RESULTS
Patterns of genetic diversity and differentiation
We identified 15,038 SNPs at a mean of 27X sequencing depth across 17 sampling locations. Using a dataset filtered by a minor allele frequency of 0.1, we find that genome-wide FST is extremely low (0.0085), and measures of genetic diversity do not vary substantially among sampling locations (Table 1). Across all of these wintering starling populations, the highest pairwise FST was 0.0106 between birds from the adjacent states of Arizona and New Mexico. Using a haplotype-based statistic of differentiation, φST among populations shows an absence of genetic structure (φST = 0.0002). Hierarchical AMOVAs reveal that 94% of the observed genetic variance is explained by variation within individuals, and the remaining variance reflects differences among individuals in the same population, with negligible variation explained at the between-population level (Figure 1C-D). Across the genome, FST and nucleotide diversity are exceptionally low (Table 1,2). Genome-wide heterozygosity is moderate at 0.339, and across loci observed heterozygosity differs significantly from expected (t = 66.6, df = 3569, P<0.001, Table 1).
Population structure
FST among populations is relatively low overall (Table 2): starlings sampled in Arizona show the highest differentiation from other sampling locations (FST = 0.008-0.011), and only a few other pairwise comparisons (NM-IL, NM-CO, and CO-WI) show an FST of 0.009 or higher. However, we do not recover clear population structure. The first two principal components each explain about 1% of variation among individuals (Figure 1A), and although STRUCTURE identified three populations at the best-supported value of K, these predicted populations do not show obvious differences in ancestry proportions (Figure 1B, Figure S1). Controlling for shared ancestry does not resolve population structure, and instead provides support for uniform gene flow among individuals (Figure S2). K-means clustering within DAPC also does not identify biologically relevant clusters.
Starling populations follow a clear pattern of isolation-by-distance (Mantel r = 0.139, P < 0.0001). Spatially explicit models of isolation-by-distance suggest a fairly uniform rate of migration range-wide, where local increases in migration rate are likely a model artifact (Supplementary Information). However, models of isolation by environment (IBE) show that the relationship between environmental and genetic distances is stronger than the simple geographic-genetic distance model. After controlling for geographic distance, we find that all bioclimatic variables tested show non-zero correlations between environmental and genetic distances (Supplementary Information). There is a strong positive relationship between genetic distance and both precipitation in the wettest quarter (BIO16, Mantelr = 0.282, P = 0.001) and elevation (Mantel r = 0.146, P = 0.001).