split haplotype

[Wan727]

Dear Author,

I would like to consult you about the haplotype separation issue in genome assembly mentioned in your paper. The original text from the paper is:
“For contigs assembly, we used hifiasm (RRID:SCR_021069) v0.18.5 to assemble contigs by integrating HiFi and Hi-C sequencing reads per species [36]. Because D. pinnata and B. alba are tetraploid plants, and hifiasm is designed to assemble a diploid genome, the resulting haplotype 1 and haplotype 2 were not properly separated. Thus, we merged 2 haplotypes, and the merged assembly contained both subgenomes and allelic genomes. For C. bipinnatus, a diploid plant, 1 haplotype can be taken as the reference assembly. Due to the heterozygous character of C. bipinnatus, we used purge_dups (RRID:SCR_021173) v1.2.6 with parameter settings “-a 0.70” and “-f 0.65” to remove haplotypic duplications [66].”

Regarding the two species, D. pinnata and B. alba, where the haplotypes were not properly separated, what specific merging method did you use? My species is expected to be a more complex pentaploid (AABBC), and I would like to split A into A_1 and A_2, and B into B_1 and B_2. Could you provide some assistance with the code? Sincerely, thank you!

[whc2]

For tetraploid plants, we just cat the assembled sequences of haplotype 1 and haplotype 2 generated by hifiasm. The merged sequences would contain tetraploid genome.

For subgenome identification, we mainly used the Hi-C signals and the phylogenetic trees of homologous chromosomes. As the methods memtioned: "For tetraploid plants D. pinnata and B. alba, the contig assembly contained 4 haplotypes, and YaHS could group contigs from 4 homologous chromosomes together. Finally, based on the Hi-C signal features and the telomere in- formation, we manually curated the scaffolds on Juicebox". The relationships of homologous chromosomes can be inferred by proteins or CDSs of genes on the chromosomes (see Figure 4C, 5C).

I think you could refer the ideas to analysis you data.

[Wan727]

Thank you for your answer, but I tried using --hap-n=2 and --hap-n=3 and found that the file size ratio of haplotype 1 and haplotype 2 and haplotype 3 does not comply with 2:2:1, and the utg size is almost equal to the ctg size.