Help with interpretation of SNaQ output network

[Brock Wooldridge]

Hello,

I've run SNaQ with two complementary datasets which both show strong support for a single hybridization edge. However, the hybrid edge differs between the two, and it's causing me to question whether I'm interpreting output correctly. Obviously bootstrapping will eventually get at the confidence in these scenarios, but what I'm wondering about at this moment is the biological history suggested by each network. The networks are attached, but here are my thoughts

Network A:

There was a major gene flow/hybridization event from the 'asi' lineage into the ancestor of the 'su + slo' lineage

Network B: 

There was a major gene flow/hybridization event from the 'asi' lineage into the ancestor of the 'po + bro + bla' lineage. 

Are these interpretations correct? What do I make of the hybrid edge stemming from the 'sister' to the 'asi' lineage in each case? If anything at all.

Thank you in advance for your help! Let me know if I can clarify.

[Cécile Ané]

I agree with your interpretation. I am not sure what you mean by "the hybrid edge stemming from the 'sister' to the 'asi' lineage". I attached another way to visualize these 2 networks, to highlight their similarity. They are similar in that they have the same unrooted topology (when we forget who are the hybrid edges, the hybrid node, and the direction of gene flow). The 2 networks also display the same trees. They differ in which node, exactly, is of hybrid origin. So this may be a case where the direction of gene flow is difficult to infer. I hope it helps!

Cécile.

[Brock Wooldridge]

Hi Cécile,

Thank you so much for the explanation and the sketches. In particular, thinking of them as the same unrooted topology really helped! I am not surprised at all if this is a case where the direction of hybridization is going to be difficult to resolve - that was beginning to be my suspicion based on some initial bootstrapping results. 

What's interesting is that one previous hypothesis based on work by others is that 'asi' itself is the hybrid lineage and the result of a hybrid speciation event. I'm surprised that I don't capture that pattern in either of these networks, although it does show up in some bootstrap replicates if I'm interpreting those results correctly. I would guess this is further evidence that the true history is going to be quite difficult to resolve.

Best,

Brock