Assessing SNP genotyping of noninvasively collected wildlife samples using microfluidic arrays

Noninvasively collected samples are a common source of DNA in wildlife genetic studies. Currently, single nucleotide polymorphism (SNP) genotyping using microfluidic arrays is emerging as an easy-to-use and cost-effective methodology. Here we assessed the performance of microfluidic SNP arrays in genotyping noninvasive samples from grey wolves, European wildcats and brown bears, and we compared results with traditional microsatellite genotyping. We successfully SNP-genotyped 87%, 80% and 97% of the wolf, cat and bear samples, respectively. Genotype recovery was higher based on SNPs, while both marker types identified the same individuals and provided almost identical estimates of pairwise differentiation. We found that samples for which all SNP loci were scored had no disagreements across the three replicates (except one locus in a wolf sample). Thus, we argue that call rate (amplification success) can be used as a proxy for genotype quality, allowing the reduction of replication effort when call rate is high. Furthermore, we used cycle threshold values of real-time PCR to guide the choice of protocols for SNP amplification. Finally, we provide general guidelines for successful SNP genotyping of degraded DNA using microfluidic technology.


Supplementary Table S6
Pairwise F ST values for European wildcats and domestic cats with microsatellite data (above the diagonal, n = 24 samples and 14 loci) and SNP data (below the diagonal, n = 35 samples and 65 loci). Note that the SNP panel we used here was designed to detect hybridization of wildcats and domestic cats (that is, maximize differentiation). All samples were collected in Germany. Potential hybrids (based on SNP data) were excluded from these analyses (n = 1 for msats, n = 2 for SNPs). Probability values were based on 999 permutations; ***p ≤ 0.001.

Supplementary Table S7
Pairwise F ST values for brown bears with microsatellite data (above the diagonal, n = 55 samples and 18 loci) and SNP data (below the diagonal, n = 55 samples and 69 loci). All samples were collected in Greece. Only groups with n > 5 were considered (16 bears from 5 locations excluded). Probability values were based on 999 permutations; n.s., not significant; *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001. Negative values were converted to zero. results for the most likely K (***), second most likely K (**) and third most likely K (*) as calculated with the Evanno method based on SNP and microsatellite genotypes and their combination. Colour-coded bars below the STRUCTURE plots correspond to the sample groupings based on sampling region (grey wolves, brown bears) or species identification (wildcats or domestic cats, based on SNP data). Figure S4 STRUCTURE plots showing results for the most likely K (***), second most likely K (**) and third most likely K (*) (upper panels) as calculated with the Evanno method (lower panels, respectively) based on SNP and microsatellite data sets. Colour-coded bars below the STRUCTURE plots correspond to the sample groupings based on sampling region (grey wolves, brown bears) or species identification (wildcats and domestic cats, based on SNP data).

Supplementary Figure S5
PCoA for wolves and wildcats showing outliers (SNP data, original data set). Further examination of these samples indicated low SNP call rates (wolf 71%; wildcats 77%, 78%). These samples were removed from the figure in the main text, Figure 3, and from further analyses.

Supplementary Figure S6
PCoA analyses for subsets of SNP and microsatellite markers used in this study to genotype grey wolves. Each point represents an individual's genotype, colour-coded to its sampling region. Subsets of loci were selected based on highest heterozygosity (H E ) for each locus.

Supplementary Figure S7
PCoA analyses for subsets of SNP markers used in this study to genotype grey wolves. Each point represents an individual's genotype, colour-coded to its sampling region. Subsets of loci were selected randomly; three times each case (a, b, c).

Supplementary Figure S8
PCoA analyses for subsets of SNP and microsatellite markers used in this study to genotype European wildcats, domestic cats and hybrids. Each point represents an individual's genotype, colour-coded to species identity. Subsets of loci were selected based on highest heterozygosity (H E ) for each locus.

Supplementary Figure S9
PCoA analyses for subsets of SNP and microsatellite markers used in this study to genotype European wildcats, domestic cats and hybrids. Each point represents an individual's genotype, colour-coded to species identity. Subsets of loci were selected based on highest F ST for each locus.

Supplementary Figure S10
PCoA analyses for subsets of SNP markers used in this study to genotype European wildcats, domestic cats and hybrids. Each point represents an individual's genotype, colour-coded to species identity. Subsets of loci were selected randomly; three times each case (a, b, c).

Supplementary Figure S11
PCoA analyses for subsets of SNP and microsatellite markers used in this study to genotype brown bears. Each point represents an individual's genotype, colour-coded to its sampling region. Subsets of loci were selected based on highest heterozygosity (H E ) for each locus.

