Curious About Gene Flow Direction in Hybrid Detection Results of HyDe

[Jhe1004]

Dear Professor,

Great software!

But i am reaching out to discuss an issue I encountered concerning the detection of hybrids. In my analysis, I used four species: A, B, C, and D, with their species tree modeled as (((A,B),C),D). I aimed to test whether species B is a hybrid of species A and C. The results indicated that B is indeed a hybrid of A and C, with significant support (P = 0; γ = 0.3).

Your paper on Hyde interprets such results as species B being formed by hybridization between A and C. However, as I understand, Hyde might also detect similar significant results if there was gene flow from species B to C in the past. I am curious to know if the Hyde software can determine the direction of gene flow (whether from B to C or C to B), or how you would interpret this scenario.

Thank you for your time and I look forward to your insights.

Best regards,

He

[pblischak]

Hi He! And thanks for using HyDe!

As you've noted, HyDe can't determine the directionality of hybridization, but I think that there are other methods that can. I would suggest trying out one of the statistics from Hibbins and Hahn (2019). They use tests based on three taxa, so you could just leave taxon 'D' out of what you've already done with HyDe. Hopefully it will prove useful for you but let me know how it goes!

Hibbins and Hahn. 2019. The Timing and Direction of Introgression Under the
Multispecies Network Coalescent. https://doi.org/10.1534/genetics.118.301831