{"paper_id":"22cd1c08-3b2c-461c-86e5-0f86e31799db","body_text":"1 \nRevisiting the genetics of Lake Constance Coregonids using \nlake-wide whole genome sequencing \n \nArne Jacobs1,†, Samuel Roch2, Barnaby Roberts2, Maria Capstick1 & Alexander Brinker2,3,† \n  \n1. School of Biodiversity, One Health and Veterinary Medicine, University of Glasgow, UK \n2. Fisheries Research Station Baden-Württemberg, Langenargen, Germany \n3. University of Konstanz, Constance, Germany \n \n† Corresponding authors: arne.jacobs@glasgow.ac.uk; alexander.brinker@lazbw.bwl.de  \n \nKeywords: Coregonids; Low -coverage whole -genome sequencing; Stock assignment; Lake \nConstance; Genomics; \n  \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 2 \nAbstract \nAnthropogenic pressures can have detrimental impacts on fish populations, with their effective \nmanagement and conservation requiring accurate monitoring tools. Yet, this is not straightforward \nfor closely-related, co-existing species that are difficult to distinguish using simple phenotypic or \ngenetic approaches. Coregonids are of cultural and economic importance across Europe but have \nfaced a multitude of pressures over the last century. Yet genomic management tools are lacking. \nIn Lake Constance, a large pre-alpine lake, stocks have drastically collapsed due to a multitude of \npressures, leading to a fishery closure. Here, we adopt a cost-effective, whole genome sequencing \napproach for lake-wide assessment of stock composition, spatial distribution and genetic diversity \nof highly admixed Lake Constance whitefish (Coregonus spp.). By sequencing 983 adult and larval \ngenomes, we show that nearly 90% of the stock is made up by one of three species, the Gangfisch \n(C. macrophthalmus ), and define the genetic relati onship between Upper and Lower Lake \nConstance whitefish stocks. We also identified strong mixing between Gangfisch and Blaufelchen \n(C. wartmanni) on traditionally specific-specific spawning grounds, and detected strong admixture \nin larvae, with potentially drastic impacts on the effectiveness of hatchery supplementation and \nstocking. Despite the collapse and admixture, species still exhibit low to  moderate levels of \ngenetic diversity, maintain ecologically -relevant genetic differences, and seem to show \ndifferences in habitat use. Overall, we present a cost -effective, translatable tool for stock -wide \nsequencing and genetically -informed fisheries management, with our results calling for the re -\nevaluation of current management practices to avoid the potential g enetic mixing between \nspecies. \n \n  \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 3 \nBackground \nAnthropogenic pressures, such as invasive species, eutrophication, and overfishing have strongly \nimpacted fish populations worldwide, leading to broad ecological changes (Kefford et al., 2023; \nReid et al., 2019; Tickner et al., 2020; Vardakas et al., 2025) . These include population size \ncollapses (Haase et al., 2025; Han et al., 2025) , genetic diversity loss (Pinsky & Palumbi, 2014) , \nincreased hybridization (Vonlanthen et al., 2012), and altered habitat use (Alexander et al., 2017). \nUnderstanding the impacts of anthropogenic pressures on fish populations requires large -scale \nmonitoring, yet this can be highly challenging when multiple species or ecotypes co -exist that \ncannot be reliably phenotypically distinguished.  \nFisheries management often relies on targeted genotyping assays (e.g. GT -seq, \nmicrosatellites) and DNA barcoding for genetic monitoring (Bernatchez et al., 2017; Friedman et \nal., 2022; Theissinger et al., 2023) . While these approaches can be powerful and cost -effective \nonce developed, their development is often expensive and labour-intensive, they only cover a small \nportion of the genome (few microsatellites to few hundred SNPs) and can be prone to \nascertainment bias if they have not been developed across all relevant populations (Beemelmanns \net al., 2024; Robledo et al., 2018) . Whole-genome sequencing (WGS) offers a powerful approach \nfor investigating fine-scale population structure and differentiation for conservation and fisheries \nmanagement (Bernatchez et al., 2017; Leitwein et al., 2020; Tengstedt et al., 2025) . Yet, WGS is \noften prohibitively expensive and requires large quantities of DNA, limiting its suitability in many \napplied contexts. The advent of low -coverage whole genome sequencing (‘lcWGS’) opens up \nexciting opportunities to rapidly and cost-effectively sequence the entire genome of an individual \nfrom relatively little DNA (Gaio et al., 2022; Lou et al., 2021; Therkildsen & Palumbi, 2017) . In \ncontrast to traditional approaches, lcWGS can be used to simultaneously investigate population \nstructure, perform stock/species assignment and also investigate fine -scale patterns of genetic \ndifferentiation and selection across the genome for a large number of individuals, even when \nlevels of genetic differentiation is weak (DeSaix et al., 2024; Lou et al., 2021; Therkildsen & \nPalumbi, 2017). However, it has been rarely applied at a large scale to directly inform fisheries \nmanagement decisions.   \n European whitefish (Coregonus spp.) are economically and culturally important fisheries \nspecies in Europe, including in Lake Constance, one of the largest pre -alpine lakes in Central \nEurope (Haase et al., 2025; Vonlanthen et al., 2012) . Originally, Upper Lake Constance (ULC) \nharboured four ecologically distinct whitefish species; the pelagic spawning, planktivorous \nBlaufelchen (C. warmanni), the generalist pelagic-littoral Gangfisch that spawned in the benthic \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 4 \nzone (C. macrophthalmus), the shallow-spawning littoral Sandfelchen (C. arenicolus), and the now-\nextinct profundal Kilch ( C. gutturosus) (Steinmann, 1950; Vonlanthen et al., 2012) . Furthermore, \nthe Lower Lake Constance (LLC), a smaller and shallower lake -basin that is connected to Lake \nConstance through a short river, contained the Unterseefelchen (or Weissfelchen), which has not \nbeen genetically characterised (De-Kayne et al., 2022; Steinmann, 1950) . Anthropogenic \neutrophication in the 20th century led to extinction of the profundal Kilch and the collapse of \nreproductive barriers between species, thereby increasing hybridization by shifting spawning and \nfeeding habits (Alexander et al., 2017; Feulner & Seehausen, 2019; Frei, Mwaiko, et al., 2023; Hirsch \net al., 2013; Jacobs et al., 2019; Numann, 1978, 1986; Schweizer, 1894; Vonlanthen et al., 2012) . \nThis led to reduced genetic differentiation between species in Lake Constance, particularly \nbetween the pelagic Blaufelchen