Geographic structuring of genetic variation differs across two contact zones in the Diglossa carbonaria superspecies

Technical Summary: Geographic Structuring of Genetic Variation in the Diglossa carbonaria Superspecies

Problem Statement
Rapid radiations often present a challenge for evolutionary biology: lineages may diversify phenotypically (e.g., plumage) while retaining minimal genetic divergence, obscuring phylogenetic relationships and species boundaries. The Diglossa carbonaria superspecies, a group of Andean flowerpiercers, exemplifies this phenomenon. It comprises taxa with striking plumage differences—ranging from all-black to dark-backed with chestnut bellies—that have been classified variously as distinct species, subspecies, or a single biological species. Previous studies using limited mitochondrial and nuclear markers suggested that D. humeralis and D. brunneiventris are polyphyletic and that plumage is a poor indicator of phylogeny. However, the mechanisms driving these patterns, particularly the nature of contact zones between parapatric taxa, remain unclear. Specifically, it is unknown whether the observed phenotypic transitions correspond to genetic breaks, whether introgressive hybridization is occurring, and how genetic variation is structured across the Andes in the absence of strong phenotypic divergence.

Methodology
The study employed Restriction site-associated DNA sequencing (RADSeq) to characterize genomic variation across the ranges of three parapatric taxa: D. humeralis atterima (all-dark), D. b. brunneiventris (dark-backed, chestnut-bellied), and D. carbonaria (dark-backed, gray-belly, chestnut crissum).

• Sampling: 72 individuals were sampled from northern Peru to central Bolivia, covering two primary contact zones: the D. h. atterima / D. b. brunneiventris boundary in Peru and the D. b. brunneiventris / D. carbonaria boundary in Bolivia. Two D. albilatera individuals served as outgroups.
• Sequencing and Filtering: Following the 3RAD protocol, libraries were sequenced on an Illumina NovaSeq 6000. Reads were aligned to a D. b. brunneiventris reference genome. Stringent filtering removed loci with low depth (<15x), missing data (>25%), non-biallelic sites, and those violating Hardy-Weinberg equilibrium or showing linkage disequilibrium. The final dataset comprised 6,675 biallelic SNPs for 69 individuals.
• Analytical Approaches:
  • Species Tree Inference: SNAPP (coalescent-based) was used on subsets of data to visualize topological uncertainty and relationships.
  • Population Structure: PopCluster (using DLK2 estimator for optimal K) and Principal Component Analysis (PCA) via SNPRelate were used to assess genetic clustering and admixture.
  • Diagnostic Alleles: Allele frequencies were analyzed to identify "diagnostic SNPs" (frequency ≥0.8 in one taxon, ≤0.2 in another) to map geographic transitions and test for introgression patterns.

Key Results

1. Low Genetic Differentiation vs. High Phenotypic Divergence: Despite distinct plumage, the three taxa showed weak overall genetic differentiation. At K=2, a single genetic break separated D. carbonaria from the other two taxa. At K=3, D. h. atterima and D. b. brunneiventris remained genetically indistinguishable in most analyses, supporting the hypothesis that plumage color is a poor phylogenetic marker in this group.
2. Contact Zone Dynamics:
   • Peru (D. h. atterima / D. b. brunneiventris): No clear genetic break or admixture signal was detected at the contact zone. The lack of differentiation prevented confirmation of hybridization. A single diagnostic SNP showed a frequency gradient that did not decay as expected from the contact zone, suggesting the pattern could result from ancestral polymorphism, drift, or introgression masked by melanin-based coloration (where black plumage obscures hybrid phenotypes).
   • Bolivia (D. b. brunneiventris / D. carbonaria): Clear evidence of introgressive hybridization was found. Clustering analyses, PCA, and diagnostic allele frequencies revealed a genetic break and a zone of admixture (~300 km wide) east of La Paz. Heterozygotes were identified specifically within this contact zone, consistent with a hybrid zone.
3. Unexpected Genetic Breaks:
   • A significant genetic break (~450 km wide) was identified within the distribution of D. b. brunneiventris in central Peru, unrelated to the phenotypic contact zones. Individuals north of this break were genetically more similar to D. h. atterima than to D. b. brunneiventris populations to the south, despite phenotypic uniformity.
   • PCA revealed a secondary axis of differentiation separating D. b. brunneiventris populations on the western slope of the Andes (arid) from those on the eastern slope (humid), suggesting the Marañón River Valley may not be the primary barrier for these western-slope populations, but other factors drive divergence.

Significance and Claims
The authors claim that their results add to the body of evidence demonstrating that rapid lineage divergence can result in dramatic phenotypic changes with minimal genomic divergence. Specifically:

• Plumage as a Poor Marker: The study reinforces that plumage color and pattern in Diglossa do not reliably reflect evolutionary relationships or species limits, as genetically distinct groups can look identical, and phenotypically distinct groups can be genetically indistinguishable.
• Hybridization Mechanisms: The paper provides evidence that introgressive hybridization is occurring in Bolivia, supporting historical observations of intermediate plumage. Conversely, the lack of genetic signal in Peru suggests that if hybridization occurs, it may be phenotypically cryptic due to melanin masking, or the taxa are not yet genetically differentiated enough to show a break.
• Biogeographic Complexity: The discovery of a genetic break within D. b. brunneiventris that does not align with known phenotypic transitions or major river valleys challenges simple biogeographic models. It suggests that barriers for western-slope, high-elevation birds may differ from those affecting eastern-slope taxa, highlighting a need for more extensive genetic sampling to understand the drivers of population divergence in the Andes.
• Leapfrog Patterns: The findings support a modified "leapfrog" hypothesis where the all-black D. h. atterima is a derived trait nested within D. b. brunneiventris, potentially spread via introgression or mutation, rather than a distinct lineage.

The paper concludes that while plumage differences are striking, they may be decoupled from the underlying genomic architecture of these taxa, necessitating genomic-level analyses to accurately define species boundaries and understand the evolutionary history of Andean radiations.