The reticulated evolution of the Bean Goose complex

Hybridization partly erased the phylogenetic patterns of three goose taxa.

Different genes tell different stories. This simple statement is probably the most important lesson that I have learned while working with genomic data. Sexual reproduction is accompanied by recombination, the shuffling of chromosomal segments during meiosis. That is why your genome is a complex mixture of your parents’ DNA. Different genomic regions can thus be traced back to your father or mother. If you extend this reasoning back across multiple generations, you will quickly understand why different genes tell different stories. Each genomic segment follows its own trajectory into the past.

On an evolutionary timescale, this process is nicely illustrated by the Bean Goose complex – a group of goose taxa that is currently classified into three species: the Taiga Bean Goose (Anser fabalis, with three subspecies), the Tundra Bean Goose (Anser serrirostris, with two subspecies) and the Pink-footed Goose (Anser brachyrhynchus, monotypic). Depending on which genes you analyze, different phylogenetic patterns arise. For example, mitochondrial DNA clusters the Taiga and Tundra Bean Goose whereas genomic data places the Pink-footed Goose next to the Tundra Bean Goose. Again, different genes tell different stories. So, which genes should we follow? Which genes show the “true” species tree?

Differentiation Islands

To find the “true” species tree, a recent study in the journal BMC Ecology and Evolution (published by yours truly) focused on highly differentiated sections in the genome. Some scientists have argued that these genomic regions of increased divergence (i.e. differentiation islands) reflect the species tree. However, several phylogenomic studies found that these regions can produce misleading results due to selection or introgression (see for example this blog post). We should thus be cautious and carefully examine the phylogenetic patterns in regions of high genetic differentiation.

To explore the reliability of these differentiation islands for phylogenetics, we constructed gene trees for highly differentiated regions across the genomes of the three goose species. As a control, we sampled random genomic regions (which mainly represent the undifferentiated sections of the genome) and again constructed gene trees. The resulting phylogenetic patterns followed our expectations.

This approach [i.e. random selection of genomic regions] did not resolve the Bean Goose complex, but resulted in a monophyletic A. brachyrhynchus clade nested within a mixed cluster of A. fabalis and A. serrirostris. In contrast, phylogenetic analyses of differentiation islands converged upon a topology of three monophyletic clades in which A. brachyrhynchus is sister to A. fabalis, and A. serrirostris is sister to the clade uniting these two species.

A closer look at the gene trees in the differentiation islands revealed one dominant phylogenetic arrangement in which the Pink-footed Goose is most closely related to the Taiga Bean Goose. It seems unlikely that species-specific selective sweeps or ancient introgression events have impacted all these differentiation islands in the same way. Hence, we are confident that we have found the “true” species tree.

Species tree for (a) a random selection of genomic windows and (b) highly differentiated genomic windows. The different goose taxa are highlighted in different colors. The gradient of colors for A. fabalis, A. serrirostris and A. brachyrhynchus in figure a indicates the mixed nature of this clade. From: Ottenburghs et al. (2023).

Introgression

But our story does not end here. In a previous study – published in Heredity and covered in this blog post – we found extensive introgression between Taiga and Tundra Bean Goose. When we tested for introgression in the current study (using D-statistics), we found that no less than 21.9% of the genetic variants showed signatures of introgression between these two species. Clearly introgressive hybridization plays an important role in the evolutionary history of the Bean Goose complex.

Putting all the phylogenetic puzzle pieces together, we came up with the following scenario. After the divergence between the Taiga and the Tundra Bean Goose, a population of Taiga Bean Goose became geographically isolated on several islands (e.g., Svalbard, Greenland or Iceland). These populations evolved into the Pink-footed Goose. Later on, extensive hybridization between the Taiga and the Tundra Bean Goose erased the phylogenetic branching pattern between these taxa, resulting in a mixed clade of the Taiga and the Tundra Bean Goose containing a monophyletic Pink-footed Goose. Differentiation islands were largely unaffected by homogenizing introgression – perhaps because they contained loci involved in reproductive isolation – and maintained the phylogenetic patterns that reflect the species tree. A plausible scenario that requires further investigation.

Within the differentiation islands the topology grouping Taiga Bean Goose (A. fabalis) and Pink-footed Goose (A. brachyrhynchus) occurred most often. The gene trees uniting Taiga Bean Goose and Tundra Bean Goose (A. serrirostris) can be partly explained by recent introgression. Asterisks (*) indicate gene trees that were not observed in the data. From: Ottenburghs et al. (2023).

Trees or Networks?

These findings raise an intriguing conundrum: How should we depict the evolutionary history of the Bean Goose complex? The answer depends on the message you want to convey. If you are focused on reconstructing the order of speciation events, you can settle for a bifurcating tree. If you want to quantify the impact of introgression, a network approach will be more suitable. The evolution of the Bean Goose complex can certainly be depicted as a simple bifurcating tree, but this would ignore the role of introgressive hybridization. Hence, we advocated that the evolutionary relationships between these taxa are best represented as a phylogenetic network. But feel free to disagree.

What is the best way to depict the evolutionary history of the Bean Goose complex: a tree or a network? From: Ottenburghs et al. (2023).

References

Ottenburghs, J., Honka, J., Heikkinen, M.E., Madsen, J., Müskens, G.J.D.M. & Ellegren, H. (2023) Highly differentiated loci resolve phylogenetic relationships in the Bean Goose complex. BMC Ecology and Evolution 23, 2.

Featured image: Taiga Bean Goose (Anser fabalis) © Marton Berntsen | Wikimedia Commons

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s