Genome-wide characterization of upper Fraser River sockeye salmon identifies a region of major differentiation between run timing groups in the Stuart River Watershed

26 Early Stuart sockeye salmon Oncorhynchus nerka undergo one of the most intensive and far-reaching 27 migrations in the Fraser River ( British Columbia, Canada) that has likely exerted selective pressure s 28 shaping their genetics and physiology. To support upper Fraser River sockeye management objectives, 29 there is a need to further understand the unique genetics of these populations. Based on neutral genetic 30 markers, Early Stuart sockeye are considered a metapopulation with little to no genetic differentiation 31 among their approximately 50 spawning streams. In this study we make use of existing resources to 32 evaluate genome-wide genetic differentiation in this metapopulation and put it into the context of other 33 nearby populations. We confirm that genetic differentiation is low or negligible in the metapopulation 34 at a whole-genome level, although we identify a slight increase in differentiation at greater distances 35 within the watershed compared to more proximal streams (e.g., FST = 0.001-0.002 compared with 36 undifferentiated). Even the greatest differentiation detected is very low, being nine times lower than 37 that between run timing groups (Early Stuart vs. Stuart-Summer; e.g., FST = 0.0091). Notably, a region 38 of major differentiation was detected between the run timing groups on Chr18 between 56.3-58.0 Mbp 39 (average FST = 0.661). This genomic region appears to contain a unique haplotype specific to the Stuart-40 Summer, Nadina River, and Stellako River populations (i.e., upstream to the Nechako River). By 41 contrast, Early Stuart and other upper Fraser River sockeye outside of the Nechako Watershed do not 42 possess this haplotype. The genomic region is within ancestral chromosome 15.2, the homeolog to a 43 chromosome arm of Chr12 (i.e., 15.1); this arm of Chr12 is known to contain a major differentiated 44 region associated with run timing and life history variation in Alaska and the Okanagan. This finding 45 provides additional evidence for the importance of this genomic region in both homeologous 46 chromosomes within different populations of the same species, and points to the uniqueness of upper 47 Fraser River sockeye in this potential major effect locus. 48 49

Early Stuart; Fraser River; genomics; major effect locus; run timing; sockeye salmon 50 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 3

51 Genetic differentiation is a key metric to provide insight into population distinctiveness (Weir and 52 Cockerham 1984) and is the substrate for applications such as genetic stock identification (GSI). 53 Population differentiation can be evaluated using 10s of neutral microsatellite markers (e.g., Beacham 54 et al. 2004) or 100s of single nucleotide polymorphism (SNP) markers for example by amplicon panels 55 (Meek and Larson 2019) , 10,000s of SNPs through reduced representation sequencing (Baird et al. 56 2008; Peterson et al. 2012) or 100,000s to millions of SNPs through whole-genome resequencing 57 (Christensen et al. 2024; Bemmels et al. 2025). These approaches can recover similar trends, although 58 power can be influenced by population coverage and genomic coverage. Population coverage can 59 include collecting sufficient sample sizes to effectively represent the population, where more subtle 60 population structure can be identifiable at higher sample sizes (Araujo et al. 2014). Genomic coverage 61 is particularly critical when differentiation is specific to small regions of the genome, while the rest of 62 the genome remains in an undifferentiated state (e.g., Barry et al. 2024). These relatively small regions 63 with significantly increased differentiation relative to the rest of the genome may involve adaptive 64 variation (Tigano and Russello 2022) and/or structural variation such as inversions (Akopyan et al. 65 2025), which can prevent local recombination, retaining haplotypes of adaptive alleles (Wellenreuther 66 and Bernatchez 2018). 67 The Fraser River watershed is the largest sockeye salmon Oncorhynchus nerka complex in 68 British Columbia (BC), Canada, and has the greatest sockeye abundance in the world for any single 69 river system (COSEWIC 2017) . Over the past century, there have been significant declines in all 70 species of salmon in the Fraser River; based on genetic evidence, sockeye salmon have had particularly 71 drastic declines (Christensen et al. 2024). Multiple salmon species show similar trends in hierarchical 72 population genetic structure in the Fraser River , with the greatest differences observed among broad 73 regional groupings of lower, middle, and upper Fraser River (greatest between the lower and middle, 74 likely due to the significant migration barrier of the Fraser Canyon; Christensen et al. 2024) . Th e 75 overall population structure of sockeye salmon in the Fraser has been documented using multiple 76 different marker types, includin g microsatellites and SNPs (Beacham et al. 2004; Rondeau 2022) . 77 Subgroupings and structure among spawning locations is generally based on geographic proximity as 78 well as ecotype and run timing (Beacham et al. 2004; Beacham and Withler 2017) . The genetic 79 diversity of Fraser River salmon collectively represents the standing genetic variation for the species 80 in the region, and holds adaptive traits and genetic diversity that is irreplaceable on timescales that are 81 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 4 relevant to humans (Waples et al. 2001; COSEWIC 2017), critical for fisheries, culture, and resilience 82 to climate change. 83 Early Stuart sockeye spawn in the upper Stuart River watershed, a major sub-basin of the Fraser 84 River watershed in central BC that is comprised of two large lakes, Takla Lake and Trembleur Lake, 85 that are connected by the wide and shallow Middle River (Bradford and Braun 2021) . Early Stuart 86 sockeye are the sole representative of the Early Stuart run timing group, one of the four run timing 87 designations in the Fraser River (COSEWIC 2017). Early Stuart sockeye have the earliest upstream 88 migration of all Fraser River sockeye (Idler and Clemens 1959; Bradford and Braun 2021). Returning 89 adults enter the Fraser River in early July and quickly migrate to spawning streams in the Stuart system, 90 where peak spawning occurs in natal streams in mid -August (Bradford and Braun 2021) . 