4.Visualizing_and_quantifying_genomic_variation_in_real_data_STUDENTS.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "tags": []
   },
   "source": [
    "# Practical 4: Visualizing and quantifying genomic variation in real data\n",
    "\n",
    "\n",
    "At the end of this practical you should:\n",
    "- Know the following bioinformatic file formats and understand by which processes they can be generated\n",
    "    - [FASTA format](https://en.wikipedia.org/wiki/FASTA_format)\n",
    "    - [BED format](https://en.wikipedia.org/wiki/BED_(file_format))\n",
    "    - [SAM/BAM format](https://en.wikipedia.org/wiki/SAM_(file_format) ) (for detailed specifications, see [here](https://samtools.github.io/hts-specs/SAMv1.pdf))\n",
    "    - [VCF format](https://en.wikipedia.org/wiki/Variant_Call_Format) (for detailed specifications, see [here](https://samtools.github.io/hts-specs/VCFv4.2.pdf))\n",
    "    \n",
    "    \n",
    "- Understand how (population) genomic data can be represented through these file formats.\n",
    "    - **genomes** are usually represented through a single (haploid) consensus sequence in FASTA format\n",
    "    - **regions of interest** (Exons, Introns, Telomeres, Transposable Elements, etc) are usually represented via annotation files in BED format\n",
    "    - **variation** data is \"called\" based on BAM files and is usually recorded in VCF files. \n",
    "    \n",
    "    \n",
    "- Know that the analysis of variation data often requires additional simplifications and/or re-classification of the data:\n",
    "    - usually **only SNPs (Single Nucleotide Polymorphisms) are analysed**, more complex variants are often ignored or discarded\n",
    "    - the phase of genotypes of diploid/polyploid samples can be known (i.e. the membership of alleles on haplotypes is known), or not. Data is therefore referred to as **phased or unphased genotypes (GTs)**    \n",
    "    - **patterns of variation can differ between genomic partitions** (CDS, introns, intergenic, synonymous/non-synonymous third codon positions, ...), which makes certain stratification of the analysis of these regions interesting and sometimes even essential. \n",
    "    \n",
    "- Be able to use common Python libraries to *parse, intersect, interrogate, and visualize population genomic data*. This will include sliding-window-analyses via simple descriptive summaries of genetic diversity $\\pi$ (see lectures 2 and 4).\n",
    "- **Understand that due to background selection genetic diversity in the genome is strongly correlated with functional constraint**."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import numpy as np\n",
    "import pysam\n",
    "import pybedtools\n",
    "import allel\n",
    "import warnings\n",
    "from collections import Counter\n",
    "from IPython.display import SVG, set_matplotlib_formats\n",
    "from matplotlib import pyplot as plt\n",
    "%matplotlib inline\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. Background"
   ]
  },
  {
   "attachments": {
    "62961793-d44a-47aa-93a7-732a23b42c5a.jpg": {
     "image/jpeg": "/9j/4AAQSkZJRgABAQAAZABkAAD/4QB0RXhpZgAATU0AKgAAAAgABAEaAAUAAAABAAAAPgEbAAUA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\n"
    }
   },
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will analyse WGS data from *Iphiclides podalirius*, the [Scarce Swallowtail](https://en.wikipedia.org/wiki/Scarce_swallowtail) butterfly. Both the [reference genome](https://academic.oup.com/g3journal/article/12/9/jkac193/6656352) and the resequencing data for seven individuals sampled throughout Europe have been recently generated by the Lohse lab.\n",
    "\n",
    "*I. podalirius* is common in open habitats in the Palaearctic, ranging from France to Western China, but is absent in Northern areas incl. the UK. Studying genome-wide variation in this species is of interest given that *I. podalirius* is separated by a narrow hybrid zone in Southern France from its closely related  sister species *I. feisthamelii* (which lives in Northern Africa and the Iberian Peninsula). Although the two species have no known differences in ecology or life history, they differ consistently in nuclear DNA, wing patterns (including UV reflectance of males) and genital morphology.\n",
    "\n",
    "The samples you will be working on are listed in the following table and map.\n",
    "\n",
    "| Sample          | Source  | Comments                 |\n",
    "|-----------------|---------|--------------------------|\n",
    "| IP_504.10X      | France  | used for genome assembly |\n",
    "| IP_502          | France  |                          |\n",
    "| IP_RVcoll14F754 | Greece  |                          |\n",
    "| IP_RVcoll14E561 | Romania |                          |\n",
    "| IP_RVcoll14H960 | France  |                          |\n",
    "| IP_1330         | France  |                          |\n",
    "| IP_RVcoll12R048 | Italy   |                          |\n",
    "\n",
    "![iphiclides_map.jpg](attachment:62961793-d44a-47aa-93a7-732a23b42c5a.jpg)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "tags": []
   },
   "source": [
    "# 2. Practical"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2.0 Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This practical analyses on real-world, genomic data. A common difficulty of working with genomic data is the size of the datasets. To keep things manageable, we will focus on chromosome 5 for most of the practical.\n",
    "\n",
    "The data has been compressed, using [`tar`](https://en.wikipedia.org/wiki/Tar_(computing)) and [`gzip`](https://en.wikipedia.org/wiki/Gzip), and deposited on Dropbox.\n",
    "\n",
    "We will first download the compressed TAR-archive from Dropbox and decompress the files. This should take about 60s."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "! wget https://www.dropbox.com/s/5s16m31zwuet1nw/p4_data.tar.gz?dl=1 -q --show-progress -O p4_data.tar.gz\n",
    "! tar zxf p4_data.tar.gz"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can now inspect the downloaded folder by running the command [`ls`](https://en.wikipedia.org/wiki/Ls)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "! ls -lh p4_data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The folder should contain the following files:\n",
    "- `iphiclides_podalirius.genome.fasta` FASTA file of the genome of *I. podalirius* (the whole genome assembly, not only chromosome 5)\n",
    "- `iphiclides_podalirius.IP_504.chromosome_5.vcf.gz` VCF file of the variants called on chromosome 5 \n",
    "- `iphiclides_podalirius.IP_504.chromosome_5.vcf.gz.tbi` The index file for the VCF file (this is needed to quickly access VCF files based on sequence and position)\n",
    "- `iphiclides_podalirius.chr5.intergenic.bed` BED file delimiting the \"intergenic\" regions along chromosome 5\n",
    "- `iphiclides_podalirius.chr5.exons.bed` BED file delimiting the \"exonic\" regions along chromosome 5\n",
    "- `iphiclides_podalirius.chr5.tcp.bed` BED file delimiting the third-codon-position sites along chromosome 5\n",
    "- `iphiclides_podalirius.chr5.introns.bed`  BED file delimiting the \"intronic\" regions along chromosome 5\n",
    "- `iphiclides_podalirius.chr5.callable.bed` BED file delimiting those regions along chromosome 5 for which we have adequate coverage from each re-sequencing sample used in the VCF \n",
    "- `iphiclides_chrom_data.csv` CSV file used in the last part of this practical"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "tags": []
   },
   "source": [
    "## 2.1.0 Working with FASTA files\n",
    "\n",
    "[FASTA](https://en.wikipedia.org/wiki/FASTA_format) files are used for storing sequence data – usually DNA, RNA, or protein sequences. In this section you will investigate the length of the sequences contained in the genome assembly using some `pandas` tricks that you have already encountered.\n",
    "\n",
    "In this section you will:\n",
    "\n",
