Genome Assembly of a bacterial genome (MRSA) sequenced using Illumina MiSeq Data

Overview
Creative Commons License: CC-BY Questions:
  • How to check the quality of the MiSeq data?

  • How to perform an assembly of a bacterial genome with MiSeq data?

  • How to check the quality of an assembly?

Objectives:
  • Run tools to evaluate sequencing data on quality and quantity

  • Process the output of quality control tools

  • Improve the quality of sequencing data

  • Run a tool to assemble a bacterial genome using short reads

  • Run tools to assess the quality of an assembly

  • Understand the outputs of tools to assess the quality of an assembly

Requirements:
Time estimation: 2 hours
Supporting Materials:
Published: Mar 24, 2021
Last modification: Feb 16, 2024
License: Tutorial Content is licensed under Creative Commons Attribution 4.0 International License. The GTN Framework is licensed under MIT
purl PURL: https://gxy.io/GTN:T00036
rating Rating: 3.7 (3 recent ratings, 7 all time)
version Revision: 27

Sequencing (determining of DNA/RNA nucleotide sequence) is used all over the world for all kinds of analysis. The product of these sequencers are reads, which are sequences of detected nucleotides. Depending on the technique these have specific lengths (30-500bp) or using Oxford Nanopore Technologies sequencing have much longer variable lengths.

Comment: Illumina MiSeq sequencing

Illumina MiSeq sequencing is based on sequencing by synthesis. As the name suggests, fluorescent labels are measured for every base that bind at a specific moment at a specific place on a flow cell. These flow cells are covered with oligos (small single strand DNA strands). In the library preparation the DNA strands are cut into small DNA fragments (differs per kit/device) and specific pieces of DNA (adapters) are added, which are complementary to the oligos. Using bridge amplification large amounts of clusters of these DNA fragments are made. The reverse string is washed away, making the clusters single stranded. Fluorescent bases are added one by one, which emit a specific light for different bases when added. This is happening for whole clusters, so this light can be detected and this data is basecalled (translation from light to a nucleotide) to a nucleotide sequence (Read). For every base a quality score is determined and also saved per read. This process is repeated for the reverse strand on the same place on the flow cell, so the forward and reverse reads are from the same DNA strand. The forward and reversed reads are linked together and should always be processed together!

For more information watch this video from Illumina

In this training we will build an assembly of a bacterial genome, from data produced in “Complete Genome Sequences of Eight Methicillin-Resistant Staphylococcus aureus Strains Isolated from Patients in Japan” Hikichi et al. 2019:

Methicillin-resistant Staphylococcus aureus (MRSA) is a major pathogen causing nosocomial infections, and the clinical manifestations of MRSA range from asymptomatic colonization of the nasal mucosa to soft tissue infection to fulminant invasive disease. Here, we report the complete genome sequences of eight MRSA strains isolated from patients in Japan.

Agenda

In this tutorial, we will cover:

  1. Galaxy and data preparation
  2. Quality Control
  3. Assembly
    1. Assembly
    2. Assembly inspection
    3. Assembly Evaluation
  4. Conclusion

Galaxy and data preparation

Any analysis should get its own Galaxy history. So let’s start by creating a new one and get the data into it.

Hands-on: History creation
  1. Create a new history for this analysis

    Click the new-history icon at the top of the history panel.

    If the new-history is missing:

    1. Click on the galaxy-gear icon (History options) on the top of the history panel
    2. Select the option Create New from the menu

  2. Rename the history

    1. Click on galaxy-pencil (Edit) next to the history name (which by default is “Unnamed history”)
    2. Type the new name
    3. Click on Save

    If you do not have the galaxy-pencil (Edit) next to the history name:

    1. Click on Unnamed history (or the current name of the history) (Click to rename history) at the top of your history panel
    2. Type the new name
    3. Press Enter

Now, we need to import the data: 2 FASTQ files containing the reads from the sequencer.

Hands-on: Import datasets
  1. Import the files from Zenodo or from Galaxy shared data libraries:

    https://zenodo.org/record/10669812/files/DRR187559_1.fastqsanger.bz2
    https://zenodo.org/record/10669812/files/DRR187559_2.fastqsanger.bz2
    
    • Copy the link location
    • Click galaxy-upload Upload Data at the top of the tool panel

    • Select galaxy-wf-edit Paste/Fetch Data
    • Paste the link(s) into the text field

    • Press Start

    • Close the window

    As an alternative to uploading the data from a URL or your computer, the files may also have been made available from a shared data library:

    1. Go into Shared data (top panel) then Data libraries
    2. Navigate to the correct folder as indicated by your instructor.
      • On most Galaxies tutorial data will be provided in a folder named GTN - Material –> Topic Name -> Tutorial Name.
    3. Select the desired files
    4. Click on Add to History galaxy-dropdown near the top and select as Datasets from the dropdown menu
    5. In the pop-up window, choose

      • “Select history”: the history you want to import the data to (or create a new one)
    6. Click on Import

  2. Rename the datasets to remove .fastqsanger.bz2 and keep only the sequence run ID (DRR187559_1 and DRR187559_2)

    • Click on the galaxy-pencil pencil icon for the dataset to edit its attributes
    • In the central panel, change the Name field to DRR187559_1
    • Click the Save button

  3. Tag both datasets #unfiltered

    1. Click on the dataset to expand it
    2. Click on Add Tags galaxy-tags
    3. Add a tag starting with #
      • Tags starting with # will be automatically propagated to the outputs of tools using this dataset.
    4. Press Enter
    5. Check that the tag appears below the dataset name

  4. View galaxy-eye the renamed file

The datasets are both FASTQ files.

