Discriminating Among Pacific Salmon, Rainbow Trout, and Atlantic Salmon Species Using Common Genetic Screening Methods Free

The five most common species of Pacific salmon and Rainbow Trout (steelhead) Oncorhynchus spp. intermingle in the North Pacific Ocean and its freshwater tributaries (Groot and Margolis 1991; Myers et al. 1996). Atlantic Salmon Salmo salar escapees from net pen aquaculture operations are also making their way into the North Pacific Ocean and its tributaries (Morton and Volpe 2002; Fisher et al. 2014). Identifying species of individual fish is usually accomplished quickly and efficiently in the field using morphological characters (Steelquist 1992; Pollard and Hartman 1997). However, there are situations where morphology provides equivocal results due to condition of the fish, life history stage, or training of the observer, or is simply not possible because only part of a fish is available.

In situations where morphology is not adequate to provide speciation, researchers have widely used genetic tests to distinguish among Pacific and Atlantic salmon species for multiple applications using a variety of methods. Applications include identification of bone fragments at archaeological sites (e.g., Yang et al. 2004; Grier et al. 2013), identification of prey items (e.g., Purcell et al. 2004), verification of seafood authenticity (e.g., Hellberg et al. 2016; Stern et al. 2017), and forensics for prosecution of illegal fishing (e.g., Withler et al. 2004). Methods include restriction fragment length polymorphism analysis (e.g., Withler et al. 1997; Russell et al. 2000; Hold et al. 2001), polymerase chain reaction (PCR) of the growth hormone gene (McKay et al. 1997), Sanger sequencing (e.g., Grier et al. 2013), DNA barcodes (e.g., Stern et al. 2017), real-time quantitative PCR (Feng et al. 2017), and single-nucleotide polymorphism (SNP) microarrays (e.g., Wenne et al. 2016).

These genetic methods have proven successful at distinguishing among salmon species, but they are not easily incorporated into fish and wildlife population monitoring analysis workflows currently used by laboratories for Pacific salmon investigations. Most of these laboratories use SNPs assayed through either 5′–3′ exonuclease (TaqMan; Livak 1999) assays or through genotyping-by-sequencing (GBS) methods (Seeb et al. 2009; Seeb et al. 2011; Campbell et al. 2015). These laboratories would benefit from methods to differentiate among salmon species that could be seamlessly incorporated into laboratory workflows. For example, incorporation of a genetic species-identification tool into a fish and wildlife population monitoring analysis workflow could be useful for evaluating the presence of nontarget species in mixture samples. This would be especially useful in cases where field species-identification–based morphology or other external characters are equivocal (e.g., juvenile collections or ocean-bright fish) or the species is unusual (e.g., Atlantic Salmon escapees or species captured outside of native range).

Here we provide methods that fisheries biologists can easily streamline into fish and wildlife population monitoring analyses, such as mixed-stock analysis applications, or run separately with just a few markers. We employ a novel method for using TaqMan assays to discriminate among species by using the scatter plot distribution to glean information beyond the variation at the targeted SNP. This method reduces the number of markers required to discriminate among species and can be assimilated into SNP analyses used commonly for describing genetic relationships among populations and mixed-stock analysis applications in Pacific salmon.

We also explore the basis for the variation in the observed scatter plot distributions (variation in florescent signals) and show that this variation is due to nucleotide diversity in and near the probe site. Because the SNPs underlying the assays developed here are all physically close to one another along the mitochondrial genome, the potential exists to develop a single DNA sequence–based assay to discriminate among salmon species. This single assay can be added to a GBS panel, such as a Genotyping in the Thousands panel (Campbell et al. 2015), to identify and exclude nontarget species from analyses.

