Hello,
I'm working on a de novo assembly of a tetraploid plant genome using Pacbio and ONT reads. Due to ressource limitations, we pooled DNA from multiple individuals of the same species to obtain sufficient material. I first tried Hifiasm to get an assembly, but the divergence among individuals result to a highly fragmented assembly (primary), with thousand of contigs (genome size estimation: 500 Mb (assembly : 750 Mb), N50 = 1.3 Mb after purging haplotigs, BUSCO score > 99%). I also used Flye but the results were not better. Do you have any suggestions for improving the assembly, either by using differents assembler or tools that are better suited to handling population level variation ? I'm also considering a pangenome-based approach, but I would still need high-quality assemblie. In that context, I was wondering whether it is possible to separate individuals directly from the pooled data and then analyze variation at the population level ?
I'm happy to provide additional details if needed. Thanks very much!
•
written
by
Anirmgls •
A tetraploid assembly is very difficult from a single sample and requires Hi-C data on top of either mid length Pacbio or long ONT reads.
A pooled sample is not going to work, as you have demonstrated.
You might be able to split reads prior to assembly using haplotyping, but seeing as you have multiple individuals trying to discern 8-12+ haplotypes would be a complete nightmare.
I think resequencing in future is your most viable option, or perhaps use of related public assemblies to try to disentangle the read sets.
Yes, we've planned to redo the sequencing using only one individual. Do you think Hi-C is necessary if I only want a collapsed assembly rather than a fully haplotype-phased assembly? Thanks for your answer!
Sequeucing with Pacbio medium reads - likely Hi-C also required for scaffolding but you did not say if there are related genomes available for scaffolding. With ONT sequencing with long high quality reads, Hi-C is not normally required these days.
You can always add in the Hi-C later.
Log in to answer this question.
