applyparastyle “fig//caption/p[1]” parastyle “FigCapt” Biological Journal of the Linnean Society, 2019, XX, 1–8. With 2 figures. Downloaded from https://academic.oup.com/biolinnean/advance-article-abstract/doi/10.1093/biolinnean/blz173/5673093 by bucknell university user on 11 December 2019 Genomic data reveal similar genetic differentiation in aquifer species with different dispersal capabilities and life histories STEVE JORDAN1*, BRIAN K. HAND2, SCOTT HOTALING3, AMANDA G. DELVECCHIA2, RACHEL MALISON2, CLARK NISSLEY1, GORDON LUIKART2 and JACK A. STANFORD2 1Biology Department, Bucknell University, Lewisburg, PA, USA 2Flathead Lake Biological Station, University of Montana, Polson, MT, USA 3School of Biological Sciences, Washington State University, Pullman, WA, USA Received 12 July 2019; revised 18 October 2019; accepted for publication 19 October 2019 Little is known about the life histories, genetic structure and population connectivity of shallow groundwater organisms. We used next-generation sequencing (RAD-seq) to analyse population genomic structure in two aquifer species: Paraperla frontalis (Banks, 1902), a stonefly with groundwater larvae and aerial (winged) adults; and Stygobromus sp., a groundwater-obligate amphipod. We found similar genetic differentiation in each species between floodplains separated by ~70 river km in the Flathead River basin of north-west Montana, USA. Given that Stygobromus lacks the above-ground life stage of P. frontalis, our findings suggest that connectivity and the magnitude of genetic structure cannot be definitively assumed from life history differences. ADDITIONAL KEYWORDS: Amphipoda – aquifer ecosystem – groundwater fauna – Plecoptera – RAD-seq. INTRODUCTION The distribution of groundwater crustaceans in alluvial aquifers is driven by hydrogeological and Gravel-bed river floodplains have been described as geomorphic processes and has little relationship to the ‘ecological nexus’ of mountain regions because of distance from the river. This means that stygobionts their important above- and below-ground communities are ubiquitous throughout the floodplain, whereas and processes (Hauer et al., 2016). Stanford & Gaufin amphibionts are most common near the river (Stanford (1974) first reported stoneflies in the alluvial aquifers et al., 1994; Ward et al., 1994). Amphibiotic stoneflies of Montana rivers. Researchers have since documented (Insecta: Plecoptera) spend 1–3 years in the aquifer diverse communities of macroinvertebrates, meiofauna before leaving the groundwater to emerge as adults and microbes in shallow aquifers worldwide, and have (Stanford et al., 1994). In contrast, as stygobionts, established the enormous ecological importance of Stygobromus (Crustacea: Amphipoda) individuals floodplain habitats (e.g. Tockner & Stanford, 2002; never leave the groundwater system. The life cycles Stanford et al., 2005; Boulton et al., 2010; Hauer of many stygobionts, including Stygobromus, are et al., 2016). These communities include insects that largely unknown, but they exhibit characteristics remain underground as nymphs but emerge above of all aquatic subterranean animals and obligate ground for reproduction (amphibionts), as well as cave dwellers that render them maladapted to life taxa (e.g. crustaceans, oligochaetes and mites) that above ground – being blind and lacking pigmentation are obligate residents of interstitial spaces in aquifers (Culver et al., 2010). Stygobromus is found in a variety (stygobionts). These animals occur over 10 m beneath of subterranean habitats including caves, phreatic riverine floodplains and up to 5 km from the main lakes, seeps and epikarst (Culver et al., 2010), as well river channels (Stanford & Ward, 1988). as alluvial aquifers, and there is no evidence that they ever voluntarily leave the groundwater. Shallow aquifers offer many challenges to resident *Corresponding author. E-mail: [email protected] organisms, including geologically bounded isolation, © 2019 The Linnean Society of London, Biological Journal of the Linnean Society, 2019, XX, 1–8 1 2 S. JORDAN ET AL. no light, variable water flow, and reduced availability TAXA AND SAMPLING of carbon, other nutrients and oxygen (Tockner et al., We sampled two taxa that exemplify major, different Downloaded from https://academic.oup.com/biolinnean/advance-article-abstract/doi/10.1093/biolinnean/blz173/5673093 by bucknell university user on 11 December 2019 2010; DelVecchia et al., 2016). Many factors, including life-history strategies: P. frontalis, an aquifer stonefly bedrock geology, soil permeability, water chemistry and with a winged adult stage, and Stygobromus sp., a quality, groundwater levels, adjacent surface flows, blind, pigmentless, groundwater-obligate crustacean. riparian