td>0.999909</td>\n",
       "      <td>10960</td>\n",
       "      <td>1</td>\n",
       "      <td>6088</td>\n",
       "      <td>...</td>\n",
       "      <td>5400</td>\n",
       "      <td>1238</td>\n",
       "      <td>12.756478</td>\n",
       "      <td>3.912233</td>\n",
       "      <td>52.740967</td>\n",
       "      <td>27.669909</td>\n",
       "      <td>4872</td>\n",
       "      <td>4.361874</td>\n",
       "      <td>1.758777</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>HG00101</td>\n",
       "      <td>GBR</td>\n",
       "      <td>EUR</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "      <td>8</td>\n",
       "      <td>0.995347</td>\n",
       "      <td>10910</td>\n",
       "      <td>51</td>\n",
       "      <td>6229</td>\n",
       "      <td>...</td>\n",
       "      <td>5358</td>\n",
       "      <td>1212</td>\n",
       "      <td>6.573118</td>\n",
       "      <td>2.887462</td>\n",
       "      <td>30.568286</td>\n",
       "      <td>25.611360</td>\n",
       "      <td>4681</td>\n",
       "      <td>4.420792</td>\n",
       "      <td>1.478031</td>\n",
       "      <td>None</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 26 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         s sa.Population sa.SuperPopulation  sa.isFemale  sa.PurpleHair  \\\n",
       "0  HG00096           GBR                EUR        False          False   \n",
       "1  HG00097           GBR                EUR         True           True   \n",
       "2  HG00099           GBR                EUR         True          False   \n",
       "3  HG00100           GBR                EUR         True          False   \n",
       "4  HG00101           GBR                EUR        False           True   \n",
       "\n",
       "   sa.CaffeineConsumption  sa.qc.callRate  sa.qc.nCalled  sa.qc.nNotCalled  \\\n",
       "0                       6        0.979017          10731               230   \n",
       "1                       6        0.993705          10892                69   \n",
       "2                       6        0.988687          10837               124   \n",
       "3                       6        0.999909          10960                 1   \n",
       "4                       8        0.995347          10910                51   \n",
       "\n",
       "   sa.qc.nHomRef            ...             sa.qc.nTransition  \\\n",
       "0           6321            ...                          5233   \n",
       "1           6152            ...                          5279   \n",
       "2           6113            ...                          5377   \n",
       "3           6088            ...                          5400   \n",
       "4           6229            ...                          5358   \n",
       "\n",
       "   sa.qc.nTransversion  sa.qc.dpMean  sa.qc.dpStDev  sa.qc.gqMean  \\\n",
       "0                 1190      4.566828       2.343927     22.506570   \n",
       "1                 1251      9.440750       3.653101     42.334833   \n",
       "2                 1229      8.240004       3.720313     37.523761   \n",
       "3                 1238     12.756478       3.912233     52.740967   \n",
       "4                 1212      6.573118       2.887462     30.568286   \n",
       "\n",
       "   sa.qc.gqStDev  sa.qc.nNonRef  sa.qc.rTiTv  sa.qc.rHetHomVar  \\\n",
       "0      22.531111           4410     4.397479          1.190760   \n",
       "1      28.494689           4740     4.219824          1.648045   \n",
       "2      28.130359           4724     4.375102          1.510096   \n",
       "3      27.669909           4872     4.361874          1.758777   \n",
       "4      25.611360           4681     4.420792          1.478031   \n",
       "\n",
       "   sa.qc.rInsertionDeletion  \n",
       "0                      None  \n",
       "1                      None  \n",
       "2                      None  \n",
       "3                      None  \n",
       "4                      None  \n",
       "\n",
       "[5 rows x 26 columns]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Peek at the data\n",
    "df.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2aab0f0c3710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plot QC metrics\n",
    "plt.clf()\n",
    "plt.subplot(1, 2, 1)\n",
    "plt.hist(df[\"sa.qc.callRate\"], bins=np.arange(.75, 1.01, .01))\n",
    "plt.xlabel(\"Call Rate\")\n",
    "plt.ylabel(\"Frequency\")\n",
    "plt.xlim(.75, 1)\n",
    "\n",
    "plt.subplot(1, 2, 2)\n",
    "plt.hist(df[\"sa.qc.gqMean\"], bins = np.arange(0, 105, 5))\n",
    "plt.xlabel(\"Mean Sample GQ\")\n",
    "plt.ylabel(\"Frequency\")\n",
    "plt.xlim(0, 105)\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2aab0ede7f50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#these metrics are correlated: call rate and mean DP\n",
    "plt.scatter(df[\"sa.qc.dpMean\"], df[\"sa.qc.callRate\"],\n",
    "            alpha=0.1)\n",
    "plt.xlabel('Mean DP')\n",
    "plt.ylabel('Call Rate')\n",
    "plt.xlim(0, 20)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZkAAAEKCAYAAADAVygjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4xLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvAOZPmwAAIABJREFUeJzsvXmspWle3/f5Pc+7n+Wec7equlVdVd093T0zDHgGmiWyEk9IcMZWFJaQCKxgsBxPIpvIioUFJFEikRCciMiyDLI1JBgmisARUpKxBCEWTEMkPGR6YBamm+7pZaprv/vZ3v19n/zxnHPrVnVV9a3uul1dVc9HurrnvOd93/Oeuzzf97eLMQaHw+FwOI4D9aAvwOFwOByPLk5kHA6Hw3FsOJFxOBwOx7HhRMbhcDgcx4YTGYfD4XAcG05kHA6Hw3FsOJFxOBwOx7HhRMbhcDgcx4YTGYfD4XAcG96DvoD3g9XVVXP+/PkHfRkPnFdeeQWA55577gFficPheBj40pe+tG2MWXsv53gsROb8+fO8+OKLD/oyHjif/OQnAXjhhRce6HU4HI6HAxG58F7P4dxlDofD4Tg2nMg4HA6H49hwIuNwOByOY8OJjMPhcDiODScyDofD4Tg2nMg4HA6H49hwIuNwOByOY8OJjMPhcDiODScyDofD4Tg2nMg4HA6H49hwIuNwOByOY8OJjMPhcDiODScyDofD4Tg2nMg4HA6H49hwIuNwOByOY8OJjMPhcDiODScyDofD4Tg2nMg4HA6H49hwIuNwOByOY8M7zpOLyK8C/y6waYz52G1eF+AfAX8VSIGfMMb8yfy1Hwf+q/mu/50x5tfn278D+DUgBn4b+LvGGHOcn8PheL/Iq4adaUFRt4SeYqUbEvn6ptev7KVsTgqMMZxYitkYxDftc6d9B0lAON9vcW7gpvfrhB670+LgmDjQpFXDLK+ZZDVpVZGXLcudkI+c6nFqmDArasZ5TVpWaKXIy4aiagh9zXovZGOY3PH67vZZ32m/ox7/Xn7e75bjOu/DyLGKDFYMfgn47B1e/yvAM/Ov7wb+CfDdIrIM/DfA84ABviQinzPG7M33+VvAH2NF5lPA7xzjZ3A8xNz6z75YRC/upmzNCprW0BqDQjCtoWoNu2lBXtWE2kMJ1K3BGMPp5Q4fPtkD4GsXR3z18h5bkxLTGsqmoW4bekHAt50Z8B1PLeMpRdUYTGuYljVXRxlZ1eBroRN4dAKfXuQR+IqybhlnFVVjONEP8bRwaTflyn7OiX7EiUEEDVzcT9nPKjaWQgadkFFW8tbOjBP9iNDXFGXNKK+Z5BXX9jMmRU038ugEHm9sT+mGPm3bcnWUMclrOoGmEwX4npAWDXvTnDD00MA3d1Le3JzRizQr/YidacFoVhJoIasbECHxfZ450WE5DhAtvL45paxrYt/n6RNdVrsRG0spH97o07SGom7BGMZ5xctXJ+RlTRJ6RJ4iCj3OLieEWoHIwe9rZ1oQeJok0FSN4fJeSif0eGNrigESX+N7igs7M4ZJQD/26YQes6K+6fd++PlCYC/vpW879+k7iOJR/96u7Ge8tTOjE3msdUNaw5HOuzj2+ihDRO4q0O+WByF+xyoyxpg/FJHzd9nl+4HPzi2RL4jIQEROAZ8E/qUxZhdARP4l8CkReQHoG2O+MN/+WeAHcCLz2LCflnzxjR3+5MIO07yhn2h6UUheN1zez1ACnghLcQDGMMor+rFPHHhgDNfHGa0x7M1qRmnB5rSgblti32M5DmiMIS0bmvlC2NQNWjQnlkKSaJ//9xXf/qPOCtK8oaVhlFYUraEfatJOzb/4sxn/6s0d/vLHTpAEPldHGaNZidKKnVFG3hrECOv9gKqBcV7ahS5U5KUVJFpDN/TodXw2L2d8+aIhjjx6nsbzPF7dnDGIC3amFetdn+laD09gXDSc6AW8uT3jm9szlrsBgyTgyiijbgxvFlMi30MruDrOubqb8cRqh0Hokbct41mF0sJeWjHJKjwNO7OC17amGDFEnqIoW7RWpHnFSi/iTy5UnF1O2JmWeJ4Qe4pct/zZpRHfekbwlLD9asHHzw7xtPDG9oyXro4IBCI/4PJ+Ri/06FQN41nJqWHCmWFMa+DlKyNWexGBZz37gSfkVcMX39jhxFJMHGimRc2b21PWehFF3ZCVilevjTmz3KEXeUzz+qbnCzERIPD0TecGa9mdHib39He5sBrf2k1Jq5pBHOApxfVxwcmliMDTdz1vXjW8vjVlb1aQ+B4IXBnl5FXDU+u9+2Zd3W9RPQrHbcm8E6eBi4eeX5pvu9v2S7fZ/jZE5NPApwHOnj17/67YcezkVcObW1Neujxic1bQtA1lbRinJVfGBTvjnLyq2E1LdictAC0Qagh8iAMPEaEbaiJfM80qWg1l1dLxfPK6xtceW7MMATBQBi1X9jO0gihUZHlLWkANJH5LO4Yoq3ijbol8hTHQGsN+WjIr7HUb01A20It8xnnJl98a8ezJHmnZMC0bJnlBi6EsG1qEP7s8ZlZWeEqRBB4CGIHIU+xmJVVjODNIUAhRoLmwO8FTwnIS0Bh45WrJyX7EpKiYli2RrzkzjNialpR1Szf05z/PFq00m+MpZQWI4CtNVjZ0Qo80rykr+1jEcHEvY7UTsttYUalaKKqawNdkZUvVGnqB4Huaom4ZdDze2p0R+RptNK2ItUa0YnNSICKcHsbMyhqAqjXUjSEINLUxdAINIoxmFZ3YoxN6jLKKk0sxxsB0fqOwYJpXFHVLHGhEhKxsSAKfsmkpG8OsrOmEPrOiph/7b3u+EJPLeynnVzs3/e35WkjL5p7/Xi/vpeympX2/vGY/LQl9K2D7acmJfnTX8+5MC9KiphP4+HPRExFmc8vjXkXvTu9xv0T1XnjQInNsGGM+A3wG4Pnnn3cxmw8YtzPb7R3qNr/38nVevjqmqO0/ZVHXVA3Udc32zFDf4ZxVAzQgeU0AbFPh+6ABrQWthcksoxXAVBhjfbGmhaJuaFsoDQR5CwLl/Lyzyu6fltA0MBHoxQEI5BVUgA/ULfOFxBD5HldHGR860WWcVuxOC7RSRL5ityyZZAU7s4rYU/Rin72sQFD0I4/roxwRIfYVW+Oc3VnJSi/EtA2zomVaVGCgaQxVY/BFISIMY59Ley2dMEAJhJ5Q1y112xJ6wnhWU5uWvG5Y6YU0TUugYZRXhJ5QNi1aCWXV4FmNYFa1aMX8ByUUtT2mrg1aW8HoBppLs4pOoKlbgzJC00IvhLSsqeqWXuhZVxnQtoZAW1citASedRfmdUMPH08JWWV/90momd2yOKdVw1LiU7cGX9vrDn3FtKhZ7YYU8xuBfP5+tz4HKybG2J/fYrG1f0OG0Lu3fKjF4m0M+J4ijjzKqmWUVqz1QrKqecfzFnVLawyevnEtnhLK2hz83N4rRd2SBDdbLO9GVO+VBy0yl4EnDj0/M992GesyO7z9hfn2M7fZ3/EQsJ+WvL414dqo4K2dKeNZxfasJC0rRnnNNC/ZmpaMZxVlA2KgwVoT93KXYIC5cUFZWQHwWkOAIa2s6BSAYNMrDfZ99KFjA3PL+WqoGzAGPM9uK8sbctcw/2cyUDeGxmuJtccsb5gUNQLEcxfFJCuZ1S1aDNpTNGa+uGnDfl6hBZQSyrolq1t8T1FUDW1rF+K41fY9mgZVCR3fQ2EX7u1phac1S4lPVjWUGASDaWUen9HEgaKuauq2JWsMntKEWiFKyMqWyNMUtSEJNHuzgtj3CCNN0xiEFi2apm1RWuh49rjQFxDrGjPKUDUNrWhEC4PEJ/D1wSKrlBAHHtOyQgSKqsGIoIQD8Vjs2w190rKhrFt8LVSNQYCNQcy0sIujr4RpXqMUDJKA/bQkr9qDO/bQUzc9B/vzPrEUU85vZBbnLuvmnu/qF4t36Cnq1rAU+1wvcyuwjY8I73je0FMoEerG4M9Fr24NSsk9i97d3uN+iOq98qBF5nPAT4rIb2ID/yNjzFUR+V3gvxeR4Xy/vwz8rDFmV0TGIvI92MD/Xwf+8QO5csc7cthaKeqGt7ZnJKHHpZ0pX724z2tbM6q6ZHtaU1U3Fur8Pl5Di7U0aKyoVIvnWKE4fI94WMhKbqYFPLEiE3oKrVqKuqVpb7ze1KA860bztea5UwmeEprGEPgK07bsTG1SQVVbqwEDZp5YgLHfk8inaSGt7IITCMzKliTURK1B5vspUfiiGHZCtFJobT+DAiKt8T1h1fcBYTcr+c6nhmS1YXdcUDQt/dBjc1qy3PFoWrsIr3QCyp7Ppd0MrYQTS7F1iTU+dQu+CthNKwRDrBSnhwmzouKZ9R57ackgCQi0Iq8axmnFd50f8m1nB4zTik7g4WlhZ1rga2E1Dihqw86sZNCxiQKJp5gVNWeGMWXdIgIf3VhiVtSkZUPoKT66scTOtCD0NNOiQpRQ1DUfObVkg/yBx+50xrDTwRjztue3isnOtDg497uJTywW70EScG2UE3iKYeKzl5WMMxuveqcA/ko3ZD+rbEzG2JhMWjYsJ/5BksJ7ZaUbcnkvBd6bqN4rx53C/BtYi2RVRC5hM8Z8AGPMP8Vmh/1V4DVsCvPfmL+2KyL/LfDF+al+bpEEAPxtbqQw/w4u6P9AOCwggF19RQ6+j7OSC7sp/chnEPu8tTNjJy25Psr549e3ubA7xRjIS1DKLvwt3NEV9m5R3BAaaa210hzafpiFyCz2uZXAA1/B2ZWEsm4oqhbTtHR8aFtoDGiBpTjgoye7BL6PaQ0fOdUlCnxeurJPN9CsdiO2pjkKoR/5hFqYFVBjiLTHcyeXKKqKN7czEHun3gmFbuSzmxZohDOrMdPcuqhCTxFoxXI3YmNJmJUNa72Q9X6EMca6oxrDuZWEsm556cqIl66OaRrDh9cDlnsxrTH0Ep9e6NEa+Nef1by1k7I5yqhaw5NrJwg9zdVxztYopW5blGi0J3zHYIXzKwlx6HFxL2Wal8S+x9nlDmeWE/qxz9NrvYPsrvMrHU4NYjbHOeO04sxyQjfULHVCe1dtDK0BXzhY9AdJcNPvIvJtIF0pYa0X0Tk7PBCiONB84tzyHZ/fKibvdZFdLN6BpznRD9ma/198+OTSHdPLbyXyNU+vdbni64Psso2l6L5ml0W+5vQwec+ieq/I41Bi8vzzz5sXX3zxQV/GA+eTn/wkAC+88MJ7Os/hLJWmbfnzqyMu7GRUbUNWtKz1QuJAkQQ+WkEn9Pi/v36Nq7szRlnDa5tjZtWNRf5Oi/p7QWEFoWpvvIcWG3OR+XO4YbEooOuDr6Gcu8V8bWMwjYGVrkfke/SjgI1hjMGgRTGrGsqqYaUbMM5K0rJGKc2/8ewaw04IBpTYZIAo0FwbpXz90thmtTWtdS8ZQ1oa1vo+nzi7zJvbMxrTEvsKjQ2ex76N3VwZ59R1y0o3JC1qRlnFmUHMmZWY1W5MVjY8MYz5+LnlmxaPy3spreHAZZRXDV+/sk9VG86tdBh2AiJfU9YtSm5eeF3NxzvzqP6MRORLxpjn38s5HrS7zPEQcjhL5euXRnz5rRG705JxYZfsF9/coRMqvuX0kKJpubiT8vXLe5R1zSgzFO3NVsT9FpgQKxCiIPJsALtubCygK4ZZAZ6G2kBPrBUSBhD7HnHgUZuGXuiDCKYFpYX1TgQKnhhEJIGmNopuqDCNYVq1jPOK1V7MrKgJfc31cYHvaYZxwEo34NJuylPrHTYGCXvTio1BTFrVXN7LCT3Fx5djPnFume9+apX9tOTPr4z4xuaUumn5ticGTIqKV65O2OhH1nLxNaaBKND4WmzAWOBEL2ClF70tNfVWV4kS4dxKFzGGXhzYAHrdvi/uk0eRhZXgeDtOZBz3TDG/272wk/HCK1vMClv8tz0ryMuGqm3ZnDSMi4a8asmqmllZsZ/df0HR8y8lsNQRPO1RVjWR7zErKuJQ0wmsBVI1hp20IC8rlIYI6yLzNYSBtVRWugG92KcX+HieTcXdSkv6kU8vULQGunHI6UFML/a4up/znasdWwy4nfLy1X0EKNvWpt4GHnGgSUIPpRTPzPe9PsrZTyvWehHPnuwzTIKDzJ9BEvA9H1rj4+eWD4rz+nHAR08NCD1F0RjSsiLxPULPWjpZWVMbg9KaWVEfFDGeHiYHd9llY9jPsoPjYt+mIG9O7LZ+7L/NffKgaiscjw5OZBy35W7m/9Y041+9us31ScEbW1PysqFsWmZVg6/AIIyzmsbkiIJZWpHV715gPKzlo7F+eq0gDIWlyFob+1kFpuXssEegNVndUMxbnJwZxvieLdgLlCH2NHXcUjctDQYx2HoGpVhOAjaWO5zshTx7osvF3ZSsavHmAfG0bBl2Az52eoleaLO3zq1o9rISY4RZWTGIA3bSiqXYY7UXUWPYnVac7IfM8pqybjnRi6hqw3In4MwwQStFWTdvC/BGvuaptS5PrXXv+vMpGgMYQs8G1uvGsDPNqZOQlUMiMUx8uqHHJCvJ6pZ+5JMEmm7oHbz/rcLxoGorHI8OTmQcN3G3thgr3ZBXro34jT++wN60ZlqUZHXNxd2MUEHZgjePg5gW8mlJ086f38M1KG7EVPIWQgW+Zxe4bhjyxHJE0ypCH5Qork8yyqYl8DRV2xL6ghIP3zcIwlJkrYT9WYnSwko/QjDsphV11aAUDDs+y92Ac0sxZ9c6GITaCCf6EeOssinOSvCVcHkv44mhEM17c1Xb9gNebwzdyGcvLVFK2J+VRPOg/JNrHbSyabqVgY2lCETeFuBe/A52psVBP7CFlXEnP39aVjbTbC4EvidUjSItq9uKxKxqwMDqXNTuJhwPqrbC8ejgRMZxwKK1xYWdKW1tGBcVF3dSktDj6n7K3qzim9sT9tOSsrVFcbOipShhAgRYa2WRRns4sH+3rDEfiH0YVze2aWyA/vSSj+8rEm2r98+udvA8RaIVRWto6pZwXs8xKSoCZftpaWWtrxO9yKYMN4rNccZyEhBoj/20YBj7+J0II3BqKebUUkAttgiubuH8SoflTsC0qPnapX2MMVzbz0jCirY1fOLsEK0UZ5cTXt2cALYW5mNPDLi4k7I1yRFguRtSNi1PL3ff0cW0cE+1BkZpgRLFqK7wtbqjmyrxPUZ1RTVPJKjn/dh6vn9bkWjbt0v+nYTjQdVWOB4dnMg4Driyn7E3K2zle9vw8qUJl/YydqcFvoaiNkxyW+hmRMjymnHVHFgpJVYYAqyoCPP6Ee5syXjzfWfzoklf2QOVAmUg8IVBGHJqGLHWDcjLlqIxPHumjxbYnBaM04aO77EzKyiaBoPw5ErMSjfi9DBmZ1owzSvSQWILE1FM8pLI92nalmdP9VnvRwRamOUNSiuUaenN61X204rQ14RauDbK6cUBWsH2pODEknB6mFA0hsRTjHJrGXVCTdN4+J6w0guRI2ZxLiwPWwfi4XuKqmmZlTXLnfC21kY/9vG1YlbWZJVNTV3thsRzcblVJNS8PucwdxKOB1Vb4Xh0cCLzmHK7mMv1kQ0AX69yvvDGLmXd8Ob2BDWviy/bht2pvbuu2xYNVLfEWjxsuUxrbqQJ360pRqwhb2wdSzAXGN+zxwOURUN3SbM1ysiKmn7ksbGUsNKNuLQ7Y7UToqRinJX0Ys2TSYcg0Kx2QtuSJK95er1P6CmenOQY4LXrE5uMUNSEvk/kaZZCn0lRsdqLwBjq1hY/bk8Kro9yTiyFtsdY5HOqH1E1LTuzgm8/b1OF+5GHpzpEs4JXr03wPc25VdtaP9A21fnKXspT6727/l4WlkfZtMRzi2XRZuVO1sZCCJY74U1CcLjTMNwQiY6vMSI3VdHfSTgeVG2F49HBicxjyOGMISVwZZTxjesTdqclKz2bbtu2MClKMIZZXRP6iqxs5w0L2wP3161LXgMEGuLWZnzVjbVwFn9oC2vHzL/njRUhhd0QeKCUIhAbK6hamFUNq92QTuQxySpGRcXlvRmBVhhj79ojT/H69pTtNGcoIYaAU0sRSik2lmwb/OVuwDe3Z6z2IkQM+7OaFkM/8nlja0YnVDx/bpnWCNdK2wp/uRsySHKqpgURnjvZox8FlLVtdXI4RTjfSzk1SGwnAWO4ul+w0g2JfVuZ/qcX9ijq9q7xlQP3lFYHLUYWbVbuZG28kxDc+tpC6I4qHC491/FecCLzGLJwybTGcH1cUNQNV/ZT3tiasvN6RdO0KITr45JRWpNXNZ4WprmhMjfHVxbLksy/GmwBow3K21qVtrqRHRbNEwQ0EGlA2f0wttlk3NEEWtvuwEVNJ1II4CnFKG042Y9JAutOevZkHwXktcHXCh/QgccwCZhmFa9dn/IXn1kFkYMGnM+d7AOGpmnpRwGeFi7szNBa8aETPeLAp2wa/sLZIVps5lldt+ylFScHEZ3Ao2pa0rKxwfs5hxf6tjXsZRUbw4hu6FPUDduTgjDUthGn8e8YX1lYJTYFOadqFK0xrHbDu7qp7iYEd3rNCYfj/cCJzGPIwiVzYTfl0l7Gq9fGNG3LzjTn4m7K5b0M31eMpxW5HW1CjbGNeLm5saSPbRiJsW3q29oWOvYjH08LRdUQeC1NbVu7e0qxn1Z22FSk8X0FDVR1QxLCSuLTGkWghbXVLpgWJTaTqxdrzq8m7M1sIFyLwtPCflawOS5AhPV+xJlhQt20jLKarWnJk6udm0SgF/l8/FwIIgi2LUmgFWVre2WdXIoJPUVa2kV9pRvyxuaEWdXMK/qF5cRn45ZFevEeK92QL7y+TaAUxhi2J7Yd53Lsc3k/p2wMYmsn3+Y+O3yddRKSlhU93ycO9CNTRe54vHAi8xgSeopJXnNlL+Py3oy2NexMC97aydieleRlwzhtqNqbG0Uu2r8k85hJ3VhhQSAJIfbtTA8QNoYx3cgn9DTjrCKvatKioTa2BX3V2u6/cSDEnkfga7qRtpMqW9vqJPI1k7wiDn18UfgadqYledXwLetLVHVLURtOLUWkRc1KN6A1wvaspKwaDIaLuzOeP78M3HxHf7jFSugpinmX3mEnYG9WMitr2x6+sr2nnlrvHbltSORrzq50rFBUDU1rWOn4bM8qIl8T+5qqaXlrN71tb6oDsbq1P5zD8RDiROYxZKUbcuHCLoHWB1bBpf2cnbRga1pSVW9vrx8yb9MyT1LylXWLabGPu1GIJ4qVjofWVoCysmU4b2zoKcWZoWZ7WnF13DIMNZ1QUxmhbWx79OWOzyit6Sce6/0IJUKaV3RDzSirGGW27f2HT/V5YrlDWtREgW3d8s2tGf3Is9ZG0VC3ttjwdmltt2ZMLbr0JmHE1f0UJQottr5m4da6VzYG8bw5pba1NaMMEVjpBsjcglpU5a90w7cJGNz/0cAOx4PAicxjSOTbhevibsrXL4/ZnmRsTgtm1dvX5IVrzPPsfJe8saIS+rDSURilUAJaKXqhR9k0rHZjitZQ1i3bs5KO57G+FHJupUNnL0WLEHiKOLRxobJuUQoE4YnlhKXEZynxGSQBZ5cTtmclnlYUZcPTaz3OriZoUdRey8l+fFAR/4XXt1FK2a68gU/dNix33572e2ugfNGl9/WtCW0LcaQYJDcaRl7ZSzFwTwv+4feIAo+8bDg9iO2wrrqlbBpb6JnX5FXztnOLiKu0dzwSOJF5jGjnszRevjLiz69OeGNrDDRcHRVkd/DILAZ6ZfWNVORQ7BRIHxt7SXxFoD3yuibQHohwsh8Q+R77M9uOfaUTkoQea/2Ibz8/ZC+t0Cj2s4L9WUkS+pxcijm1FKGVTZE+uRSxFPtcG+UUdUPka1t7Y2xngTPD5CDV+cm1Lm/tpXY2C/NeZlHAaifg4m76NjfXQgQWqdy7s5KqNmwMY+Lgxr+Fr4XLe3Zhv9cFf/Eep7EuuZ1pQVY1BFpxcim2llpZ0O3Fbzv3pd0ZT97STsZV2jseRpzIPCbkVTNvbCkUdcMoK/jKpZFdgI/g8l/s4jNvfd8JiANNMK+zMDQoEdb7ISDszyo+9kTCsBNQlC2nhzGhrzi1tAQGhklEN7J/fpujnDi089rNfCZL2RgmecUkrwk9xVLsU9YGg+FkPzqwMvy5+y7yNc/M4yYGCLQiCTSbk4LQU7e1QG5t/qiVcGkv44nlGxZK1dgBYf6hsbhw7wv+Te6zQ7Upie/d9twi4irtHY8E7i/2MWFnWqBE7CjYxrA9Lrm4PWN72ty1WHKBxsZluiH0YmXb5gceO7OKcV4yKxsaY5iVFVvTnKw2dAKfD6312RjGrHRDuqHPIPFJy5pA2wwywY6Z7YTeXEha8qpFgHFuI0Pr/YhhJ5xntgm70+KgLf3hppIbg5jlTsCppZiTSxGjvDo4XuYuukU1/eJnsnBJyTwzDQybkxwzd+OVdcOJpXg+j/4G97rgL6waNZ94uJjZ0o/92557vRce1OIcvpb7NSXR4Xi/OFaREZFPicgrIvKaiPzMbV4/JyK/JyJfFZEXROTMfPu/KSJfPvSVi8gPzF/7NRF589BrHz/Oz/AokFfN3GVkFy0wvHp9xG5WH2kSZYgdPdyNQGlF7CnysmUvLannbqzAUxgMaWHmPbQMm+OC2Bc+fnZAUbW2eNFTPHuyh1IQejYt99xKTD2f3T6IPfLKpgl3fM2ZudUR+ZqTSxGBJ4zz6mCRPhwTWSzkVdPy5taUizspvnq7lbDI1irmFe+Hjz8zTKjr9iYh2BjE92XBX1zfU2vdg2tf6d5eTDaGyW1FyQX9HQ8bx+YuExEN/DLwfcAl4Isi8jljzEuHdvtF4LPGmF8Xke8FfgH4MWPM54GPz8+zjB3P/P8cOu7vG2N+67iu/VEirxre2JzMYw62Yv+PvrHJVy7tkx/R26OVTQBAQBshLRoQKBqP5U7AWj8GBA/DpGrmrqiWQcejG/sMkpBnTwjRPNYReooPrfcOdR2OubKXcnE3ZZxXDDshTwzj+XvfuA+yi3LEWi+6ayzEGMOZ5Q5xoClrw7VRxsklm4p82AK5XfNHrRRnVzpvO/9xtVY5SrW+w/Ewc5wxme8CXjPGvAEgIr8JfD9wWGQ+Cvy9+ePPA//nbc7zw8DvGGPSY7zWR5Yreym7aUU3sjGMWVHz+Zc32S+OdvyR9/y0AAAgAElEQVR6RzHLW1oDeQmB12AEIqUwraFsDFpZF09u7ALeDTyGnYBBHDDOa84IbAyTgxnvtxL5mo1hcpDBtYhZjPOKvGoOJjcepTnjYRfYsBNybZQfuNhWe9FNx99L88fjbK3i2rY4HmWO0112Grh46Pml+bbDfAX4ofnjHwR6IrJyyz4/AvzGLdt+fu5i+4ci4pzUdyCvGr5+ecT1ccbFnYy6tdllV0bZOx6rgPVESAKfKBACDwLfFjB2Ah/taWLPI/Y1u7OKji80rQ2SR4HmuVN9Qs+jE+qDOpDWQBLog/k0eXXDlLo1PhJ4in7kEwXePbmMDrvAIt8O6tqc5Hzh9W2+8tYe5aH3vFOcxLmkHI77x4POLvsp4JdE5CeAPwQuc6jnooicAr4V+N1Dx/wscA3bUf4zwE8DP3friUXk08CnAc6ePXs8V/8BZpE5VdQtW+OMGiGv7GjkrGwPWuzfisJmkCURnF/v46EoOgFKCZsTaxUkkcYTTVU3rHUjpmXJXl4zSEJ8BZ7WBJ6mbluqqj3SdMU7DceqmntzGR12geVVw7VxTlk3PLneZb0XsZNWmM0JT83ddc6KcDiOl+MUmcvAE4een5lvO8AYc4W5JSMiXeDfN8bsH9rlPwT+D2NMdeiYq/OHhYj8M6xQvQ1jzGewIsTzzz9/L4MZHwne3JzwjetTXr0+ZntacmYYU7eGomrI5uriY1vsHyZQ85ku2DTg1STESEDgafqxzze3ZsRa4/uaIPYJfEVffDtyOPLwtWLQsY0nF2byUaYr3hofyauGzXFOM+9AfNS+XYddYHuzgnFW4Wtt+5N5Nk16Nq+NceLicBw/xykyXwSeEZEnseLyI8BfO7yDiKwCu8aYFmuh/Oot5/jR+fbDx5wyxlwVEQF+APizY7r+h5Zro4zPv7LJtGwwxpBXNV+6sEtVNxhzw1SU2xwbeLDaCegnAZ4Iq4MI2pbGQCdQjLMGJQZfhGEnwPeEThATjnN6kc9Taz2SUDGIQ0ZZYYd9HWG64mFxaNqWS3spIJwZxgfutaO4sg4H0kdZhQAnB7Zb8uYkp6ga2tbQDTxOD9/DD9nhcByJYxMZY0wtIj+JdXVp4FeNMV8XkZ8DXjTGfA74JPALImKw7rK/szheRM5jLaE/uOXU/5uIrGHXyC8D/+lxfYaHkbxq+P/e2CHwFPm0IPE94tDjyn5G2dguwwsOu8s6HiSB0I1CVroB55YT+nHAWicAhFlVo4GPnIKvXdpHK1jr+SixVfjf8/Qq06KmNS2+sq6ylW7IUhIcKcB+WBwu7WeEvma9F90kKke1Pg67wK7sZ9SNYS8tCTyFp4Syhb3UNtp08ReH43gRc8SxsA8zvV7PfMd3fMeDvoz3haJuGWfV3IJpbapx1ZCVNaPLryEI0YknaQ0HrfsVoLUQaNs63zZvtG35RWReLd9SNrbyvWkMVdvOK/TNvNWLwVOKOLCdlEUErYTQU0S+PmhpY4x9zdeCktvZUlYo1aIX/gJj3+NeRKE1hqxqyEp7voPr1IpA24JSV0HvcNyZP/iDP/iSMeb593KOBx34d9xHWmMO3EFlY5tOZmVD2bTU8yZfBjvjXQw0xqCMHcwFoJXt22/3tDNc6tbQGmOzykIrOk3T0gJ10x4IStMamralqO3YYCVCY2zb/tYYlAihd3tRuRUROZhbs8DMt98LSsS21a/ttRkRfG8uMCK0j8ENlsPxoHksROa5557jhRdeeNCXcawsssl205L9WcHnX95CKUiLigu7Kd/YnPDmr/8MrYEnf+wXMIA2oHzh9KBDL/bwRNO0LRtLIR89PSAJPb719IBTg5jXtyZM85pu5PH0Wo/dacGfX5+QlzWtsdX7u2lJUzecGCScGcYMkgAlcpAafK+fZVEzM8lrtiY5wyS46+jiO3F5L71pfgxgOz/f43U5HI8b93pjdzucr+ARYZEmvBT5XBvnnFwKGaUlW9OS3VlJP9IYbAPKtLRdirWn2Bgk9EKfpoHVrs9yErDcDVntRbZtithzr/dinjvZZ70XH8w+6YWayPcYdkJ6kYeeu7iWYu+gwv5wG5ejcrh+ZS+t2J7krPcihp3gtjU278SdWre4PmAOx/HzWFgyjwPF/M58Ly3p+j7x0GM/q7i8bwPuWhRagTHWDVXV0O1oluKAQeRRtXB2ucMg8UkinzPDhEHic3U/48xy5201LvtZyaRoUErw5v3BAq2oG0VW3hCVd9s5+CB4v5fSDb33NFflnVq3OByO48OJzCNC6Cm+uTNjmtek8zt2A/TCEE8r0tyORVYC/dhaHFoJedlw8mSPZ072ePaEbcMvAqcGMWXdHgTp86phPy0p6pZAC1opPLHbtdjuyFpD6NvkATMP9L9TG5h34ig1NkfBFV06HA8GJzKPCJ3Q48L2jCTQKAwvX5+wMy3wVUulFFFgXWSiBK00UaA5uxxjDFzYSfl3PnaKWWFb658ZJgcupfVeyDSv2ZnZFODY12RlQ902PLGSsDMtGWUVGMOZ5YRu4JHXzX2zGI5SY+NwOD64OJF5yFlMdnxrZ0bgKbamJXHgsdoNCJXh8n5L3dYM+jb+YDB0I81qNyKcV/HHgaZuDCudgFFacmFndhDgj3zNn17YtZaLEurGDg5b60VogVNLEedWOjfVvzx9qMPye+Vemlg6HI4PHk5kHmIOZ2EpJQd3+91QkxWaKvSJo4bVXsC0aA7Sdte7EWdWEjaWE0wDG8OIQSck0MLp5RuCsYh7DDoheVnfNDo49BTpfG79e4l1LETy1vHIC1w8xeF4uHEi8xBzU+diYHNSsjcrefX6mPVeiGlblmKPSVqSlw3e3FW23g/RSpjMKrRSTLKK169P+PCp/m0D7P3Iuyn4DjYFeFFo+W6tilvHH986HnnBBzWe8k4C6XA4nMg81CyC4nnVMC5qtic5o1mJAEXVzidUCq0Rzq4kDDshRd0Qhz5JoDAGVnsBwyRknJdc3k8573UPFspFgH1jEB+Ly+pu3ZkX4wGKup2nxNnXPiiL+VEF0uF43HHR04eYRVB8b1aSlQ3LnZDKGHwl+ErYT2vq1nB21QqMp20FfOTbyv1+5HNupcuT611WehGTvGY/LQ/OvwiwH9fclVvHH4MVsXFWHRRQKoEro5wr+ylKeFd1MsfB7ebfBJ5mZ3rEaXAOx2OCE5mHmEWR4d6s4JvbU/KqZhAHrPcjOrFH1TTsz2za8aKtDIBCUPO5MFXTMkoruqFHWTfM8vq2BYu3m0//XlmI5GGqxpBW9cECPsoqOqFHJ/DZT6sPzGJ+J4G818JTh+NRx4nMQ4ydeR8yyW1QPvI1Z1cSGgNlbTjZD9BawAhZ2dgml3VL4Atl2xIozSSvyOuG3WnJMAkIA/W+TYm8UyV+EvgHC3hRW5efp4WysQv4B2Exv5NALlKrF+60N7amHwjLy+F4ULiYzEPOrKj5yMYSTWvYzyqSwENLySgtGXZDljsRWitmRUVaNLZNzKyiE2r2s5K+8RhlNo6z2gv5xLllBknwvlz7nTLHdqbFQW1M6ClrhRnbUQA+GHUyd0utdvEah+MGzpJ5iMmrhou7KbuzgqUkIC0q/uzSPhe2ZwSe5kPrPT66sYQSbPAcQ6CFtU7I+ZUOVW14bWvKOKs5u9JhpRuxMy3e17vu27nhDls4S7HPrKiZlRWDxP/A9B27W5zKxWscjhs4S+YhZT8teenKiCt7KdOyZpzVZEXNuZUO+1lFWTdsTwuGcchTa102xyWeVvhaM+wGlLVh2A2IQsWpfsypgV0wF4vhg0wZPmzhVAY2liIQoTXgvw9uvHu9zlu5X61wHI5HAScyDyF51fDylRFNazAIb22njPKKfuKzl5acWrIpx5vjklFW89Rql6ptCbQ6GD28OS3sEK+2RStb93JyKfrALIYf1NqYo+Ba4TgcNzjWv3oR+ZSIvCIir4nIz9zm9XMi8nsi8lUReUFEzhx6rRGRL8+/Pndo+5Mi8sfzc/5zEXl/AggfIHamBcbAtVHGNzbH7GYVk7RiNCvZnlZc2k8xQFnVpEXN1jRn2PFRSmgxbE9L20DTQBx49OOAk0t21LFbDN87brSAw3GDY1tNREQDvwz8FeCjwI+KyEdv2e0Xgc8aY74N+DngFw69lhljPj7/+vcObf8fgH9ojPkQsAf8zeP6DB9UirqlMYYvv7WPQljp+HQjj2nekFcVRdWwFPt0o4D1fkgn8GnmaczG2HkvT612We0EbAwTzq90CD3lFsP7xHHVFTkcDyPHecv6XcBrxpg3jDEl8JvA99+yz0eB358//vxtXr8JsWPavhf4rfmmXwd+4L5d8UNC6Cl2ZiXak/m8es8WZabF3EqxA8kiTzFMQpJAsT0rwRi0EpoW8rqhG/lEWjHKS/bS6q6LoUvJvTeOo67I4XgYOU6ROQ1cPPT80nzbYb4C/ND88Q8CPRFZmT+PRORFEfmCiCyEZAXYN8bUdznnI89KNyQrG84NE9KyYWeWE/nCIA6YZBW704Kl0OdjTyzRDT320woF9CKfpdjnI6f7eEqzvhTST3zWezGBlju2a1kITGsgCfQHpure4XB88HnQgf+fAn5JRH4C+EPgMrBYuc4ZYy6LyFPA74vI14DRUU8sIp8GPg1w9uzZ+3rRD5rI1zy52uHKfsberEIUtK1BKcWJpZCzKx1mZYOvFb1IMStsirKa9/9Ky4Yk1EyyhvV+eNA77E5ZZXfrMfawBucdDsf7w3FaMpeBJw49PzPfdoAx5oox5oeMMZ8A/sv5tv3598vz728ALwCfAHaAgYh4dzrnoXN/xhjzvDHm+bW1tfv2oT4ofPhUn2Hic24lYTn2wYAWYWOQsNYLAcPlvYy6bTm7knBmGNNiYzJF1aCBWVkx7Ni8ibtV0d/vFirO9eZwPD4cp8h8EXhmng0WAD8CfO7wDiKyKiKLa/hZ4Ffn24ciEi72Af4i8JIxxmBjNz88P+bHgf/rGD/DB5bI1wwTn7d2U752eYSvhL9wdkA/Dnh9a0bTtgwSn28/t2wHjCllK+bF1mUWTcvGMD5wj90tq+ydWqjcC8715nA8XhybyMzjJj8J/C7wMvC/G2O+LiI/JyKLbLFPAq+IyKvACeDn59s/ArwoIl/Biso/MMa8NH/tp4G/JyKvYWM0/8txfYYPKnnV8NLVEW/uZCzFPieXYhC4uJvRGsOJXsQgCcHAlf2McV5zaXeGmbdmeeZEj06gWUnCI6XY3s+UXFcN73A8XhxrTMYY89vAb9+y7b8+9Pi3uJEpdnifPwK+9Q7nfAObufbYcmU/463tGbHvEcwtmq9fHZPMB4stdwKu7GfUrcHTwtPrPTwlVE2LQUgCzSfOLTMr6iNNm7yf0yldNbzD8XjxoAP/jnfB9VGGpxRKwbSoiQOPc4OEvbzk2qigF/kME59e5LE5KTiz3KEf+4RzUVgE6++lEeb9qsAPPcU0r5mV9cFEyU7gEQcuxdfheBRxIvMQIiL4nmJvVjJIbNqyaMiqhn7ok5Y1/aUIXylUIOzNSk4N4gNX1xtb07tOmFyMFR5nFWlVkwQ+/ci7LxMpO6HHq9fGdEIfEcOl3YxRVvKJc0NW5uMKHA7Ho4PrH/IQst4L8bW1CDwxRL4wyRraxjDs+ES+Zm9W8ub2lKKuubJvF/KibhG5e8B9EZjPyoZRVtE0MEqtm+x+BOhnRc2Z5Q4Gw6W9HE8L51c7TPPaJQA4HI8gzpJ5iFhYGEVjbHFlbK2WcVbz9HqHYTIkrxuqBuq2oW5axmlNYyqujTOa1hAGeh5wv32tyyIwvzsrCH2NrxVV3TIrala64XuujSnqll7kkZY151c7+FphjCGrmg9EB2iHw3F/cZbMQ8Lh1N9h4rPaD1Fi0CL0Qo8Pn+pzejkh9n1WuwEn+xEAjYFhJ0BaW7B5mNvVuixqYhYTKYGDqZT3YyLlIh368PnrxhBo9YGYeOlwOO4vTmQeEg6n/hZ1y6xoOLvc5dmTPU4MIq7t5yiBj53uU7ctr29OqVrD+lLI6WHC2bUOCHbK5Jzb1bosROBgIiU3ROB+dGhepEOLQNW0VHVL2TQMO4HrAO1wPIK4/+iHhMNV9/tpSeAp4kBT1S2r3QiD4dpezvVxzua4oKhbnlzpEGjN5rjAV4KAnUFzl1qXhQh0Ao+iakiLmqKu6YTefenQvMhSW0kCxllF3bac6EcoEdcB2uF4BDmSyMznvvzb88exiPSO97Ict3K46n7hapoVNaOswleKjUHEtXHGS1fGLCcBwyRgZ1rStDbYPylqIl+/Y/v5hQjEgWYp9tEalpKQJND3rZtw5GueWu/xPU+vsjGIaQ2uHb7D8YjyjoF/Eflb2EaTy8DT2H5h/xT4t4730hyHWemGB9lX+7OSy2XN7qxk2PG5PsnZm+WcGiQg0Ak8NpZjXr024fqoYGMQEWkPpYTY0zy11r3rey2E5vTweD/Twzz90uFwHI2jWDJ/B9s7bAxgjPkGsH6cF+V4O5GvWemGbE9yWmBnUnJ5L+Xizoy8rEjzlqZt8QSyoqEfBXzs9BLLnYAo0PQSj9BTB52YHQ6H4/3gKCnMhTGmlPniNO+AbO5+iOM4mBU1vdhnZ1ailBD6QlnDa9entjhzs8D3FKcGCVXdokRYXwpZToL5pEYnMA6H4/3lKCLzByLyXwCxiHwf8LeBf3G8l+W4lbxqeOnKmD+5sENVG7KmIdTCOKvIy4on17sEviLLW4qiZj8rqBpDN/Io65adaUFrjBMah8PxvnIUkfkZ4G8CXwP+E+C3jTG/cqxX5biJvGp4fWvKN7cnTLKaJNSMJyWhrxlnFb7WTIqGc52Q1a4m8T1mWc1T6116cYCv5aA2ZZEifFDYOd92P1rGOBwOx60cJSbznxljfsUY8x8YY37YGPMrIvJ3j/3KHAfsTAvSokaLIg41RWXoBB7jrGJ3ViJiWE58Ak/z3Mk+T611KFtDLw5uaqmvxIqNm+nicDjeL44iMj9+m20/cZ+vw3EXirqlNYYwUJwZxGS1bdEf+5pu6FFULQY7GXOUVry+NWWSVzTtzdXzAhhj3EwXh8PxvnFHd5mI/Cjw14AnReTwRMsesHvcF+a4wSIrrB8H7GcVy4nPRASDsBQHxIEm9jVX92cUdUta1Kz1Q97YmvL0eu/ADWawHZzdTBeHw/F+cbeYzB8BV4FV4H86tH0CfPU4L8pxMyvdkP2souMrilLNZ8kI/STg/ErCoBPwtUt7jLKKU4OEZ9aXqFvDS1cnbE8KnjnRoxv51hry1EFh56JJJrz7ccoOh8NxN+4oMsaYC8AF4F97tycXkU8B/wjQwP9sjPkHt7x+DvhVYA1rHf1HxphLIvJx4J8AfaABft4Y88/nx/wa8JeA0fw0P2GM+fK7vcaHgcjXPL3WJfY1Ws1Iy4ZTg5hBJ2Qp9uhFPk8Mu3zLhsdqL2JzknN9VBB5wiSvycqGvGrwtbWIFoWdwEFSQFk3rjDS4XDcd45S8f89wD8GPgIEWMGYGWP673CcBn4Z+D7gEvBFEfmcMealQ7v9IvBZY8yvi8j3Ar8A/BiQAn/dGPMNEdkAviQiv2uM2Z8f9/fno5sfGyLfVuo/tdZlPy156coIAyS+Jgk8RCDyFJvjgnFe0Q01VS2kVcPpZVsj07QGT8l9HafscDgcd+Mo/pFfAn4U+AYQA/8xVjzeie8CXjPGvGGMKYHfBL7/ln0+Cvz+/PHnF68bY16ddxbAGHMF2MRaO489i9Tj9V5EJ9BkZcP2JOdD61320goB2hYwtuPyiX7Eflria8GYGzW0C6F5aq3rBMbhcBwbR3LCG2NeA7QxpjHG/DPgU0c47DRw8dDzS/Nth/kK8EPzxz8I9ERk5fAOIvJdWAvq9UObf15Evioi/1BEbtu2V0Q+LSIvisiLW1tbR7jch4NFZlgv9jk1SHj6RI8zyx36kU8S2mYMbWso6oZ+4nNyKaKoW6rGIK4Q0+FwvM8cRWRSEQmAL4vI/ygi//kRjzsKPwX8JRH5U2yc5TI2BgOAiJwC/lfgbxhjFvm4Pwt8GPhObNPOn77diY0xnzHGPG+MeX5t7dExgg63/F/g///t3XuUpHV95/H3p6q6+jI992kuc5OLREVjQGeRSDReCbIEjO4xcAjixpX1uLirq7uHRJclxJwc1+zlsKvmaIJ4QZHFNc6Jo6gRY3aPKCOX4WKAAQnMxZlhLsz09KVu3/3jeWp4qKnuKaCfqurm8zqnTj/1ey71q6effr79/J7f8/0VBRIvHhtl5eJBXrRqhKUjZY5fnKTQl6BSqx+1nplZ3jp54v8ykqByJfBhYB3wzg7W254u27Q2LTsibQp7B4CkUeCdzfsukpYA3wY+FhG3Z9bZmU5OS/oCSaB6wWj2DGtEgwMTFaZrSSr/lSNlVi8fIYByqcja5Q32HJrm4GSV9StGWO3cZWbWA8cMMmkvM4Ap4E8AJJ0DbD3GqncAp0k6mSS4XEzy3M0RklYB+9KrlD8i6WlGeuX0TZJOAbe0rHNiROxU0vbzduC+Y32HhWTl6CCP7j7EvokqI+UiJYmJao3JdNji5g39asDqZcNOF2NmPTXbw5hF4F0k91G+GxH3SboA+GOSDgBnzrbhiKhJuhK4laRH2vURcb+ka4HNEbEReAPw55IC+DHJsAKkn/t6YKWk96Rlza7KN0oaI3mA/W7g/c/+a89fQwNFhsolBit1ao1kWOR1KxZRkNg7Pp2OA+OuyGbWH2a7kvlrkuaunwHXSdoBbACuioi/6WTjEbEJ2NRSdnVm+hbgqK7IEfEV4CszbPNNnXz2QpNNaLn74BQnLB1iuPz0ry8i/MS+mfWd2YLMBuCVEdGQNAT8Cjg1IvZ2p2rW1ExoWS4VGSkXKRbEtv2TrFvxdNdjP7FvZv1otrNSpdmjKyKmgEcdYHqjNaHlcUuGgGD3oSkigkqtQaVWZ+Vo297cZmY9M9uVzEslNXOUCTg1fZ8m841X5l47AzgqoeXQQJG1y0fYeWDST+ybWV+bLci8rGu1sFm1S2hZLBRYv3KRb/KbWV87VoJM6wPZhJb1RvL8y+HpGutXjDBVrfsKxsz6lu8UzwPNPGPVeoPHnjwMgpNWLWKgVPSIlmbW1zp54t/6wNBAkcFSgVOPW0y5pRdZ8/kYM7N+4yuZeWSmvGXTtcYMa5iZ9dZsT/zfSzJi71GzcO+ynmjtADBVrbP70BS1WoPBUsEpZMys78zWXHZB12phHWntAPDInnEOTVZZuajMjgOTHJissmbZMIena0w78JhZH3Dvsnkgm1JGQLXe4Il9ExyarHLC0mEWDZYYn65xz+P7+PsHq6xZNsLxSwZZOTrE1P4JP0NjZj0zW3PZIWZvLpt1+GWbG9mUMgXBnsMVDk/V2Ds+zeplw4wODTBdS5rNdj01RUFipFxkz6EK1XpwwtJhdwwws56Z7UpmcTcrYu01U8pMVmo8vPsQjUhu9u8dn2agVKA8UOSpiSpTlTrFoigVi5RLRSQxWWkwPlWlWPA4MmbWGx13YZZ0HDDUfB8Rj+dSI3uG6VqDguChXeOUigUWlQrUGgGIyUqNveMV6o0GlXTwsqUjAwCUCmKyXmeiWuc4J840sx455tlH0oWSHgZ+Cfw98BjwnZzrZanBUoE949MIGEoTZAqxetkQ5WKByWqNRj0IYGx0mNHyALVGUK03aDSStk0nzjSzXunkSuZPgbOBH0TEmZLeCPxBvtWyppWjgzy86xClgqjWGxRUoFKvc8KyYQ5N1VgyXGKkPMDIU5NEBMMDRQ5OVTk0WWNs8SCnr17qm/5m1jOdtKNU0xT/BUmFiLiNZKyZY5J0nqQHJW2VdFWb+S+S9HeStkj6kaS1mXmXS3o4fV2eKX+1pHvTbV6XDsO8YA0NFFm/chGLhkqMT9epNRqMLR6k0Uiucl6+ehkvO3EJv3nqKk5aNQrA0uEyr3rRcl5z6iqWjZR7/A3M7IWskyuZA5JGSYZHvlHSbuDwsVZKh2/+NPBWYBtwh6SNEfFAZrG/AL4UEV+U9Cbgz4HLJK0A/jNJMAvg5+m6+4HPAu8Dfkoy6uZ5LPDmu9XLhplM85MdmKjw2J5xkHjxcaPsHZ8+8izMKWOjnDI22uPampk9rZMrmYuACeDDwHeBR4Df7WC9s4CtEfFoRFSAm9JtZZ0O/DCdvi0z/3eA70fEvjSwfB84T9KJwJKIuD0iAvgS8PYO6jKvTVXr7Hlqir2Hp6nW6jRCrFhUZtlImUbgJJlm1rdmDDKSXizpnIg4HBGNiKhFxBeBO4FlHWx7DfBE5v22tCzrHuAd6fTvAYslrZxl3TXp9GzbXFCmqnUe2PEUI4MlTlk1ysjQAMODRUbKJQ5MVCmXCpRLRfaOT/e6qmZmR5ntSuZ/AAfblD+VzpsLHwV+W9JdwG8D24E5+Zdc0hWSNkvavGfPnrnYZE/sHZ8mgOFy8uxLRNLLbKJSp1JPEmM6SaaZ9avZgszxEXFva2FadlIH294OrMu8X5uWZbe1IyLeERFnAh9Lyw7Msu72dHrGbWa2/bmI2BARG8bGxjqobn+arjUYGShSqyfJF8ppN+bJ6TrlYvLrq9aDQT8LY2Z9aLYz02xNYsMdbPsO4DRJJ0sqAxcDG7MLSFolqVmHPwKuT6dvBc6VtFzScuBc4NaI2AkclHR22qvs3cC3OqjLvDVYKjA6NMCh6Srb908wPlXlsScPc7hWZdnIAJVag0qt7mdhzKwvzRZkNkt6X2uhpH8F/PxYG46IGnAlScD4BXBzRNwv6VpJF6aLvQF4UNJDwPHAn6Xr7iN5PueO9HVtWgbwAeCvgK0knRAWdM+ylaODVOoNarU6kaaSWzI8wLLBMpPVJBuAE2CaWb+arQvzh4BvSrqUp4PKBqBMcpP+mCJiE0k342zZ1ZnpW4BbZlj3ep6+ssmWbwZe0cnnLwRDA0UKEUzVg0qtwehgiVnB57cAABHcSURBVJeeMMJwuXQkwJiZ9avZEmTuAl6bPuHfPKl/OyJ+ONM6NvemqnV2HZrmhCVDDBSTvGX7015l1XY5ss3M+sgxH8ZMn/C/rQt1sTb2jk+zaLCEEJKODL+8Z3ya1Us7uTVmZtY77pLU56ZrSRqZSr1OtdYgIohGcHiq9qxu9k9V60zXGkfGp/HDm2bWDQ4yfW6wVKBYKHDC0mEkmKzWqUewfuWijm/2NwMLQEFylgAz6xoHmT63cnSQSq1OQeKEpUOcuHSYFYvKrF7WeVNZc+AzCRDOEmBmXdPxoGXWG0MDRdYsH2Hv+DQTlTqDpcIxuyxPVevsHZ9mutZgsFTg4FSN5elgZk0DRTFR8ZWMmeXLQWYeaAaaY5mq1tlxYJLH9x5m0VCJsdFBGgEHDk9TahmC2VkCzKwbfJZZIJr3XfaOT7NkaIBSocCug9M0Ilg8PMADO59islKnUm1wcLLqLAFm1hUOMgtE875LAAOlAgPFAuVSgV8dnGR8qsrSwQGKBdGI4MlDU0fGoDEzy5ODzAIxXWswUBTlYuFIMs1SQTx5sEJBBZaPlimXCgyVi6xdsYjD07Ue19jMXggcZBaIwVKBaj1Yvqh85Jmaar1BrdGgEfGMYZg9NICZdYuDzAKR7ep8/JIhao3k3svYkkFWtTSN+aa/mXWLe5ctENmuztWA1cuGj9zY375/gkqtAQEBVGp1J9Y0s65wkJkHWp97memm/UxdnZvBpxGBJA8NYGZd4yDT55pdk8ulIiPlItV6sH3/xJFg8myCT3OeA4yZdYsb5vtcs2tyc9jlZkqYHfsn2L5/gkbASLnofGRm1pccZPpcs2ty1kBR7D7UPvg4H5mZ9ZNcg4yk8yQ9KGmrpKvazF8v6TZJd0naIun8tPxSSXdnXg1JZ6TzfpRusznvuDy/Q681uyZnVetBRLQNPu6abGb9JLcgI6kIfBp4G3A6cImk01sW+zhwc0ScCVwMfAYgIm6MiDMi4gzgMuCXEXF3Zr1Lm/MjYnde36EfNLsmV9KxZCq1BpVaneOXDrcNPu6abGb9JM8z0lnA1oh4NCIqwE3ARS3LBLAknV4K7GiznUvSdV+QmjftC4KJSp2Ckt5iq5cNtw0+zkdmZv0kz95la4AnMu+3Aa9pWeYa4HuSPggsAt7SZju/z9HB6QuS6sA3gE9ExIIe7f5YXZM7HQLAzKzbet22cglwQ0SsBc4HvizpSJ0kvQaYiIj7MutcGhG/DrwufV3WbsOSrpC0WdLmPXv25PcNeqgZfE4ZG3WAMbO+lGeQ2Q6sy7xfm5ZlvRe4GSAifgIMAasy8y8GvpZdISK2pz8PAV8laZY7SkR8LiI2RMSGsbGx5/E1zMzsucozyNwBnCbpZEllkoCxsWWZx4E3A0h6GUmQ2ZO+LwDvInM/RlJJ0qp0egC4ALgPMzPrS7ndk4mImqQrgVuBInB9RNwv6Vpgc0RsBD4CfF7Sh0k6Abwnc3/l9cATEfFoZrODwK1pgCkCPwA+n9d36HedppsxM+uVXNPKRMQmYFNL2dWZ6QeAc2ZY90fA2S1lh4FXz3lF56HZ0s040JhZv+j1jX97jmZKN+Mn/s2snzjIzFMzpZvxE/9m1k8cZOapmdLN+Il/M+snTvU/T60cHWT7/gkguYKp1qPtYGTZzgHtrn7MzPLkf3vnqZnSzWRv+jc7BzSHA4Ckmc3DAZhZt/hKZh6bKd1MU7ZzAIAEBcTe8WkPv2xmXeErmQWsXfOY0nIzs25wkFnA2nUOiLTczKwbfLZZwFrHoomARoSHAzCzrnGQWcBaOwdAchXjjABm1i2+8b/AZTsHuJnMzLrNZx0zM8uNg4yZmeXGzWXzhNP6m9l85CuZeaD1yf1GwPb9E35y38z6noPMPOC0/mY2XznIzANO629m81WuQUbSeZIelLRV0lVt5q+XdJukuyRtkXR+Wn6SpElJd6evv8ys82pJ96bbvE7Sgk8r3Hxyf6paZ+eBSf5p72Ge2DfR62qZmR1TbkFGUhH4NPA24HTgEkmntyz2ceDmiDgTuBj4TGbeIxFxRvp6f6b8s8D7gNPS13l5fYd+sXJ0kINTVZ7Yd5hGIygVkquYqUrN92XMrK/leSVzFrA1Ih6NiApwE3BRyzIBLEmnlwI7ZtugpBOBJRFxe0QE8CXg7XNb7f4zNFBkuFRgcKBILYJiQaxbMcLi4bLvy5hZX8uzC/Ma4InM+23Aa1qWuQb4nqQPAouAt2TmnSzpLuAg8PGI+Id0m9tatrmm3YdLugK4AmD9+vXP/Vv0C4l1y0fItg5GxJF0MWZm/ajXN/4vAW6IiLXA+cCXJRWAncD6tBnt3wNflbRklu0cJSI+FxEbImLD2NjYnFe82zzcspnNR3meobYD6zLv16ZlWe8FbgaIiJ8AQ8CqiJiOiL1p+c+BR4BfS9dfe4xtLkitGZUrtQaVWt0Zlc2sr+UZZO4ATpN0sqQyyY39jS3LPA68GUDSy0iCzB5JY2nHASSdQnKD/9GI2AkclHR22qvs3cC3cvwOfaOT4ZbNzPpNbvdkIqIm6UrgVqAIXB8R90u6FtgcERuBjwCfl/Rhkk4A74mIkPR64FpJVaABvD8i9qWb/gBwAzAMfCd9vSAca7hlM7N+k2vusojYBGxqKbs6M/0AcE6b9b4BfGOGbW4GXjG3NTUzszz4rrGZmeXGQcbMzHLjIGNmZrlxkDEzs9w4yJiZWW4cZMzMLDcOMmZmlptcn5OxuTVVrbN3fJrpWoPBUoGVo4N+4t/M+pqvZOaJqWqd7fsnaASMlIs0Arbvn/B4MmbW1xxk5om949OUS0XKpQKSKJcKlEtFjydjZn3NQWaemK41GCg+c6TpgWIyQqaZWb9ykJknPJ6Mmc1HPkPNEx5PxszmIweZecLjyZjZfOQuzPOIx5Mxs/nGVzJmZpabXIOMpPMkPShpq6Sr2sxfL+k2SXdJ2iLp/LT8rZJ+Lune9OebMuv8KN3m3enruDy/g5mZPXe5NZdJKgKfBt4KbAPukLQxHQ2z6ePAzRHxWUmnk4yieRLwJPC7EbFD0itIhnBek1nv0nSETDMz62N5XsmcBWyNiEcjogLcBFzUskwAS9LppcAOgIi4KyJ2pOX3A8OS3I3KzGyeyTPIrAGeyLzfxjOvRgCuAf5A0jaSq5gPttnOO4E7IyL7aPsX0qay/yRJbdYxM7M+0Osb/5cAN0TEWuB84MuSjtRJ0suBTwL/OrPOpRHx68Dr0tdl7TYs6QpJmyVt3rNnT25fwMzMZpZnkNkOrMu8X5uWZb0XuBkgIn4CDAGrACStBb4JvDsiHmmuEBHb05+HgK+SNMsdJSI+FxEbImLD2NjYnHwhMzN7dvIMMncAp0k6WVIZuBjY2LLM48CbASS9jCTI7JG0DPg2cFVE/L/mwpJKkppBaAC4ALgvx+9gZmbPQ25BJiJqwJUkPcN+QdKL7H5J10q6MF3sI8D7JN0DfA14T0REut6LgatbuioPArdK2gLcTXJl9Pm8voOZmT0/uT7xHxGbSG7oZ8uuzkw/AJzTZr1PAJ+YYbOvnss6mplZfnp949/MzBYwBxkzM8uNg4yZmeXGQcbMzHLjIGNmZrlxkDEzs9w4yJiZWW4cZMzMLDcOMmZmlhslWVwWNkmHgAd7XY8OrCIZsK3fzYd6zoc6gus511zPufWSiFj8fDaQa1qZPvJgRGzodSWORdJm13NuzIc6gus511zPuSXpeY9A7OYyMzPLjYOMmZnl5oUSZD7X6wp0yPWcO/OhjuB6zjXXc24973q+IG78m5lZb7xQrmTMzKwHFlSQkXSepAclbZV0VZv5g5K+ns7/qaSTuly/dZJuk/SApPsl/bs2y7xB0lOZEUGvbretLtT1MUn3pnU4qoeJEtel+3KLpFf1oI4vyeynuyUdlPShlmV6sj8lXS9pt6T7MmUrJH1f0sPpz+UzrHt5uszDki7vQT0/Jekf09/rN9Ph0NutO+sx0oV6XiNpe+Z3e/4M6856XuhCPb+eqeNjku6eYd2u7M+ZzkO5HZ8RsSBeQBF4BDgFKAP3AKe3LPMB4C/T6YuBr3e5jicCr0qnFwMPtanjG4C/7YP9+Riwapb55wPfAQScDfy0D37/vwJe1A/7E3g98CrgvkzZfwGuSqevAj7ZZr0VwKPpz+Xp9PIu1/NcoJROf7JdPTs5RrpQz2uAj3ZwXMx6Xsi7ni3z/ytwdS/350znobyOz4V0JXMWsDUiHo2ICnATcFHLMhcBX0ynbwHeLEndqmBE7IyIO9PpQ8AvgDXd+vw5dhHwpUjcDiyTdGIP6/Nm4JGI+Kce1uGIiPgxsK+lOHv8fRF4e5tVfwf4fkTsi4j9wPeB87pZz4j4XkTU0re3A2vz+vxOzbA/O9HJeWHOzFbP9FzzLuBreX1+J2Y5D+VyfC6kILMGeCLzfhtHn8CPLJP+ET0FrOxK7VqkTXVnAj9tM/s3Jd0j6TuSXt7Vij0tgO9J+rmkK9rM72R/d9PFzPzH2w/7E+D4iNiZTv8KOL7NMv22X/+Q5Iq1nWMdI91wZdqsd/0MzTv9tD9fB+yKiIdnmN/1/dlyHsrl+FxIQWbekDQKfAP4UEQcbJl9J0mTz28A/xP4m27XL/VbEfEq4G3Av5H0+h7V45gklYELgf/dZna/7M9niKTtoa+7dkr6GFADbpxhkV4fI58FTgXOAHaSNEX1s0uY/Sqmq/tztvPQXB6fCynIbAfWZd6vTcvaLiOpBCwF9naldilJAyS/2Bsj4v+0zo+IgxExnk5vAgYkrepmHdPP3p7+3A18k6TZIauT/d0tbwPujIhdrTP6ZX+mdjWbFNOfu9ss0xf7VdJ7gAuAS9MTzlE6OEZyFRG7IqIeEQ3g8zN8fr/szxLwDuDrMy3Tzf05w3kol+NzIQWZO4DTJJ2c/md7MbCxZZmNQLM3xL8AfjjTH1Ae0jbZvwZ+ERH/bYZlTmjeJ5J0FsnvqNuBcJGkxc1pkhvB97UsthF4txJnA09lLrW7bcb/EPthf2Zkj7/LgW+1WeZW4FxJy9Pmn3PTsq6RdB7wH4ELI2JihmU6OUZy1XIP8Pdm+PxOzgvd8BbgHyNiW7uZ3dyfs5yH8jk+8+7J0M0XSY+nh0h6k3wsLbuW5I8FYIikSWUr8DPglC7X77dILkG3AHenr/OB9wPvT5e5ErifpBfM7cBre7AfT0k//560Ls19ma2ngE+n+/peYEOPfueLSILG0kxZz/cnSdDbCVRJ2q3fS3L/7++Ah4EfACvSZTcAf5VZ9w/TY3Qr8C97UM+tJO3uzWO02SNzNbBptmOky/X8cnrsbSE5QZ7YWs/0/VHnhW7WMy2/oXlMZpbtyf6c5TyUy/HpJ/7NzCw3C6m5zMzM+oyDjJmZ5cZBxszMcuMgY2ZmuXGQMTOz3DjImHVIUkj6SuZ9SdIeSX+b8+feIOmXaWqchyR9SdLazPxm9t4tkr4n6YQ862P2bDjImHXuMPAKScPp+7fSvafH/0MkqXFeAtwF/DB9uLDpjRHxSmAz8MddqpPZMTnImD07m4B/nk4/I9NA+tT29ZJ+JukuSRel5SdJ+gdJd6av16blb5D0I0m3KBm/5cZjZQWPxH8nSWD4tjaL/Bh48Rx8T7M54SBj9uzcBFwsaQh4Jc/Mov0xklRFZwFvBD6VpgjZDbw1kuSHvw9cl1nnTOBDJON5nAKc02E97gRe2qb8ApKn4M36QqnXFTCbTyJiS5oe/RKSq5qsc4ELJX00fT8ErAd2AP9L0hlAHfi1zDo/izSflZIRE08C/m8HVWm94rlNUp0kVcjHO/0+ZnlzkDF79jYCf0Ey6mZ2PCIB74yIB7MLS7oG2AX8BknrwVRm9nRmuk7nf5NnkuSZanpjRDzZ4bpmXePmMrNn73rgTyKitVnqVuCDmazPZ6blS4GdkaSkv4xkSODnJM16/W9JhtD97nPdjlm3OMiYPUsRsS0irmsz60+BAWCLpPvT9wCfAS6XdA/JfZTDz+FjP5Wu/xDwz0iuXCrPYTtmXeUszGZmlhtfyZiZWW4cZMzMLDcOMmZmlhsHGTMzy42DjJmZ5cZBxszMcuMgY2ZmuXGQMTOz3Px/O+eLDSpRNhUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2aab0ede78d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Removing outliers from the dataset will generally improve association results. \n",
    "#We can draw lines on the above plot to indicate outlier cuts. \n",
    "#We’ll want to remove all samples that fall in the bottom left quadrant.\n",
    "plt.scatter(df[\"sa.qc.dpMean\"], df[\"sa.qc.callRate\"],\n",
    "            alpha=0.1)\n",
    "plt.xlabel('Mean DP')\n",
    "plt.ylabel('Call Rate')\n",
    "plt.xlim(0, 20)\n",
    "plt.axhline(0.97, c='k')\n",
    "plt.axvline(4, c='k')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "After filter, 843/1000 samples remain.\n"
     ]
    }
   ],
   "source": [
    "#Filtering is simpler when use these cutoff values:\n",
    "vds = vds.filter_samples_expr('sa.qc.dpMean >= 4 && sa.qc.callRate >= 0.97')\n",
    "print('After filter, %d/1000 samples remain.' % vds.num_samples)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Next is genotype QC. To start, we’ll print the post-sample-QC call rate. It’s \n",
    "#increased since the initial summary - dropping low-quality samples disproportionally removed missing genotypes.\n",
    "call_rate = vds.query_genotypes('gs.fraction(g => g.isCalled)')\n",
    "print('pre QC call rate is %.3f' % call_rate)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "#It’s a good idea to filter out genotypes where the reads aren’t where they should be: if we find \n",
    "#a genotype called homozygous reference with >10% alternate reads, a genotype called \n",
    "#homozygous alternate with >10% reference reads, \n",
    "#or a genotype called heterozygote without a ref / alt balance near 1:1, it is likely to be an error.\n",
    "filter_condition_ab = '''let ab = g.ad[1] / g.ad.sum() in\n",
    "                         ((g.isHomRef && ab <= 0.1) ||\n",
    "                          (g.isHet && ab >= 0.25 && ab <= 0.75) ||\n",
    "                          (g.isHomVar && ab >= 0.9))'''\n",
    "vds = vds.filter_genotypes(filter_condition_ab)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "post QC call rate is 0.955\n"
     ]
    }
   ],
   "source": [
    "post_qc_call_rate = vds.query_genotypes('gs.fraction(g => g.isCalled)')\n",
    "print('post QC call rate is %.3f' % post_qc_call_rate)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Struct{\n",
      "     rsid: String,\n",
      "     qual: Double,\n",
      "     filters: Set[String],\n",
      "     pass: Boolean,\n",
      "     info: Struct{\n",
      "         AC: Array[Int],\n",
      "         AF: Array[Double],\n",
      "         AN: Int,\n",
      "         BaseQRankSum: Double,\n",
      "         ClippingRankSum: Double,\n",
      "         DP: Int,\n",
      "         DS: Boolean,\n",
      "         FS: Double,\n",
      "         HaplotypeScore: Double,\n",
      "         InbreedingCoeff: Double,\n",
      "         MLEAC: Array[Int],\n",
      "         MLEAF: Array[Double],\n",
      "         MQ: Double,\n",
      "         MQ0: Int,\n",
      "         MQRankSum: Double,\n",
      "         QD: Double,\n",
      "         ReadPosRankSum: Double,\n",
      "         set: String\n",
      "     }\n",
      " }\n"
     ]
    }
   ],
   "source": [
    "#Variant QC is a bit more of the same: we can use the \n",
    "#variant_qc method to produce a variety of useful statistics, plot them, and filter.\n",
    "pprint(vds.variant_schema)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "#The cache (Hail's version of persist) is used to optimize some of the downstream operations\n",
    "vds = vds.variant_qc().cache()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Struct{\n",
      "     rsid: String,\n",
      "     qual: Double,\n",
      "     filters: Set[String],\n",
      "     pass: Boolean,\n",
      "     info: Struct{\n",
      "         AC: Array[Int],\n",
      "         AF: Array[Double],\n",
      "         AN: Int,\n",
      "         BaseQRankSum: Double,\n",
      "         ClippingRankSum: Double,\n",
      "         DP: Int,\n",
      "         DS: Boolean,\n",
      "         FS: Double,\n",
      "         HaplotypeScore: Double,\n",
      "         InbreedingCoeff: Double,\n",
      "         MLEAC: Array[Int],\n",
      "         MLEAF: Array[Double],\n",
      "         MQ: Double,\n",
      "         MQ0: Int,\n",
      "         MQRankSum: Double,\n",
      "         QD: Double,\n",
      "         ReadPosRankSum: Double,\n",
      "         set: String\n",
      "     },\n",
      "     qc: Struct{\n",
      "         callRate: Double,\n",
      "         AC: Int,\n",
      "         AF: Double,\n",
      "         nCalled: Int,\n",
      "         nNotCalled: Int,\n",
      "         nHomRef: Int,\n",
      "         nHet: Int,\n",
      "         nHomVar: Int,\n",
      "         dpMean: Double,\n",
      "         dpStDev: Double,\n",
      "         gqMean: Double,\n",
      "         gqStDev: Double,\n",
      "         nNonRef: Int,\n",
      "         rHeterozygosity: Double,\n",
      "         rHetHomVar: Double,\n",
      "         rExpectedHetFrequency: Double,\n",
      "         pHWE: Double\n",
      "     }\n",
      " }\n"
     ]
    }
   ],
   "source": [
    "pprint(vds.variant_schema)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2aaafb46fdd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "variant_df = vds.variants_table().to_pandas()\n",