    "1. parse the FASTA file of the *I. podalirius* genome assembly using the [`pysam`](https://pysam.readthedocs.io/en/latest/) library and\n",
    "2. visualise the distribution of sequence lengths"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.1.1 Parsing the FASTA file of the *I. podalirius* genome assembly\n",
    "Let's start by creating a [`pysam.Fastafile`](https://pysam.readthedocs.io/en/latest/api.html#pysam.FastaFile) object from the *I. podalirius* reference genome sequence contained in the file `~/p4_data/iphiclides_podalirius.genome.fasta` and store it in the variable `fasta`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "fasta = pysam.Fastafile(\"p4_data/iphiclides_podalirius.genome.fasta\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.1.2 Visualising the distribution of sequence lengths"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-warning\">\n",
    "    <b>Python recipe: pd dataframe refresher</b>\n",
    "    \n",
    "    \n",
    "We have so far dealt with python lists and arrays. When analyzing real data, we often want to store many pieces of information together. A [pandas](https://pandas.pydata.org/docs/reference/index.html) dataframe (see practical 2) lets us do just that (if you've used `R` before, you will be familiar with dataframes)! Dataframes make further analysis and plotting with [seaborn](https://seaborn.pydata.org/api.html) much easier.\n",
    "    \n",
    "Let's use the <code>pysam.Fastafile</code> object stored in the variable <code>fasta</code> to get the names (<code>fasta.references</code>) and lengths (<code>fasta.lengths</code>) of the sequences contained in the FASTA file and use them to construct a dict from which the dataframe is build. This creates dataframe with two columns: <code>'chrom'</code> and <code>'lengths'</code>:\n",
    "</div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>chrom</th>\n",
       "      <th>lengths</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>iphiclides_podalirius.IP_504.chromosome_Z</td>\n",
       "      <td>21051824</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>iphiclides_podalirius.IP_504.chromosome_2</td>\n",
       "      <td>17990613</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>iphiclides_podalirius.IP_504.chromosome_3</td>\n",
       "      <td>17426007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>iphiclides_podalirius.IP_504.chromosome_4</td>\n",
       "      <td>16757360</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>iphiclides_podalirius.IP_504.chromosome_5</td>\n",
       "      <td>15345845</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>255</th>\n",
       "      <td>iphiclides_podalirius.IP_504.contig_226</td>\n",
       "      <td>5021</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>256</th>\n",
       "      <td>iphiclides_podalirius.IP_504.contig_227</td>\n",
       "      <td>4857</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>257</th>\n",
       "      <td>iphiclides_podalirius.IP_504.contig_228</td>\n",
       "      <td>4599</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>258</th>\n",
       "      <td>iphiclides_podalirius.IP_504.contig_229</td>\n",
       "      <td>4132</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>259</th>\n",
       "      <td>iphiclides_podalirius.IP_504.contig_230</td>\n",
       "      <td>2926</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>260 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                         chrom   lengths\n",
       "0    iphiclides_podalirius.IP_504.chromosome_Z  21051824\n",
       "1    iphiclides_podalirius.IP_504.chromosome_2  17990613\n",
       "2    iphiclides_podalirius.IP_504.chromosome_3  17426007\n",
       "3    iphiclides_podalirius.IP_504.chromosome_4  16757360\n",
       "4    iphiclides_podalirius.IP_504.chromosome_5  15345845\n",
       "..                                         ...       ...\n",
       "255    iphiclides_podalirius.IP_504.contig_226      5021\n",
       "256    iphiclides_podalirius.IP_504.contig_227      4857\n",
       "257    iphiclides_podalirius.IP_504.contig_228      4599\n",
       "258    iphiclides_podalirius.IP_504.contig_229      4132\n",
       "259    iphiclides_podalirius.IP_504.contig_230      2926\n",
       "\n",
       "[260 rows x 2 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.DataFrame({'chrom': fasta.references, 'lengths': fasta.lengths})\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Pandas also allows running simple computations on the dataframe columns:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(\"=> Total length of sequences in FASTA is: \", df['lengths'].sum(), \"bases\")\n",
    "print(\"=> Total number of sequences in FASTA is: \", df['lengths'].count())\n",
    "print(f\"=> The longest sequence is {df['lengths'].max()} b, the shortest is {df['lengths'].min()} b\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The min and max suggest that there is a big range of sequence lengths contained in the fasta file. Let's plot this distribution!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "p = sns.histplot(data=df, x='lengths', log_scale=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-success\">    \n",
    "    <b>Exercise 1:</b>  \n",
    "Can you guess why this distribution is bi-modal? \n",
    "    \n",
    "**Hint:** Look at the names of the sequences in the fasta file.\n",
    "    </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#SOL"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's add a new column `'label'` to the dataframe to split up sequences in the fasta file by length."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-warning\">\n",
    "    <b>Python recipe: conditional expressions</b> \n",
    "    \n",
    "We will use the <a href=\"https://numpy.org/doc/stable/reference/generated/numpy.where.html\"> <code>numpy.where()</code> </a> function to populate the new <code>'label'</code> column based on the conditional expression <code> < 1_000_000 </code>. If we execute this expression on its own, we get a boolean series composed of True/False values for each row of the dataframe:\n",
    "    </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0      False\n",
       "1      False\n",
       "2      False\n",
       "3      False\n",
       "4      False\n",
       "       ...  \n",
       "255     True\n",
       "256     True\n",
       "257     True\n",
       "258     True\n",
       "259     True\n",
       "Name: lengths, Length: 260, dtype: bool"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['lengths'] < 1_000_000"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-warning\">        \n",
    "We will use the conditional expressions to populate the new column <code>df['label']</code> with strings using the numpy function <code>np.where(condition, x, y)</code>:\n",
    "    </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['label'] = np.where(df['lengths'] < 1_000_000, '< 1 Mb', '>= 1Mb')\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-success\">    \n",
    "    <b>Exercise 2:</b>  \n",
    "Colour the histogram of sequence length we've plotted above using the newly generated column.\n",
    "    \n",
    "**Hint:** Check out the <a href=\"https://seaborn.pydata.org/tutorial/distributions.html\"> <code>sns.displot()</code> </a> documentation, the parameter `hue` will do the trick.\n",
    "    </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# SOL"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-success\">    \n",
    "    <b>Exercise 3:</b>\n",
    "<ul>\n",
    "<li>    \n",
    "Subset <code>df</code> so that it only contains chromosomes and store this new dataframe in the variable <code>chrom_df</code>.</li>\n",
    "<li>Calculate the total length and number of chromosomes.</li>\n",
    "<li>How many chromsomes does <i>I. podalirius</i> have?</li>\n",
    "<li>What is the total length of chromosome scaffolds?</li>\n",
    "    </ul>\n",
    "\n",