91 Approximately half of the Early Stuart run timing group passes Hell’s Gate by July 14 th, whereas the 92 early summer, summer, and late run timing groups correspond to Aug. 6 th, Aug. 17th, and Sept. 21 st, 93 respectively (COSEWIC 2017) . The Early Stuart sockeye and other neighbouring populations, 94 including the Stuart-Summer sockeye, must travel great distances and overcome significant river force 95 (Wright et al. 2025) to reach the spawning grounds . This migration is thought to have shaped Early 96 Stuart sockeye genetics and physiology through long-exerted evolutionary forces based on energetic 97 requirements and local environments of the route (Hinch and Rand 1998; Rand et al. 2006; Crossin et 98 al. 2004; Eliason et al. 2011). Early Stuart sockeye have smaller and more fusiform body shapes than 99 other deeper bodied morphologies that occur in sockeye populations with easier routes, higher somatic 100 energy density, fewer carried eggs and exhibit more efficient travel (Gilhousen 1980; reviewed by 101 Crossin et al. 2004) , are capable of higher aerobic scope and cardiac capacity (Crossin et al. 2004; 102 Eliason et al. 2011). 103 Early Stuart sockeye are considered to exist in a metapopulation, or a collection of streams that 104 all comprise one large population with some extent of gene flow expected between streams (Bradford 105 and Braun 2021). The sockeye populations that spawn most temporally and geographically proximal 106 to the Early Stuart sockeye are from the genetically distinct Stuart-summer run timing group (Beacham 107 et al. 2004; Rondeau 2022) . Of the over 50 Early Stuart spawning sites within the Stuart River 108 watershed (COSEWIC 2017), 36 streams have been characterized in detail by Bradford and Braun 109 (2021), who report that 12 streams contain good and stable habitat, are occupied in more than 90% of 110 evaluated years, and contain at least 85% of the spawners of the metapopulation. The other 24 streams 111 exhibit poorer habitat, are not persistently abundant with spawners, and are periodically recolonized 112 by dispersers from the 12 strong streams, which are therefore considered to be the key to the resiliency 113 of the system (Bradford and Braun 2021). Genetic characterization with high numbers of samples and 114 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 5 microsatellite or SNP markers have not identified significant population structure in the 115 metapopulation (Beacham et al. 2004; Rondeau 2022) , and so the current expectation for the 116 metapopulation is panmixia. However, this type of analysis has yet to be conducted at a genome-wide 117 level. Evaluating differentiation in the metapopulation at this scale will provide valuable information 118 to support management and conservation objectives. 119 The primary purpose of this work is to determine whether any neutral or putatively functional 120 genetic variation , including that within punctuated and relatively small genomic regions, are 121 differentiated between sockeye spawning in individual tributaries used by Early Stuart sockeye salmon. 122 A secondary objective is to determine whether unique characteristics exist in the genome of the Early 123 Stuart sockeye relative to Stuart-Summer or other nearby populations that can aid in our understanding 124 of differences between these populations. By using the recently published genome-wide SNP data for 125 Fraser River sockeye salmon (Christensen et al. 2024), we directly address these two objectives in the 126 aim to guide conservation and management of these important populations. 127 128

129 Samples and genotypes 130 Multilocus genotypes were obtained as a VCF file from a recent study on salmonids of the Fraser River 131 (see Data Availability; Christensen et al. 2024). The VCF file was subset with bcftools (Danecek et al. 132 2021) to retain sockeye collections from the mid- or upper-Fraser River (Table 1). Annotations were 133 added to the VCF file and SNPs were filtered to only retain those with minor allele frequency (MAF) 134 > 0.01. An additional dataset was generated for use in population genetic analyses by filtering for MAF 135 > 0.05 and removing SNPs in high linkage (maximum r2 = 0.5 in 50 kb windows) using bcftools. 136 137 Population genetics 138 The population genetic dataset (MAF > 0.05 and LD -filtered) was used as an input to a principal 139 components analysis (PCA) including all samples. Region-specific analyses were also conducted, with 140 datasets filtered for (1) all waters that flow into the Nechako River (i.e., Early Stuart, Stuart-Summer, 141 Nadina River, and Stellako River); (2) all waters that flow in to the Quesnel River (i.e., Quesnel and 142 Horsefly); and (3) all waters that flow into the Chilcotin River (i.e., Chilko Lake, Chilko River, Taseko 143 River). The region-specific analyses used PCA to identify general trends, and discriminant analysis of 144 principal components (DAPC) to identify loci contributing to the separation of the most differentiated 145 groups in each dataset. Loading values for these loci were plotted on sockeye salmon chromosomes 146 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 6 (GCF_034236695.1; Christensen et al. 2020) using fastman (Paria et al. 2022) . To investigate 147 substructure within the Early Stuart sockeye, a dataset was generated containing only the Early Stuart 148 collections, and a DAPC was conducted and loading values plotted across the chromosomes. Genetic 149 differentiation was evaluated using FST (Weir and Cockerham 1984) calculated as the average of all 150 per-locus FST values between specific contrasts calculated by VCFtools (Danecek et al. 2011). 151 152 Outlier region characterization 153 To explore a detected region of increased differentiation, the full dataset with all SNPs (MAF > 0.01) 154 was limited to only Chr18 and only Early Stuart and Stuart-Summer samples (i.e., the two genetically 155 most proximal populations that exhibited the elevated differentiation on Chr18. Monomorphic and low 156 MAF loci (MAF ≤ 0.01) in the subset dataset were removed and per-locus FST was calculated for all 157 loci on Chr18 using pegas (Paradis 2010) in R (R Core Team 2025) . To understand the geographic 158 context of the differentiated region, the section of the chromosome with FST > 0.5 was identified and 159 SNPs from this region for all individuals in the study were selected using bcftools and read into R 160 using vcfR (Knaus and Grünwald 2017). Genotypes were converted to allele dosage and plotted in a 161 heatmap in R, allowing dendrogram clustering of samples , but not SNPs, using heatmap function 162 defaults. 