Only one pair of homologous chromosomes were not phased

[apple-orange-banana]

Hi~
SubPhaser is a great piece of software, I have suffered some problems when I use this software to phase my diploid genome.

After the previous hic scaffolding, I got 22 superscaffolds, then I want to divide these scaffolds into 2 parts(2n=2x=22).

so I used this SubPhaser(-k 17 -q 50 -f 1.5), then 20 superscaffolds were phased and only 1 pair of homologous scaffolds(scaffold_9 and scaffold_10) were not phased.

k17_q50_f1.5.kmer_freq.pdf
k17_q50_f1.5.kmer_pca.pdf
k17_q50_f1.5.ltr.insert.density.pdf

How can I solve this problem?
Looking forward to your reply!
Yang

[zhangrengang]

The evidence is quite weak and I do not think the results are reliable. The two parents may be too closely related to be well distinguished.

If you want to force the phasing, you may adjust the parameters by varying -k and -q. If it do not work well, you may adjust the phasing manually by editing the subgenome assignments (edit and rename the *chrom-subgenome.tsv file) and then feeding it to SubPhaser via the -sg_assigned option. This means to assign scaffold_9 and scaffold_10 by yourself.

[apple-orange-banana]

Thank you for your reply!
Our genome with about 1.2% heterozygosity is sequenced by PacBio HiFi and Hi-C, I found there is less specific kmer from 2 subgenomes by using SubPhaser, Maybe 2 parents are too close, In this situation, how can I phase it into 2 parts? or maybe I should just assemble one genome?

[zhangrengang]

I prefer to not to phase it unless the data of two parents are available. But assembling two haplotypes is okay.

[apple-orange-banana]

At first I don't know which chromosomes belong to one subgenome, so I use this SubPhaser. I will try to finish the switch error analysis to test my assambly. And proceed with downstream analysis and get back to you if there is further questions.

Thanks!
yang

[zhangrengang]

I prefer to not to phase it unless the data of two parents are available. But assembling two haplotypes is okay.