We selected tissue samples from the five most common species of Pacific salmon, Rainbow Trout (representing steelhead) O. mykiss, and Atlantic Salmon from the archives of the Alaska Department of Fish and Game Gene Conservation Laboratory. We subsampled tissues from Chinook Salmon Oncorhynchus tshawytscha, Sockeye Salmon Oncorhynchus nerka, and Chum Salmon Oncorhynchus keta from four sites each that encompassed the full Pacific Rim range (Table 1). We subsampled tissues from Coho Salmon Oncorhynchus kisutch that we selected from three sites that encompassed the eastern Pacific Ocean range. We subsampled tissues from Pink Salmon Oncorhynchus gorbuscha, Atlantic Salmon, and Rainbow Trout that we selected from one or two sites from a small geographical range. We included 13 to 14 individuals from each species for a combined total of 95 individuals.

We extracted genomic DNA from tissue samples using a DNeasy^® 96 Blood and Tissue Kit (Qiagen). Development of TaqMan assays followed the methods described by Smith et al. (2005). Briefly, we aligned all mitochondrial DNA sequences available in GenBank for the species of interest (as of December 2005) using SeqMan software (DNASTAR), and submitted the consensus sequence and informative SNP positions to Assays-by-Design (Applied Biosystems). This resulted in five functional TaqMan assays distributed across four mitochondrial regions: the NADH dehydrogenase subunit 3 gene, the 16S ribosomal RNA gene, cytochrome c oxidase III, and cytochrome b (Table 2). We used these assays to genotype the extracted genomic DNA following a protocol developed by the manufacturer and commonly applied in fisheries laboratories employing this technology (e.g., Von Bargen et al. 2015). We performed each PCR in 384-well plates in a 5-μL volume consisting of 6–40 ng/μL of DNA, 2× TaqMan^® GTXpress™ Master Mix (Applied Biosystems), and Custom TaqMan SNP Genotyping Assay (Applied Biosystems). We performed thermal cycling on a dual 384-well GeneAmp^® PCR System 9700 (Applied Biosystems) as follows: an initial “hot-start” denaturation of 95°C for 10 min followed by 50 cycles of denaturation at 92°C for 1 s and annealing at 60°C for 1 min, with a final “cool-down” hold at 10°C. The plates were scanned after amplification and analyzed using the Life Technologies QuantStudio™ 12K Flex (Applied Biosystems).

The process described above resulted in two measurements of florescence for every sample for every assay, one measurement corresponding to each of the two alleles at the SNP underlying the assay. We viewed these data as scatter plots, with each point representing an individual fish, and the x-axis and y-axis representing the magnitude of florescence corresponding to each allele. We considered assays resulting in scatter plots with a single, distinct cluster for one of our species useful for distinguishing that species. We determined the clustering by selecting each species separately within the software and noting where the data points fell on the scatter plots. After we tracked all species, we generated a figure with annotations to designate which clusters represented which species.

We quantified extracted DNA using a NanoDrop^® ND-1000 spectrophotometer (NanoDrop 130 Technologies) and normalized to 20 ng/μL. We used PCR to isolate and amplify the sequence targeted by TaqMan assays discussed above using a very low concentration of nonlabeled M13 (Messing 1983) tailed oligonucleotide primers to facilitate Sanger sequencing without PCR cleanup. Sequences and names for these forward and reverse tailed primers were: TGTAAAACGACGGCCAGT [M13F(-21)] and CAGGAAACAGCTATGAC [M13R(-27)], respectively. We used the tailed primers to increase throughput and sequencing length of the loci to allow for forward and reverse read assembly of the relatively small amplicons (<200 base pairs) targeted in this study. Reaction conditions for each locus consisted of approximately 40 ng template DNA, 1× Taq PCR buffer (100 mM KCl, 100 mM Tris-HCl, pH 9.0), 0.2 mM dNTPs, 1.5 mM MgCl₂, 0.01 μM each primer (Table 2), and 0.5 U Taq DNA polymerase (New England Biolabs) in a final volume of 10 μL. Thermal cycling conditions consisted of 94°C for 1 min; 5 cycles of 94°C for 30 s, 50°C for 30 s, and 72°C for 1 min; followed by 35 cycles of 94°C for 30 s, 54°C for 30 s, and 72°C for 1 min; with a final 72°C for a 10-min extension.