Question
  1. What are the 4 main features of each read in a FASTQ file.
  2. What does the _1 and _2 mean in your filenames?
  1. The following:

    • A @ followed by a name and sometimes information of the read
    • A nucleotide sequence
    • A + (optional followed by the name)
    • The quality score per base of nucleotide sequence (Each symbol represents a quality score, which will be explained later)
  2. Forward and reverse reads, by convention, are labelled _1 and _2, but they might also be _f/_r or _r1/_r2.

Quality Control

During sequencing, errors are introduced, such as incorrect nucleotides being called. These are due to the technical limitations of each sequencing platform. Sequencing errors might bias the analysis and can lead to a misinterpretation of the data. Adapters may also be present if the reads are longer than the fragments sequenced and trimming these may improve the number of reads mapped. Sequence quality control is therefore an essential first step in any analysis.

Assessing the quality by hand would be too much work. That’s why tools like NanoPlot or FastQC are made, as they generate a summary and plots of the data statistics. NanoPlot is mainly used for long-read data, like ONT and PACBIO and FastQC for short read, like Illumina and Sanger. You can read more in our dedicated Quality Control Tutorial.

Before doing any assembly, the first questions we should ask about the input reads include:

  • What is the coverage of my genome?
  • How good are my reads?
  • Do I need to ask/perform for a new sequencing run?
  • Is it suitable for the analysis I need to do?
Hands-on: Quality Control
  1. FastQC ( Galaxy version 0.74+galaxy0) with the following parameters:
    • param-files “Short read data from your current history”: both DRR187559_1 and DRR187559_2
    1. Click on param-files Multiple datasets
    2. Select several files by keeping the Ctrl (or COMMAND) key pressed and clicking on the files of interest

  2. Inspect the webpage outputs

FastQC combines quality statistics from all separate reads and combines them in plots. An important plot is the Per base sequence quality.

DRR187559_1 DRR187559_2
FastQC plot showing reads that mostly stay in the green. Same as previous plot, but the beginning of the reads are slightly better quality.

Here you have the reads sequence length on the x-axes against the quality score (Phred-score) on the y-axis. The y-axis is divided in three sections:

  • Green = good quality,
  • Orange = mediocre quality, and
  • Red = bad quality.

For each position, a boxplot is drawn with:

  • the median value, represented by the central red line
  • the inter-quartile range (25-75%), represented by the yellow box
  • the 10% and 90% values in the upper and lower whiskers
  • the mean quality, represented by the blue line

For Illumina data it is normal that the first few bases are of some lower quality and how longer the reads get the worse the quality becomes. This is often due to signal decay or phasing during the sequencing run.

Before doing any assembly, the first questions we should ask about the input reads include:

  • What is the coverage of my genome?
  • How good are my reads?
  • Do I need to ask/perform for a new sequencing run?
  • Is it suitable for the analysis I need to do?

Depending on the analysis it could be possible that a certain quality or length is needed. In this case we are going to trim the data using fastp (Chen et al. 2018):

  • Trim the start and end of the reads if those fall below a quality score of 20

    Different trimming tools have different algorithms for deciding when to cut but trimmomatic will cut based on the quality score of one base alone. Trimmomatic starts from each end, and as long as the base is below 20, it will be cut until it reaches one greater than 20. A sliding window trimming will be performed where if the average quality of 4 bases drops below 20, the read will be truncated there.

  • Filter for reads to keep only reads with at least 30 bases: Anything shorter will complicate the assembly

Hands-on: Quality improvement
  1. fastp ( Galaxy version 0.23.2+galaxy0) with the following parameters:
    • “Single-end or paired reads”: Paired
      • param-file “Input 1”: DRR187559_1
      • param-file “Input 2”: DRR187559_2
    • In “Filter Options”:
      • In “Length filtering Options”:
        • Length required: 30
    • In “Read Modification Options”:
      • In “Per read cuitting by quality options”:
        • Cut by quality in front (5’): Yes
        • Cut by quality in front (3’): Yes
        • Cutting window size: 4
        • Cutting mean quality: 20
    • In “Output Options”:
      • “Output JSON report”: Yes
  2. Edit the tags of the fastp FASTQ outputs to
    1. Remove the #unfiltered tag
    2. Add a new tag #filtered

fastp generates also a report, similar to FASTQC, useful to compare the impact of the trimming and filtering.