163 To inspect the genomic context of the differentiated region, the identified region and its 164 flanking sections were inspected for annotated genes in NCBI and for indications of inversions based 165 on manual inspections of short read alignments, respectively. The SNP positions from the genome used 166 by Christensen et al. (2024) were converted to the latest genome available in NCBI (i.e., 167 GCF_034236695.1; Christensen et al. 2020) using SNPlift (Normandeau et al. 2023) . The region of 168 interest was confirmed to be in a similar position based on the positions of the SNPs delimiting the 169 region, as well as re -running an FST analysis for the Stuart River system samples only with the 170 converted dataset. Inspections for structural variation was explored using two representative samples 171 from each haplotype group (i.e., unique haplotype to upstream to Nechako represented by Kuzkwa 172 River, and the haplotype more typical of the rest of the Fraser River from the Driftwood River). These 173 samples were obtained as fastq data from NCBI using SRA -toolkit. The samples were trimmed for 174 quality using cutadapt (Martin 2011), and aligned against Chr18 of the reference genome using bwa 175 mem (Li 2013). All alignments or discordant alignments were inspected using the Integrated Genomics 176 Viewer (IGV; Robinson et al. 2011) with a focus on the delimiters of the elevated differentiation region 177 for indications of inversions. 178 179 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 7

180 Population genetic structure in upper Fraser River sockeye salmon 181 Upper Fraser River sockeye salmon samples sourced from Christensen et al. (2024) that are included 182 in the present analysis are presented in Table 1 and include the following regions: Stuart (Early Stuart, 183 Stuart-Summer), Nadina/Francois, Francois/Fraser, Bowron, Quesnel and Horsefly, and Chil ko. For 184 each collection included, year of sampling was obtained from Christensen et al. (2024), and run timing 185 designations were sourced from Rondeau (2022). 186 Population genetic trends were explored using linkage filtered SNPs (n = 627,769 ; MAF > 187 0.05). A PCA clustering all samples by SNPs shows four main clusters in PC1 and PC2 (Figure 1A; 188 PVE = 3.6% and 1.8%, respectively) . These clusters include the populations using tributaries to the 189 Nechako River as natal habitat (i.e., both run timing groups from the Stuart River watershed, Nadina, 190 and Stellako), the Bowron River, Quesnel/Horsefly, and the Chilcotin Watershed (Figure 1A). 191 Populations from Quesnel, Horsefly, and Chilcotin regions have a similar position on PC1 but are 192 separated across PC2. A genetic dendrogram shows similar groupings, with the largest difference in 193 the data being between the populations upstream to the Nechako from the other collections (Figure 194 S1). The second grouping contains Bowron as the most external population to the grouping, followed 195 by Chilcotin region populations, and a grouping containing Quesnel and Horsefly populations, which 196 are grouped together but in separate branches (Figure S1). Some indication of unexpected clustering 197 due to low sample size is suggested by the dendrogram, with Felix (n = 5) and Takla (n = 4) collections 198 grouping closely and separately from other collections , as well as poor clustering of Middle River (n 199 = 7) and Pinchi (n = 4). Notably, Driftwood (n = 7) also clusters on the outside of the other Early Stuart 200 collections. 201 Within-region variation was inspected for populations upstream to the Nechako River (n = 96 202 samples, 12 collection sites, and 620,725 SNPs). A PCA groups all Early Stuart collections separately 203 from Stuart-Summer, Nadina, and Stellako on PC1 (PVE = 2.7% ; Figure 1B) . Nadina and Stellako 204 collections grouped generally with the Stuart -Summer for some samples along PC1, but other 205 individual samples within these collections cluster distantly on both PC1 and PC2. Applying a DAPC 206 separates three groups: (1) Early Stuart; (2) Stuart-Summer and Stellako; and (3) Nadina (Figure 2A). 207 Loci contributing the most to this separation were clustered in a region on Chr18, showing a peak that 208 is significantly elevated relative to the rest of the genome (Figure 2B; see Major differentiating region 209 on Chr18 below). 210 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 8 Populations upstream to the Chilcotin River (n = 45 samples, 611,274 SNPs) evaluated by PCA 211 found that all Chilko Lake and Chilko River samples (run timing: Summer) grouped together and 212 separately from the Taseko River collection (run timing: Early Summer ; Figure S2A ; PC1 PVE = 213 5.3%). These two groups were strongly separated by DAPC (Figure S2B) and the loci separating the 214 groups came from throughout the genome without any clear outlier regions (Figure S2C). 215 Populations upstream to the Quesnel River (n = 59 samples, 612,528 SNPs) evaluated by PCA 216 found overlap in groupings within some Quesnel collections, and separate clustering of the Horsefly 217 collections (i.e., Horsefly and McKinley; Figure S3A; PC1 PVE = 5.4%). A DAPC recovered similar 218 trends (Figure S3B) , and loci contributing to the separation o f Quesnel and Horsefly came from 219 throughout the genome, with no clear indication of outlier genomic regions (Figure S3C). 220 221 Major differentiating region on Chr18 222 The major differentiating region on Chr18 observed between the Early Stuart and Stuart -Summer, 223 Nadina and Stellako populations (Figure 2B) was explored further. The approximate genomic 224 coordinate boundaries of the outlier region , based on these populations, were identified as being 1.7 225 Mbp in size and spanning from 56,293,719 - 57,982,347 bp, a region with average (± s.d.) FST of 0.661 226 ± 0.133 (Figure 3A). 