Genetic variation is a fundamental component of biodiversity, and studying population structure, gene flow, and demographic history can help guide conservation strategies for many species. Like other aerial insectivores, the purple martin (Progne subis) is in decline, and yet their genetic background remains largely unknown. To address this knowledge gap, we assessed population structure in the nominate eastern subspecies (P. s. subis) with relation to natal dispersal and examined historical genetic patterns in all three subspecies (P. s. subis, P. s. arboricola, P. s. hesperia) across their North American breeding range by estimating effective population sizes over time. We used next-generation sequencing strategies for genomic analyses, integrating whole-genome resequencing data with continent-wide band encounter records to examine natal dispersal. We documented population structure across P. s. subis, with the highest differentiation between the northern (Alberta) and more southern colonies and following patterns of isolation-by-distance. Consistent with spatial patterns of genetic differentiation, we also found greater longitudinal than latitudinal natal dispersal distances, signifying potential latitudinal constraints on gene flow. Earlier contractions in effective population sizes in the western P. s. arboricola and P. s. hesperia compared to the eastern P. s. subis subspecies suggest these subspecies originated from two different glacial refugia. Together, these findings support latitudinal distinction in P. s. subis, and elucidate the origin of subspecies differentiation, highlighting the importance to conserve populations across the range to maximize genetic diversity and adaptive potential in the purple martin.

This dataset contains the unfiltered and filtered single-nucleotide-polymorphisms (SNPs) for 71 purple martins and corresponding code for the filtering process and population analyses. DNA was sequenced through skimSeq low-coverage whole-genome sequencing. The processed dataset has been filtered for quality (QUAL > 20, MQ > 20, MAF > 0.05, missingness < 20%), HWE, bi-allelic, autosomes, and LD-pruned with r²=0.2. Raw sequence reads are available on the Sequence Read Archive (SRA).

Additionally included is code for analyzing bird band encounter records (data was obtained from North American Bird Banding Program in May 2019).

The filtered dataset includes four outliers that were removed from downstream population analyses. See the metadata file for details on sample information. This study also looked at 4 additional samples for demographic history, and raw sequence data for these samples are available in the same SRA BioProject. See README.txt for more details.