When to use Genome assembly and when to use short read alignment for consensus genration??

Question

[arriyaz.nstu]

Let's say I have Illumina short-reads for a virus and its reference sequence is also available in the Database.

In this case what I should use to get a full-length genome from my Illumina short-reads?

• Genome assembly
• Or, short-read alignment for consensus generation.

Many times I get confused about when I should use which option.

Thanks in advance.

[Asaf]

You can do a reference based de-novo assembly. See this review for more info: https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-017-1911-6

[Joe]

I would say this boils down mostly to what you intend to do downstream, and how good the reference is.

If the reference is a complete genome with lots of validation, and you only care about variant identification, I'd say just mapping the reads is the way to go.

If the reference isn't that good, or you intend to do further analysis, there's no harm in assembling the genome.

Of course, you can do both pretty quickly.

[young_bioinformatician]

You can also consider that pipeline: A combined de novo assembly approach increases the quality of prokaryotic draft genomes

Answer 1

[shelkmike]

If you want a full-length genome, you should do a de novo assembly. The reason is that if there are long indels or structural mutations between your genome and the reference, your reads may not align to those regions.

[GenoMax]

long indels or structural mutations

Unless one is dealing with really large viruses those should not be applicable?

[shelkmike]

I'm not a specialist in viral genomics and don't know what is considered very large, but structural mutations difinitely happen in viruses sometimes, for example https://pubmed.ncbi.nlm.nih.gov/20018465/