Skip to content

Phylo Task

Warning

Be sure to use the phylo_task Codespace when initiating VM for this practical. This task you will also need to wget the data wget -O - http://genomics.lshtm.ac.uk/phylo_task.tar.gz | tar -xvz

Introduction

Welcome to the Phylo analysis practical, where you will reconstruct evolutionary relationships between Mycobacterium tuberculosis isolates starting from raw sequencing data.

This workflow mirrors real outbreak genomics pipelines used to: - track transmission - identify clusters of infection - understand lineage structure - compare genomic relationships between isolates

You will move from raw reads to a final phylogenetic tree and visualise it using iTOL.

Goal

By the end of this task you will produce: - consensus genomes for each isolate - a core-genome alignment - a phylogenetic tree - an annotated iTOL visualisation

Input Data

You are given paired-end sequencing files :

  • *_1.fastq.gz
  • *_2.fastq.gz

Each pair represents a TB isolate.

You are also given a reference genome:

  • Mycobacterium tuberculosis reference (tb.fasta)

And given an annotation file to use in Itol at the end

  • itol.txt

These will need to be downloaded using

  • wget -O - http://genomics.lshtm.ac.uk/phylo_task.tar.gz | tar -xvz

Workflow Overview

You will complete the following stages:

  1. Read alignment
  2. Variant detection
  3. Consensus genome generation
  4. Multi-sample genome collection
  5. Phylogenetic reconstruction
  6. Tree visualisation and annotation

Stage 1 — Read Alignment

Each sample must be aligned to the reference genome.

What you should think about:

  • ensuring correct pairing of reads
  • generating a sorted alignment file per sample
  • indexing the alignment for downstream analysis

Output expected:

  • one BAM file per sample

Stage 2 — Variant Detection

From each alignment, identify differences relative to the reference genome.

What you should think about:

  • detecting SNPs and small variants
  • filtering low-confidence calls
  • producing a variant file per sample

Output expected:

  • VCF file per sample

Stage 3 — Consensus Genome Construction

Use variant information to reconstruct each isolate’s genome.

What you should think about:

  • Use bcftools on the vcfs to create a consensus with bcftools consensus
  • generating a full-length FASTA per sample
  • ensuring consistent naming across samples

Output expected:

  • one consensus FASTA per isolate

Stage 4 — Combine Genomes

Merge all consensus genomes into a single dataset. Here we will use Parsnp, you can read the manual here

What you should think about:

  • move the generated fastas into one dir
  • running parsnp on that one directory

Output expected:

  • multi-FASTA file containing all isolates

Stage 5 — iTOL Visualisation

Upload your tree to iTOL (Interactive Tree of Life). We have provided an annotation file and a script to generate the annotation file if interested.

What to look for:

  • clustering patterns
  • lineage structure
  • potential transmission clusters
  • outliers or unusual branches

Key Hints

Tips for Mtb Phylo Analysis
  • Keep sample names consistent across all files
  • Low-quality samples may break downstream clustering
  • Might need to remove some samples
  • The tree reflects genetic distance, not just lineage
  • Lineage labels are metadata overlays, not the result itself
  • Parsnp builds a core-genome alignment, not just SNPs
  • Interpretation is more important than execution