Aim: Shaped by both climate change and sea-level rise, tidal salt marshes represent ephemeral systems that are home to only a few, highly specialized species. The dynamic ecological histories and spatial complexities of these habitats, however, render it challenging to reconstruct the complete biogeographic histories of their endemic taxa. Here, we leverage three species of North American Ammospiza sparrows that inhabit tidal marshes ( Ammospiza caudacuta, A. maritima, and A. n. subvirgatus) and closely related freshwater species to demonstrate the utility of whole-genome data in resolving demographic and evolutionary history as it relates to divergence and dispersal events in ephemeral ecosystems. We employ a combination of demographic and biogeographic reconstructions to shed new light on the colonization history of freshwater-saline environments in this system.

Location: North America

Taxon: Ammospiza Sparrows

Methods: We sequenced whole genomes from Ammospiza sparrows to address our objectives. We conducted phylogenomic analyses and reconstructed the demographic and biogeographic history of this clade based on 21 million SNPs from 54 total individuals.

Results: Phylogenies based on several million SNPs supported several well-resolved clades that predominantly corresponded to their prior species designations. Phylogenetic and biogeographic reconstructions suggest a series of saltwater to freshwater colonization events within this group, with some endemic taxa exhibiting associations with tidal marsh habitat over longer evolutionary time scales and some habitat transitions occurring as recently as 5,000 years ago. 

Main Conclusions: Our reconstructions support a biogeographic hypothesis with fewer vicariance and dispersal events among Ammospiza sparrows that is in contrast with the currently supported evolutionary scenario. Biogeographic reconstructions further suggest saltwater to freshwater transitions in A. n. subvirgatus as opposed to the long hypothesized freshwater origin. Our results highlight the fact that reconstructing biogeographic and evolutionary dynamics in ephemeral systems poses challenges given the propensity for high turnover. This reinforces the importance of a multifaceted approach to biogeographic reconstructions in historically dynamic ecosystems.

We prepared individually indexed libraries using 1 µg of DNA following the TruSeq DNA PCR free library preparation kit protocol (San Diego, CA, USA), with an insert size of 350 bp. We pooled 24 libraries using concentrations of adapter-ligated DNA determined through digital PCR. The pooled libraries were sequenced on three Illumina NextSeq500 lanes at the Cornell Institute for Biotechnology core facility. Each lane contained 24 individuals, as some samples were run with samples for a different study. The quality of individual libraries was assessed using FastQC (version 0.11.5; http://www.bioinformatics.babraham.ac.uk/projects/fastqc).

Reads were aligned to the A. caudacuta reference genome (Walsh et al. 2019a). We performed sequence trimming, adapter removal, and quality filtering with AdapterRemoval version 2.1.1 (Lindgreen 2012). We allowed a minimum threshold quality score of 10 (increased from the default value of 2) and merged overlapping paired-end reads. Filtered reads were aligned to the reference using bwa 0.7.4 (Li and Durbin 2009) with default settings. Alignment statistics were obtained using qualimap version 2.1.1 (Okonechnikov et al. 2015). 

            BAM files were sorted and indexed using SAMtools version 1.3 (Li et al. 2009). We realigned around indels and fixed mate-pairs using GATK version 3.5 (McKenna et al. 2010). SNP variant discovery and genotyping for the 54 individuals was performed using BCFTools (-mpileup and -call). For an initial filtering step, variants that were out of Hardy-Weinberg equilibrium, had a minor allele frequency less than 2%, had a mean depth less than 2 or greater than 50, and more than 20% missing data were filtered from the data set using VCFtools version 0.1.14 (Danecek et al. 2011). This initial filtering step resulted in 21,622,990 SNPs. 

Individuals are named by species abbreviation and a number.

LCSP: A. leconteii

Nelsoni: A. n. nelsoni

SALS: A. caudacuta