vegetation, and climate, can influence the We used mechanical pumping of seven permanent spatial distribution and ecology of resident organisms wells to collect these species in June 2011 and 2012 (e.g. Johns et al., 2015; Korbel & Hose, 2015). Previous (Fig. 1). We extracted DNA from 96 individuals of each genetic studies have identified widespread, long- species and confirmed species identifications based on term barriers to dispersal by groundwater species, cytochrome c oxidase subunit I (COI) gene barcoding even within drainages, despite potentially linking before proceeding with RAD-seq. floods (Lefebure et al., 2006; Cooper et al., 2007, 2008; Finston et al., 2007). For example, groundwater amphipods in Western Australia showed strong SNP CALLING AND FILTERING genetic differentiation between discrete pockets of We prepared RAD-seq libraries following standard groundwater, despite repeated Quaternary floods that protocols using the restriction enzyme SbfI, an covered the entire area (Cooper et al., 2007). 8-base cutter, and unique 6-bp barcodes (Miller The ability of groundwater organisms to actively et al., 2012). We sequenced 192 individuals on two disperse within and between adjacent habitats lanes of an Illumina HiSeq 2500 sequencer with remains unclear. Life history probably plays a role in 100-bp, single-end chemistry. Raw sequences were population connectivity. For example, the retention of filtered with 90% of the bases required to have a ephemeral, winged life stages by some amphibionts quality score ≥20 using the FASTQ Quality Filter offers dispersal advantages over taxa that never (http://hannonlab.cshl.edu/fastx_toolkit/). Reads for leave the groundwater. Current genomic tools can Stygobromus were trimmed to 80 bp to maximize facilitate the study of local adaptation to atypical quality. We used the process_radtags script in Stacks environments (Luikart et al., 2003) and resolve fine- v.1.19 (Catchen et al., 2013) to demultiplex reads scale differentiation in aquatic insects (Hotaling et al., by barcode, removing any with uncalled bases, and 2018). A better understanding of dispersal in these to call single nucleotide polymorphisms (SNPs) systems would greatly benefit biological understanding with a read depth (-m) of 5 and a maximum of two of connectivity along the river corridor, a major theme mismatches (-n) per locus and between catalogue loci in river ecology (e.g. Stanford et al., 2005). (-N). Additional SNP calling details are provided in Here we address these issues by using RAD-seq the Supporting Information. (restriction site associated DNA sequencing) datasets To investigate the influence of data scale (SNP for two co-occurring groundwater species with vastly number) and missing data on our results, we constructed different life histories from two floodplains of the four datasets for downstream analyses. All filtering Flathead River in north-west Montana. We hypothesized steps applied to the two species and all datasets that the groundwater-obligate Stygobromus would received the baseline filtering described above. Dataset have fewer opportunities for gene flow and thus vastly 1: we removed all individuals with >50% missing data, higher genetic differentiation between floodplains than and all loci genotyped in <60% of individuals. We the amphibiont Paraperla frontalis (Banks, 1902). also removed all loci that were not present in at least 50% of individuals in each population. For datasets 2–4, we removed any individuals with more than one standard deviation of missing data above the mean in METHODS the raw SNP dataset. Dataset 2: no additional filters. STUDY SITES Dataset 3: we only included loci genotyped in >25% of We sampled two Flathead River sites ~70 km apart: individuals. Dataset 4: we only included loci genotyped the Nyack and Kalispell Floodplains (Fig. 1). These in >75% of individuals. floodplains have been the focus of long-term research (e.g. Helton et al., 2014), and their aquifers are PAIRWISE DIFFERENTIATION AND POPULATION probably not connected in the subsurface because STRUCTURE the floodplains are bounded by bedrock knickpoints (Hauer et al., 2007). Each aquifer is known to contain We calculated pairwise FST values using GENEPOP a diverse array of meiofauna and macroinvertebrates (Rousset, 2008) for all sample–well pairs using (Stanford et al., 1994; Gibert et al., 1994; DelVecchia dataset 1. We pooled samples with pairwise FST = 0 et al., 2016). for subsequent analyses of population structure © 2019 The Linnean Society of London, Biological Journal of the Linnean Society, 2019, XX, 1–8 GROUNDWATER INVERTEBRATE ECOLOGICAL GENOMICS 3 Downloaded from https://academic.oup.com/biolinnean/advance-article-abstract/doi/10.1093/biolinnean/blz173/5673093 by bucknell university