Question

Looking at fastp HTML report

  1. How did the average read length change before and after filtering?
  2. Did trimming improve the mean quality scores?
  3. Did trimming affect the GC content?
  4. Is this data ok to assemble? Do we need to re-sequence it?
  1. Read lengths went down more significantly:
    • Before filtering: 190bp, 221bp
    • After filtering: 189bp, 219bp
  2. It increased the percentage of Q20 and Q30 bases (bases with quality score above 20 and 30 respectively)
  3. No, it did not. If it had, that would be unexpected.
  4. This data looks OK. The number of short reads in R1 is not optimal but assembly should partially work but not the entire, closed genome.

Assembly

Now that the quality of the reads is determined and the data is filtered and/or trimmed, we can try to assemble the reads together to build longer sequences.

There are many tools that create assembly for short-read data, e.g. SPAdes (Prjibelski et al. 2020), Abyss (Simpson et al. 2009). In this tutorial, we use Shovill. Shovill is a SPAdes-based genome assembler, improved to work faster and only for smaller (bacterial) genomes.

Assembly

Comment: Results may vary

Your results may be slightly different from the ones presented in this tutorial due to differing versions of tools, reference data, external databases, or because of stochastic processes in the algorithms.

Hands-on: Assembly using Shovill
  1. Shovill ( Galaxy version 1.1.0+galaxy1) with the following parameters:
    • “Input reads type, collection or single library”: Paired End
      • param-file “Forward reads (R1)”: fastp Read 1 output
      • param-file “Reverse reads (R2)”: fastp Read 2 output
    • In “Advanced options”:
      • “Estimated genome size”: 2914567

*Shovill generates 3 outputs:

  • A log file
  • A FASTA file with contigs, i.e. contiguous length of genomic sequences in which bases are known to a high degree of certainty.

    Question

    Inspect the Contigs file

    1. How big is the first contig?
    2. What is the coverage of your biggest (first) contig?

    Both of these can be found in the header line of the Contigs produced by Shovill

    The results can differ from this example, because Shovill can differ a bit per assembly

    1. 589,438 bp, almost 6 times lower than the estimated genome size
    2. 17.7
  • A file with the contig graph or assembly graph

    The assembly graph is more information-rich because it not only contains the sequences of all assembled fragments (including the ones shorter than the threshold length defined for inclusion of the fragments into the FASTA output), but also indicates the relative average coverage of the fragments by sequenced reads and how some of the fragments could potentially be bridged after resolving ambiguities manually.

Assembly inspection

Th assembly graph format takes some getting used to before you can make sense out of the information it provides. This issue can be alleviated through the use of Bandage (Wick et al. 2015), a package for exploring assembly graphs through summary reports and visualizations of their contents.

Hands-on: Assembly inspection
  1. Bandage Info ( Galaxy version 2022.09+galaxy2) with the following parameters:
    • param-file “Graphical Fragment Assembly”: Shovill Contig Graph

Let’s look at the assembly statistics

Question
  1. What is the number of dead end?
  2. What is the number of nodes?
  1. Only 2
  2. 131: a bit disapointing as it should be encoded in one contig

Let’s now visualize the graph.

Hands-on: Assembly inspection
  1. Bandage Image ( Galaxy version 2022.09+galaxy4) with the following parameters:
    • param-file “Graphical Fragment Assembly”: Shovill Contig Graph

Bandage output showing a mess of a genome graph.

Question

First read this page in the Bandage wiki to help understand what the graph means.

What do you think of this assembly? Is it useful? Is it good enough?

This is a very messy assembly, with a lot of potential paths through the sequence. We cannot feel confident in the output FASTA file (as it is much smaller than the expected 2.9Mbp). In real life we might consider doing a hybrid assembly with Nanopore or other long read data to help resolve these issues.

Assembly Evaluation

To evaluate the assembly, we use also Quast (Gurevich et al. 2013) (QUality ASsessment Tool), a tool providing quality metrics for assemblies. This tool can be used to compare multiple assemblies, can take an optional reference file as input to provide complementary metrics, etc

Hands-on: Quality Control of assembly using Quast
  1. Quast ( Galaxy version 5.2.0+galaxy1) with the following parameters:
    • Assembly mode?: Co-assembly
      • “Use customized names for the input files?”: No, use dataset names
        • param-file “Contigs/scaffolds file”: Shovill Contigs output

QUAST outputs assembly metrics as an HTML file with metrics and graphs.

This fasta file contains

Question
  1. How many contigs is there?
  2. What is the total length of all contigs?
  3. What is you GC content?
  1. 32 contigs, meaning the chromosome is separated over multiple contigs. These contigs can also contain (parts of) plasmids.
  2. 2,911,349 (Total length (>= 0 bp)). Not far from the estimated genome size: 2,914,567
  3. The GC content for our assembly was 32.77%. For comparison MRSA Isolate HC1335 Strain GC% is 32.89%.

The total length and the GC content of the assembly are coherent with expectations.

Conclusion

In this tutorial, we prepared short reads, assembled them, and inspect the produced assembly for its quality. The assembly, even if uncomplete, is reasonable good to be used in downstream analysis, like AMR gene detection