    "**Hint:** Use the boolian array <code>df['label']=='>= 1Mb'</code> and inside a python index to subset\n",
    "    <code>df</code>\n",
    "        </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# SOL"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For most of the rest of the practical you will focus on chromosome 5 of the *I. podalirius* genome. Let's declare the two variables `chrom_name` and `chrom_length` for this chromosome:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "chrom_name = chrom_df['chrom'][4]\n",
    "chrom_length = 15345845"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "tags": []
   },
   "source": [
    "## 2.2.0 Working with VCF files\n",
    "\n",
    "- [VCF](https://en.wikipedia.org/wiki/VCF_format) files are used for storing information about genomic variation \n",
    "  - along one (or more) linear reference sequences and\n",
    "  - for one or more samples for which genotypes were called.\n",
    "- A detailed description of the file format is available [here](https://samtools.github.io/hts-specs/VCFv4.2.pdf) and in the Appendix at the end of the notebook.\n",
    " \n",
    "In this section you will:\n",
    "1. Parse the VCF file of one of the *I. podalirius* chromosomes using the [scikit-allel](https://scikit-allel.readthedocs.io/en/stable/) library.\n",
    "2. Create a mask array which includes **only biallelic SNPs** that passed the filters. These will be the variants we will use for subsequent analysis.\n",
    "4. Categorize variants based on genotypes (HOMREF, HOMALT, HET, MISS) by sample."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "tags": []
   },
   "source": [
    "### 2.2.1 Parsing a VCF file using scikit-allel"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A VCF file contains a lot of information about the variants and is grouped into `fields`. Let's first define the fields we want to parse from the VCF file (we don't need all of them):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "query_fields = ['samples', 'calldata/GT', 'variants/CHROM', 'variants/POS', 'variants/NUMALT', 'variants/is_snp']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The VCF file we will be parsing is:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "vcf_f = \"p4_data/iphiclides_podalirius.chr5.vcf.gz\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The function `allel.read_vcf()` will load the entire VCF file into a python [dictionary](https://docs.python.org/3/tutorial/datastructures.html#dictionaries) of arrays/lists:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "# reading the file takes a few seconds to read the file, be patient ...\n",
    "vcf_dict = allel.read_vcf(vcf_f, fields=query_fields)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Using the `keys()` method of the dict, we can check that the keys of `vcf_dict` are `query_fields` we have specified:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['samples', 'calldata/GT', 'variants/CHROM', 'variants/NUMALT', 'variants/POS', 'variants/is_snp'])"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vcf_dict.keys()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Data can be explored using the keys:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(\"[+] samples \", vcf_dict['samples'], vcf_dict['samples'].dtype)\n",
    "print(\"[+] genotypes (shape, dtype) \", vcf_dict['calldata/GT'].shape, vcf_dict['calldata/GT'].dtype)\n",
    "print(\"[+] CHROM (shape, dtype) \", vcf_dict['variants/CHROM'].shape, vcf_dict['variants/CHROM'].dtype)\n",
    "print(\"[+] POS (shape, dtype) \", vcf_dict['variants/POS'].shape, vcf_dict['variants/POS'].dtype)\n",
    "print(\"[+] NUMALT (shape, dtype) \", vcf_dict['variants/NUMALT'].shape, vcf_dict['variants/NUMALT'].dtype)\n",
    "print(\"[+] is_snp (shape, dtype) \", vcf_dict['variants/is_snp'].shape, vcf_dict['variants/is_snp'].dtype)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We often want to filter variants before analysing them. Since no simple mutation models exist for indels and complex variants, population genomic analyses are generally restricted to SNPs. For the same reason, analyses of WGS data are also often limited to biallelic SNPs which can be interpreted assuming the infinite sites mutation model (see mini lecture 1 & 2).\n",
    "\n",
    "How do we filter vcf files? We can construct a boolean `mask_array` by chaining several logical expressions together with the `&` operator. Let's define a mask for biallelic SNPs:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "numalt_array = vcf_dict['variants/NUMALT']\n",
    "is_SNP_array = vcf_dict['variants/is_snp']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- `is_SNP_array` is a boolean array of which we want only those values which are `True` (i.e. those variants that are SNPs)\n",
    "- `numalt_array` is an integer array of which we want only those values which are `1` (i.e. those variants that are bi-allelic)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "Counter(numalt_array)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "mask_array = (numalt_array == 1) & (is_SNP_array == True)\n",
    "print(mask_array)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This boolean `mask_array` can be used to index (and thus subset) other arrays, e.g. the genotype array `gt_array` and the `pos_array`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "snp_gts = vcf_dict['calldata/GT'][mask_array]\n",
    "snp_pos = vcf_dict['variants/POS'][mask_array]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-success\">    \n",
    "    <b>Exercise 3b:</b>\n",
    "<ul>\n",
    "<li>    \n",
    "What fraction of variants passes both filtes? Which of the two filtering criteria matters more? \n",
    "    </ul>\n",
    "\n",
    "**Hint:** <code>Counter</code> and/or <code>len()</code> are useful.\n",
    "        </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "#SOL"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-danger\">        \n",
    "From here onwards we will only use the biallelic SNP genotypes (<code>snp_gts</code>) and their positions (<code>snp_pos</code>).\n",
    "    </div>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.2.3 Categorize variants by genotype and sample\n",
    "- Scikit-allel's `GenotypeArray` class is a convenient way of interacting with genotype data.\n",
    "- A full description of its methods can be found [here](https://scikit-allel.readthedocs.io/en/stable/model/ndarray.html#genotypes)\n",
    "- Genotypes are represented as a 3-dimensional numpy array of integers. By convention, \n",
    "    - the first dimension corresponds to the variants genotyped, \n",
    "    - the second dimension corresponds to the samples genotyped, and \n",
    "    - the third dimension corresponds to the ploidy of the samples.\n",
    "- Each integer within the array corresponds to an allele index, where \n",
    "    - `0` is the reference allele, \n",
    "    - `1` is the first alternate allele, \n",
    "    - `2` is the second alternate allele, \n",
    "    - ... \n",
    "    - `-1` (or any other negative integer) is a missing allele call. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print('[+] Genotype data is parsed as numpy arrays by scikit-allel: \\n%s with shape %s' % (type(snp_gts), snp_gts.shape))\n",
    "print('[+] The first variant has the following genotype data in the VCF file: \\n%s' % snp_gts[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-warning\">\n",
    "    <b>Python recipe:</b> We can use the class   \n",
    "    <a href=\"https://scikit-allel.readthedocs.io/en/stable/model/ndarray.html#allel.GenotypeArray\"><code>allel.GenotypeArray</code></a> to interact more easily with genotype data. This class also has access to all methods available for <a href=\"https://scikit-allel.readthedocs.io/en/stable/model/ndarray.html#genotypes\"><code>allel.Genotypes</code></a>.\n",
    "    </div>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Typing the variable name gives the `shape` and prints the head and tail of the array:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "snp_ga = allel.GenotypeArray(snp_gts)\n",
