Copy number variants outperform SNPs to reveal genotype-temperature association in a marine species

Abstract

Copy number variants (CNVs) are a major component of genotypic and phenotypic variation in genomes. Yet, our knowledge on genotypic variation and evolution is often limited to single nucleotide polymorphism (SNPs) and the role of CNVs has been overlooked in non-model species, partly due to their challenging identification until recently. Here, we document the usefulness of reducedrepresentation sequencing data (RAD-seq) to detect and investigate copy number variants (CNVs) alongside SNPs in American lobster (Homarus americanus) populations. We conducted a comparative study to examine the potential role of SNPs and CNVs in local adaptation by sequencing 1141 lobsters from 21 sampling sites within the southern Gulf of St. Lawrence which experiences the highest yearly thermal variance of the Canadian marine coastal waters. Our results demonstrated that CNVs accounts for higher genetic differentiation than SNP markers. Contrary to SNPs for which no association was found, genetic-environment association revealed that 48 CNV candidates were significantly associated with the annual variance of sea surface temperature, leading to the genetic clustering of sampling locations despite their geographic separation. Altogether, we provide a strong empirical case that CNVs putatively contribute to local adaptation in marine species and unveil stronger spatial signal than SNPs.Our study provides the means to study CNVs in non-model species and underlines the importance to consider structural variants alongside SNPs to enhance our understanding of ecological and evolutionary processes shaping adaptive population structure.

Thermal adaptation rather than demographic history drives genetic structure inferred by copy number variants in a marine fish

“…Both DNA extractions and library preparations were performed following protocols fully described in Cayuela et al (2020). Libraries were size-selected using a BluePippin prep (Sage Science), amplified by PCR and sequenced on the Ion Proton P1v2 chip (single-end sequencing).…”

“…The HDplot approach proposed by McKinney et al (2017) allows identifying CNVs based on two characteristics: the proportion of heterozygous individuals within a population and allelic ratios within heterozygous individuals. We used a refined version of this approach (Dorant et al 2020) that discriminates "singleton" SNPs (i.e., non-duplicated) from "duplicated" SNPs (i.e., combining duplicated SNPs and SNPs with a high coverage) using 5 parameters: (1) proportion of heterozygotes (PropHet), (2) Fis, (3) median of allele ratio for heterozygotes (MedRatio), (4) median of SNP read depth for heterozygotes (MedDepthHet) and (5) median of SNP read depth for homozygotes (MedDepthHom). The parameters were calculated using a custom python script (available at https://github.com/enormandeau/stacks_workflow) parsing the filtered VCF file.…”

“…However, with the exception of targeted candidate genes studies, CNV genotyping remained expensive for a long time as it required whole genome sequencing data with deep coverage (Pirooznia et al 2015, Makałowski et al 2019, precluding the use of these markers in population genomic studies that involve thousands of samples from dozens of populations. This situation has recently changed through the development of a novel methodological approach allowing the identification of CNVs in a cost-effective way using RAD-seq data (McKinney et al 2017, Dorant et al 2020. Using RAD-seq data, Dorant et al (2020) recently showed that CNVs can be more efficient than SNPs at revealing genotypetemperature associations in marine invertebrates.…”

“…This situation has recently changed through the development of a novel methodological approach allowing the identification of CNVs in a cost-effective way using RAD-seq data (McKinney et al 2017, Dorant et al 2020. Using RAD-seq data, Dorant et al (2020) recently showed that CNVs can be more efficient than SNPs at revealing genotypetemperature associations in marine invertebrates. Yet, at present time, the usefulness of CNVs from RAD-seq data as valuable DNA markers for population genomic studies still needs to be more broadly exemplified.…”

“…More precisely, we first used the method developed by McKinney et al (2017) and refined by Dorant et al (2020) to detect reliable CNVs. Second, we investigated how CNVs normalized read depth, a robust proxy of putative copy number (Dorant et al 2020), correlates with the gonadosomatic index, a commonly used fitness proxy in fishes (Brewer et al 2008, Ressel et al 2019. Third, we examined the putative role of CNVs in thermal adaptation by assessing correlations between normalized read depth and water temperature in beach spawning sites.…”