We sequenced the PCR products using the BigDye^® Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems). Forward and reverse sequencing reaction mixtures contained 0.75× BigDye Buffer, 0.75 μL BDX64 (BigDye^® enhancing buffer, MCLAB), 0.32 μM primer, 0.25 μL BigDye v3.1 (Applied Biosystems), and 1.5 μL of PCR product in a final volume of 10 μL. We used the primer M13F (-21) to generate forward sequences and primer M13R (-27) to generate reverse sequences. We purified cycle sequencing reactions with the Montage SEQ96 Sequencing Reaction Cleanup Kit (EMD Millipore) and sequenced them on an ABI 3730xl Genetic Analyzer (Applied Biosystems).

We visually examined the sequences examined and manually edited them in Geneious (v10.1; Kearse et al. 2012), and removed low-quality sequences that did not produce clear electropherograms. We assembled the forward and reverse reads of each sample in Geneious (when possible) to produce consensus sequences at each locus for each individual. Finally, we aligned sequences, primers, and probes for each locus in BioEdit using the ClustalW module, and identified and examined variable sites and to determine if they fell within the primer or probe sequences. We did not report SNPs that were only found within one individual because these SNPs could have been caused by sequencing errors or poor sequencing quality and are not useful for species identification even if they represent real variation.

Variation in priming regions and in sequence surrounding the probe-binding site can impact amplification and binding efficiency, and thus fluorescence. For each fish, we compared DNA sequence information for the TaqMan amplicon to the location of the individual on the TaqMan scatter plot to determine what sequence variation might account for the scatter plot variation. If the location on the TaqMan scatter plot was consistent among individuals within a species, but different from individuals from another species, then we judged the TaqMan assay to be useful in distinguishing between these two species. The one caveat to this rule was that if the location on the TaqMan scatter plot was indistinguishable from a failed amplification for either species, then we judged the TaqMan assay as not useful for species identification because failed amplification can occur with any species. Consistent sequence variation among species in the TaqMan amplicon can be useful for species identification using GBS methods, regardless of whether this variation resulted in differences in the TaqMan scatter plot. We evaluated the sequenced markers to determine which ones could be used individually in a GBS application to differentiate among the five most common species of Pacific salmon, Rainbow Trout, and Atlantic Salmon.

All 95 samples successfully amplified for all five TaqMan assays. The single nontemplate control was at the origin on all five scatter plots indicating that significant contamination was not present. Sequencing was difficult with the short amplicons derived from the TaqMan assays. The addition of common universal DNA sequencing primer (M13) tails increased the amplification success. All assays worked well for the forward sequence, the reverse sequence, or both. We observed dropout of some species for certain assays (see below). We have provided consensus amplicons for each species for each assay (Table S1, Supplemental Material) and FASTA files (a common text format used to describe genetic data; Lipman and Pearson 1985) for each assay (Text S1–S5, Supplemental Material). Finally, we provided a description of the results for each assay below, including a summary of variation in primer and probe regions (Figure 1).

OKI1-OKI

We produced sequences in at least one direction for 94 of the 95 individuals. Sequence in only the reverse direction was available for six individuals. We produced consensus sequences (successful matching sequences in both directions) for 88 of the 95 individuals. The sequence was 113 bp long and contained four variable sites (Figure 1; Table S1, Supplemental Material). The probe extended from position 58 for the probe targeting the “T” allele or 59 for the probe targeting the “C” allele to position 76. The target TaqMan SNP was at position 67 (C/T). We found all variable sites in the region overlapped by the probe. The probe encompassed four SNPs (66, 67, 74, and 75). The reverse primer extended from position 71 to position 95 and overlapped with the probe for SNP sites at positions 74 and 75. There were no SNPs in the forward primer positions.