and the Gangfisch (Frei, Mwaiko, et al., 2023) . Although \nhybridization led to increased allelic richness within species (Gum et al., 2014; Vonlanthen et al., \n2012), an overall loss in genetic diversity was observed in all whitefish species over the last century \n(Frei, Mwaiko, et al., 2023) . Over the last decades, a combination of overfishing, environmental \nchange, and invasive species (Baer, Spiessl, et al., 2022; Dahms et al., 2024; Haase et al., 2025; \nRösch et al., 2018)  have contributed to a collapse of the whitefish populations, especially the \nBlaufelchen, a key fisheries target (Haase et al., 2025), leading to the closure of the commercial \nwhitefish fishery in Lake Constance in 2024 (Baer et al., 2016; Haase et al., 2025) . Despite \nextensive stocking over the last 150 years, larval recruitment has been low (Baer et al., 2023; Haase \net al., 2025) , raising questions on the efficiency of stocking efforts. To better understand the \ncurrent situation and inform effective management strategies, more detailed information on the \ncontemporary composition and habitat use of European whitefish species across  all of Lake \nConstance, from larvae to adults, is required.  \nFisheries management strategies often rely on historical knowledge of species, including \ntheir genetic relationships, levels of genetic diversity and habitat preferences. Yet, these factors \ncan rapidly change in the face of environmental and population change. Here, we integrate large-\nscale sampling and phenotypic analysis of European whitefish in Upper and Lower Lake Constance \nacross several years, seasons and developmental stages with novel, cost -effective low-coverage \nwhole genome sequencing (lcWGS) of n early 1,000 individuals to reconstruct the contemporary \ngenetic composition of European whitefish in Lake Constance, investigate the spatial distribution \nof individual species, and map genomic levels of genetic diversity and differentiation across \nspecies. Using this comprehensive approach, we demonstrate the suitability of lcWGS for large -\nscale cost -effective fisheries management and show that historical patterns of species \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 5 \ndistribution are not reliable indicators of species composition, identify the mixing of species on \nspawning grounds, genetically identify Unterseefelchen as Sandfelchen ( C. arenicolus ), and \nhighlight low levels of genetic diversity and differentiation. Thus, these results question the \ndirection and effectiveness of current management strategies.  \n \n \nMethods \n  \nSampling  \nAdult individuals of European whitefish species occurring in Lake Constance were targeted using \nbenthic and pelagic gillnets during sampling campaigns from 2021 to 2023 (Table S1, Fig. 1). We \ntargeted all three species in the Upper Lake Constance (ULC), sampling “Blaufelchen” ( C. \nwartmanii) and “Gangfisch” ( C. macrophthalmus) during the annual spawning fisheries in the \npelagic and littoral zone, respectively. The littoral “Sandfelchen” ( C. arenicolus) was repeatedly \ntargeted and sampled over a longer timeperiod in the littoral zone of the lake. Whitefish from the \nLower Lake Constance (“Unterseefelchen” in LLC) were sampled in various locations. Whitefish \nthat migrate in September and October into the Alpine Rhine to spawn (“Alpenrheinfelchen”) were \ntargeted by local anglers (n=1 sample). Furthermore, adult whitefish were caught as part of a lake-\nwide fishing campaign in September 2024, during which the entire lake was systematically fished \nusing 391 randomized benthic and pelagic gillnets in varying water depths (Fig . 1; Vonlanthen et \nal., in preparation). A total of 34 whitefish in the Lower Lake and 70 whitefish in the Upper Lake \nwere caught (Table 1 and S2). Fish were stunned with a blo w on the head, killed with a cut at the \ngills and total length (0.1 cm) and wet weight (0.1 g) were measured. All individuals from the initial \nfishing campaign and most individuals from the lake-wide fishing campaign were photographed \nlaterally using a dig ital camera (Pentax K3 II with 18 –135 mm lens and fixed focal length). For \ngenetic analysis, fin clips (~0.5 cm2) were preserved in absolute ethanol. \nLarval whitefish were captured in the pelagic and littoral zone of ULC (Fig . 1) using light \ntraps developed at the Fisheries Research Station. Larvae are attracted toward a pipe via a light \nand subsequently suctioned into the pipe and retained through a pump. The capture of larvae in \nthe light traps is almost always non -lethal. In 2021 and 2022, five light -traps were deployed at \n>20 m depth. In 2022, five additional traps were deployed in the deeper littoral zone (~5 m) of \nthe lake. Light traps were deploye d from February to May in both years. In the shallower littoral \nzone larvae were captured by hand in 2021/22, either on foot or onboard a boat between March \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 6 \nand June in 2021 and 2022. Larvae were attracted using headlamps or spotlights and captured \nusing aquarium hand-nets. A total of 744 larvae were caught (Table 1 and S3) and preserved in \nabsolute ethanol.  \nAll individuals included in the current study were caught by licensed personnel with \npermission of the local fisheries administration (Regierungspräsidium Tübingen) and according \nto German animal protection legislation (§4) and the ordinance on the slaught er and killing of \nanimals (Tierschutzschlachtverordnung §13). \n \n \nFigure 1 - Sampling information Overview over the study area and sampling locations for adult \nand larval whitefish in Lake Constance, southwest Germany. The lower relief map shows the \nsampling locations for each fishing campaign, with explanations in the legend. Red dots in the \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 7 \nmain map show the locations of targeted fishing campaigns, whereas the yellow and orange \nsquares show the locations of all benthic and pelagic gill nets used for the lake -wide fishing \ncampaign. Crosses in the insert map show the sampling locations of larva e in Upper Lake \nConstance. More Details can be found in Tables S1-S3. \n \n \n \nPhenotypic analysis  \nTo identify phenotypic differences between whitefish species in Lake Constance, we performed \nlandmark-based body shape analyses and analyses of linear measurements (see Supplementary \nMethods for details). We compared body shape among the genetically identified whitefish species \nusing 17 homologous landmarks (Fig. S1 A, Table S5) using the geomorph V4.0.10 (Baken et al., \n2021). Generalized Procrustes Analysis was performed to remove size, position and orientation \ndifferences, and