227 To understand the geographic context of the differentiated region, a heatmap was constructed 228 including the 10,563 SNPs within the genomic region (MAF > 0.01) from all samples in the study (n 229 = 210; Figure 3B). The heatmap shows a distinct haplotype with low heterozygosity specific to the 230 grouping of Stuart-Summer, Nadina, and Stellako, whereas the rest of the upper Fraser River 231 collections (including Early Stuart) do not have this haplotype. The top differentiated SNPs had FST 232 values of 0.975 and were at positions 57,673,269 bp, 57,712,980 bp, and 57,982,347 bp. These three 233 SNPs show near fixation between the more derived grouping ( i.e., Stuart-Summer, Nadina, Stellako) 234 and the rest of the samples (Additional File S1). For example, SNP Chr18:57,712,980 is a homozygous 235 alternate genotype (A/A) in 49 of 50 samples and a heterozygote in one sample (from Tachie) in the 236 Stuart-Summer/Stellako/Nadina grouping, and is homozygous reference (G/G) in 158 of 160 samples 237 and heterozygous in two samples (both from Horsefly River) in the rest of the upper Fraser River 238 populations analyzed. The other two SNPs show the same pattern, or exhibit three heterozygotes in the 239 Horsefly River (for SNP Chr18:57,712,980). Visualizing read alignments from representatives of each 240 grouping of the major differentiation region with a focus on discordant alignments in flanking genomic 241 regions did not provide any clear evidence of an inversion (data not shown). 242 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 9 The annotated gene content in the outlier region was inspected, finding 48 annotated putative 243 genes (RefSeq annotation; Additional File S2). Putative genes within this region included leucine-rich 244 repeat-containing protein 9-like (lrrc9-like; 56.33-56.36 Mbp), fibroblast growth factor receptor-like 245 1 (56.48-56.56 Mbp), and homeobox protein OTX2 (57.88-57.89), among others. 246 247 Metapopulation differentiation in Early Stuart sockeye 248 To determine whether the whole-genome data indicates genetic distinctiveness or separation between 249 tributaries within the Early Stuart metapopulation, a limited dataset with the four populations with the 250 greatest sample sizes was analyzed (i.e., between 7-10 samples per site; total n = 37 samples, 611,155 251 SNPs). The four collections were sourced from throughout the two-lake system, including Driftwood 252 River from the northern end of Takla Lake (n = 7), Dust Creek from the western arm of Takla Lake (n 253 = 10), Bivouac Creek from the southern end of Takla Lake (n = 10), and Paula Creek from the western 254 end of Trembleur Lake (n = 10). All of these samples were collected in 2005 (Table 1). 255 Many of the contrasts within this system showed 95% confidence intervals (CI) for FST that 256 overlapped zero, suggesting no discernable differentiation at a genome -wide scale between the 257 tributaries (Table 2). Two of the contrasts had very low, non-zero FST values: Driftwood River-Bivouac 258 Creek and Driftwood River-Paula Creek 95% CI FST of 0.0017-0.0021 (mean = 0.0019) and 0.0008-259 0.0012 (mean = 0.0010) , respectively. A PCA shows a general overlap of all ellipses around each 260 collection (Figure S4A), suggesting low to negligible FST. A DAPC recovers similar trends to the FST 261 contrasts, where Driftwood is projected furthest from Bivouac Creek and Paula Creek, although some 262 overlap exists between Bivouac Creek and Driftwood River (Figure 4). Therefore, the most distant 263 locations geographically ( i.e., Paula Creek or Bivouac Creek with Driftwood River) were the most 264 distant genetically, albeit with very low genetic difference. Notably, there were no genomic regions 265 showing elevated contributions to the separation of the Early Stuart collections by DAPC (Figure S4B), 266 suggesting equal differentiation or lack thereof throughout the genome. 267 To put the slight non-zero genetic distance observed between the Driftwood River and Paula 268 Creek into context, contrasts between an Early Stuart collection (Paula Creek) and populations from 269 outside the Early Stuart group were conducted. Paula Creek compared with Kuzkwa River (Stuart -270 Summer) shows an average genome-wide FST of 0.0091 (95% CI: 0.0089-0.0093) based on this dataset, 271 which is 9.1x higher than Driftwood River vs. Paula Creek. Paula Creek compared with Bowron River 272 shows an average genome -wide FST of 0.0355 (95% CI: 0.0353 -0.0358), or 35.5x higher than 273 Driftwood vs. Paula. 274 275 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 10

276 The present study took an in -depth and region -specific investigation into genome-wide genetic 277 variation within the upper Fraser River sockeye salmon through the re -use of existing genomic 278 resources generated as part of a three-species analysis of the Fraser River salmonids (Christensen et al. 279 2024). General population genetic trends were similar to those observed using other resources 280 including microsatellites (Beacham et al. 2004) or SNPs (Rondeau 2022) , with some notable 281 exceptions, as described below. 282 283 Major region of differentiation on Chr18 in sockeye from the Nechako Watershed 284 A main finding of this work was the identification of a major differentiating region between the Early 285 Stuart sockeye and the Stuart-Summer/Nadina/Stellako sockeye. This genomic region may be involved 286 in phenotypic variation in run-timing between these groupings. The Nadina and Stellako populations 287 are considered to be Early Summer and Summer run timing groups , respectively (COSEWIC 2017). 288 The unique haplotype identified here appears to be in a derived form with low heterozygosity in the 289 Stuart-Summer/Nadina/Stellako samples, as the Early Stuart samples share alleles with samples from 290 other populations within the upper Fraser River. Interestingly, some variants within this haplotype are 291 nearly fixed between the Stuart-Summer/Nadina/Stellako grouping and the rest of the upper Fraser 292 River. At the most differentiated loci, there are very few heterozygous genotypes in either grouping . 293 This suggests that these populations have very low gene flow between the haplotype groups. There 294 was no clear evidence of an inversion in the area, but this remains inconclusive as the current available 295 data was comprised of short reads, whereas a long -read analysis could provide a more definitive 296

