Inquiry About Incorporating a Haploid Genome into Pangenie Pangenome

[HuangYihang1222]

Dear Jana,

I am currently constructing a pangenome using Pangenie, with the Chinese genome (CN) as the reference. My dataset includes three ragweed genomes:

• CN (China, reference genome)
• NA (North America, haploid)
• EU.1 & EU.2 (Europe, diploid, two haplotypes)

Since Pangenie requires haplotype-resolved genomes, I was able to successfully integrate CN and EU into the pangenome. However, I am facing a challenge with the NA genome, as it does not have haplotype information.

I would like to know:

1. Is there a recommended way to incorporate a haploid genome (NA) into a Pangenie-based pangenome?
2. Would it be valid to pseudo-diploidize the NA genome by duplicating its sequence, or is there a better approach?
3. Alternatively, would you recommend aligning NA to the existing pangenome VCF and analyzing presence/absence variations (PAVs) separately?

I appreciate any suggestions you may have on how to best integrate the NA genome while maintaining the integrity of the pangenome model.

Thank you for your time and insights!

Best regards,
Yihang Huang

[eblerjana]

Hi Yihang Huang,

in principle, PanGenie is designed to work with diploid input genomes. However, since in your case only one sample in the input is haploid, I think pseudo-diploidizing the NA genome should be reasonable.

[HuangYihang1222]

Thank you again for your helpful advice.

Following your suggestion, I performed pseudo-diploidization by duplicating the NA genome and used it as input for the pangenome-graph-from-assemblies pipeline.

The workflow ran successfully, and the graph-filtered.vcf file looks like this:

As expected, the GTs for the NA assembly are all homozygous. I’m wondering whether this could significantly affect downstream genotype analysis, particularly in the context of variant frequency estimation or population-level comparisons.

Would you consider this a potential source of bias, or is it generally acceptable when working with a single assembly?

Looking forward to your thoughts, and thank you again for your time and guidance.

[eblerjana]

I'd expect that duplicating the haplotype of a single sample should not effect the results much, however, since your input panel is very small (just four haplotypes in total), it's difficult to estimate how big the effect really is. In general, genotyping results get better the larger the panel is, as the panel more accurately reflects the allele frequencies.