Admixture and hybridization are common in domesticated yeasts. I need to run TreeMix to detect or rule out (there isn't much evidence for it in our strains at the moment, as very few alleles are shared with other groups) possible admixture to generate an admixture graph in addition to a classical phylogenetic approach. I want to take an approach similar to the study by Fay et al. (2019) however I have questions regarding the validity of using TreeMix on polyploid strains. Yeasts have various ploidy and I am quite sure our strains are tetraploid, and therefore I made tetraploid variant calls. On the other hand, I can see that Fay et al. used sample_ploidy=1 in their pipeline.
The standard treemix workflow uses VCF-tools and won't work on tetraploid data. So here are my questions:
- Do I need to re-call all my variants and which ploidy setting should I use (or does it not matter)?
- Do I need to select and call variants from a large panel of other strains or can I simply combine our calls with an existing panel VCF file from another study?
- How large should and can such a panel be to rule out (if that's at all possible) admixture? There are thousands of strains sequenced now.
Comments anyone? ...
Looking at their panel, they combined variants from samples of different ploidy and even variants from assemblies, so it might not matter that much.
Mensur Dlakic would you mind having a look?
For background, this is additional work we think we need to do for a current manuscript in addition to more sequencing.
Cheers and thank you
Michael
This is outside of my expertise. Your manuscript seems interesting, good luck getting it published.