    "\n",
    "plt.clf()\n",
    "plt.subplot(2, 2, 1)\n",
    "variantgq_means = variant_df[\"va.qc.gqMean\"]\n",
    "plt.hist(variantgq_means, bins = np.arange(0, 84, 2))\n",
    "plt.xlabel(\"Variant Mean GQ\")\n",
    "plt.ylabel(\"Frequency\")\n",
    "plt.xlim(0, 80)\n",
    "\n",
    "plt.subplot(2, 2, 2)\n",
    "variant_mleaf = variant_df[\"va.qc.AF\"]\n",
    "plt.hist(variant_mleaf, bins = np.arange(0, 1.05, .025))\n",
    "plt.xlabel(\"Minor Allele Frequency\")\n",
    "plt.ylabel(\"Frequency\")\n",
    "plt.xlim(0, 1)\n",
    "\n",
    "plt.subplot(2, 2, 3)\n",
    "plt.hist(variant_df['va.qc.callRate'], bins = np.arange(0, 1.05, .01))\n",
    "plt.xlabel(\"Variant Call Rate\")\n",
    "plt.ylabel(\"Frequency\")\n",
    "plt.xlim(.5, 1)\n",
    "\n",
    "plt.subplot(2, 2, 4)\n",
    "plt.hist(variant_df['va.qc.pHWE'], bins = np.arange(0, 1.05, .025))\n",
    "plt.xlabel(\"Hardy-Weinberg Equilibrium p-value\")\n",
    "plt.ylabel(\"Frequency\")\n",
    "plt.xlim(0, 1)\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-12-20 11:18:40 Hail: INFO: Running LD prune with nSamples=843, nVariants=9085, nPartitions=4, and maxQueueSize=257123.\n",
      "2017-12-20 11:18:40 Hail: INFO: LD prune step 1 of 3: nVariantsKept=8478, nPartitions=4, time=378.749ms\n",
      "2017-12-20 11:18:41 Hail: INFO: LD prune step 2 of 3: nVariantsKept=8478, nPartitions=12, time=1.160s\n",
      "2017-12-20 11:18:41 Hail: INFO: Coerced sorted dataset\n",
      "2017-12-20 11:18:42 Hail: INFO: LD prune step 3 of 3: nVariantsKept=8478, time=429.164ms\n"
     ]
    }
   ],
   "source": [
    "#Let’s do a GWAS!\n",
    "#First, we need to restrict to variants that are:\n",
    "\n",
    "# 1) common (we’ll use a cutoff of 1%)\n",
    "# 2) uncorrelated (not in linkage disequilibrium)\n",
    "\n",
    "common_vds = (vds\n",
    "              .filter_variants_expr('va.qc.AF > 0.01')\n",
    "              .ld_prune(memory_per_core=256, num_cores=4))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(843L, 8555L)"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "common_vds.count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-12-20 11:18:42 Hail: INFO: Running linear regression on 843 samples with 1 covariate including intercept...\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Struct{\n",
      "     rsid: String,\n",
      "     qual: Double,\n",
      "     filters: Set[String],\n",
      "     pass: Boolean,\n",
      "     info: Struct{\n",
      "         AC: Array[Int],\n",
      "         AF: Array[Double],\n",
      "         AN: Int,\n",
      "         BaseQRankSum: Double,\n",
      "         ClippingRankSum: Double,\n",
      "         DP: Int,\n",
      "         DS: Boolean,\n",
      "         FS: Double,\n",
      "         HaplotypeScore: Double,\n",
      "         InbreedingCoeff: Double,\n",
      "         MLEAC: Array[Int],\n",
      "         MLEAF: Array[Double],\n",
      "         MQ: Double,\n",
      "         MQ0: Int,\n",
      "         MQRankSum: Double,\n",
      "         QD: Double,\n",
      "         ReadPosRankSum: Double,\n",
      "         set: String\n",
      "     },\n",
      "     qc: Struct{\n",
      "         callRate: Double,\n",
      "         AC: Int,\n",
      "         AF: Double,\n",
      "         nCalled: Int,\n",
      "         nNotCalled: Int,\n",
      "         nHomRef: Int,\n",
      "         nHet: Int,\n",
      "         nHomVar: Int,\n",
      "         dpMean: Double,\n",
      "         dpStDev: Double,\n",
      "         gqMean: Double,\n",
      "         gqStDev: Double,\n",
      "         nNonRef: Int,\n",
      "         rHeterozygosity: Double,\n",
      "         rHetHomVar: Double,\n",
      "         rExpectedHetFrequency: Double,\n",
      "         pHWE: Double\n",
      "     },\n",
      "     linreg: Struct{\n",
      "         beta: Double,\n",
      "         se: Double,\n",
      "         tstat: Double,\n",
      "         pval: Double\n",
      "     }\n",
      " }\n"
     ]
    }
   ],
   "source": [
    "#These filters removed about 15% of sites (we started with a bit over 10,000). \n",
    "#This is NOT representative of most sequencing datasets! \n",
    "#We have already downsampled the full thousand genomes dataset to \n",
    "#include more common variants than we’d expect by chance.\n",
    "#In Hail, the association tests accept sample annotations for the sample phenotype and covariates. \n",
    "#Since we’ve already got our phenotype of interest (caffeine consumption) in the dataset, we are good to go:\n",
    "gwas = common_vds.linreg('sa.CaffeineConsumption')\n",
    "pprint(gwas.variant_schema)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Looking at the bottom of the above printout, you can see the linear regression adds \n",
    "#new variant annotations for the beta, standard error, t-statistic, and p-value.\n",
    "def qqplot(pvals, xMax, yMax):\n",
    "    spvals = sorted(filter(lambda x: x and not(isnan(x)), pvals))\n",
    "    exp = [-log(float(i) / len(spvals), 10) for i in np.arange(1, len(spvals) + 1, 1)]\n",
    "    obs = [-log(p, 10) for p in spvals]\n",
    "    plt.clf()\n",
    "    plt.scatter(exp, obs)\n",
    "    plt.plot(np.arange(0, max(xMax, yMax)), c=\"red\")\n",
    "    plt.xlabel(\"Expected p-value (-log10 scale)\")\n",
    "    plt.ylabel(\"Observed p-value (-log10 scale)\")\n",
    "    plt.xlim(0, xMax)\n",
    "    plt.ylim(0, yMax)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2aab0f19e210>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#make a Q-Q (quantile-quantile) plot.\n",
    "qqplot(gwas.query_variants('variants.map(v => va.linreg.pval).collect()'),\n",
    "       5, 6)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2017-12-20 11:18:47 Hail: INFO: Running PCA with 5 components...\n"
     ]
    }
   ],
   "source": [
    "#The observed p-values drift away from the expectation immediately. \n",
    "#Either every SNP in our dataset is causally linked to caffeine consumption (unlikely), or there’s a confounder.\n",
    "#Sample ancestry was actually used to simulate this phenotype. \n",
    "#This leads to a stratified distribution of the phenotype. \n",
    "#The solution is to include ancestry as a covariate in our regression.\n",
    "#The linreg method can also take sample annotations to use as covariates. \n",
