Does population size affect genetic diversity? A test with sympatric lizard species

Sampling and choice of species

We sampled four species: zebra-tailed lizards (Callisaurus draconoides), western banded geckos (Coleonyx variegatus), chuckwallas (Sauruomalus ater), and desert iguanas (Dipsosaurus dorsalis). Similar to U. stansburiana, C. draconoids and C. variegatus are small, insectivorous and very abundant in MNP (Persons and Nowak, 2007). In contrast, S.ater and D. dorsalis are large, mainly herbivorous and less numerous species (Persons and Nowak, 2007). In addition, our study site is centrally located in the ranges of all species (Jones and Lovich, 2009), so the reduction in diversity associated with range-limiting effects is unlikely range or marginal habitat is a problem. Additionally, sampling all four species at the same location in the MNP increases the likelihood that all populations have had roughly coincidental biogeographical histories. With two similar species in each experimental group, our population survey at MNP is a replicated analysis of variation in levels of molecular diversity.

We are confident in the designations of these species as high or low density populations. The results of the single herpetofauna survey (Persons and Nowak, 2007) are consistent with our classification, as is our own experience capturing these lizards at MNP over the past 19 years. Three of these four species, in addition to U. stansburiana, are diurnal lizards that spend most of their active time basking on rocks, with occasional forays to feed. A typical search for lizards involves walking a transect while scanning rocks and ground for basking lizards. During these searches, we usually come across 15 to 20 U. stansburiana and C. draconoids for each individual of D. dorsalis Where S.ater, even in optimal habitat for less common species. The night C. variegatus are sampled by driving on the main tarmac road through the MNP and sampling the lizards as they cross the roads. C. variegatus is by far the most common reptile species seen on the road, and in normal years a sample of 40 lizards would easily be collected over several nights. Unfortunately, our study was conducted during a drought year, which seemed to depress C. variegatus activity, as well as that of other species.

We randomly collected MNP individuals from the region of the Cima volcanic field (Figure 1, data available in the Dryad digital repository: The samples consisted of 35 S.ater, 21 D. dorsalis, 35 C. draconoids and 21 C. variegatus. Diurnal lizards were captured using a slipknot. Night C. variegatus were captured by hand. A 0.2 cm tail tip tissue sample was taken from all individuals and preserved in 95% ethanol for genetic analysis.

Figure 1

Satellite imagery of the Cima Volcanic Field region of the Mojave National Preserve with collection locations. The star on the inset map indicates the location of the maps.

Given the limited geographical distribution of our samples, we are confident that these samples come from a single population for each species. The greatest geographic distances between two individuals in our sample were found in C. draconoids (12km) and C. variegatus (18 km), but most samples were taken from a much more localized area and suitable habitat was continuous between individuals. Some samples for C. draconoids (4 lizards) and C. variegatus (3 lizards) were somewhat geographically separated from the majority of collected individuals, but reanalysis of the data excluding these individuals had no effect on the genetic parameter estimates described below.

Choice of locus and molecular methods

For consistency with the U. stansburiana study, we interviewed the cytb and MC1R genes in the four species. For a better representation of the permanent level of genetic diversity, we analyzed an additional autosomal gene: the recombination activation gene-1 (RAG1). RAG1 has been used successfully as a marker in higher-level phylogenetic studies, including the determination of basal divergences in squamate reptiles (Townsend et al., 2004), and therefore should not be highly variable at the population level.

Whole genomic DNA was extracted from frozen tissues using the Quick-gDNA MiniPrep Kit (Zymo Research, Irvine, CA, USA). Primers for the cytb mitochondrial locus and the MC1R and RAG1 autosomal loci were obtained from previously published studies and developed by aligning sequences from closely related species accessible via GenBank (Supplementary Information 1). Target loci were amplified with 20 µl PCR reactions (Supplementary Information 1) using AccuPower PyroHotStart Taq PCR PreMix (Bioneer, Alameda, CA, USA). Each PCR reaction included 18 µl of PCR water, 0.5 µl of each primer and 1 µl of genomic DNA. The PCR product was cleaned up using ExoSAP-IT (Affymetrix, Santa Clara, CA, USA) and the purified product was sequenced back and forth by Elim Biopharmaceuticals, Hayward, CA, USA. Sequences were aligned and edited in Geneious 4.8.5 (Biomatters, available at To avoid confusing poor sequence data with high genetic diversity, sequence chromatograms with indistinct peaks or suspicious base changes were resequenced. Heterozygous sites in autosomal loci were identified by visual inspection and confirmed in both directions of sequencing.

The linking phase of the haplotypes was deduced by calculation with the PHASE program (Stephens et al., 2001). PHASE can sometimes have difficulty resolving low-frequency alleles, and omission of these alleles from analyzes can lead to artifactual reductions in molecular diversity estimates in population studies (Garrick et al., 2010). Individuals with alleles that PHASE could not reliably infer at the 90% confidence level were resolved by cloning and sequencing the PCR product from that locus. The PCR product from heterozygous individuals was cloned using a TOPO TA cloning kit for sequencing (Life Technologies, Grand Island, NY, USA). Several clones were sequenced with the same methods used for sequencing PCR products from genomic templates to determine the true haplotypes for each individual.

Intraspecific genetic diversity

Population genetic analyzes have been performed with Arlequin 3.5 (Excoffier and Lischer, 2010). Each species was treated as a single population. For each species, we tested for recent population expansion using a pairwise shift distribution (Rogers and Harpending, 1992). We tested significance using the irregularity index (r; Harpending, 1994) and the sum of squared deviations test (Rogers and Harpending, 1992). We also tested deviations from neutral expectations using Tajima D (D; Tajima, 1989).

To assess molecular diversity, we estimated haplotype number and haplotype diversity (h) at each locus (Nei, 1987). We also estimated nucleotide diversity with Watterson’s method θ (θ; Watterson, 1975) and Nei θ (π; Watterson, 1975; Tajima, 1983). For each locus, we performed pairwise comparisons of haplotype diversity between species. Differences in haplotype diversity between species were considered significant if the 95% confidence intervals of the two estimates did not overlap by more than half of a one-sided error bar (Cumming and Finch, 2005). We also compared haplotype diversities using a z– score test proposed by Nei (1987). To better visualize the genetic diversity at each locus, we constructed haplotype networks for each locus using TCS 1.21 (Clement et al., 2000). Closed loops in the haplotype network were resolved by comparison to a maximum likelihood tree estimated in PAUP* 4.0 (Swofford, 2003).

Comments are closed.