    "snp_ga"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The genotype array class has a whole bunch of useful methods many of which are self-explanatory: "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(dir(snp_ga))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-warning\">\n",
    "    <b>Python recipe:</b> We can use the functions \n",
    "    \n",
    "- <a href=\"https://scikit-allel.readthedocs.io/en/stable/model/ndarray.html#allel.Genotypes.count_missing\"><code>allel.Genotypes.count_missing()</code></a>,\n",
    "- <a href=\"https://scikit-allel.readthedocs.io/en/stable/model/ndarray.html#allel.Genotypes.count_hom_ref\"><code>allel.Genotypes.count_hom_ref()</code></a>,\n",
    "- <a href=\"https://scikit-allel.readthedocs.io/en/stable/model/ndarray.html#allel.Genotypes.count_hom_alt\"><code>allel.Genotypes.count_hom_alt()</code></a>, and\n",
    "- <a href=\"https://scikit-allel.readthedocs.io/en/stable/model/ndarray.html#allel.Genotypes.count_het\"><code>allel.Genotypes.count_het()</code></a> \n",
    "\n",
    "to count those four possible genotypes for each sample. \n",
    "    </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "count_missing = snp_ga.count_missing(axis=0)\n",
    "count_hom_ref = snp_ga.count_hom_ref(axis=0)\n",
    "count_hom_alt = snp_ga.count_hom_alt(axis=0)\n",
    "count_het = snp_ga.count_het(axis=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's turn these arrays of counts into columns of a dataframe:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[+] samples  ['IP_RVcoll14H960' 'IP_RVcoll14F754' 'IP_502' 'IP_1330' 'IP_RVcoll12R048'\n",
      " 'IP_504.10X' 'IP_RVcoll14E561']\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sample</th>\n",
       "      <th>MISS</th>\n",
       "      <th>HOMREF</th>\n",
       "      <th>HOMALT</th>\n",
       "      <th>HET</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>IP_RVcoll14H960</td>\n",
       "      <td>2286</td>\n",
       "      <td>395855</td>\n",
       "      <td>44273</td>\n",
       "      <td>97540</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>IP_RVcoll14F754</td>\n",
       "      <td>2507</td>\n",
       "      <td>385604</td>\n",
       "      <td>60453</td>\n",
       "      <td>91390</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>IP_502</td>\n",
       "      <td>2075</td>\n",
       "      <td>405837</td>\n",
       "      <td>41623</td>\n",
       "      <td>90419</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>IP_1330</td>\n",
       "      <td>6271</td>\n",
       "      <td>203691</td>\n",
       "      <td>90480</td>\n",
       "      <td>239512</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>IP_RVcoll12R048</td>\n",
       "      <td>4281</td>\n",
       "      <td>428711</td>\n",
       "      <td>57904</td>\n",
       "      <td>49058</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>IP_504.10X</td>\n",
       "      <td>17</td>\n",
       "      <td>456263</td>\n",
       "      <td>98</td>\n",
       "      <td>83576</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>IP_RVcoll14E561</td>\n",
       "      <td>2384</td>\n",
       "      <td>387195</td>\n",
       "      <td>57351</td>\n",
       "      <td>93024</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            sample  MISS  HOMREF  HOMALT     HET\n",
       "0  IP_RVcoll14H960  2286  395855   44273   97540\n",
       "1  IP_RVcoll14F754  2507  385604   60453   91390\n",
       "2           IP_502  2075  405837   41623   90419\n",
       "3          IP_1330  6271  203691   90480  239512\n",
       "4  IP_RVcoll12R048  4281  428711   57904   49058\n",
       "5       IP_504.10X    17  456263      98   83576\n",
       "6  IP_RVcoll14E561  2384  387195   57351   93024"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "samples = vcf_dict['samples']\n",
    "print(\"[+] samples \", vcf_dict['samples'])\n",
    "df_GT_counts = pd.DataFrame({'sample': samples, 'MISS': count_missing, 'HOMREF': count_hom_ref, 'HOMALT': count_hom_alt, 'HET': count_het})\n",
    "df_GT_counts"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-success\">\n",
    "    <b>Exercise 4:</b> Use these counts to compute the per base heterozygosity for each sample. How does the mean heterozygosity across samples compare to your estimate of genetic diversity for 4D sites in <i>D. melanogaster </i> (exercise 14 from practical 2) and other insects <a href=\"https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.1001388\"> (see Figure 1 in Leffler et al)</a>\n",
    "    </div"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "#SOL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can now also look where the different genotypes are located on chromosome 5, as opposed to just counting them.\n",
    "For this we can use the `plot_int_array_by_samples()` function we have used in practical 3."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Note this function is a tool that allows visualising genotype arrays. \n",
    "#It is mainly plotting code. \n",
    "#you do not need to understand this function (unless you are interested in python plotting).\n",
    "import matplotlib as mpl\n",
    "def plot_int_array_by_samples(array, labels, samples, title=None):\n",
    "    # set figure height depending on number of samples\n",
    "    fig, ax = plt.subplots(figsize=(20, .4 * array.shape[1]))\n",
    "    # create colour palette\n",
    "    palette = sns.color_palette('colorblind')\n",
    "    # subset palette based on unique values in array\n",
    "    cmap = mpl.colors.ListedColormap([palette[i] for i in range(len(labels))])\n",
    "    # plot\n",
    "    ax.pcolormesh(array.T, cmap=cmap, vmin=np.min(array), vmax=np.max(array))\n",
    "    # tidy up axes\n",
    "    ax.set_yticks(np.arange(array.shape[1]) + .5)\n",
    "    ax.set_yticklabels(samples)\n",
    "    ax.set_ylabel('Samples')\n",
    "    ax.set_xlabel('Variant')\n",
    "    # make legend based on labels (MUST BE IN CORRECT ORDER!)\n",
    "    patches = [mpl.patches.Patch(color=cmap(i)) for i in range(len(labels))]\n",
    "    box_pos = ax.get_position()\n",
    "    plt.legend(patches, labels, bbox_to_anchor=(1.1, 0.5))\n",
    "    if title:\n",
    "        ax.set_title(title)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We first need to create the input for `plot_int_array_by_samples()`:\n",
    "- `array` should contain a digit for each of the four genotypes, i.e.: 0 (=HOMREF), 1 (=HET), 2 (=HOMALT), 3 (=MISSING)\n",
    "- `labels` should contain the labels we want to see in the legend `['HOMREF', 'HET', 'HOMALT', 'MISSING']`\n",
    "- `samples` should contain the sample names\n",
    "\n",
    "\n",
    "Generating this plot will take a few seconds so be patient!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "array = np.zeros((snp_gts.shape[0], snp_gts.shape[1]))\n",
    "array = np.where(allel.GenotypeArray(snp_gts).is_hom_ref(), 0, \n",
    "                              np.where(allel.GenotypeArray(snp_gts).is_het(), 1,\n",
    "                                    np.where(allel.GenotypeArray(snp_gts).is_hom_alt(), 2, 3)))\n",
    "labels = ['HOMREF', 'HET', 'HOMALT', 'MISSING']\n",
    "samples = vcf_dict['samples']\n",
    "plot_int_array_by_samples(array, labels, samples)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-warning\">\n",
    "    <b>Python recipe:</b> In order to plot the counts of genotypes in <code>df_GT_counts</code>, we are going to turn this dataframe into a <a href=\"https://en.wikipedia.org/wiki/Pivot_table\">Pivot table</a> using <a href=\"https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.pivot_table.html#pandas.DataFrame.pivot_table\"><code>pandas.DataFrame.pivot_table</code></a>. This will index the dataframe by the sample column and allow plotting functions to access values based on sample name.\n",
