I would like to perform de novo genome assembly on a diploid microalgal strain.

I have two datasets:

  • PacBio CCS/HiFi reads, low coverage.
  • Illumina PE 2x150 (standard shotgun)

Does anybody have any software recommendations for how best to use both in a hybrid de novo assembly?

Most options out there are set up for Pacbio CLR reads, not Pacbio CCS/HiFi, for using in hybrid assembly. Perhaps I should assemble short read contigs first then combine with wengan, or go back to the PacBio subreads instead of CCS to use in a hybrid method?

  • $\begingroup$ Please clarify your specific problem or provide additional details to highlight exactly what you need. As it's currently written, it's hard to tell exactly what you're asking. $\endgroup$
    – Community Bot
    Oct 23 '21 at 14:39

If you have reasonable coverage with HiFi (30X per haplotype), then there usually isn't any reason to also use Illumina. My understanding is that Illumina can make the assembly worse if you try to polish with it, because it can erroneously induce haplotype switching errors.

I would use HiCanu or HiFiAsm, depending on whether you expect significant heterozygosity, and also assuming that you have non-trivial HiFi coverage.

If your HiFi coverage is very low (<15X per haplotype, maybe), you might consider going back to subreads, assembling those, and polishing with Pilon and your Illumina reads, or one of the workflows that you have found. But otherwise I don't think the Illumina will be very useful. If you had Hi-C reads you could use them directly in HiFiAsm to assist phasing and scaffolding, but I'm assuming that your Illumina is shotgun.

I know it's a PB promo material, but you see that PB promotes HiFi as not needing polishing, as HiFi has accuracy >99%, so shotgun Illumina is not going to do much. There is some argument for still polishing with your original long reads.

I can recommend any of the linked papers for ideas or further explanation.

  • $\begingroup$ Unfortunately, my HiFi coverage isn't that great, (unexpectedly large genome). However, my shotgun short read dataset is large and good quality, hence the question. $\endgroup$
    – bishopia
    Oct 24 '21 at 14:57
  • $\begingroup$ @bishopia Hmmmm I see what you mean. In that case, subread assembly+Illumina polishing might be your best bet. It would still be helpful to know more specifics (e.g. coverage, read length profile of each dataset: CCS, subread, ILMN). I am pretty skeptical of the usefulness of Illumina contigs as the basis of assemblies of large/complex/heterozygous genomes, compared to even low-coverage PB data. $\endgroup$ Oct 24 '21 at 18:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.