We observed four clusters in the TaqMan scatter plot (Figure 2). Coho Salmon formed a monospecific cluster and were diagnostic at position 67 (C; Figure 1). Chum Salmon formed a monospecific cluster and were diagnostic at position 66 (G). However, the Chum Salmon cluster was at the point of origin on the scatter plot. A failed PCR could result in a false positive identification as a Chum Salmon if this was the only assay performed. Chinook Salmon and Rainbow Trout formed a bispecific cluster and shared a haplotype between positions 66 and 74 (TTCG) that was unique to these two species. The other three species formed one loose cluster. Inspection of this cluster indicated some segregation but at levels that were not deemed adequate for species determination. Sequence variation at position 74 may explain some of the cluster segregation (Atlantic Salmon were fixed for the “A” allele, Sockeye Salmon were fixed for the “T” allele, and Pink Salmon had both “A” and “T” alleles). However, individual Pink Salmon did not cluster with the species that shared their allele (i.e., Pink Salmon with a “T” allele did not cluster with Sockeye Salmon, and Pink Salmon with an “A” allele did not cluster with Atlantic Salmon).

SSA2-SSA

We produced sequences in at least one direction for 77 of the 95 individuals and consensus sequences for 58 of the 95 individuals. We found most of the dropouts in Coho Salmon, Chinook Salmon, and Sockeye Salmon, with fewer dropouts observed in Rainbow Trout and Pink Salmon. The sequence was 198 bp long and contained 32 variable sites (Figure 1; Table S1, Supplemental Material). The probe extended across positions 125–141. The target TaqMan SNP was at position 133 (C/T). The probe encompassed three SNPs (132, 133, and 138). There were no SNPs in the forward or reverse primer positions.

We observed four clusters in the TaqMan scatter plot (Figure 3). Atlantic Salmon formed a monospecific cluster and were diagnostic at positions 132 and 133 haplotype (CC; Figure 1). Rainbow Trout formed a monospecific cluster and were diagnostic at positions 133 and 138 haplotype (TC). Sockeye Salmon formed a monospecific cluster and were diagnostic at positions 132 and 133 haplotype (AC). All the remaining species (Chinook, Chum, Coho, and Pink salmon) clustered together and had the same haplotype at the probe site (CTT).

OKESSA1-OKE

We produced sequences in at least one direction for all individuals and consensus sequences for 85 of the 95 individuals. The sequence was 174 bp long and contained 24 variable sites (Figure 1; Table S1, Supplemental Material). The probe extended across positions 49–65. The target TaqMan SNP was at position 59 (C/T). The probe encompassed three SNPs (50, 59, and 63). There were no SNPs in the forward or reverse primer positions.

We observed four clusters in the TaqMan scatter plot (Figure 4). Chinook Salmon formed a monospecific cluster and were diagnostic at position 50 (C; Figure 1). Sockeye Salmon formed a monospecific cluster and were diagnostic at position 50 (G). However, the Sockeye Salmon cluster was at the point of origin on the scatter plot. A failed PCR reaction could result in a false positive identification as Sockeye Salmon if this was the only assay performed. Atlantic Salmon and Chum Salmon formed a bispecific cluster and shared the same haplotype throughout the probe site (ACC). Within this cluster, there was an indication of clustering by species, but we did not deem it adequate for reliable differentiation among the two species using the scatter plot. These two species had a different haplotype from the other species at position 68 (T, which is just outside the probe site), which could explain this difference in scatter plot migration. Coho Salmon, Pink Salmon, and Rainbow Trout made up the last cluster. Ignoring the within-species SNP for Coho Salmon, all of these species had the same haplotype at the probe site (ATC).

ONEOGO1-ONE

We produced sequences in at least one direction for 92 of the 95 individuals and consensus sequences for 64 of the 95 individuals. The sequence was 141 bp long and contained 10 variable sites (Figure 1; Table S1, Supplemental Material). The probe extended across positions 52–68. The target TaqMan SNP was at position 57 (A/G). The probe encompassed two SNPs (57 and 60). There were two SNPs in the probe region. An indel was found at position 71. There were no SNPs in the forward or reverse primer positions.

We observed four clusters in the TaqMan scatter plot (Figure 5). Pink Salmon formed a monospecific cluster and were diagnostic for haplotype for positions 57 and 60 (GT; Figure 1). Sockeye Salmon formed a monospecific cluster and were diagnostic for haplotype for positions 57 and 60 (GC). Atlantic Salmon formed a more dispersed cluster and were diagnostic at position 84 (T). Chinook, Chum, and Coho Salmon and Rainbow Trout make up the last cluster and all had the same haplotype for positions 57 and 60 (AT).