    "    </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "pivot_table = pd.pivot_table(df_GT_counts, index='sample', aggfunc=np.sum)\n",
    "pivot_table"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-warning\">\n",
    "    <b>Python recipe:</b> All Pandas DataFrames support plotting via <a href=\"https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.plot.html#pandas-dataframe-plot\"><code>pandas.DataFrame.plot()</code></a>. \n",
    "    </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "pivot_table.plot(kind='barh', stacked=True)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " Another useful plot is <a href=\"https://seaborn.pydata.org/generated/seaborn.barplot.html\"><code>seaborn.barplot()</code></a>. This plots the heterozygosity of each sample:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "heterozygosity = snp_ga.count_het(axis=0) / chrom_length\n",
    "sns.barplot(x=heterozygosity, y=samples)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-success\">\n",
    "    <b>Exercise 5:</b> Look at the two plots above. \n",
    "    \n",
    "- Are there clear outliers in terms of HET counts? Given the map of sample locations what could explain them? \n",
    "    \n",
    "- Does any sample stand out in terms of HOMALT? \n",
    "\n",
    "- Adapt the code cell below to remove the outlier sample from the dataset. This will overwrite the variable <code>snp_ga</code>. How is the mean heterozygosity affected?\n",
    "    \n",
    "- Redo the barplot to make sure you get the same plot as before, with the exception of the removed outlier sample.\n",
    "    </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# SOL\n",
    "# code for subsetting samples\n",
    "samples = df_GT_counts['sample']\n",
    "samples_new = [sample for sample in samples if sample != 'outlier']\n",
    "# load vcf again\n",
    "vcf_dict_new = allel.read_vcf(vcf_f, fields=query_fields, samples=samples_new)\n",
    "# create the genotype array of biallelic SNPs (using the mask_array)\n",
    "snp_ga_new = allel.GenotypeArray(vcf_dict_new['calldata/GT'])[mask_array]\n",
    "# estimate heterozygosity\n",
    "heterozygosity_new = snp_ga_new.count_het(axis=0) / chrom_length\n",
    "#what is the mean heterozygosity after removing the outlier?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "# CAUTION: Only once you are happy with the removal of the outlier sample, execute this cell!\n",
    "snp_ga = snp_ga_new\n",
    "samples = samples_new"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.2.5 Visualising measures of diversity along the chromosome"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-warning\">\n",
    "    <b>Python recipe:</b> We can convert the   \n",
    "    <a href=\"https://scikit-allel.readthedocs.io/en/stable/model/ndarray.html#allel.GenotypeArray\"><code>allel.GenotypeArray</code></a> into an <a href=\"https://scikit-allel.readthedocs.io/en/stable/model/ndarray.html#allelecountsarray\"><code>allel.AllelCountsArray</code></a> using the method\n",
    "    <a href=\"https://scikit-allel.readthedocs.io/en/stable/model/ndarray.html#allel.GenotypeArray.count_alleles\"><code>allel.GenotypeArray.count_alleles()</code></a>. Allele counts can be used to compute many <a href=\"https://scikit-allel.readthedocs.io/en/stable/stats.html#\">population genetic summary statistics</a> using scikit-allel methods.\n",
    "    </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "snp_ac = snp_ga.count_alleles()\n",
    "print(snp_ac)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-success\">\n",
    "    <b>Exercise 6:</b> Use the <code>allel</code> function <a href='https://scikit-allel.readthedocs.io/en/stable/stats/diversity.html'> <code>sequence diversity</code></a> to estimate pairwise genetic diversity (Nei's $\\pi$) in <i>I. podalirius</i>. \n",
    "    \n",
    "<b>Hint</b>: The method requires two inputs: \n",
    "  - the position of each variant <code>snp_pos</code>\n",
    "  - the allele counts for each variant <code>snp_ac</code> and\n",
    "    </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# SOL"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-success\">\n",
    "    <b>Exercise 7:</b> \n",
    "    \n",
    "- What could explain the fact that the estimate of pairwise diversity is higher than the average heterozygosity you've estimated in the previous section? \n",
    "\n",
    "- Can you quantify this differrence?\n",
    "    </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "#SOL"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-info\">\n",
    "    <b>PopGen:</b> We can quantify the difference between expected heterozygosity (under Hardy-Weinberg equilibrium) and the observed heterozygosity via the fixation index:\n",
    "    $$ F_{IS} = 1 - \\frac{H_O}{H_E}$$\n",
    "    </div>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-warning\">\n",
    "    <b>Python recipe:</b> A more granular view of how genetic diversity varies along the chromosome can be gained through the     \n",
    "    <a href=\"https://scikit-allel.readthedocs.io/en/stable/stats/diversity.html#allel.windowed_diversity\"><code>allel.windowed_diversity()</code></a> method.\n",
    "</div>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For example, diversity measured in 100kb windows varies over a factor of three:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The maximum diversity is 0.011199650072150073\n",
      "The SD of pi is 0.0013755248145731977\n"
     ]
    }
   ],
   "source": [
    "# plotting along the genome\n",
    "window_size = 100_000\n",
    "pi, windows, bases, counts = allel.windowed_diversity(snp_pos, snp_ac, size=window_size, step=int(window_size/2))\n",
    "print(\"The maximum diversity is\", max(pi))\n",
    "print(\"The SD of pi is\", np.std(pi))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It helps to store the windowise $\\pi$ estimates and the coordinates in a dataframe:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "win_coord = windows[:,1] - (window_size / 2)\n",
    "win_len = ['100kb']*(len(pi))\n",
    "win_pi_df = pd.DataFrame({'win_coord':win_coord,'pi':pi,'win_len': win_len})\n",
    "fig, ax = plt.subplots(figsize=(18,6))\n",
    "sns.lineplot(x='win_coord', y='pi', hue = 'win_len', data = win_pi_df)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-success\">\n",
    "    <b>Exercise 8:</b> Explore how the heterogeneity in $\\pi$ depends on the window size. Change the <code>window_size</code> variable. Go crazy!!!\n",
    "    \n",
    "<b>Hint:</b> Use <code>pd.concat([], ignore_index =True)</code> (see practical 2) to concatenate new dataframes to  <code>win_pi_df</code> and plot $\\pi$ trajectories for different window sizes in a single plot.\n",
    "    </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "#SOL "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "win_coord = windows[:,1] - (window_size / 2)\n",
    "win_len = ['100kb']*(len(pi))\n",
    "win_pi_df = pd.DataFrame({'win_coord':win_coord,'pi':pi,'win_len': win_len})\n",
    "fig, ax = plt.subplots(figsize=(18,6))\n",
    "sns.lineplot(x='win_coord', y='pi', hue = 'win_len', data = win_pi_df)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2.3.0 Working with BED files\n",
    "- The *I. podalirius* genome has been annotated using the gene prediction software [BRAKER](https://github.com/Gaius-Augustus/BRAKER) and RNAseq read data. \n",
    "- This yielded an \"gene annotation\" describing the position of each gene, mRNA, exon, intron, and start/stop-codon.\n",
    "- These BED files are in the `p4_data/` folder you have downloaded at the beginning of the practical:\n",
    "\n",
    "|          feature type          |                               filename                               |\n",
    "|--------------------------------|----------------------------------------------------------------------|\n",
    "| exons                          | `p4_data/iphiclides_podalirius.chr5.exons.bed`          |\n",
