Illumina and ONT reads comparison

Question

[magatam]

Hi, so I've got this question about hybrid genome assembly. We assembled phage genomes with illumina reads, and then also had a hybrid approach with illumina+ont. Now for some isolates we see different results in both assemblies (like different phages). When running dnadiff on those they appeared as having an amount of unaligned bases similar to taking mismatching isolates (like 10-20), but with high identity. I compared illumina and ont reads with Mash and it showed 0.14 distance, which might indicate that the reads are from different isolates. But then I also checked this for samples when I had no such problems and it showed also 0.1 distance- so again quite high, now I'm not so sure that it's a read mismatch problem, but maybe it can be explained with phage-specific issues? I'm not sure what can be checked further. Any thoughts?

[GenoMax]

had a hybrid approach with illumina+ont

What is the expected size of the genomes? What was the reason behind adding nanopore reads? Have you tried nanopore only assemblies?

we see different results in both assemblies (like different phages)

There is more than one type/genome? If so, this is a metagenomic assembly.

Answer 1

[colindaven]

I would not use hybrid assembly (ONT+illumina), but rather assemble with ONT only, then use illumina short reads afterwards for polishing using a tool like hypo.

Recent ONT reads are getting to be very high quality these days.

You might want to also try using dorado correct to improve ONT read quality if your reads are recent enough.