Shared and nonshared genomic divergence in parallel ecotypes of Littorina saxatilis at a local scale

Abstract

Parallel speciation occurs when selection drives repeated, independent adaptive divergence that reduces gene flow between ecotypes. Classical examples show parallel speciation originating from shared genomic variation, but this does not seem to be the case in the rough periwinkle (Littorina saxatilis) that has evolved considerable phenotypic diversity across Europe, including several distinct ecotypes. Small 'wave' ecotype snails inhabit exposed rocks and experience strong wave action, while thick-shelled, 'crab' ecotype snails are larger and experience crab predation on less exposed shores. Crab and wave ecotypes appear to have arisen in parallel, and recent evidence suggests only marginal sharing of molecular variation linked to evolution of similar ecotypes in different parts of Europe. However, the extent of genomic sharing is expected to increase with gene flow and more recent common ancestry. To test this, we used de novo RAD-sequencing to quantify the extent of shared genomic divergence associated with phenotypic similarities amongst ecotype pairs on three close islands (<10 km distance) connected by weak gene flow (Nm ~ 0.03) and with recent common ancestry (<10 000 years). After accounting for technical issues, including a large proportion of null alleles due to a large effective population size, we found ~8-28% of positive outliers were shared between two islands and ~2-9% were shared amongst all three islands. This low level of sharing suggests that parallel phenotypic divergence in this system is not matched by shared genomic divergence despite a high probability of gene flow and standing genetic variation.

Beyond the Coral Triangle: high genetic diversity and near panmixia in Singapore's populations of the broadcast spawning sea starProtoreaster nodosus

“…Applying a minor allele frequency filter of requiring alleles to be present in at least two or more individuals (MAF > 0.06), however, resulted in a shift in peak away from F IS  = 0 and increase in the number of loci at F IS  = 1 (electronic supplementary material, figure S6 a–c ), which was similar to the effect of using the rxstacks catalogue correction module in STACKS (preliminary trials, data not presented). A similar effect was also reported in [59]. All in all, 1431–6140 SNP loci were called across the various parameter combinations tested (electronic supplementary material, figure S3, shows SNP numbers for mismatch threshold sensitivity analysis), but results were similar across all combination sets.…”

Lineage divergence, local adaptation across a biogeographic break, and artificial transport, shape the genetic structure in the ascidian Pyura chilensis

“…In spite of the possible caveats, 3.5% (81) of the total retained SNPs (2,332) showed low and significant q-values and higher levels of F ST for 90 individuals, allowing us to distinguish individuals genetically between sampling locations at a relatively small geographic scale, and suggesting that significant differences could be due to positive selective pressures49 that differ north/south of 30°S. The percentage of outlier loci detected is proportionally consistent with other studies that show high variation between local populations1922235051. Further studies are needed to evaluate the genes involved in the putatively adaptive genetic structure detected here, and determine if putatively selected loci are actually under positive selection due to different environmental conditions north and south of the biogeographic boundary at 30°S in the southeast Pacific.…”

Recent advances in conservation and population genomics data analysis

“…A simple visualization of expected and observed frequencies of homozygote genotypes across single nucleotide polymorphisms (SNPs) can be effective in identifying data problems (Figure 1). A simple model for estimating the heterozygote miscall (dropout) rate was applied to 12 publicly available RAD‐seq datasets (Fernández et al., 2016; Hecht, Matala, Hess, & Narum, 2015; Laporte et al., 2016; Larson et al., 2014; Le Moan, Gagnaire, & Bonhomme, 2016; Portnoy et al., 2015; Prince et al., 2017; Puritz, Gold, & Portnoy, 2016; Ravinet et al., 2016; Swaegers et al., 2015). While a few had low genotyping error rates (<5%), in others, allelic dropout, low read depth, PCR duplicates, erroneous assembly, and/or poor filtering resulted in much higher estimated error rates, with between 5% and 72% of heterozygotes apparently being miscalled as homozygotes.…”