    "| third-codon-positions of exons | `p4_data/iphiclides_podalirius.chr5.tcp.bed` |\n",
    "| introns                        | `p4_data/iphiclides_podalirius.chr5.introns.bed`        |\n",
    "| intergenic                     | `p4_data/iphiclides_podalirius.chr5.intergenic.bed`    |\n",
    "| **callable-regions BED file**  | `p4_data/iphiclides_podalirius.chr5.callable.bed`                                 |\n",
    "\n",
    "- The **callable-regions BED file** contains regions of the genome where all samples displayed sufficient, but not excessive, sequencing depth. \n",
    "  - This means that for every base in **callable-regions BED file** we can be sure that it is either:\n",
    "    - accurately described as **variant if present in the VCF** or\n",
    "    - accurately described as **invariant if not present in the VCF**\n",
    "\n",
    "\n",
    "\n",
    "- One troublesome part in Bioinformatics is that BED files have a different coordinate system than GFF/GTF/VCF files.\n",
    "  - BED files are 0-based (the first base of the chromosome is considered position 0) and the feature does not include the stop position.\n",
    "  - VCF, GFF, and GTF files are 1-based (the first base of the chromosome is considered position 1) and the feature includes the stop position.\n",
    "\n",
    "\n",
    "  \n",
    "- The python library [pybedtools](https://daler.github.io/pybedtools/) is a convenient way of interacting with annotation files (BED/GFF/GTF/VCF)\n",
    "  - It takes care of the complicated steps involving coordinate-system-operations\n",
    "  - For bigger datasets it is often easier to use the underlying [BedTools](https://bedtools.readthedocs.io/en/latest/) program directly in the commandline.\n",
    "\n",
    "In this section you will focus on the following genomic regions described in BED files:\n",
    "- **exons**: all genomic positions located within exons of genes\n",
    "- **tcp**: all genomic positions located within the third codon positions of exons of genes\n",
    "- **introns**: all genomic positions located within introns of genes\n",
    "- **intergenic**: all genomic positions that are not part of genes\n",
    "\n",
    "For each of these genomic regions you will:\n",
    "1. parse the BED file describing the genomic regions\n",
    "2. Intersect the intervals with the **callable-regions BED file** which delimits those regions that obtained sufficient sequencing depth in all samples (i.e. a depth filter), this will yield:\n",
    "   - **exons_callable**\n",
    "   - **tcp_callable**\n",
    "   - **introns_callable**\n",
    "   - **intergenic_callable**\n",
    "3. Subset the variants to only those contained within each genomic region\n",
    "4. Quantify variation within each of these four types of genomic regions "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.3.1 Parsing BED files"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's first define variables for each of the paths to the BED files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "exon_f = 'p4_data/iphiclides_podalirius.chr5.exons.bed'\n",
    "tcp_f = 'p4_data/iphiclides_podalirius.chr5.tcp.bed'\n",
    "intron_f = 'p4_data/iphiclides_podalirius.chr5.introns.bed'\n",
    "intergenic_f = 'p4_data/iphiclides_podalirius.chr5.intergenic.bed'\n",
    "callable_f = 'p4_data/iphiclides_podalirius.chr5.callable.bed'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's load the different BED files of regions of interest into variables using `pybedtools`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "exon_intervals = pybedtools.BedTool(exon_f)\n",
    "tcp_intervals = pybedtools.BedTool(tcp_f)\n",
    "intron_intervals = pybedtools.BedTool(intron_f)\n",
    "intergenic_intervals = pybedtools.BedTool(intergenic_f)\n",
    "callable_intervals = pybedtools.BedTool(callable_f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's take a look at the exons ..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iphiclides_podalirius.IP_504.chromosome_5\t2290\t2546\tg13519.t1.exon1\t.\t+\n",
      " iphiclides_podalirius.IP_504.chromosome_5\t7206\t7439\tg13519.t1.exon2\t.\t+\n",
      " iphiclides_podalirius.IP_504.chromosome_5\t8853\t9251\tg13520.t1.exon1\t.\t-\n",
      " iphiclides_podalirius.IP_504.chromosome_5\t10882\t11147\tg13520.t1.exon2\t.\t-\n",
      " iphiclides_podalirius.IP_504.chromosome_5\t12260\t13178\tg13520.t1.exon3\t.\t-\n",
      " iphiclides_podalirius.IP_504.chromosome_5\t29786\t29812\tg13521.t1.exon1\t.\t+\n",
      " iphiclides_podalirius.IP_504.chromosome_5\t29947\t30014\tg13521.t1.exon2\t.\t+\n",
      " iphiclides_podalirius.IP_504.chromosome_5\t30609\t30868\tg13521.t1.exon3\t.\t+\n",
      " iphiclides_podalirius.IP_504.chromosome_5\t31636\t31825\tg13521.t1.exon4\t.\t+\n",
      " iphiclides_podalirius.IP_504.chromosome_5\t32651\t32772\tg13521.t1.exon5\t.\t+\n",
      " "
     ]
    }
   ],
   "source": [
    "exon_intervals.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chrom  :  iphiclides_podalirius.IP_504.chromosome_5\n",
      "start  :  2290\n",
      "end    :  2546\n",
      "name   :  g13519.t1.exon1\n",
      "Fields :  ['iphiclides_podalirius.IP_504.chromosome_5', '2290', '2546', 'g13519.t1.exon1', '.', '+']\n"
     ]
    }
   ],
   "source": [
    "example_interval = exon_intervals[0]\n",
    "print(\"chrom  : \", example_interval.chrom)\n",
    "print(\"start  : \", example_interval.start)\n",
    "print(\"end    : \", example_interval.end)\n",
    "print(\"name   : \", example_interval.name)\n",
    "print(\"Fields : \", example_interval.fields)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.3.2 Intersecting annotation intervals with callable-regions BED file\n",
    "Let's start by defining a function which will help us to sum up the lengths of intervals in pybedtools objects:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_bed_length(bed_intervals):\n",
    "    return sum([(bed_interval.end - bed_interval.start) for bed_interval in bed_intervals])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-success\">\n",
    "    <b>Exercise 9:</b> Use <code>get_bed_length</code> to compute the total length of <code>callable_intervals</code>. How does this compare to the total <code>chrom_length</code> of chromosome 5?\n",
    "    </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "#SOL"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-warning\">\n",
    "    <b>Python recipe: Intersecting bed intervals </b>\n",
    "    \n",
    "We often need to find the genomic intervals of sequence meeting several filters or features we have defined as bed files. For example, we might be interested in all exons that are callable. Let's use <code>.intersect()</code> to find these and assign them to a new variable <code>exons_callable</code>:\n",
    "</div>"
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![image.png](attachment:image.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "(Fig from the [bedtools documentation](https://bedtools.readthedocs.io/en/latest/content/tools/intersect.html))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The total length of callable exon sequence on iphiclides_podalirius.IP_504.chromosome_5 is 636108\n",
      "This is 0.9924239114833282 of exonic sequence\n"
     ]
    }
   ],
   "source": [
    "exons_callable = exon_intervals.intersect(callable_intervals)\n",
    "print('The total length of callable exon sequence on', chrom_name, 'is',  get_bed_length(exons_callable))\n",
    "print('This is',  get_bed_length(exons_callable)/get_bed_length(exon_intervals), 'of exonic sequence') "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-success\">\n",
    "    <b>Exercise 10:</b> Generate analogous intervals for the callable intronic <code>intron_callable</code>, intergenic sequence <code>intergenic_callable</code> and third codon positions <code>tcp_callable</code>. How does the fraction of callable sequence for these partitions compare to that of exons?\n",