regarding structural variation. 297 The putatively derived state of the haplotype in the Stuart-Summer/Nadina/Stellako groupings 298 poses interesting questions around the emergence of this haplotype, as the putatively ancestral form is 299 present in the group with the unique run timing designation, i.e., the Early Stuart sockeye. If the unique 300 haplotype is related to phenotypic run timing variation, it therefore follows that the derived phenotype 301 would be that of the Stuart-Summer/Nadina/Stellako grouping. If this genomic region underlies run 302 timing variation, this may indicate that the early run for Early Stuart sockeye is the ancestral run timing 303 form, and the later runs to the region (e.g., Stuart -Summer) would be an adaptation allowing for 304 expanded timing of spawning resource access (as discussed by Barry et al. 2024). However, it is also 305 possible that the unique region of differentiation is not related to run timing but is rather only coincident 306 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 11 to the run timing difference between the analyzed groups. However, other details around the major 307 differentiated region on Chr18 suggest involvement in run timing or other life history variation. 308 The region of major differentiation on Chr18 at 56.3 – 58.0 Mbp falls within the sockeye 309 salmon chromosome arm that corresponds to the ancestral chromosome 15. 2 (see Sutherland et al. 310 2016). The ancestral designation is related to the post -whole-genome duplication chromosome 311 correspondence where each of 25 ancestral chromosomes has two copies in the present-day salmonid 312 genome, and where the 25 chromosomes roughly correspond to the present -day northern pike Esox 313 lucius genome (Rondeau et al. 2014). Importantly, a major differentiated section on the chromosome 314 arm of Chr12 of sockeye salmon and Chr10 of pink salmon has been identified as being involved in 315 run timing in Alaska (Barry et al. 2024). These two chromosomes both contain 15.1, the homeolog (or 316 ohnolog) to 15.2 (Sutherland et al. 2016; Christensen et al. 2020; Barry et al. 2024) . Furthermore, a 317 major differentiating region of Chr12 is associated with life history variation in Okanagan O. nerka 318 (Tigano and Russello 2022). Recently, this region has been proposed as a master regulatory region in 319 multiple salmonid species including pink salmon O. gorbuscha, chum salmon O. keta, and sockeye 320 salmon in Alaska (Barry et al. 2024; Barry et al. 2025). Secondary regions of high differentiation have 321 been identified on the homeolog 15.2 in chum salmon and coho salmon O. kisutch related to run timing 322 (ancestral 15.2 or chum Chr29; Barry et al. 2025), but not yet in sockeye salmon until the present work. 323 A gene of interest in the genomic region of differentiation previously identified and highlighted 324 is leucine-rich repeat -containing protein 9 (lrrc9) (Barry et al. 2024) . The region of major 325 differentiation in the Stuart sockeye identified here contains a gene annotated as lrrc9-like, the 326 homeologous gene to lrrc9 highlighted previously. However, in the sockeye salmon genome, this gene 327 is annotated as a putative pseudogene by the NCBI RefSeq annotation. Interestingly, Barry et al. (2025) 328 propose that the region of differentiation is likely driving phenotypic effects through transcription 329 regulation activity rather than changes in protein coding sequences due to the lack of non-synonymous 330 mutations identified and the prevalence of transcription factors in the region. If lrrc9-like is a 331 pseudogene, this would support the lack of direct involvement of this gene in any downstream 332 phenotype. There are other genes in the major differentiated region identified here in sockeye salmon 333 that may be of interest regarding a migratory timing phenotype, most notably homeobox protein OTX2 334 (orthodenticle homeobox 2), a transcription factor involved in brain and sensory organ development 335 (Beby and Lamonerie 2013) . OTX2 is involved in the formation of the pituitary gland in zebrafish 336 (Bando et al. 2020) , and the pituitary -gonadal axis activity in salmonids is connected to season and 337 directly related to homing adults, return migration and spawning preparation (Onuma et al. 2009; Ueda 338 2011). Furthermore, the role of OTX2 in the development of olfactory system, including imprinting 339 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 12 and homing has been demonstrated in the anemonefish Amphiprion percula (Veilleux et al. 2013), of 340 interest given the important role of the olfactory system in homing migration in sockeye salmon (Bett 341 et al. 2018). 342 The present study therefore identifies new evidence that points to potential involvement of the 343 proposed master control region on ancestral chromosome 15 in a completely different population to 344 that originally identified, and that has taken an alternate homeologous chromosome path to that which 345 was previously identified in sockeye salmon (Tigano and Russello 2022; Barry et al. 2024) . This 346 supports the conclusions of Barry et al. (2025), and points to unique genetics and evolutionarily 347 significant adaptation route of the upper Fraser River sockeye salmon. As functional analysis of the 348 major differentiated region here has not been conducted, the role of this region in run timing remains 349 putative. Although other run timing differences are reported in the populations analyzed in this study, 350 most notably the two run timings of the Chilcotin Watershed sockeye (i.e., Chilko as Summer and 351 Taseko as Early Summer), there were no similar regions of major differentiation observed in these 352 other populations using the available samples. 353 Considering the important role of ancestral chromosome 15.1 and 15.2, it is worth briefly 354 noting other features of these chromosomes in the evolution of salmonids. First, these chromosomes 355 are part of ancestral and conserved fusion events; ancestral 15.1 is proposed to have been fused to 4.1 356 prior to the divergence of genera Salmo and Oncorhynchus, although lineage -specific fissions and 357 fusions occurred in the lineage leading to the pink, chum, sockeye branch (Sutherland et al. 2016) . 358 Ancestral 15.2 was part of a conserved fusion to 25.2 that occurred prior to the diversification of 359 Oncorhynchus; this fusion remains in all Oncorhynchus species. Second, the ancestral chromosome 360 15.1 is within the chromosome that holds the master sex determining gene in Arctic charr Salvelinus 361 alpinus and brook charr S. fontinalis (Sutherland et al. 2017) . Although these features may not be 362 related to the specific life history and run timing variation described here, given the proposed role of 363 this genomic region as a master control region for multiple salmonid species, continued understanding 364 of these chromosomes may yield valuable insights in the evolution of phenotypic variation in the 365 salmonids. 