Adaptive Potential
Finally, we performed a preliminary screen for the potential signature
of multi-locus adaptation and IBE using redundancy analysis (RDA;
Forester et al. 2018; Capblancq and Forester 2021). We evaluated
Genotype-Environment Associations (GEA) and Genotype-Phenotype
Association (GPA) to identify sets of candidate loci that might be
adaptively differentiated in response to both sets of predictors. We
then estimated whether these sets of loci overlap to determine any
signal of concerted selection or adaptation linking ecomorphological
divergence (Zhang et al. 2019; Carvalho et al. 2021; Valette et al.
2023). While the lack of an annotated genome and identification of the
GBS loci limits subsequent inference, the genome-wide nature of the
markers should capture at least some variation related to adaptive
ecological and phenotypic differentiation, if present. We used the 9
climatic variables from the GDM analysis for the 49 samples with low
missing data, and 14 phenotypic measurements for 40 of those samples
that matched between the two datasets. We dropped IN, TL, and ES from
the trait data, as they had ~0 loadings in the LDA
between populations (Fig. 2).