Chromosomal rearrangements are often associated with local adaptation and speciation because they suppress recombination, and as a result, rearrangements have been implicated in disrupting geneflow. Although there is strong evidence to suggest that chromosome rearrangements are a factor in genetic isolation of divergent populations, the underlying mechanism remains elusive. Here we applied an integrative cytogenetics and genomics approach testing whether chromosomal rearrangements are the initial process, or a consequence, of population divergence in the dwarf goanna, Varanus acanthurus. Specifically, we tested whether chromosome rearrangements are indicators of genetic barriers that can be used to identify divergent populations by looking at geneflow within and between populations with rearrangements. We found that geneflow was present between individuals with chromosome rearrangements within populations, but there was no geneflow between populations that had similar chromosome rearrangements. Moreover, we identified a correlation between reduced genetic variation in populations with a higher frequency of homozygous submetacentric individuals. These findings suggest that chromosomal rearrangements were widespread prior to divergence and because we found populations with higher frequencies of submetacentric chromosomes were associated with lower genetic diversity, this could indicate that polymorphisms within populations are early indicators of genetic drift.

SNP analysis

We used Diversity Arrays Technology (DArT, Bruce, ACT, Australia) for genotyping 49 individuals including the 34 individuals that were karyotyped (Table 1). DArT is a genome-wide SNP typing technology that utilizes complexity reduction and Illumina sequencing (Kilian et al. 2012). DNA samples were digested with restriction enzymes to ~ 500bp size fragments; these fragments were cloned, amplified and sequenced using Illumina. The target libraries were generated from all individuals and the allele differences for each sequence were characterized by 0, 1 or 2 (homozygous, heterozygous, homozygous for opposite allele). These data were compared to the reference clone from Illumina Barcoding (Kilian et al. 2012).

Data sorting and Analysis

The data was analyzed using the dartR package, which was developed specifically for Diversity Arrays Technology output (Gruber et al. 2018). Briefly, the data were initially filtered for call rate by locus and individuals separately with a threshold of 0.8 and 0.75 respectively. Following call rate filtering, we filtered by reproducibility at a threshold of 0.99 and removed monomorphic loci from the whole dataset for initial analysis. We then performed a scree plot of eigenvalues that indicated informative axes (Unmack et al. 2019) and performed principal coordinate analysis (PCA) (Pearson 1901; Hotelling 1933; Jollife 2002; Jollife and Cadima 2016) with no a priori population assignments (Table 4a and Table 4b). Fixed differences and Fst analysis were performed for each population to further characterize the distribution and frequencies of alleles between populations (Table 4). To analyze each population independently we filtered additional monomorphic loci that were population specific. We then used (PCA) to assess genetic distances for each individual within populations from each locality and assign the karyotypes to individual genotypes within the PCA to observe the distribution of karyotype morphologies with Euclidean genetic distance measures (Figure 3a, 3c, 3e and 3g). This allowed for within-population genetic distance measures between individuals with different karyotype morphologies.

To determine geneflow between individuals with different karyotype morphologies within populations, we used both PCA and isolation by distance using a dissimilar measure of unshared alleles (1 – proportion of shared alleles). We assigned karyotypes to each individual and plotted the analyses (Figure 3b, 3d, 3f and 3h). In Figure 3 we also used a neighbor-joining tree to demonstrate the relationships between populations based on SNP data and karyotypes.

To investigate genetic differentiation and estimate geneflow between 221 populations, we performed an isolation-by-distance analysis based on the mantel test (Rousset 1997). First, we analyzed each population individually as defined by locality and PCA ordination (north, south, east and west) pairwise with unshared alleles (1 – proportion of shared alleles) for each individual versus distance (Figure 4c). Next, we analyzed these four populations pairwise for F_st values (Figure 4d) versus distance. Following these analyses, we performed the same analysis with the karyotypes as populations to test for geneflow between karyotypes using both metrics, unshared alleles (1 – proportion of shared alleles) versus distance (Figure 4e) and Fst vs distance (Figure 4f).

R

Rstudio