OTSOKI1-OKI

We produced consensus sequences for 94 of the 95 individuals. We sequenced the remaining individual only in the reverse direction. The sequence was 180 bp long and contained 19 variable sites (Figure 1; Table S1, Supplemental Material). The probe extended across positions 52–66. The target TaqMan SNP was at position 57 (T/C). The probe encompassed four SNPs (54, 57, 60, and 66). There were no SNPs in the forward or reverse primer positions.

We observed five clusters in the TaqMan scatter plot (Figure 6). Pink Salmon formed a monospecific cluster and were diagnostic at the haplotype for positions 54, 57, 60, and 66 (CTTA; Figure 1). Atlantic Salmon formed a monospecific cluster and were diagnostic at position 54 (T). Chum Salmon formed a monospecific cluster and were diagnostic at the haplotype for positions 54, 57, 60, and 66 (CCTA). Rainbow Trout and Sockeye Salmon formed a bispecific cluster and shared the same haplotype throughout the probe site (CTCA). Within this cluster, there was an indication of separation by species, but we did not deem the separation adequate for identification of the two species using the scatter plots. Rainbow Trout and Sockeye Salmon had a different haplotype at position 84 (which is 18 bases outside the probe site), which might explain the difference in scatter plot migration. Chinook and Coho Salmon formed a bispecific cluster and, with one exception, shared the same haplotype throughout the probe site (CCCA). The one exception was for Chinook Salmon where one fish had a G SNP in the last position (66), but this individual remained in the same cluster as the other Chinook Salmon.

Four of the five markers examined allowed determination of species among the five most common species of Pacific salmon, Rainbow Trout, and Atlantic Salmon with at least two redundant signals for each species pair. The four markers with multiple redundancies were SSA2-SSA, OKESSA1-OKE, ONEOGO1-ONE, and OTSOKI1-OKI. If redundancy wasn't required, species identification could be accomplished using any one of the following four marker pairs:

1. OKI1-OKI and SSA2-SSA

2. OKI1-OKI and OKESSAI-OKE

3. OKI1-OKI and OTSOKI1-OKI

4. OKESSAI-OKE and OTSOKI1-OKI

We identified four marker pair combinations of the five TaqMan assays that could discriminate among the five most common species of Pacific salmon, Rainbow Trout, and Atlantic Salmon. This redundancy allows researchers to either select a pair for a specific question (e.g., it is known the samples are either Coho Salmon or Chinook Salmon), or screen multiple markers to increase confidence in species discrimination. Additional markers may be advisable in cases where within-species variation might exist but was not represented in our reference samples.

Oligonucleotides, such as primers and probes used in an assay, will rarely, if ever, account for the full range of biological variation that exists in nature. The accuracy of inferences made based on any assay can thus only be inferred in terms of a specific reference sample. In the context of the current assays, we feel that the samples analyzed here make us confident of the use of these assays to infer species identity in fisheries samples from Alaska. We recommend further examination of how well the variation we observed in our samples represents the total variation present in the species those samples are being used to represent. This is of utmost importance for those species where our samples originated from a limited range (e.g., Coho Salmon, Pink Salmon, Atlantic Salmon, and Rainbow Trout). In any case, consideration of a genetic assay as “diagnostic” requires an assumption that the samples used to develop and test the assay are representative of existing variation in both groups for populations of interest. This assumption is unlikely to ever be strictly met in the context of natural populations. We recommend that these assays be thoroughly tested in the specific populations of interest (and sample-tissue quality) prior to being applied as a predictive tool on samples that might have originated from those populations.

Individual samples may not fluoresce (i.e., appear near the origin) either because the DNA quality or quantity were low, or because the sequences near the probe or priming sites were too divergent for binding of the probes or primers. Because we cannot distinguish between these two causes, markers that produced clusters near the origin in some species (Chum Salmon for OKI1-OKI, Sockeye Salmon for SSA2-SSA and OKESSA1-OKE) should not be considered useful for discriminating between these species and others.