366 367 Metapopulation genetics in Early Stuart sockeye 368 Using genome-wide data we found some evidence of increased genetic differentiation with distance 369 within the Early Stuart spawning tributaries, specifically that only the spawning sockeye from the most 370 distant tributaries showed non-zero FST. However, this level of differentiation was still very low, being 371 9-times lower than that which occurs between run timing groups. Further, in this genome -wide 372 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 13 analysis, we did not identify any regions with elevated differentiation relative to the rest of the genome, 373 suggesting that the results from panels of neutral markers should be similar to that obtained with 374 genome-wide data. In the absence of large genetic differences among spawning tributaries of Early 375 Stuart sockeye, i t is possible that any phenotypic variation observed between streams could be 376 associated with plasticity, environmental effects, or epigenetic differences between populations 377 (Lamka et al. 2022). The greatest differentiation identified was still much lower than the FST 0.01 cutoff 378 proposed as the lower limit of what is possible for genetic stock identification (GSI) (Araujo et al. 379 2014). Therefore, these results support previous work indicating an inability to discriminate between 380 populations using GSI (Beacham et al. 2004; Rondeau 2022) , and support the connectivity and 381 characteristics of a metapopulation (Bradford and Braun 2021), but do provide some evidence for slight 382 spatial differentiation that is worthwhile considering as not completely panmictic. Although low 383 sample size appears to have resulted in misplaced populations in the genetic dendrogram (most 384 noticeable with sample size under n = 5), and therefore raises the potential concern that this could drive 385 the non-zero FST observed between Driftwood and other more distant streams, the observation that 386 Driftwood (n = 7 ) compared to Dust Creek (n = 10) , the more proximal sampling location, had an 387 estimated zero differentiation indicates that the conclusion of no differentiation could be determined at 388 this sample size. It should also be noted that the samples used to represent the Early Stuart sockeye in 389 this study were from 2005 , and if large scale genetic changes occurred for salmon spawning in these 390 tributaries since 2005, these results may not reflect current day genetic signatures. 391 Further investigation of the extent of gene flow in this metapopulation could be conducted by 392 parentage-based tagging (i.e., PBT) or other tagging methods that track returning adults to determine 393 the extent of flow between tributaries. Furthermore, epigenetic analyses could be conducted to 394 determine the level of differentiation epigenetically between streams with different environmental 395 parameters (Venney et al. 2021; Lamka et al. 2022). 396 397

398 In the present study we made use of existing genomic resources to characterize upper Fraser River 399 sockeye salmon populations from waterbodies that flow into the Nechako (including Nadina, Stellako, 400 and Stuart River systems), the Bowron River, Quesnel and Horsefly, and the Chilcotin Watershed. This 401 analysis would not have been possible without the generation and publication of large -scale, open-402 source genomic resources and metadata. The finding of a major differentiating region on Chr18 403 (ancestral 15.2) between two conservation units (Early Stuart and Stuart-Summer) with differing run 404 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 14 timings points to the potential for this region to contain adaptive variation underlying the run timing 405 phenotype, although interestingly the more unique run timing (i.e., Early Stuart) was in the grouping 406 with the putatively more ancestral form of the haplotype on Chr18. Convergence upon the Chr18 (15.2) 407 chromosome near the lrrc9-like gene fits with other work that has proposed that this region, in both 408 duplicate chromosomes, is a major control region of run timing, evolving independently in multiple 409 species. This region and the genes within it, including otx2, merit further investigation. The genome-410 wide data on the Early Stuart metapopulation supports previous work indicating differentiation being 411 too low for GSI, but does identify a very minor increase in differentiation with increased geographic 412 distance within the Early Stuart spawning tributaries . These results collectively provide new insights 413 on upper Fraser River sockeye salmon that will be useful for conservation and management. 414 415

416 Thanks to Drs. Ben Koop , Eric Rondeau , and Kris Christensen for discussion on the major 417 differentiation region, and to Linda Stevens of UFFCA for technical support on the project. Thanks to 418 the UFFCA for supporting this work in the interests of Upper Fraser First Nations. 419 420 Funding 421 Funding for this work was made available via the Pacific Salmon Strategy Initiative in support of the 422 Tuzist'ol T'ah process. Tuzist'ol T'ah is a collaboration between the Upper Fraser Fisheries 423 Conservation Alliance, Binche Whut’en, Nak’azdli Whut’en, Takla Nation, Tl’azt’en Nation, 424 Yekooche Nation, Province of British Columbia, and Fisheries and Oceans Canada tasked with 425 developing a recovery strategy for Early Stuart Sockeye. 426 427 Competing Interests 428 BJGS is affiliated with Sutherland Bioinformatics. The authors have no competing financial interests to 429 declare. 430 431 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 15 Data Availability 432 SNP data were downloaded from FigShare: https://doi.org/10.25387/g3.25705428.v1 433 The following analytic repositories supported this analysis: 434 Code and README for project: https://github.com/bensutherland/ms_oner_estu 435 Population genetic functions: https://github.com/bensutherland/simple_pop_stats 436 437 Additional Materials 438 A Supplemental Results section is provided with supplemental figures and tables. 439 Additional File S1. Genotypes of top differentiated loci in major differentiation region. 440 Additional File S2. Annotated gene content in major differentiation region. 441 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 16

