An approach for identifying cryptic barriers to gene flow that limit species' geographic ranges

Abstract

Species' geographic range limits are most often not demarcated by obvious dispersal barriers. Poor-quality habitat at the edge of a species' range can prevent range expansion by preventing outward migration or through reducing adaptive potential resulting from decreased genetic diversity. We identified habitat variables that constrain gene flow across the entire geographic range of an endemic salamander (Ambystoma barbouri) in the eastern United States, and we tested whether increased resistance resulting from these variables provides cryptic dispersal barriers at the range edges. Using polymorphic microsatellite loci, we first identified three genetic clusters that are separated by the Ohio and Kentucky rivers. Through a combination of landscape genetic analyses and generalized dissimilarity modelling, we then classified variables that (i) restrict gene flow in each of the genetic clusters across the geographic distribution of A. barbouri and (ii) become more common towards the peripheries of the distribution. A decrease in limestone availability and an increase in growing season precipitation were correlated with high resistance to gene flow across the range, and both became more common at the edges of the species' distribution. However, other landscape variables were more important for explaining variation in geneflow rates in different portions of the range, such as increased mean annual temperature and frost-free period in the south vs. growing season precipitation in the north. Taken together, these results suggest that there are both range-wide and regionally specific cryptic habitat barriers preventing geographic range expansion. Species 'geographic range limits are probably governed by a set of ecological and evolutionary factors, and our landscape genetic approach could be applied to gain additional insight into many systems.

Landscape features along migratory routes influence adaptive genomic variation in anadromous steelhead (Oncorhynchus mykiss)

“…We measured landscape variables both at natal sites and along the migration path for each population using the Spatial Analyst Toolbox in ArcGIS (ESRI, Redlands, CA). For natal‐site variables, we measured each landscape value as the mean of a 5‐km buffer around the collection site to account for natal habitat variation (Micheletti & Storfer, ). Migration path measurements, on the other hand, encompass a long geographic extent, only some of which is shared between populations.…”

“…To avoid type I error in downstream analyses, we removed landscape measurements with a pairwise correlation >0.75 (Asuero, Sayago, & Gonzalez, ), retaining biologically relevant variables per previous salmonid studies (Hecht et al., ; Olsen et al., ). This initial filter was intended to remove highly correlated variables and retain putatively informative variables (McGaughran, Morgan, & Sommer, ; Micheletti & Storfer, ) for further analyses with subsequent identification of autocorrelation. After eliminating highly correlated variables, we ultimately retained 45 measurements across 19 unique landscape variables (Table ).…”

Mixed support for gene flow as a constraint to local adaptation and contributor to the limited geographic range of an endemic salamander

“…For example, if dispersal is limited due to harsh habitat conditions or fragmentation, genetic drift may reduce genetic diversity below a threshold required for adaptation (Polechová, 2018;Polechová & Barton, 2015). While there is considerable empirical evidence supporting the first two hypotheses (Bridle et al, 2009;Hargreaves et al, 2013;Micheletti & Storfer, 2015, 2017Sexton et al, 2011), empirical tests of the third hypothesis have lagged behind theoretical considerations, largely due to constraints of adequately sampling species' geographic ranges (Eckert et al, 2008;Sexton et al, 2009). Further, such studies necessitate a dual-pronged approach that requires estimation of range-wide gene flow and overcoming the challenges of conducting reciprocal transplant experiments across a large geographic extent.…”

“…Gene flow between fish and fishless populations resulted in overactive and less camouflaged larvae from fish populations, making them more susceptible to fish predation (Storfer et al, 1999;Storfer & Sih, 1998). On a larger spatial scale, edge populations tend to have smaller effective population sizes in A. barbouri and range-wide environmental conditions between the centre and edges differ (Micheletti & Storfer, 2015, 2017.…”

Combining Harvest and Genetics to Estimate Reproduction in Wolves

“…Under these circumstances, genetic monitoring can be an efficient approach to obtaining reliable demographic information (De Barba et al , Stansbury et al ). Genetic data can help identify species and individuals, provide estimates of population parameters, and offer insights into space use and connectivity (Schwartz et al , Paetkau et al , Mumma et al , Micheletti and Storfer ). Measures of relatedness and genetic diversity can be used to reconstruct pedigrees, gain greater understanding of mating systems, assess population viability, and track quantitative traits (Thomas and Hill , DeWoody , Lucia and Keane , Putnam and Ivy , Gooley et al ).…”