SNP identification and validation in two invasive species: zebra mussel (Dreissena polymorpha) and Asian clam (Corbicula fluminea)

SNP identification and validation in two invasive species: zebra mussel (Dreissena polymorpha) and Asian clam (Corbicula fluminea). The development of affordable massive parallel sequencing (MPS) has reduced both time and costs of SNP identification for use in conservation and population genetic studies. After MPS, a second validation is usually required. High resolution melting analysis (HRMA) is a fast and simple method for mutation scanning, and thus a suitable validation protocol, particularly in non–model species. We present a set of nine novel polymorphic SNPs identified by MPS and validated with HRMA in two invasive species (the zebra mussel Dreissena polymorpha and the Asian clam Corbicula fluminea). These SNPs can be used in genetic studies to accurately assess and understand past and future invasion events.

Rapid developments in massive parallel sequencing (MPS) technologies have facilitated the use of single nucleotide polymorphisms (SNPs) in population genetic studies (Morin et al., 2009).SNPs have many advantages, including low-scoring error rates, high abundance, functional relevance, easy high-throughput genotyping (Liu et al., 2005), and more accurate estimates of population differentiation (Morin et al., 2009).
SNP identification in non-model species can be performed using MPS technologies without a reference genome (Everett et al., 2011).However, after the SNP calling step, a subsequent validation is usually required.High resolution melting analysis (HRMA) is a relatively new and inexpensive technology.It is based on highly precise and accurate measures of melting temperatures (Tm) of PCR-amplified DNA achieved by recording the fluorescence of a saturating DNA dye (Wittwer et al., 2003;Reed et al., 2007).As differences in melting curve profiles are diagnostic of SNPs, homozygote and heterozygote genotypes can be distinguished (Montgomery et al., 2007).This technique has been successfully used for SNP validation in several species, such as swordfish (Smith et al. 2010) and chum salmon (Seeb et al., 2011).
We selected 46 of these putative SNPs in D. polymorpha and 40 SNPs in C. fluminea for validation with short amplicon (SA) HRMA assays (Smith et al., 2013) in up to 96 individuals collected for previous studies (Peñarrubia et al., 2015a(Peñarrubia et al., , 2015b)).All the selected positions harbored a single SNP and the length of the amplicon after primer design was limited to less than 65 bp.Melting temperatures of the putative PCR products were pre-checked using uMELT SM v2.0 (https://www.dna.utah.edu/umelt/umelt.html).BLASTN analysis (https://blast.ncbi.nlm.nih.gov/Blast.cgi)was run to identify possible homologies producing non-specific amplifications in the PCR.
HRMA amplifications were conducted in 10 μL reactions containing 25-100 ng of genomic DNA, 1× EconoTaq Plus Master Mix (Lucigen), 1× LC-Green+ (Idaho Technology), and 0.2 μM of each prim-er.Thermal cycling was performed on a LightCycler 480 Real-Time PCR system (Roche Diagnostics) with an initial denaturation of 10 min at 95 ºC followed by 35 cycles denaturing for 10 s at 95 ºC, annealing at 60 ºC for 30 s, and extension for 10 s at 72 ºC.Reactions were overlaid with 15 μL of mineral oil to ensure that evaporative losses did not affect ionic strength which may affect melting uniformity across samples (Smith et al., 2010).Twenty-five HRMA data acquisitions per ºC were collected with a ramp rate of 0.02 ºC/s between 60 and 95 ºC.All melting curve patterns were analyzed using the LightCycler 480 Gene Scanning Software v. 1.5.0SP1 (Roche Diagnostics).
SA-HRMA characterization of the 46 putative positions in D. polymorpha indicated that five were monomorphic, seven produced patterns that were not consistent with homo-and hetero-duplex curves (e.g., three or more melting peaks, potential primer dimers, etc.), 23 produced heteroduplex curves (i.e., double peaks) indicative of the heterozygous condition in every individual in the sample (n = 15) tested, six failed to amplify even after repeated attempts to optimize PCR (not shown), and five loci (10,87 %) displayed polymorphic SNPs melting curves (table 1).In C. fluminea, four were identified as monomorphic, 10 produced non-scorable melting patterns, 18 generated double-peaks in every individual tested, four failed to amplify, and four SNPs (10 %) produced melting curves of polymorphic SNPs.While 10 % may be considered a low success rate in polymorphic yield (validated/polymorphic loci), our results suggest that a complete validation of the initial 783 and 446 sequence variants in both species would generate an increase in the number of SNPs to 40 and 80 respectively for C. fluminea and D. polymorpha.
Those positions with a heterozygous condition in all the first analyses (23 for D. polymorpha and 18 for C. fluminea) were genotyped in 96 individuals of each species, and all of them displayed the same result.This pattern may be explained by paralogous sequence variants (PSV) (Smith et al., 2005).PSV are a common trait of the genomes of mollusks because of the highly abundant cryptic repetitive genomic DNA (McInerney et al., 2011).
All the validated SNPs (five in D. polymorpha and four in C. fluminea) and their flanking regions are available in GenBank (accession numbers KT220181-85 for D. polymorpha, and KT220186 and KT220188-90 for C. fluminea, table 1).They are the first validated SNPs in the two species and they add to the limited number of currently available loci (Peñarrubia et al., 2015a(Peñarrubia et al., , 2015b)).They can thus be used in further studies with alternative genotyping techniques.Interestingly, even such a limited number of SNPs can be useful to describe population structure.In the invasive mosquito fish (Gambusia holbrooki), 5 SNPs were used to further characterize the genetic structure in European populations (Vidal et al., 2012), and in swordfish (Xiphias gladius) the same number of SNPs was able to detect genetic differentiation in Atlantic and Mediterranean samples (Smith et al., 2013).