structural variant discovery from highly mutated viral population using nanopore

Question

[rezaeir75]

I have a population with hundreds of viruses from the same species with different insertions and deletions. I'd like to use MinION from nanopore to sequence this sample and compare it to the reference genome to get a good picture of the amount of variation in this population. My question is if there is any specific long-read pipeline for structural variant calling from a viral population of the same viral species and if there is a tutorial or pipeline explanation on how to use it.

[samuel.a.odonnell]

Just to clarify, you would sequence a single sample in order to try discover mutations that occur at a low frequency in this population? Or sequence each mutant separately?

[rezaeir75]

I would sequence a single sample to try to find all the mutations (low and high frequency) in this population.

[rezaeir75]

one more thing is that these mutations are not necessarily more than 50bp changes (according to structural variant definition) but can be from single mutations to less a 100bp changes (deletion/insertion)

Answer 1

[Brunox13]

I am looking for the same thing. For short reads, a tool I have used with success is breseq in the --polymorphism-prediction mode and I would like to see if there is anything similar for long reads. Better yet, it would be great to have a tool that would use a combination of short and long reads (as this post suggests, for example).

As far as I know, sniffles would be a tool to do this (doi.org/10.1038/s41592-018-0001-7); however, according the publication, it sounds like it is only suitable for clonal samples (in the context of bacterial or viral genomes), equivalent to breseq's consensus (default) mode:

Variants below the minimum allele frequency (default: below 30%) are considered unreliable, variants with high allele frequency (default: above 80%) are considered homozygous, and variants with an intermediate allele frequency are considered heterozygous. Note that Sniffles does not currently consider higher ploidy, although this will be the focus of future work.