The sequence data were useful in providing explanations for the cluster variation observed in the TaqMan assays. Although we cannot be certain that the sequence variation is the cause of the cluster variation, there are reasons why this is likely the case. First, we examined variation at or near the probe site as the sequence variation that is most likely to affect clustering in the scatter plot. Second, the variation in the probe area or near the probe area explains the scatter plot patterns for all markers.

Based on the interspecies variation observed, four of the five markers could be used to differentiate these species as part of a GBS panel (SSA2-SSA, OKESSA1-OKE, ONEOGO1-ONE, and OTSOKI1-OKI). During data analysis, researchers should identify nontargeted species and eliminated them before analyses begin on the target species. This step would not only eliminate nontarget species from the analysis, but would also allow the researcher to understand why the sample was eliminated from analysis by providing determination of the incorrect species. Because GBS is based on sequence reads whereas TaqMan assays are based on probe binding, a GBS approach could provide additional information (e.g., sequence differences in regions surrounding the probe-binding sites). If these markers are used for GBS applications, the redundant diagnostic sites should provide highly robust species determinations even with the use of as few as one or two of these markers. These redundancies would provide a measure of confidence, especially for individuals that originate from outside the geographic areas covered in this analysis. Additionally, redundancies would allow for the detection of lower-frequency alleles that are shared among species.

Here we provide methods to discriminate among the five most common species of Pacific salmon, Rainbow Trout (steelhead), and Atlantic Salmon that researchers can seamlessly add to standard mixed-stock analysis methods. Researchers can integrate these methods into mixed-stock analysis programs that use TaqMan assays or GBS methods. By adding these methods into existing programs, species identification can be obtained cost effectively.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Table S1. Mitochondrial DNA haplotype for each sequence position for five assays used for differentiating species across five species of Pacific salmon (Chinook Salmon Oncorhynchus tshawytscha, Sockeye Salmon Oncorhynchus nerka, Coho Salmon Oncorhynchus kisutch, Chum Salmon Oncorhynchus keta, Pink Salmon Oncorhynchus gorbuscha), Rainbow Trout Oncorhynchus mykiss (representing steelhead), and Atlantic Salmon Salmo salar. Haplotypes shared between the five Pacific salmon species and Rainbow Trout, and Atlantic salmon are shown as “*,” variation within species are shown with both alleles divided with a slash, indels are show with an allele and a “-”.

Found at DOI: https://doi.org/10.3996/052018-JFWM-038.S1 (30 KB XLSX).

Text S1. FASTA file (a common text format used to describe genetic data; Lipman and Pearson 1985) for assay OKESSA1-OKE. The first eight lines contain the sequences for the forward and reverse primers and probes for the assay. Following the assay information are alternating lines, with the first line identifying the individual salmon and the second line providing the consensus sequence for that individual. A string of characters identifies the individual salmon; the first character(s) identify the species: K = Chinook Salmon Oncorhynchus tshawytscha; S = Sockeye Salmon Oncorhynchus nerka; CO = Coho Salmon Oncorhynchus kisutch; CH = Chum Salmon Oncorhynchus keta; P = Pink Salmon Oncorhynchus. gorbuscha; RT = Rainbow Trout Oncorhynchus mykiss (representing steelhead); and AT = Atlantic Salmon Salmo salar. Following the species identifier is a code for the sampling location and year (see Table 1) followed by the individual number.

Found at DOI: https://doi.org/10.3996/052018-JFWM-038.S2 (19 KB TXT).

Text S2. FASTA file (a common text format used to describe genetic data; Lipman and Pearson 1985) for assay OKI1-OKI. The first eight lines contain the sequences for the forward and reverse primers and probes for the assay. Following the assay information are alternating lines, with the first line identifying the individual salmon and the second line providing the consensus sequence for that individual. A string of characters identifies the individual salmon; the first character(s) identify the species: K = Chinook Salmon Oncorhynchus tshawytscha; S = Sockeye Salmon Oncorhynchus nerka; CO = Coho Salmon Oncorhynchus kisutch; CH = Chum Salmon Oncorhynchus keta; P = Pink Salmon Oncorhynchus. gorbuscha; RT = Rainbow Trout Oncorhynchus mykiss (representing steelhead); and AT = Atlantic Salmon Salmo salar. Following the species identifier is a code for the sampling location and year (see Table 1) followed by the individual number.