    "#We already annotated our samples with reported ancestry, \n",
    "#but it is good to be skeptical of these labels due to human error. \n",
    "#Genomes don’t have that problem! \n",
    "#Instead of using reported ancestry, \n",
    "#we will use genetic ancestry by including computed principal components in our model.\n",
    "#The pca method produces sample PCs in sample annotations, and can also produce variant loadings and global eigenvalues when asked.\n",
    "\n",
    "pca = common_vds.pca('sa.pca', k=5, eigenvalues='global.eigen')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{u'eigen': {u'PC1': 56.34707905481801,\n",
      "            u'PC2': 37.81090030103982,\n",
      "            u'PC3': 16.91974301822238,\n",
      "            u'PC4': 2.707349935634391,\n",
      "            u'PC5': 2.0851252187821228}}\n"
     ]
    }
   ],
   "source": [
    "pprint(pca.globals)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "pprint(pca.sample_schema)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2aab0f234fd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Now that we’ve got principal components per sample, we may as well plot them! \n",
    "#Human history exerts a strong effect in genetic datasets. \n",
    "#Even with a 50MB sequencing dataset, we can recover the major human populations.\n",
    "pca_table = pca.samples_table().to_pandas()\n",
    "colors = {'AFR': 'green', 'AMR': 'red', 'EAS': 'black', 'EUR': 'blue', 'SAS': 'cyan'}\n",
    "plt.scatter(pca_table[\"sa.pca.PC1\"], pca_table[\"sa.pca.PC2\"],\n",
    "            c = pca_table[\"sa.SuperPopulation\"].map(colors),\n",
    "            alpha = .5)\n",
    "plt.xlim(-0.6, 0.6)\n",
    "plt.xlabel(\"PC1\")\n",
    "plt.ylabel(\"PC2\")\n",
    "legend_entries = [mpatches.Patch(color=c, label=pheno) for pheno, c in colors.items()]\n",
    "plt.legend(handles=legend_entries, loc=2)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Now we can rerun our linear regression, controlling for the first few principal components and sample sex.\n",
    "pvals = (common_vds\n",
    "        .annotate_samples_table(pca.samples_table(), expr='sa.pca = table.pca')\n",
    "        .linreg('sa.CaffeineConsumption', covariates=['sa.pca.PC1', 'sa.pca.PC2', 'sa.pca.PC3', 'sa.isFemale'])\n",
    "        .query_variants('variants.map(v => va.linreg.pval).collect()'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2aab0fa3d410>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "qqplot(pvals, 5, 6)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "pvals = (common_vds\n",
    "        .annotate_samples_table(pca.samples_table(), expr='sa.pca = table.pca')\n",
    "        .linreg('sa.CaffeineConsumption',\n",
    "                covariates=['sa.pca.PC1', 'sa.pca.PC2', 'sa.pca.PC3', 'sa.isFemale'],\n",
    "                use_dosages=True)\n",
    "        .query_variants('variants.map(v => va.linreg.pval).collect()'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2aab0f6adc10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "qqplot(pvals, 5, 6)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Rare variant analysis\n",
    "#Hail doesn’t yet have rare variant kernel-based methods, but we have linear and logistic burden tests.\n",
    "#But, one can use the expression language to group and count by any arbitrary properties in variant \n",
    "#or sample annotations.\n",
    "\n",
    "\n",
    "kt = (vds.genotypes_table()\n",
    "         .aggregate_by_key(key_expr=['pop = sa.SuperPopulation', 'chromosome = v.contig'],\n",
    "                           agg_expr=['n_het = g.filter(g => g.isHet()).count()']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "+---+----------+-----+\n",
      "|pop|chromosome|n_het|\n",
      "+---+----------+-----+\n",
      "|EUR|        14|16380|\n",
      "|SAS|        17|  511|\n",
      "|EUR|         5|30717|\n",
      "|AFR|         7|11889|\n",
      "|EAS|         9|23951|\n",
      "|AFR|        21| 3529|\n",
      "|EAS|         X| 7403|\n",
      "|EAS|         1|49375|\n",
      "|EUR|        19|13483|\n",
      "|AMR|        15|18935|\n",
      "|AMR|         7|31527|\n",
      "|EUR|        13|17321|\n",
      "|EUR|        12|26134|\n",
      "|EUR|        15|15807|\n",
      "|EUR|         6|33910|\n",
      "|EAS|        20|17466|\n",
      "|SAS|        11|  901|\n",
      "|AFR|         3|15829|\n",
      "|EAS|         2|45384|\n",
      "|AMR|        18|18982|\n",
      "+---+----------+-----+\n",
      "only showing top 20 rows\n",
      "\n"
     ]
    }
   ],
   "source": [
    "kt.to_dataframe().show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "kt2 = (vds.genotypes_table()\n",
    "          .aggregate_by_key(key_expr=['''maf_bin = if (va.qc.AF < 0.01) \"< 1%\"\n",
    "                                                   else if (va.qc.AF < 0.05) \"1%-5%\"\n",
    "                                                   else \"> 5%\" ''',\n",
    "                                     'purple_hair = sa.PurpleHair'],\n",
    "                           agg_expr=['mean_gq = g.map(g => g.gq).stats().mean',\n",
    "                                     'mean_dp = g.map(g => g.dp).stats().mean']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "+-------+-----------+------------------+-----------------+\n",
      "|maf_bin|purple_hair|           mean_gq|          mean_dp|\n",
      "+-------+-----------+------------------+-----------------+\n",
      "|   > 5%|       true| 36.09305651197578|7.407450459057423|\n",
      "|   < 1%|       true| 22.68197887434976|7.374254453728496|\n",
      "|   < 1%|      false|22.986128698357074|7.492131714314245|\n",
      "|   > 5%|      false|36.341259980753755|7.533399982371768|\n",
      "|  1%-5%|       true|24.093123033233528|7.269552536649012|\n",
      "|  1%-5%|      false|  24.3519587208908|7.405582424428774|\n",
      "+-------+-----------+------------------+-----------------+\n",
      "\n"
     ]
    }
   ],
   "source": [
    "kt2.to_dataframe().show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#stop Hail context\n",
    "hc.stop()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#stop Spark context\n",
    "sc.stop()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:sparkhpc.sparkjob:\r\n"
     ]
    }
   ],
   "source": [
    "!sparkcluster stop 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}