Landscape genomics of the streamside salamander: Implications for species management in the face of environmental change

Understanding spatial patterns of genetic differentiation and local adaptation is critical in a period of rapid environmental change. Climate change and anthropogenic development have led to population declines and shifting geographic distributions in numerous species. The streamside salamander, Ambystoma barbouri, is an endemic amphibian with a small geographic range that predominantly inhabits small, ephemeral streams. As A. barbouri is listed as near-threatened by the IUCN, we describe range-wide patterns of genetic differentiation and adaptation to assess the species’ potential to respond to environmental change. We use outlier scans and genetic-environment association analyses to identify genomic variation putatively underlying local adaptation across the species’ geographic range. We find evidence for adaptation with a polygenic architecture and a set of candidate SNPs that identify genes putatively contributing to local adaptation. Our results build on earlier work that suggests that some A. barbouri populations are locally adapted despite evidence for asymmetric gene flow between the range core and periphery. Taken together, the body of work describing the evolutionary genetics of range limits in A. barbouri suggest that the species may be unlikely to respond naturally to environmental challenges through a range shift or in situ adaptation. We suggest that management efforts such as assisted migration may be necessary in the future.

Methods
The dataset contains sequencing data for 112 individuals of the streamside salamander (Ambystoma barbouri). Sample DNA was processed using double-digest restriction site-associated DNA sequencing (ddRADseq) using the restriction enzymes EcoRI and PstI. Due to the large genome size of A. barbouri, we included only 30 individuals in each of four final libraries to attempt to increase sequencing depth across individuals; two individuals from each of four sampling locations were duplicately sequenced in different libraries. The four libraries were sequenced on an Illumina Hiseq 2000 sequencing system at the University of Oregon Genomics Core Facility (gc3f.uoregon.edu), using single-end 100bp reads. Samples were demultiplexed using the Stacks version 2.52 module process_radtags with options -c -q -r enabled.