Found at DOI: https://doi.org/10.3996/052018-JFWM-038.S3 (14 KB TXT).

Text S3. FASTA file (a common text format used to describe genetic data; Lipman and Pearson 1985) for assay ONEOGO1-ONE. The first eight lines contain the sequences for the forward and reverse primers and probes for the assay. Following the assay information are alternating lines, with the first line identifying the individual salmon and the second line providing the consensus sequence for that individual. A string of characters identifies the individual salmon; the first character(s) identify the species: K = Chinook Salmon Oncorhynchus tshawytscha; S = Sockeye Salmon O. nerka; CO = Coho Salmon O. kisutch; CH = Chum Salmon O. keta; P = Pink Salmon O. gorbuscha); RT = Rainbow Trout O. mykiss (representing steelhead); and AT = Atlantic Salmon Salmo salar. Following the species identifier, is a code for the sampling location and year (see Table 1) followed by the individual number.

Found at DOI: https://doi.org/10.3996/052018-JFWM-038.S4 (15 KB TXT).

Text S4. FASTA file (a common text format used to describe genetic data; Lipman and Pearson 1985) for assay OTSOKI1-OKI. The first eight lines contain the sequences for the forward and reverse primers and probes for the assay. Following the assay information are alternating lines, with the first line identifying the individual salmon and the second line providing the consensus sequence for that individual. A string of characters identifies the individual salmon; the first character(s) identify the species: K = Chinook Salmon Oncorhynchus tshawytscha; S = Sockeye Salmon Oncorhynchus nerka; CO = Coho Salmon Oncorhynchus kisutch; CH = Chum Salmon Oncorhynchus keta; P = Pink Salmon Oncorhynchus. gorbuscha; RT = Rainbow Trout Oncorhynchus mykiss (representing steelhead); and AT = Atlantic Salmon Salmo salar. Following the species identifier is a code for the sampling location and year (see Table 1) followed by the individual number.

Found at DOI: https://doi.org/10.3996/052018-JFWM-038.S5 (20 KB TXT).

Text S5. FASTA file (a common text format used to describe genetic data; Lipman and Pearson 1985) for assay SSA2-SSA. The first eight lines contain the sequences for the forward and reverse primers and probes for the assay. Following the assay information are alternating lines, with the first line identifying the individual salmon and the second line providing the consensus sequence for that individual. A string of characters identifies the individual salmon; the first character(s) identify the species: K = Chinook Salmon Oncorhynchus tshawytscha; S = Sockeye Salmon Oncorhynchus nerka; CO = Coho Salmon Oncorhynchus kisutch; CH = Chum Salmon Oncorhynchus keta; P = Pink Salmon Oncorhynchus. gorbuscha; RT = Rainbow Trout Oncorhynchus mykiss (representing steelhead); and AT = Atlantic Salmon Salmo salar. Following the species identifier, is a code for the sampling location and year (see Table 1) followed by the individual number.

Found at DOI: https://doi.org/10.3996/052018-JFWM-038.S6 (16 KB TXT).

We thank Jim and Lisa Seeb for coordinating international tissue exchanges and domestic tissue collections used in these analyses. We thank Nataly Varnavskaya for Russian samples, John Barr for Washington samples, Shunpei Sato for Japanese samples, and the many Alaska Department of Fish and Game staff from both divisions of Commercial and Sport fisheries who collected tissue samples from throughout the state. We thank the Federal Office of Subsistence Management project 04-507 and Alaska Sustainable Salmon Fund projects 45335 and 45420 for funding the collection of many of the samples used in this analysis and Alaska Energy Authority for funding sequencing. Finally, we thank Penny Crane and three anonymous reviewers for providing review and increasing the clarity of the manuscript.

Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.