Supplementary Figure S12
PCoA analyses for subsets of SNP markers used in this study to genotype brown bears. Each point represents an individual's genotype, colour-coded to its sampling region. Subsets of loci were selected randomly; three times each case (a, b, c).

Mitochondrial DNA sequencing
Brown bear hair samples were checked macroscopically for species identification, in order to avoid wild boar hairs. Grey wolf scats and cat hairs were checked for species identity using mtDNA sequencing in order to avoid samples from other species (mainly, fox, dog or domestic cat). PCR  Table S1), 0.2 µl of Taq DNA polymerase (5 U/µl) (New England BioLabs) and 6.1 µl of molecular grade water. PCRs were performed in a T1 plus Thermocycler (Biometra). Initial denaturation was at 95 °C for 3 min, followed by 35 cycles of 94 °C for 30 s, 54 °C for 30 s, and 72 °C for 1 min and a final extension at 72 °C for 10 min. PCR products were purified with 2 µl Exonuclease I and FastAP™ Thermosensitive Alkaline Phosphatase mixture (1:2; Thermo Scientific) at 37 °C for 15 min, followed by 80 °C for 15 min and diluted 1:20 (scats) or 1:40 (hairs). Sequencing was performed using the BigDye Terminator 3.1 Cycle Sequencing Kit (Applied Biosciences) using a cycling protocol which involved an initial denaturation step at 95 °C for 60 s, followed by 30 cycles of 10 s at 96 °C, 10 s at 50 °C and 2 min at 60 °C. The products were purified using ABI-XTerminator beads (Applied Biosystems) and separated on an ABI 3730 DNA Analyzer (Applied Biosystems). Sequences of wolves and wildcats were aligned with Geneious v7.1.8 1 and aligned to our laboratory reference samples to identify haplotypes.

Microsatellite genotyping
Unlinked autosomal microsatellite data for grey wolves and European wildcats were obtained as part of the regular genetic monitoring conducted in our laboratory. Brown bear microsatellite genotyping data was obtained from collaborators in Greece 5 . The markers and laboratory procedures are described elsewhere (wolves, 6 ; wildcats, 7 ; brown bears, 8 ). Briefly, a multiple-tubes approach was applied for wolves and wildcats, as is common practice for noninvasive samples, including three  Figure S1 Overview of numbers of samples and loci included in each analysis based on quality criteria.

Supplementary Table S3
Overview of individuals that were represented by multiple samples in SNP and microsatellite data sets. One mismatch at one locus was accepted to consider two genotypes as belonging to the same individual. Note that brown bear samples had been individualized using microsatellites in the course of a previous study and consequently no matching genotypes were found in this study. n, number of samples; f.a., sample failed in microsatellite amplification; NA, matching not available due to failed microsatellite amplification of one of the samples.

Supplementary Table S5
Pairwise F ST values for grey wolves with microsatellite data (above the diagonal, n = 30 samples and 13 loci) and SNP data (below the diagonal, n = 35 samples and 85 loci). All samples were collected in Germany. Only groups with n > 5 were considered (5 wolves from 2 locations excluded). Probability values were based on 999 permutations; *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001.

Supplementary Table S6
Pairwise F ST values for European wildcats and domestic cats with microsatellite data (above the diagonal, n = 24 samples and 14 loci) and SNP data (below the diagonal, n = 35 samples and 65 loci). Note that the SNP panel we used here was designed to detect hybridization of wildcats and domestic cats (that is, maximize differentiation). All samples were collected in Germany. Potential hybrids (based on SNP data) were excluded from these analyses (n = 1 for msats, n = 2 for SNPs). Probability values were based on 999 permutations; ***p ≤ 0.001.

Supplementary Table S7
Pairwise F ST values for brown bears with microsatellite data (above the diagonal, n = 55 samples and 18 loci) and SNP data (below the diagonal, n = 55 samples and 69 loci). All samples were collected in Greece. Only groups with n > 5 were considered (16 bears from 5 locations excluded). Probability values were based on 999 permutations; n.s., not significant; *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001. Negative values were converted to zero. PCoA analyses for subsets of SNP markers used in this study to genotype grey wolves. Each point represents an individual's genotype, colour-coded to its sampling region. Subsets of loci were selected randomly; three times each case (a, b, c).

Supplementary Figure S12
PCoA analyses for subsets of SNP markers used in this study to genotype brown bears. Each point represents an individual's genotype, colour-coded to its sampling region. Subsets of loci were selected randomly; three times each case (a, b, c).