    "    </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "#SOL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "#SOL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "#SOL"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.3.3 Subsetting variants by genomic partition\n",
    "We are often interested in how genetic diversity (and other popgen summaries) varies between genomic partitions. We can use the bed intervals we have for each genome partition to subset variants. The function below which takes a bed object and generates a numpy array of all positions is useful:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "def bed_to_array(bed):\n",
    "    big_pos_list = []\n",
    "    for interval in bed:\n",
    "        big_pos_list.extend(np.arange(interval.start, interval.end))\n",
    "    return np.asarray(big_pos_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "exon_pos = bed_to_array(exons_callable)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can use the [`numpy.isin()`](https://numpy.org/doc/stable/reference/generated/numpy.isin.html) function to return a boolean array of all positions in `snp_pos` (all biallelic SNPs) that are in exons, i.e. in `exons_callable`)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[False False False ... False False False]\n"
     ]
    }
   ],
   "source": [
    "variant_in_exons_callable_mask = np.isin(snp_pos, exon_pos, assume_unique=True)\n",
    "print(variant_in_exons_callable_mask)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can now use this mask to subset the genotype array in `snp_ga`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "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>samples</th>\n",
       "      <th>het_exons</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>IP_RVcoll14H960</td>\n",
       "      <td>0.003756</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>IP_RVcoll14F754</td>\n",
       "      <td>0.003089</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>IP_502</td>\n",
       "      <td>0.003303</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>IP_RVcoll12R048</td>\n",
       "      <td>0.001641</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>IP_504.10X</td>\n",
       "      <td>0.003106</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>IP_RVcoll14E561</td>\n",
       "      <td>0.003081</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           samples  het_exons\n",
       "0  IP_RVcoll14H960   0.003756\n",
       "1  IP_RVcoll14F754   0.003089\n",
       "2           IP_502   0.003303\n",
       "3  IP_RVcoll12R048   0.001641\n",
       "4       IP_504.10X   0.003106\n",
       "5  IP_RVcoll14E561   0.003081"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "snp_ga_exons = snp_ga[variant_in_exons_callable_mask]\n",
    "heterozygosity_exons = snp_ga_exons.count_het(axis=0) / get_bed_length(exons_callable)\n",
    "het_df = pd.DataFrame({'samples': samples, 'het_exons': heterozygosity_exons})\n",
    "het_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(\"mean heterozygosity = \", het_df['het_exons'].mean())\n",
    "# plot heterozygosity\n",
    "sns.barplot(x=heterozygosity_exons, y=samples)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can use analogous boolean masks as arguments in `allel.sequence_diversity` to compute diversity at exonic sites. Note that `allel.sequence_diversity` expects a mask for the whole chromosome which contains a boolean entry for each position of the chromosome. The function `bed_to_boolian_chrom()` generates a boolean mask based on a bed file for a whole chromosome."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [],
   "source": [
    "def bed_to_boolian_chrom(bed,chrom_end):\n",
    "    big_pos_list = bed_to_array(bed)\n",
    "    mask = np.isin(np.arange(0, chrom_end), big_pos_list, assume_unique=True)\n",
    "    return mask"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "exon_mask = bed_to_boolian_chrom(exons_callable,chrom_length)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "ac = snp_ga.count_alleles()\n",
    "pi_exon = allel.sequence_diversity(snp_pos, ac, is_accessible=exon_mask)\n",
    "print(\"Genetic diversity in exons is:\", pi_exon)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-success\">\n",
    "    <b>Exercise 11:</b>\n",
    "<ul>\n",
    "How does genetic diversity differ between 3rd codon exon positions, introns and intergenic sequence? Based on your estimates, which genomic partition is least affected by background selection?\n",
    "<li><b> Hint:</b> Subset the analogous masks for\n",
    "    </li>\n",
    "    <ul>\n",
    "   <li> intron_callable</li>\n",
    "   <li> intergenic_callable</li>\n",
    "   <li> tcp_callable</li>\n",
    "    </ul>\n",
    "<li>\n",
    "   Calculate the mean heterozygosity for each of the genomic partitions for each sample</li>\n",
    "    <ul>\n",
    "   <li> make sure to subset using the correct mask</li>\n",
    "   <li> make sure to divide by the correct length</li>\n",
    "   <li> add your heterozygosity estimates for all samples and genomic partitions to the dataframe <code>het_df</code> </li>\n",
    "   </ul>\n",
    "</ul>\n",
    "   </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "#SOL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "#SOL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "#SOL"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-success\">\n",
    "    <b>Exercise 11b (optional python practice):</b>\n",
    "<ul>\n",
    "   Calculate the mean heterozygosity for each of the genomic partitions for each sample</li>\n",
    "    <ul>\n",
    "   <li> make sure to subset using the correct mask</li>\n",
    "   <li> make sure to divide by the correct length</li>\n",
    "   <li> add your heterozygosity estimates for all samples and genomic partitions to the dataframe <code>het_df</code> </li>\n",
    "   </ul>\n",
    "   </div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "#SOL\n",
    "intron_pos = bed_to_array(intron_callable)\n",
    "variant_in_introns_callable_mask = np.isin(snp_pos, intron_pos, assume_unique=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "#SOL\n",
    "snp_ga_introns = snp_ga[variant_in_introns_callable_mask]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [],
   "source": [
    "#SOL\n",
    "heterozygosity_introns = snp_ga_introns.count_het(axis=0) / get_bed_length(intron_callable)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "het_df = pd.DataFrame({'samples': samples,'het_exons': heterozygosity_exons,\n",
    "                       'het_introns': heterozygosity_introns,\n",
    "                       'het_intergenic': heterozygosity_intergenic,\n",
    "                       'het_tcp': heterozygosity_tcp})\n",
    "print(\"mean heterozygosity intron = \", het_df['het_introns'].mean())\n",
    "print(\"mean heterozygosity intergenic= \", het_df['het_intergenic'].mean())\n",
    "print(\"mean heterozygosity tcp = \", het_df['het_tcp'].mean())\n",
    "het_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plot heterozygosity\n",
    "sns.barplot(x=heterozygosity_introns, y=samples)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 3. Exploring variation in genetic diversity between chromosomes "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For this section we will look at how heterozygosity in *I. podalirius* differs between e chromosomes. For this, we have collated at CSV which contains, for each chromosome, the (callable) lengths for each region of interest and the estimated mean (per site) heterozygosity. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [],
   "source": [
    "chrom_data_df = pd.read_csv('p4_data/iphiclides_chrom_data.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, let's take a look at the dataframe:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-success\">\n",
    "    <b>Exercise 12:</b>\n",
    "Check whether the mean heterozygosity and the size of the genomic partitions you computed for chromosome 5 in section 2 are correct. \n",
