8  Assemble the Genome

The SPAdes software (Prjibelski et al. (2020)) is an excellent, multipurpose sequence assembly package.

• SPAdes documentation
• Protocols for SPAdes

It performs well in assembling large and small genomes, using short-read, long-read, or hybrid (both short- and long-read) inputs. SPAdes can handle single- or paired-end reads and, although specialised tools may perform better, it can also carry out metagenome assembly.

Warning

The generality of SPAdes is an advantage, but this also means that, to get the best performance in any particular situation - such as assembling a eukaryote genome from long read data - careful attention should be paid to parameter choices.

Shovill is an assembly pipeline that uses SPAdes as the assembler tool. It is optimised for assembly of small genomes using paired-end, short-read data.

• Shovill documentation

The Shovill pipeline

Shovill does not simply use SPAdes with a couple of changes to the settings. There is a more involved set of steps that we would otherwise need to perform ourselves to obtain a high-quality genome assembly.

1. Estimate genome size and read length from reads (unless –gsize provided)
2. Reduce FASTQ files to a sensible depth (default –depth 100)

• These two steps are needed because there is an optimal read coverage level for accurate assembly

3. Trim adapters from reads (with –trim only)

• We have already done this with trimmomatic (Chapter 7) so will skip the step

4. Conservatively correct sequencing errors in reads
5. Pre-overlap (“stitch”) paired-end reads
6. Assemble with SPAdes/SKESA/Megahit with modified kmer range and PE + long SE reads
7. Correct minor assembly errors by mapping reads back to contigs
8. Remove contigs that are too short, too low coverage, or pure homopolymers

• This avoids including some obviously poorly-assembled sequence in our final output

9. Produce final FASTA with nicer names and parseable annotations

• This is a convenience but, if you are working with the output computationally, it’s a very helpful thing

Good practice

When reporting how you assembled your sequence data in a manuscript or dissertation, you should always state:

1. The software tool you used, with its version number and a citation of the paper describing it (if available; provide a URL to the software if there is no paper)
2. The parameters used when running the tool (if default parameters were used, state this)
3. The number of contigs or scaffolds that were assembled, a measure of the average length (such as N50), and the G+C% content

8.1 Using Shovill

1. Navigate to the shovill tool using the Tools sidebar in Galaxy
2. Select the shovill tool
3. Make sure that you have selected Paired End as the Input reads type
4. Choose the (R1 paired) output from trimmomatic as Forward reads (R1)
5. Choose the (R2 paired) output from trimmomatic as Reverse reads (R2)
6. Select No for Trim reads (we already did this in Chapter 7)
7. Make sure Spades is selected in Assembler to use
8. Click Run Tool

Video: Assembling the SARS-CoV-2 genome using Shovill

Caution

Shovill can take a few minutes to run to completion.

8.2 Shovill Output

Shovill produces three output files.

• A contig file, containing the genome assembly as one or more contig sequences
• A contig graph file, describing the assembly as a graph, with assembled contigs as nodes, and edges linking them in the way the assembler thinks they might be connected
• A log file, describing the progress of the assembly run

Figure 8.1: Shovill output files in Galaxy. From the top: the contig graph file, contig file, and the log file.

As with other Galaxy tools, you can inspect the contents of these files by clicking on the filename in the History sidebar, and also by clicking on the corresponding eye icon to view the trimmed reads in the workspace.

Questions

Open the contig file output from Shovill and inspect the data.

1. How many contigs (sequences) were assembled?
2. How long are the assembled sequences?

Answers

1. A single contig sequence is assembled.
2. The length of the contig is 29878 bases.

8.3 Visualising the Assembly

The Bandage software package (Wick et al. (2015)) can take assembly output from tools such as SPAdes and Shovill, and visualise them as a graph. This is a useful step in assessing the quality of an assembly and can help identify poorly-assembled regions and potential sequencing of co-cultures of related strains, rather than an axenic isolate.

• Bandage documentation
• Bandage website

To use Bandage to visualise your genome assembly

1. Navigate to the Bandage Image tool using the Tools sidebar in Galaxy
2. Select the Bandage Image tool
3. Make sure that you have selected the Contig Graph file as the Graphical Fragment Assembly input
4. Click Run Tool

Video: Visualising the SARS-CoV-2 assembly using Bandage Image

Tip

Inspect the assembly graph by clicking on the eye icon for the Assembly Graph Image output.

8.4 Next steps

With a complete genome assembly, we can begin to annotate genomic features such as gene sequences, and you will do this using Prokka in Chapter 9.

Prjibelski, Andrey, Dmitry Antipov, Dmitry Meleshko, Alla Lapidus, and Anton Korobeynikov. 2020. “Using SPAdes DE Novo Assembler.” Curr. Protoc. Bioinformatics 70 (1): e102.

Wick, Ryan R, Mark B Schultz, Justin Zobel, and Kathryn E Holt. 2015. “Bandage: Interactive Visualization of de Novo Genome Assemblies.” Bioinformatics 31 (20): 3350–52.