442 443 Akopyan, M., A. Tigano, A. Jacobs, A.P. Wilder, and N.O. Therkildsen, 2025 Genetic differentiation is 444 constrained to chromosomal inversions and putative centromeres in locally adapted populations 445 with higher gene flow. Molecular Biology and Evolution. 446 Araujo, H.A., J.R. Candy, T.D. Beacham, B. White, and C. Wallace, 2014 Advantages and challenges of 447 genetic stock identification in fish stocks with low genetic resolution. Transactions of the 448 American Fisheries Society 143 (2):479-488. 449 Baird, N.A., P.D. Etter, T.S. Atwood, M.C. Currey, A.L. Shiver et al., 2008 Rapid SNP discovery and 450 genetic mapping using sequenced RAD markers. PLOS ONE 3 (10):e3376. 451 Bando, H., P. Gergics, B.L. Bohnsack, K.P. Toolan, C.E. Richter et al., 2020 Otx2b mutant zebrafish have 452 pituitary, eye and mandible defects that model mammalian disease. Human Molecular Genetics 453 29 (10):1648-1657. 454 Barry, P., N. Howe, G.L. Owens, D. Baetscher, K. D’Amelio et al., 2025 Diversification of large-effect 455 loci in a duplicated genomic region leads to complex phenotypes. 456 bioRxiv:2024.2003.2030.587279. 457 Barry, P.D., D.A. Tallmon, N.S. Howe, D.S. Baetscher, K.L. D’Amelio et al., 2024 A major effect locus 458 involved in migration timing is shared by pink and sockeye salmon. 459 bioRxiv:2024.2003.2030.587279. 460 Beacham, T.D., M. Lapointe, J.R. Candy, B. McIntosh, C. MacConnachie et al., 2004 Stock identification 461 of Fraser River sockeye salmon using microsatellites and major histocompatibility complex 462 variation. Transactions of the American Fisheries Society 133 (5):1117-1137. 463 Beacham, T.D., and R.E. Withler, 2017 Population structure of sea-type and lake-type sockeye salmon 464 and kokanee in the Fraser River and Columbia River drainages. PLOS ONE 12 (9):e0183713. 465 Beby, F., and T. Lamonerie, 2013 The homeobox gene Otx2 in development and disease. Experimental 466 Eye Research 111:9-16. 467 Bemmels, J.B., S. Starko, B.L. Weigel, K. Hirabayashi, A. Pinch et al., 2025 Population genomics reveals 468 strong impacts of genetic drift without purging and guides conservation of bull and giant kelp. 469 Current Biology 35 (3):688-698.e688. 470 Bett, N.N., S.G. Hinch, K.H. Kaukinen, S. Li, and K.M. Miller, 2018 Olfactory gene expression in 471 migrating adult sockeye salmon Oncorhynchus nerka. Journal of Fish Biology 92 (6):2029-2038. 472 Bradford, M.J., and D.C. Braun, 2021 Regional and local effects drive changes in spawning stream 473 occupancy in a sockeye salmon metapopulation. Canadian Journal of Fisheries and Aquatic 474 Sciences 78 (8):1084-1095. 475 Christensen, K.A., A.-M. Flores, D. Sakhrani, C.A. Biagi, R.H. Devlin et al., 2024 Revealing the 476 evolutionary history and contemporary population structure of Pacific salmon in the Fraser River 477 through genome resequencing. G3 Genes|Genomes|Genetics 14 (10). 478 Christensen, K.A., E.B. Rondeau, D.R. Minkley, D. Sakhrani, C.A. Biagi et al., 2020 The sockeye salmon 479 genome, transcriptome, and analyses identifying population defining regions of the genome. 480 PLOS ONE 15 (10):e0240935. 481 COSEWIC, 2017 COSEWIC assessment and status report on the sockeye salmon Oncorhynchus nerka, 482 24 Designatable Units in the Fraser River Drainage Basin, in Canada, pp. xli + 179, edited by 483 Committee on the Status of Endangered Wildlife in Canada. 484 Crossin, G.T., S.G. Hinch, A.P. Farrell, D.A. Higgs, A.G. Lotto et al., 2004 Energetics and morphology of 485 sockeye salmon: effects of upriver migratory distance and elevation. Journal of Fish Biology 65 486 (3):788-810. 487 Danecek, P., A. Auton, G. Abecasis, C.A. Albers, E. Banks et al., 2011 The variant call format and 488 VCFtools. Bioinformatics 27 (15):2156-2158. 489 Danecek, P., J.K. Bonfield, J. Liddle, J. Marshall, V . Ohan et al., 2021 Twelve years of SAMtools and 490 BCFtools. GigaScience 10 (2). 491 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 17 Eliason, E.J., T.D. Clark, M.J. Hague, L.M. Hanson, Z.S. Gallagher et al., 2011 Differences in thermal 492 tolerance among sockeye salmon populations. Science 332 (6025):109-112. 493 Gilhousen, P., 1980 Energy sources and expenditures in Fraser River sockeye salmon during their 494 spawning migration. Int. Pac. Salmon Fish. Comm. Bull. 22:1-80. 495 Hinch, S.G., and P.S. Rand, 1998 Swim speeds and energy use of upriver-migrating sockeye salmon 496 (Oncorhynchus nerka): role of local environment and fish characteristics. Canadian Journal of 497 Fisheries and Aquatic Sciences 55 (8):1821-1831. 498 Idler, D., and W. Clemens, 1959 The energy expenditures of Fraser River sockeye salmon during the 499 spawning migration to Chilko and Stuart Lakes in International Pacific Salmon Fisheries 500 Commission Progress Report. International Pacific Salmon Fisheries Commission, New 501 Westminster, BC, Canada. 502 Knaus, B.J., and N.J. Grünwald, 2017 vcfr: a package to manipulate and visualize variant call format data 503 in R. Molecular Ecology Resources 17 (1):44-53. 504 Lamka, G.F., A.M. Harder, M. Sundaram, T.S. Schwartz, M.R. Christie et al., 2022 Epigenetics in 505 ecology, evolution, and conservation. Frontiers in Ecology and Evolution Vo l u m e 1 0 - 2022. 506 Li, H., 2013 Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM, pp. 507 arXiv:1303.3997. 508 Martin, M., 2011 Cutadapt removes adapter sequences from high-throughput sequencing reads. 509 EMBnet.journal 17 (1):3. 510 Meek, M.H., and W.A. Larson, 2019 The future is now: Amplicon sequencing and sequence capture usher 511 in the conservation genomics era. Molecular Ecology Resources 19 (4):795-803. 512 Normandeau, E., M. de Ronne, and D. Torkamaneh, 2023 SNPLift: Fast and accurate conversion of 513 genetic variant coordinates across genome assemblies. bioRxiv:2023.2006.2013.544861. 514 Onuma, T.A., S. Sato, H. Katsumata, K. Makino, W. Hu et al., 2009 Activity of the pituitary–gonadal axis 515 is increased prior to the onset of spawning migration of chum salmon. Journal of Experimental 516 Biology 212 (1):56-70. 517 Paradis, E., 2010 pegas: an R package for population genetics with an integrated–modular approach. 518 Bioinformatics 26 (3):419-420. 519 Paria, S.S., S.R. Rahman, and K. Adhikari, 2022 fastman: A fast algorithm for visualizing GWAS results 520 using Manhattan and Q-Q plots. bioRxiv:2022.2004.2019.488738. 521 Peterson, B.K., J.N. Weber, E.H. Kay, H.S. Fisher, and H.E. Hoekstra, 2012 Double digest RADseq: an 522 inexpensive method for de novo SNP discovery and genotyping in model and non-model species. 523 PLOS ONE 7 (5):e37135. 524 R Core Team, 2025 R: A Language and Environment for Statistical Computing. R Foundation for 525 Statistical Computing, Vienna, Austria. 526 Rand, P.S., H.S. G., M. J., F.M.G. G., M.M. J. et al., 2006 Effects of river discharge, temperature, and 527 future climates on energetics and mortality of adult migrating Fraser River sockeye salmon. 528 Transactions of the American Fisheries Society 135 (3):655-667. 529 Robinson, J.T., H. Thorvaldsdóttir, W. Winckler, M. Guttman, E.S. Lander et al., 2011 Integrative 530 genomics viewer. Nature Biotechnology 29 (1):24-26. 