Primary haploid assembly from Simplex reads?

[Matteo-Paini]

Hi Verkko team,

I am assembling a diploid genome (estimated haploid size ~2.7Gbp) using only corrected ONT Simplex reads (N50: 37kbp, corrected using hifiasm v0.25.0 --write-ec), without Hi-C/Pore-C data.
I ran verkko v2.3.2 using the following command: verkko -d ./asm --hifi hifiasm_corr.fq --local-cpus 70 --cleanup

The assembly finished successfully, and the resulting assembly.fasta has a total length of ~5.2Gbp, which is consistent with the combined size of both haplotypes.

Since my final goal is to obtain a haploid (collapsed) assembly, is there any way to generate one directly with Verkko from these data, or is purging (with a tool like purge_haplotigs) after the assembly the only recommended approach?

[skoren]

Yes, verkko is currently designed to expect some kind of phasing data. The output is always phased sequences so you could run purge_haplotigs but the sequences you would end up with would likely be quite short (depending on your genome's heterozygosity rate) since it will stop at every homozygous region to avoid introducing haplotype switches. So while you can do it, I probably wouldn't recommend it. Getting Hi-C/Pore-C/trio would likely make a big difference.

It's on our future wish list to add a primary assembly (see #157) and I'll tag @Dmitry-Antipov if he has any comments or suggestions to approximate a primary assembly with the current code/version.