    "</div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "chrom_data_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can now explore the dataframe using plotting functions. For example the relationship between `total_length` and the number of `genes` for each chromosome. \n",
    "\n",
    "For this we will use the function [`seaborn.scatterplot()`](https://seaborn.pydata.org/generated/seaborn.scatterplot.html#seaborn-scatterplot), by providing `x` and `y` values, and by specifying `hue` to be based on the `chrom` column by using the colourmap `nipy_spectral` as the palette. Since by default, `seaborn` tries to position the legend on the plot, we need a special call to [`matplotlib.legend()`](https://matplotlib.org/stable/api/legend_api.html#module-matplotlib.legend) which formats the legend slightly differently so it doesn't obscure the plot."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "chrom_data_df['exon_length']/chrom_data_df['total_length']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "sns.scatterplot(x=chrom_data_df['total_length'], y=chrom_data_df['exon_length']/chrom_data_df['total_length'], hue=chrom_data_df['chrom'], palette='nipy_spectral')\n",
    "plt.legend(ncol=2, bbox_to_anchor=(1, 1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div class=\"alert alert-block alert-success\">\n",
    "    <b>Exercise 13:</b>\n",
    "Investigate the relationship between chromosome length (i.e. <code>total_length</code>) and heterozygosity (i.e. <code>total_het</code>)\n",
    "</div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "#SOL"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Appendix"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### A.1. Linear coordinate systems used in genomics:\n",
    "- a genome is traditionally represented as one or more (haploid) consensus sequences which are stored in a text files (e.g. FASTA files).\n",
    "- a coordinate system is needed to delimit start and end positions of features (e.g. genes, exons, variants, telomeres, stop-codons, etc) along the genome.\n",
    "- in genomics, two different coordinate systems are used depending on the file format:\n",
    "\n",
    "#### A.1.1. <b>zero-based</b> indexing\n",
    "\n",
    "- <b>numbers the spaces</b> between bases\n",
    "- also called interval-based indexing\n",
    "- used in BED files\n",
    "- length of a feature $length = end - start$\n",
    "- used in many programming languages: C, Perl, Python, Java, Ruby, JavaScript, etc.\n",
    "\n",
    "<code>\n",
    "          A T A T G G T T A C\n",
    "         | | | | | | | | | | |\n",
    "         0 1 2 3 4 5 6 7 8 9 10\n",
    "             [-----)   [-)\n",
    "              [2,5)   [7,8)\n",
    "</code>\n",
    "\n",
    "#### A.1.2 <b>one-based</b> indexing\n",
    "\n",
    "- <b>numbers the bases</b>\n",
    "- also called character-based\n",
    "- length of feature $length = (end + 1) - start$\n",
    "- used in GFF3/GTF/VCF files\n",
    "- used in many programming languages: FORTRAN, SASL, MATLAB, R, Mathematica, Smalltalk, etc.\n",
    "\n",
    "<code>\n",
    "         A T A T G G T T A C\n",
    "         | | | | | | | | | |\n",
    "         1 2 3 4 5 6 7 8 9 10\n",
    "             [---]    [-]\n",
    "             [3,5]   [8,8]\n",
    "</code>\n",
    "\n",
    "#### A.1.3 Coordinate transformations between 0-based and 1-based systems\n",
    "\n",
    "<code>\n",
    "     <b>0-based to 1-based</b>                      <b>1-based to 0-based</b>\n",
    "\n",
    "   0 1 2 3 4 5 6 7 8 9 10                   1 2 3 4 5 6 7 8 9 10\n",
    "   | | | | | | | | | | |                    | | | | | | | | | |\n",
    "    A T A T G G T T A C                     A T A T G G T T A C\n",
    "    | | | | | | | | | |                    | | | | | | | | | | |\n",
    "    1 2 3 4 5 6 7 8 9 10                   0 1 2 3 4 5 6 7 8 9 10\n",
    "</code>\n",
    "<code>\n",
    "    <b>1_start = 0_start + 1</b>                  <b>0_start = 1_start - 1</b>\n",
    "    1_end   = 0_end                        0_end   = 1_end\n",
    "</code>\n",
    "\n",
    "#### A.1.4 Further reading\n",
    "- https://tidyomics.com/blog/2018/12/09/2018-12-09-the-devil-0-and-1-coordinate-system-in-genomics/\n",
    "- Casey M. Bergman (2012) A proposal for the reference-based annotation of de novo transposable element insertions, Mobile Genetic Elements, 2:1, 51-54, DOI: 10.4161/mge.19479\n",
    "- https://wiki.c2.com/?WhyNumberingShouldStartAtZero\n",
    "- https://wiki.c2.com/?WhyNumberingShouldStartAtOne"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### A.2 VCF files \n",
    "- In brief, VCF files are composed of \n",
    "  - a header section (indicated by \"#\" at the beginning of the line) explaining the data types listed in each row\n",
    "  - rows containing information in columns (see following table) about the genetic variation across samples for a given position in the genome\n",
    "\n",
    "| Column | Fieldname |                                     Description                                     |\n",
    "|--------|-----------|-------------------------------------------------------------------------------------|\n",
    "| 1      | CHROM     | Chromosome name                                                                     |\n",
    "| 2      | POS       | 1-based position. For an indel, this is the position preceding the indel.           |\n",
    "| 3      | ID        | Variant identifier. Usually the dbSNP rsID.                                         |\n",
    "| 4      | REF       | Reference sequence at POS involved in the variant. For a SNP, it is a single base.  |\n",
    "| 5      | ALT       | Comma delimited list of alternative seuqence(s).                                    |\n",
    "| 6      | QUAL      | Phred-scaled probability of all samples being homozygous reference.                 |\n",
    "| 7      | FILTER    | Semicolon delimited list of filters that the variant fails to pass.                 |\n",
    "| 8      | INFO      | Semicolon delimited list of variant information.                                    |\n",
    "| 9      | FORMAT    | Colon delimited list of the format of individual genotypes in the following fields. |\n",
    "| 10+    | Sample(s) | Individual genotype information defined by FORMAT.                                  |\n",
    "\n",
    "- **VCF files** are generated by providing BAM files (one for each sample, composed of reads of that sample mapped to the genome) to a variant caller. To map reads of resequenced individuals to the *I. podalirius* reference, we have used [freebayes](https://github.com/freebayes/freebayes)\n",
    "- The **VCF files** you will be working on have been processed using [bcftools](https://samtools.github.io/bcftools/bcftools.html), and a variant was kept if:\n",
    "  - it was not too bad bad (`QUAL>1`)\n",
    "  - supported by reads on both strands (`SAF > 0 & SAR > 0`) \n",
    "    - `SAF`: Number of alternate observations on the forward strand\n",
    "    - `SAR`: Number of alternate observations on the reverse strand\n",
    "  - supported by both ends of reads (`RPR > 0 & RPL > 0`)\n",
    "    - `RPL`: Reads Placed Left: number of reads supporting the alternate balanced to the left (5') of the alternate allele\n",
    "    - `RPR`: Reads Placed Right: number of reads supporting the alternate balanced to the right (3') of the alternate allele\n",
    "- An additional **callable-regions BED file** was created from the BAM files using [mosdepth](https://github.com/brentp/mosdepth). \n",
    "  - This BED file list the regions of the genome which have been covered to a certain depth $d \\in [8, 2*mean]$) by each dataset. \n",
    "  - This file will be relevant when working with BED files "
   ]
  }
 ],
 "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": "3.8.3"
  },
  "toc-autonumbering": false,
  "toc-showcode": false,
  "toc-showmarkdowntxt": false,
  "toc-showtags": false
 },
 "nbformat": 4,
 "nbformat_minor": 4
}