531 Rondeau, E.B., 2022 Sockeye salmon SNP panel genetic baseline for the Fraser River, Year 2 in Southern 532 Endowment Fund Annual Reports, edited by Pacific Salmon Commission. 533 Rondeau, E.B., D.R. Minkley, J.S. Leong, A.M. Messmer, J.R. Jantzen et al., 2014 The genome and 534 linkage map of the northern pike (Esox lucius): conserved synteny revealed between the salmonid 535 sister group and the Neoteleostei. PLoS One 9 (7):e102089. 536 Sutherland, B.J.G., T. Gosselin, E. Normandeau, M. Lamothe, N. Isabel et al., 2016 Salmonid 537 chromosome evolution as revealed by a novel method for comparing RADseq linkage maps. 538 Genome Biology and Evolution 8 (12):3600-3617. 539 Sutherland, B.J.G., C. Rico, C. Audet, and L. Bernatchez, 2017 Sex chromosome evolution, 540 heterochiasmy, and physiological QTL in the salmonid brook charr Salvelinus fontinalis. G3-541 Genes Genomes Genetics 7 (8):2749-2762. 542 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 18 Tigano, A., and M.A. Russello, 2022 The genomic basis of reproductive and migratory behaviour in a 543 polymorphic salmonid. Molecular Ecology 31 (24):6588-6604. 544 Ueda, H., 2011 Physiological mechanism of homing migration in Pacific salmon from behavioral to 545 molecular biological approaches. General and Comparative Endocrinology 170 (2):222-232. 546 Veilleux, H.D., L. Van Herwerden, N.J. Cole, E.K. Don, C. De Santis et al., 2013 Otx2 expression and 547 implications for olfactory imprinting in the anemonefish, Amphiprion percula. Biology Open 2 548 (9):907-915. 549 Venney, C.J., B.J.G. Sutherland, T.D. Beacham, and D.D. Heath, 2021 Population differences in Chinook 550 salmon (Oncorhynchus tshawytscha) DNA methylation: Genetic drift and environmental factors. 551 Ecology and Evolution 11 (11):6846-6861. 552 Waples, R.S., R.G. Gustafson, L.A. Weitkamp, J.M. Myers, O.W. Jjohnson et al., 2001 Characterizing 553 diversity in salmon from the Pacific Northwest. Journal of Fish Biology 59 (sA):1-41. 554 Weir, B.S., and C.C. Cockerham, 1984 Estimating F-statistics for the analysis of population structure. 555 Evolution 38 (6):1358-1370. 556 Wellenreuther, M., and L. Bernatchez, 2018 Eco-evolutionary genomics of chromosomal inversions. 557 Trends in Ecology & Evolution 33 (6):427-440. 558 Wright, M.J., M. Hurson, K.A. Robinson, D.A. Patterson, and J.G. Venditti, 2025 A typology of potential 559 hydraulic barriers to adult salmon migration in a bedrock river. Canadian Journal of Fisheries 560 and Aquatic Sciences 82:1-19. 561 562 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 19 Figures 563 564 565 Figure 1. (A) All upper Fraser River samples clustered by principal components analysis (PCA) using 566 MAF and LD -filtered SNPs (n = 627,769) shows main groupings of study, including the Nechako 567 Watershed (negative PC1), the Bowron River (positive PC1), the Chilcotin Watershed (negative PC2), 568 and Horsefly and Quesnel regions (positive PC1 and PC2). (B) PCA with only the populations from 569 the Nechako Watershed shows Early Stuart grouping separately from Stuart-Summer, which were more 570 genetically proximal to populations from the Nadina River and Stellako River. 571 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 20 572 573 Figure 2. (A) Discriminant analysis of principal components (DAPC) analysis of populations from the 574 Nechako Watershed including Early Stuart, Stuart -Summer, Nadina River, and Stellako River shows 575 separation across discriminant function 1 ( DF1) from the Early Stuart and Stuart -Summer (and 576 Stellako), and Nadina River grouping separately. (B) The loadings of SNP loci contributing to the 577 separation on DF1 indicate a major contributing genomic region on Chr18. 578 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 21 579 Figure 3. (A) Genetic differentiation by FST within the Nechako Watershed for Chr18 between the 580 Early Stuart and Stuart summer populations. (B) Heatmap of the differentiated genomic region based 581 on allelic dosage of alternate alleles (0, 1, 2) coloured as grey, yellow, and blue, respectively. All 582 samples from the study (n = 210) were included and allowed to cluster based on genotypes. Samples 583 are coloured on the vertical axis based on the source collection site as indicated in the legend. 584 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 22 585 Figure 4. Disciminant analysis of principal components (DAPC) of collections from Early Stuart 586 spawning tributaries that contained at least seven samples each. 587 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 23 Tables 588 Table 1. Samples available for the analysis from Christensen et al. (2024) genotyped by whole-genome 589 resequencing. Collection names and dates were provided in sample metadata from the original study, 590 full names were determined from BC Geographical Names (Government of BC ), and run -timing 591 designations were obtained from Rondeau (2022). 592 593 Region Run-Timing Collection Full Name Sample size Year(s) sampled Stuart Early-Stuart Bivouac Bivouac Cr. 10 2005 Driftwood Driftwood R. 7 2005 Dust Dust Cr. 10 2005 Felix Sidney (Felix) Cr. 5 2005 Paula Paula Cr. 10 2005 Takla Takla Lk. 4 unkn. Summer Kuzkwa Kuzkwa R. 9 2001 Middle Middle R. 7 2001 Pinchi Pinchi Cr. 4 2005 Tachie Tachie R. 11 2012, unkn. Nadina/Francois Early summer Nadina Nadina R. 9 2000 Francois/Fraser Summer Stellako Stellako R. 10 2011 Bowron Early summer Bowron Bowron R. 10 2001 Quesnel Summer BlueLead Blue Lead Cr. 5 2001 McKinley McKinley Cr. 9 2001 Mitchell Mitchell Bay 8 2001 Wasko Wasko Cr. 5 2001 Horsefly* Horsefly R. and Lk. 24 2001, unkn Quesnel Quesnel R. and Lk. 8 2013 Chilcotin Summer Chilko* Chilko R. and Lk. 35 2001, 2008, unkn Early Summer Taseko Taseko R. 10 2011 TOTAL 210 *includes various sub-naming 594 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint v.0.1.2 24 Table 2. (A) Average FST between Early Stuart collections showing non -weighted estimate in the 595 bottom and weighted estimate in the top section; or (B) 95% confidence interval of FST showing the 596 upper limit in the top and lower limit in the bottom. 597 598 (A) 599 Bivouac Driftwood Dust Paula Bivouac NA 0.0056 0.0010 0.0008 Driftwood 0.0019 NA 0.0022 0.0046 Dust -0.0011 -0.0009 NA 0.0016 Paula -0.0011 0.0010 -0.0006 NA 600 (B) 601 Bivouac Driftwood Dust Paula Bivouac NA 0.0021 -0.0009 -0.0010 Driftwood 0.0017 NA -0.0007 0.0012 Dust -0.0013 -0.0011 NA -0.0004 Paula -0.0013 0.0008 -0.0007 NA 602 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.04.657877doi: bioRxiv preprint