Duplicated chrX in male parental assemblies with hifiasm 0.16.1

[projectoriented]

Hello 👋 ,

We've encountered multiple instances with our non-trio male assemblies, where both haplotypes contain contigs that map to chrX. Here's an example with HG00731 (pacbio ccs reads, 33X coverage):

hifiasm HG00731.asm HG00731.fq.gz

The fasta output is aligned to T2Tv2 with minimap2 and then visualized with SafFire. Figure below is a view of chrX in both haplotypes for HG00731. Complementary contigs are unique contigs that should theoretically be aggregated along with the duplicated ones on one haplotype.

Dotplot below is one of the duplicated contigs that show 100% sequence identity.

Below figure is what the chrX of HG00731 should look like.

Do you know why this is happening? Please let me know if there is anything I can do to resolve the problem. Thanks!

[chhylp123]

Do you use the dual assemblies or Hi-C phased assemblies?

[projectoriented]

I'm not sure what you mean by dual assembly. This is a standalone assembly for HG00731 with HiFi reads and no additional read support from other technologies or family members. Does that answer the question? 😅

[chhylp123]

Thanks. We called it the dual assembly. Actually, it is possible that hifiasm may put contigs of chrX into two haplotypes. In this mode, hifiasm could only utilize the similarity among contigs to cluster/phase, but the similarity is not as reliable as Hi-C or trio. Tuning the parameter -s might be helpful, but the final results are still by chance. Probably you can utilize both haplotypes at the same time...?