individuals with abnormal body shape (e.g. due to body arching) were identified \nusing the plotOutliers function and excluded if necessary, leaving 158 individuals (Table 1). We \ntested for differences in shape between species using a permutation -based procrustes ANOVA \nwith residual randomization with the following model: coords ~ log(size) * species. We did  not \nperform a size correction, as allometric ef fects were different across species. Instead, we post -\nhoc compared species using the RRPP package with the null model: coords ~ log(size) to account \nfor allometric effects (Collyer & Adams, 2018)  and adjusted p -values were adjusted using the \nBonferroni-Holm method. Shape differences were illustrated using a PCA in geomorph and \nwireframe models. A canonical variate analysis (CVA) was used to examine the accuracy of species \nclassification based on body shape. \nFurthermore, we investigated phenotypic differences using 15 linear measurements for the \nsame 158 individuals (Fig. S1B, Table S6) (see Supplementary methods). Highly correlated linear \nmeasurements (spearman rank coefficients >0.9) were excluded, and we tested different size \ncorrection approaches. We used Recursive Feature Elimination (RFE) to identify linear \nmeasurements that disti nguished the species. Measurements with the highest classification \naccuracy were used for PCA in MorphoTools2 V1.0.2.1 (Šlenker et al., 2022) and we performed a \nK-nearest-neighbor classificatory discriminant analysis to estimate the accuracy of linear \nmeasurements for species classification in MorphoTools2. \n \n \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 8 \nWhole genome sequencing  \nDNA was extracted from fin clips for larval (n=744) and adult (n=239) samples using a magnetic \nbead-based protocol ( dx.doi.org/10.17504/protocols.io.b46bqzan), concentrations quantified \nusing a Qubit Flex (1X BR dsDNA assay), and quality determined using agarose gels. Whole \ngenome sequencing (WGS) libraries were prepared using a modified HackFlex protocol (Gaio et \nal., 2022) (see Supplementary File 1 for protocol) from 10ng of DNA per individual. Compared to \nthe original protocol, we reduced the amount of Dimethylformamide (DMF) in the tagmentation  \nbuffer to 20% from 50% to minimise the amount of this highly toxic chemical without impacting \nlibrary quality, and we performed a double-sided bead clean up on the final libraries. Final libraries \nwere quantified using the Qubit Flex (HS dsDNA assay) and the fragment size distribution for a \nsubset was checked on the Tapestation (D5000 assay), with an expected peak around 400 -\n600bp. We pooled libraries at equal molarities. We sequenced up to 288 libraries per pool using \n150PE reads on NovaSeq X Plus 10B lanes at Novogene UK to an average of 1.6 Gb per individual. \nSome individuals were sequenced on an additional NovaSeq X Plus 25B lane to increase coverage. \nFurthermore, we downloaded public whole genome data for adult individuals with confirmed \nspecies assignment from ENA for Gangfish, Sandfelchen, and Blaufelchen (Frei et al. 2023, 2022) \n(Table S4).  \n \nBioinformatic processing and variant identification \nRaw sequencing reads were processed following best practices for lcWGS data (Lou et al., 2021; \nLou & Therkildsen, 2022) (see Supplementary methods for detailed parameters) using fastp (Chen \net al., 2018) . Processed reads were mapped to the European whitefish (‘Balchen’) reference \ngenome (AWGv2; (De-Kayne & Zoller, 2020) ) using bwa mem  (Li & Durbin, 2009) , replicated \nsamples merged and duplicated reads removed using sambamba v.0.8.2 (Tarasov et al., 2015) , \nand overlapping reads clipped using the bamUtils clipOverlap  \n(https://github.com/statgen/bamUtil). Bam files were indexed using sambamba before indel \nrealignment using GATK3.8. We estimated the depth of coverage per bam file using mosdepth \nv.0.3.11 (Pedersen & Quinlan, 2018). \nDue to the low coverage of the sequencing data, we performed genotype likelihood-based \n(GL) analyses with ANGSD v0.938 (Korneliussen et al., 2014) . We created two different SNP \ndatasets to address distinct questions. For the first list, we identified SNPs across all 250 adult \nsamples (‘adult SNP dataset’) using the samtools genotype likelihood model, a minimum SNP \nvalue of 1e-6 and determined major/minor alleles based on allele frequencies. We removed sites \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 9 \nbelow a total sequencing depth of 250 (number of individuals x mean sequencing depth) and \nabove 1300 (2 x number of individuals x mean sequencing depth), and removed sites with missing \ndata in more than 20% of individuals, minimum base quality below 30, a minimum individual \ndepth below 1, multi-mapping reads, unmapped or duplicated reads and those without a matching \npair. We also adjusted the mapping quality for excessive mismatches and only kept SNPs with a \nminor allele frequency of 5%. Lastly, we removed SNPs falling within potentially problematic \ngenomic regions that can impact analyses accuracy. We removed SNPs with excess heterozygosity \npotentially due to collapsed paralogs, estimated using the ngsParalog calcLR (Dallaire et al. 2023), \nexcluded all sites within likely collapsed regions in assembly (Frei et al. 2023), and removed sites \nwithin regions without unique mappability (Wang et al. 2024), as inferred using genmap (with 150-\nmers) (Pockrandt et al. 2020).  \nFor the second list, we created an additional SNP dataset for a subset of adult reference \nsamples (published data and individuals from targeted sampling) that were used for species \nassignments of larvae (‘larval SNP dataset’). We identified variant sites a cross adult reference \nsamples (Table S2) using the same filtering strategy as above using adjusted filtering values. Since \nlarvae had ultra-low sequencing coverage, we did not use them for SNP identification but instead \ninferred genotype likelihoods for larvae based on the filtered ‘larval SNP dataset’ with the -sites \nfunction We did not apply any other filters at this step, such as depth filters, due to the very low \ncoverage (DeSaix et al. 2024).  \n \nPopulation structure \nWe investigated population structure across adults using PCAngsd (Meisner & Albrechtsen, 2018), \nusing published reference individuals to identify genetic clusters corresponding to species, refine \nthe species-assignment of adults caught during the targeted fishing campaign based on their \nclustering PCA space, and assign individuals caught across the  upper and lower lake in 2024 to \ntheir respective species. Across adult and larval samples, we estimated the pairwise IBS matrix \nacross all individuals based on randomly sampled single reads for each site ( -doIBS 1) for MDS \nanalysis. The random sampling approach is less biased to variation in coverage across samples \n(Lou et al., 2021)  and better suited for samples with ultra -low coverage and varying sequencing \ndepth. Processing and plotting of PCA and MDS results were performed in R.   \nWe investigated the correlation between morphology (based on landmarks and linear \nmeasurements) with genetic variation to determine if genetically intermediate individuals are also \nphenotypically intermediate. We used composite principal component scores across PC1 and PC2 \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 10 \nas a measure for genetic and phenotypic variation (Verta & Jones, 2019), which we estimated for \nthe genetic PCA, landmark-based PCA and linear-measurement PCA independently, by summing \nup PC1 and PC2, weighting each by their eigenvalues. We used linear mixed models ( lmer in the \nlme4 R-package) to estimate the correlation between genetic variation and phenotypic variation, \nwith species as a random factor.  \n  \nLarval Species assignment  \nWe assigned larvae to their respective species using a probabilistic framework in \nWGSassign (DeSaix et al. 2024). Assignment accuracy is heavily biased by differences in sample \nsize and sequencing depth between reference populations (DeSaix et al., 2023, 2024) . To \nminimise biases, we normalised the effective sample size (ESS) of reference populations by \nsubsampling an equal number of adult individuals per species with similar sequencing depths \n(DeSaix et al., 2024). We tested our assignment accuracy based on adult samples of known origin \nthat were not part of the reference populations and used sambamba view to subsample four test \nsamples to mean and minimum sequencing depths observed for larval samples (0.5X, 0.1X and \n0.01X) to test the effect of sequencing depth on assignment accuracy. As we had 100% accuracy \nfor downsampled samples, we assumed similar assignment accuracy for larvae. Since admixed \nindividuals have mosaic genomes that cannot be completely assigned to one species, and \nbecause posterior probabilities of assignment are suggested to be unreliable for lcWGS data, we \nsplit up the SNP dataset into 10 equal sets of ~240,000 SNPs, performed the species assignment \nfor each subset separately, to estimate the consistency of assignment across sections of the \ngenome (DeSaix et al., 2023, 2024). We determined an assignment as accurate if 8 of 10 genomic \nsubsets (0.8) assigned an individual to the same species. \n \nPopulation genomics adults \nTo investigate the contemporary genomic diversity of whitefish in Lake Constance, we performed \nGL-based population genomic analyses in adult whitefish. First, we investigated patterns of \nLinkage disequilibrium (LD) decay across the genome by species using ngsLD (Fox et al., 2019). \nWe estimated LD between all SNPs within 4Mb and used these LD estimates to infer patterns of \nLD decay and reconstruct recent population trends over the last few hundred generations in SNeP \n(Barbato et al., 2015). Second, we inferred genome-wide genetic diversity (nucleotide diversity [π] \nand Tajima’s D) across all filtered sites (variant and invariant) using ANGSD, compared levels of \nπ between species using an ANOVA and Tukey’s Posthoc test in R, and tested if levels of Tajima’s \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 11 \nD were different from zero using a Wilcoxon rank sum test. Third, we inferred the landscape of \ndifferentiation and selection between species across the genome. We estimated Fst between all \nspecies pairs for SNPs and in 50kb sliding windows with 25kb steps (based on LD decay rates) \nusing ANGSD and realSFS. We defined outlier regions as Fst values above the 99% of the Fst \ndistribution. Furthermore, we estimated the population branch excess statistic (PBE) for each \nspecies from windowed Fst data (see Supplemen tary methods for details). We identified outlier \nwindows as those with PBE values in the top 0.1% of the genome -wide distribution. Lastly, we \nidentified genomic outlier regions indicative of low recombination regions and structural variants \nusing a modified LOSTRUCT approach (Li & Ralph, 2019) (see Supplementary Methods). \n \nResults \n  \nPopulation structure and species composition  \nUsing the adopted HackFlex protocol, we cost -effectively generated low whole -genome \nsequencing data for 239 adult whitefish (mean depth of coverage = 2.6X; range 0.01X to 8.83X ; \nFig.S2), moderate PCR-duplications rate (mean=8%, range=1.7%-11.1%), and identified 4,206,794 \nfiltered SNPs across all adults. The PCA separated the three species in ULC into three clusters \nalong PC1 and PC2 (Fig. 2A, Fig.S 3-S4). However, 15 adults that caught during the targeted \ncampaign on spawning grounds were assigned to the wron g species in the field, with 13 \nBlaufelchen assigned as Gangfisch and two Gangfisch as Blaufelchen (Fig.S 3). Species \nassignments were corrected based on the genomic data. Unterseefelchen from Lower Lake \nConstance were genetically indistinguishable from Sandfelchen. The single individual caught in \nthe rhine inflow (‘Alpenhein’) genetically clustered with Sandfe lchen in PCA space. Furthermore, \nthe species assignment of the 105 adults caught during the lake-wide survey in 2024 identified 7 \nBlaufelchen (6.7%% o f individuals) and 63 Gangfisch (60%). As we couldn’t distinguish \nSandfelchen and Unterseefelchen, we assigned individuals from that genetic cluster caught in \nULC as Sandfelchen (12; 11.4%) and those caught in LLC as Unterseefelchen (23; 21.9%).  \n \nLow genetic diversity and population declines \nEstimates of genetic diversity and population history indicated low levels of genetic \ndiversity across all species, with significant differences between species (ANOVA: F(3)=1697, p<2e-\n16), and values of π ranging from 0.00230 Blaufelchen to 0.00234 in Sandfelchen and 0.00237 \nUnterseefelchen and 0.00282 in Gangfisch (Fig.2B). Genome -wide levels of Tajima’s D were \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 12 \nskewed toward positive values in all populations and statistically different from zero (Wilcox test, \np<0.05), indicative of sudden population contraction (Fig. 2C).   \nLD-based analyses indicated declines in effective population sizes (‘Ne’) over the last \n1,000 generations in all species (Fig. 2D,E). Blaufelchen showed the lowest contemporary Ne of \n62, the lowest levels of genetic diversity and highest levels of LD (Fig.  2D,E). In contrast, \nSandfelchen and Unterseefelchen had estimated Ne’s of 144 and 169, respectively, and Gangfisch \nhad the highest estimated Ne of 255. While Blaufelchen showed a continuous steep decline in Ne \nover the last 1,000 generations, the other species showed steeper declines in Ne in the last 100 \n- 200 generations (Fig.2D).   \n \n \nFigure 2 - Genetic diversity and population history. (A) Principal components analysis showing the \npopulation structure of adult whitefish in Lake Constance. Individuals are coloured by species and shapes \nare filled in by catch location. The shape represents the dataset. See legend for details. ( B, C) Nucleotide \ndiversity (B) and Tajima’s D (C) differed between species in Upper Lake Constance but not between \nSandfelchen and Unterseefelchen. Statistically different groups are highlighted by letters above the  violin \nplots. See text for statistical results. ( D) Changes in effective population size (Ne) over the last 1,000 \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 13 \ngenerations by species as estimated from linkage disequilibrium (LD) data with SNeP. ( E) LD decay by \nspecies as estimated with SNeP. \n \n \nLandscape of genetic differentiation and selection \nGenome-wide pairwise Fst between species was overall low, ranging from 0.006 between \nSandfelchen and Unterseefelchen to 0.043 between Blaufelchen and Unterseefelchen (Fig. 3A, \nFig. S5, Table S7). Genome-wide scans of differentiation revealed peaks of elevated Fst between \nspecies across the genome (Fig.2A, Fig.S 5), although overall Fst values were still low (< 0.5). \nComparing outlier Fst windows (Fst > 0.99%) to known adaptive trait QTL, we detected elevated \ndifferentiation Gangfisch and Sandfelchen around t he edar gene on chr23, which has been \nassociated with variation in gill raker count in the Alpine whitefish radiation (Fig.3B) (De-Kayne et \nal., 2022). These species are known to differ in gill raker count, with Sandfelchen having fewer gill \nrakers than the other species (Hirsch et al., 2013; Vonlanthen et al., 2012) . The same region did \nnot show strong differentiation between the other whitefish species but showed signatures of \nselection in Blaufelchen based on population branch excess (PBE > 0.999%; Fig.3B, Fig.S 6). \nGenomic signatures of selection across the genome were species-specific, without shared outlier \nwindows (PBE > 0.999%) across species (Fig.S 6). Furthermore, local PCA identified 83 outlier \nregions across the genome (>4 SD from the mean per MDS axis), which showed distinct patterns \nof population structure compared to the rest of the genome, potentially due to reduced \nrecombination rates and/or structural genomic variation (Li & Ralph, 2019) (Fig. S7 and S8). Of \nthese local PCA outlier windows, between 31 and 143 overlapped with Fst outlier windows in at \nleast one comparison, including the Fst outlier region on chr23 associated with gill raker count. \nMost outlier regions showed PCA patterns indicative of low -recombination regions, i.e. without \nthree distinct clusters, except the outlier regions on chr8 and chr24 (Fig. S9) (Mérot et al., 2021). \nHowever, these regions did not distinguish species and did not overlap Fst outlier regions. \n \n \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 14 \n \nFigure 3 - Genetic differentiation (A) Manhattan plots showing the levels of genetic differentiation in \n50kb sliding windows (25kb steps) across the genome between species in Upper Lake Constance. SF = \nSandfelchen; BF = Blaufelchen; GF = Gangfisch.  (B) The plot on the left shows SNP-based Fst along chr23, \nhighlighting increased differentiation between SF and GF around a SNP previously associated with variation \nin gill raker count in European whitefish. The plot on the right shows the PBE value in Blauf elchen for the \nsame genomic region, with the gill raker count associated region highlighted in red. \n \n \nPhenotypic results \nWe assessed phenotypic variation between adult whitefish based on genetic species assignment \n(Table S8). Body size had a statistically significant, species-dependent effect on body shape based \non landmarks (Z [effect size] = 1.790, p = 0.038), with shape differing between species (Z = 1.687, \np = 0.046). Shape differed between Blaufelchen and Sandfelchen (Z = 2.528, p = 0.024), \nBlaufelchen and Unterseefelchen (Z = 4.014, p = 0.001), and Gangfisch and Unterseefelchen (Z = \n3.187, p = 0.006) (Fig.4A). However, there was considerable overlap in shape between species \n(Fig.4A), with changes in body shape along PC1 primarily being explained by the degree of ventral \ncurvature and head position and changes along PC2 mainly being confined to body height and \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 15 \nhead length (Fig . 4B). Overall classification accuracy was 65.8% (Table S7), with the highest \naccuracy for Unterseefelchen (75.0%) and the lowest for Blaufelchen (33.3%). \nFor linear measurements, we excluded PostA and PostD due to high correlations of PreA \nand PostA (r = -0.91) and PreD and PostD (r = -0.94) (Fig.4C,D) and we identified the following \nCD, BD, SN, PrePc, PreD, ED, HL, PostO, PrePl as optimal features for dist inguishing whitefish \nspecies (Fig.4E). For the PCA, we only used the selected measurements listed above, and similar \nto the landmark data, species showed a considerable overlap in phenotype (Fig . 4C). Species \nseparated largely along PC1, which was explained by variation in Caudal peduncle depth (CD) and \nbody depth (BD) (Fig. 4D). The importance of CD and BD for species differentiation was supported \nby mean decrease accuracy values derived from the RFE (Fig. 4D). Overall classification accuracy \nbased on K nearest neighbour classificatory discriminant analysis (K = 7) was 62.7% (Table S9), \nwith the highest accuracy again for Unterseefelchen (83.6%) and lowest for Blaufelchen (0%). \nPhenotypic variation was significantly correlated with genetic variation (compound PC1 + \nPC2) for landmarks (Linear mixed model: beta = 1.46, 95% CI [0.70, 2.22], t(137) = 3.81, p < .001), \nbut not for linear measurements (Linear mixed model: beta = -22.46, 95% CI [ -103.70, 58.78], \nt(137) = -0.55, p = 0.585) (Fig.S10). Thus, overall body shape seems to vary with genetic variation, \nwith admixture potentially leading to intermediate body shapes.  \n \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 16 \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \nFigure 4 - Phenotypic variation (A) Scatter plot showing the first two principal components (PCs) of the \nphenotypic variation of identified whitefish species, based on 17 homologous landmarks. The first two PCs \nare explaining 44.8% of the variance of the data. Each point represents one ind ividual, which is coloured \nbased on their genetic assignment. (B) Wireframe graphs of the shape comparing the extremes along the \nPC axes in the landmark PCA. ( C) Scatter plot showing the first two principal components (PCs) of the \nphenotypic variation of identified whitefish species, based on selected linear measurements. The first two \nPCs are explaining 63.2% of the variance of the data. (D) Loading plot for the linear measurement PCA. (E) \nMean decrease accuracy values based on a Random Forest model, showing importance of selected linear \nmeasurements for species differentiation. \n \n \nSpecies assignment of whitefish larvae \nWe sequenced 744 larvae to a mean depth of 0.5X (range 0.01X to 3.55X; Fig.5A) and inferred \ngenotype likelihoods at a maximum of 2,446,925 SNPs Due to the low sequencing depth and \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 17 \nbreadth, larvae had on average sufficient sequencing data at 38% of all SNPs (ranging from 0.7% \nto 95%), which is sufficient for inferring population structure and population assignment (DeSaix \net al., 2024). In comparison, adults covered ~90% of all sites. Standardisation of effective sample \nsize (ESS) for each reference group resulted in 10 individuals per species, with ESS from 6.5 to \n6.9 (Fig.5B). The assignment accuracy of test samples of known species w as 100% for all \nsubsampled coverages (0.5x, 0.1x and 0.01x) with 100% assignment consistency across all ten \nSNP subsets, indicating sufficient power to accurately assign larvae to their species down to 0.01x \ncoverage. Species assignment of larvae using lcW GS showed that 91.3% of individuals were \nGangfisch (679 out of 744), with 7.3% being assigned as Blaufelchen (54 out of 744), and only \n1.5% were assigned as Sandfelchen (11 out of 744) (Fig.5B). Gangfisch had a marginally lower ESS \n(6.5) compared to Blaufelchen (6.9) and Sandfelchen (6.8), suggesting that this large assignment \nbias between species is not driven by differences in ESS (DeSaix et al. 2023). Three individuals \nwere not consistently assigned to the same species, with assignment consistencies below 0.8. Of \nthese, two individuals showed assignment consistencies of 0.5 to Gangfish and Blaufelchen, \nsuggesting recent admixture of these individuals. T he third individual showed an assignment \nconsistency of 0.7 to Gangfisch and 0.3 to Blaufelchen. These three individuals were also \nintermediate in their position in the MDS plot between Blaufelchen and Gangfisch, suggesting \nthat these are potentially recen t hybrids (Fig.5A). Like the species assignment, most larvae \nclustered genetically with adult Gangfisch in the MDS plot (Fig.5A). Larval whitefish showed a \nbroader spread in the genomic space compared to adults, with many individuals clustering \nintermediate between species, especially Gangfisch and Blaufelchen. While this could potentially \nbe explained by the lower coverage for many samples, we found that downsampled test samples \nwith coverages down to 0.01x still clustered closely with the other adult samp les (Fig.S11), \nsuggesting that coverage alone does not explain the broader spread of larval samples in the MDS \nspace. \n  \n \n \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 18 \n \n \nFigure 5 - Larval assignment. (A) MDS plot based on filtered SNPs with larvae (crosses) coloured by \nspecies assignment. Reference individuals are shown as dots. Three larval individuals with inconsistent \nassignment (consistency < 0.8) are highlighted as magenta diamonds. (B) Species assignment results for \nlarvae. The upper panel shows the reference populations with their respective effective sample sizes (ESS). \nThe large grey circle shows the total number of larvae, and the lower coloured circles show the number of \nlarvae as signed to each s pecies, coloured by species. Circle sizes are proportional to the number of \nindividuals assigned. Numbers in brackets show the number of individuals with assignment consistencies \nbelow 0.8. \n \n \nSpatial distribution of species in Lake Constance \nTo determine how species are spatially distributed across Lake Constance at different life stages, \nwe investigated differences in spatial and depth distribution across species and lake parts. Almost \nall adult Gangfisch and Blaufelchen were caught in the Up per Lake, yet one Blaufelchen and ten \nGangfisch were caught in the lower lake. We couldn’t genetically distinguish Sandfelchen and \nUnterseefelchen, suggesting that Sandfelchen and Unterseefelchen are one population (Fig. 6A). \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 19 \n The depth distribution of adult whitefish caught during the lake -wide fishing campaign \nsuggests the presence of depth preferences between species (Fig.6B). The majority of Blaufelchen \nwere caught in deeper nets between 35 and 50m (71.4%; n=5), with smalle r proportions in 20 -\n35m (14.3%, n=1) and 6 -12m (14.3%, n=1). The majority of Gangfisch (60.3%; n=38) and \nSandelchen (83,3%; n=10) were caught between 20 -35m, although Gangfisch seemed to be \npresent across the entire depth distribution down to 150m (n=1 Gangfisch in a deep net).  However, \nthis could potentially be explained by the larger sample size of Gangfisch. Unterseefelchen, which \nare genetically Sandfelchen, were mostly caught in shallow pelagic nets below 6m (43.5%; n=10) \nand between 20-35m (26.1%; n=5). Similar to the adults, we detected subtle differences in habitat \ndistribution between species as larvae. While most larvae for all species were caught in the \nshallow-littoral zone (43.4% to 79.1% of larvae per species; Fig. 6C), a higher proportion of \nBlaufelchen and Sandfelchen larvae (35.8% and 36.4%, respectively) were present in the \nnearshore pelagic zone compared to Gangfish larvae (8.1%). While we did not detect any \nSandfelchen larvae in deeper pelagic or littoral zones, we found both Blaufelchen a nd Gangfisch \nin these zones. These results suggest that Sandfelchen larvae are largely present in the shallow, \nnearshore zone, while both Blaufelchen and Gangfisch are present throughout all zones. However, \nit has to be noted that sampling bias can also impact these results, as most larvae were caught in \nthe shallow littoral zone (72% of all larvae, n=531).  \n \n \n \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 20 \n \n \nFigure 6 - Habitat usage of adult and larval Lake Constance whitefish. (A) Spatial distribution of adult \nwhitefish caught during the lake -wide fishing campaign in September 2024. Points show the location of \nindividual nets, with the size indicating the number of individuals for each net and the colour the genetic \nspecies assignment. (B) The proportion of adults caught during the lake-wide fishing campaign in 2024 in \nnets set in different depth zones and different locations/habitats. ( C) The proportion of larvae caught in \ndifferent habitat zones by species. The numbers below the abbreviated species names in (B) and (C) show \nthe sample size. BF = Blaufelchen, GF = Gangfisch, SF = Sandfelchen, UF = Unterseefelchen.   \n \n \n \n \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 21 \n \nDiscussion \nHere we use large-scale low- to ultra-low coverage whole-genome sequencing of adult and larval \nLake Constance whitefish to characterise the contemporary genetic diversity, population trends, \nand species composition across entire Lake Constance at unprecede nted spatial and genomic \nresolution. Beyond European whitefish, our results provide a general framework for genomic \nmonitoring of fisheries-relevant populations with large population sizes and ongoing gene flow. \nBy adapting cost-effective sequencing and analytical approaches for vertebrate-sized genomes, \nwe demonstrate how population assignment, species composition, and genomic diversity can be \nassessed simultaneously, even under weak genetic differentiation. Our findings have important \nimplications for current management practices in Lake Constance, as discussed below. \n  \nSuitability of lcWGS for fisheries management \nGenetic stock assignment remains challenging in systems with strong gene flow and \nlimited genotyping tools (Benestan et al., 2015; Bernos et al., 2024; DeSaix et al., 2019; Theissinger \net al., 2023). Reduced-representation sequencing approaches have been widely adopted but are \nlabour-intensive, target only a small fraction of the genome, and can be difficult to integrate across \nstudies, limiting their value for long-term monitoring (Theissinger et al., 2023). Here, we show that \n(ultra-)low coverage WGS enables accurate species assignment, spatial analyses, and genome -\nwide inference of diversity and differentiation in weakly structured whitefish populations. Despite \nlow genome -wide differentiation, individuals  could be reliably assigned to species down to \nsequencing depths of 0.01×, despite being genotyped at only ~1% of total SNPs (DeSaix et al., \n2023, 2024). This was enabled by a rapid and cost-effective library preparation protocol (Gaio et \nal., 2022), which we adapted to large genomes ( Coregonus spp. genome: ~2.5Gb) (De-Kayne et \nal., 2020). This enabled us to prepare WGS libraries for around £2.5 per library (after some minor \nstart-up costs) in less than 1 day (for 96 samples) and sequence them to 1x coverage for \napproximately £7.5 per sample (sequencing cost in 2024), making it significant ly cheaper and \nquicker than most reduced-representation sequencing approaches (e.g. RADseq) (Lou et al., 2021). \nImportantly, the use of whole -genome data without targeted markers facilitates seamless \nintegration of future samples, as demonstrated by combining our dataset with previously \npublished whitefish genomes  (Frei et al., 2022) . These results highlight the broad potential of \nlcWGS for fisheries management, including in highly connected systems, such as hybridizing \nspecies or marine organisms (Jacobs et al., 2018; Quintela et al., 2020; Seljestad et al., 2024).  \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 22 \n \nGenetic mixing of species across habitats and life-stages \nEffective fisheries management, including hatchery supplementation, relies on the \naccurate species identification of spawning adults, often inferred from spawning location or \nmorphology in the field. However, environmental change can erode prezygotic barriers and \nundermine such assumptions, as previously shown for Lake Constance whitefish (Vonlanthen et \nal. 2012; Frei et al. 2023; Frei et al. 2022; Eckmann and Rösch 1998; Numann 1978) . Targeted \nsampling revealed extensive mixing of Blaufelchen and Gangfisch on traditional spawning \ngrounds, which are often used to determine the species during broodstock fisheries, which has \ntherefore strong implications for hatchery practices that aim to create species-pure crosses (Baer \net al., 2023) . Notably, ~70% of individuals caught on Blaufelchen spawning grounds were \ngenetically Gangfisch, raising the possibility that hatchery supplementation unintentionally \ncontributes to genetic mixing between species. Larval data further support this interpre tation, \nshowing nearly continuous genetic variation between Blaufelchen and Gangfisch, particularly at \nearly life stages. The reduced frequency of genetically intermediate adults relative to larvae \nsuggests possible selection against admixed individuals later in life (‘viability selection’) (Blain et \nal., 2024; Moser et al., 2016; Schluter et al., 2025), although targeted experimental and longitudinal \nstudies are needed to confirm this. Viability selection against strongly admixed individuals could \nbe one explanation for the maintenance of distinct whitefish species despite strong gene flow.  \nDespite historical stocking of non -endemic species, we found no evidence for genetic \ncontributions from such introductions, consistent with previous work in Lake Constance (Baer, \nSchliewen, et al., 2022; Frei, Mwaiko, et al., 2023) . Contemporary samples consistently cluster by \nspecies, indicating that current genetic patterns largely reflect endemic lineages. \n Species misidentification is further compounded by weak morphological differentiation. \nTraditional phenotypic traits have long been recognised as unreliable, particularly between \nBlaufelchen and Gangfisch since eutrophication (Eckmann & Rösch, 1998; Hartmann & Knöpfler, \n1977; Numann, 1986) . Our results confirm that even advanced morphometric analyses fail to \nreliably distinguish these species at the individual level, and that body shape correlates with \ngenetic admixture. While gonadal maturity and life-stage differences likely influence body shape \n(Helland et al. 2009), we could still distinguish Sandfelchen/Unterseefelchen from Gangfisch, \nsuggesting that morphology reflects broad habitat use in some instances (Etheridge et al., 2012; \nHarrod et al., 2010). Overall, morphology and spawning habitat alone are currently unreliable for \nspecies identification in Lake Constance, underscoring the need for genomic tools and analyses.  \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 23 \n \nSkewed stock composition and habitat use \nOur lake-wide species assignment of adults and larvae revealed a highly skewed stock \ncomposition, with Gangfisch comprising ~87% of individuals. This dominance may reflect the \nbroader ecological niche and extended spawning period of Gangfisch, which could buffer \nrecruitment against unfavourable conditions (Frei, Reichlin, et al., 2023; Jacobs et al., 2019; \nSchweizer, 1894; Steinmann, 1950) . In contrast, Blaufelchen - the primary commercial target - \naccounted for only ~7% of sampled individuals, consistent with recent stock collapses (Haase et \nal., 2025). Although Sandfelchen were rare in Upper Lake Constance (~3%), all individuals in Lower \nLake Constance clustered genetically with Sandfelchen with minimal genetic differentiation, \nindicating that the Sandfelchen/Unterseefelchen form a single population, a nd should likely be \nmanaged as a single stock. Furthermore, the identification of a Sandfelchen individual in the \nAlpenrhein further supports regular river use by this species (Steinmann, 1950). Interestingly, \nBlaufelchen and Gangfisch were also detected in Lower Lake Constance during lake-wide survey \nin 2024, although their spawning there remains unresolved. \nHabitat use seemed to differ across species and life stages, at least during the snapshot \nthat the sampling across a few timepoints provided. Adult distributions during feeding (September \n2024) were more complex than traditional classifications suggest (Steinmann, 1950), with the \npelagic-planktivorous Blaufelchen frequently caught in deeper benthic nets and the littoral \nSandfelchen in pelagic nets at medium depth. Gangfisch consistently occupied the broadest \ndepth range, also below 100m, which is consistent with the hypothesised niche expansion due to \nintrogression from the extinct Kilch (Frei et al., 2022; Frei, Reichlin, et al., 2023; Hirsch et al., 2013; \nJacobs et al., 2019). However, to fully understand contemporary habitat use differences between \nspecies it will be crucial to follow genetically-confirmed individuals throughout the year, e.g. using \nacoustic telemetry, , as habitat use can differ throughout and across years (Thomas et al. 2010).  \nWhile larvae were mostly present in the shallow nearshore habitat, larval distributions \naligned more closely with expected spawning habitats (Steinmann 1950), although sampling bias \n- particularly limited coverage of deep pelagic zones - may underestimate the abundance of \nBlaufelchen larvae. Larval Blaufelchen are thought to prefer deep pelagic habitats due to the lack \nof anti-predator mechanisms, which are less important in the deep pelagic zone (Ros et al., 2019). \nHowever, the low proportion of adult Blaufelchen in the catch still suggests a low census \npopulation size.   \n \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 24 \n \nGenetic diversity and genomic conservation tools \nConsistent with previous work, we detected low effective population sizes and low genetic \ndiversity (Robinson et al., 2016) across all species, with pronounced declines in Blaufelchen (Frei, \nMwaiko, et al., 2023). While lower sample sizes for Blaufelchen could influence estimates, sample \nsizes likely reflect true census abundances (Haase et al., 2025). Given the importance of neutral \nand adaptive genetic diversity for population recovery and resilience (Exposito-Alonso et al., 2022; \nKardos et al., 2021; Steffen et al., 2015), these findings raise concerns about the capacity of Lake \nConstance whitefish - particularly Blaufelchen - to respond to future environmental change.  \nDespite extensive genome -wide homogenisation, we identified genomic regions with \nelevated differentiation, including a region on chr23 overlapping the edar gene associated with \ngill raker variation (De-Kayne et al., 2022). Although allele frequencies have converged over time \n(Frei, Mwaiko, et al., 2023), this region remains differentiated between Sandfelchen and Gangfisch \nand shows signatures of selection in Blaufelchen, indicating the persistence of functional genomic \nvariation. Many differentiated regions overlapped putative low-recombination regions rather than \nstructural variants, suggesting that reduced recombination facilitates the persistence of adaptive \ndivergence despite gene flow. Further work is needed to resolve the genomic architecture \nunderlying adaptive variation in this system (De-Kayne et al., 2022; Frei et al., 2022; Frei, Reichlin, \net al., 2023; Jacobs et al., 2019), which can form the basis for novel genomic monitoring tools of \nadaptive genetic variation and contribute to effective management decisions (Garner et al., 2016; \nHendricks et al., 2018; Pearse, 2016).  \n \nConclusions & Limitations \nBy integrating lake -wide genomic data across life stages and lake basins, we show that Lake \nConstance whitefish stocks are highly skewed toward one species, show mixing of species on \nspawning grounds, and characterised by low genetic diversity and differentiation between species. \nImportantly, we show that the commercially important Blaufelchen makes up only a small \nproportion of the stock. These findings have direct implications for current hatchery practices and \nmanagement strategies, which rely on accurat e species identification. To fully understand the \nimpact of mixed spawning grounds on hatchery supplementation and natural admixture, future \nstudies should directly compare the genetic composition of hatchery and wild juveniles. While our \nlow-coverage approach limits inference of fine -scale inbreeding patterns, it provides a powerful \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 25 \ntool for lake -wide genomic monitoring of stock composition. Overall, our results call for a re -\nevaluation of current fisheries management strategies in Lake Constance. \nAcknowledgements  \nThis project was funded by the Ministry of Food, Rural Area, and Consumer Protection (Baden \nWurttemberg; Germany) to AB. AJ was supported by a NERC Independent Research Fellowship \n(NE/W008963/1). \n \nAuthor contributions \nConceptualisation: AB and AJ; Sample collection: AB and BR; Laboratory work: AJ and MC; \nGenomic Analysis: AJ; Phenotypic analysis: SR; Writing first draft: SR and AJ; Final draft: All \nauthors.  \n \nConflicts of interest \nThe authors declare no conflicts of interest. \n \nData availability statement \nAll data will be made available upon acceptance of the manuscript. The raw sequencing files are \naccessible on ENA under the BioProject XXXXX. Additional data are available on xxxx.   \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 26 \nReferences: \nAlexander, T. J., Vonlanthen, P., & Seehausen, O. (2017). Does eutrophication -driven evolution \nchange aquatic ecosystems? Philosophical Transactions of the Royal Society of London. \nSeries B, Biological Sciences, 372(1712). https://doi.org/10.1098/rstb.2016.0041 \nBaer, J., Eckmann, R., Rösch, R., Arlinghaus, R., Alexander, & Brinker. (2016). Managing Upper Lake \nConstance fishery in a multi-sector policy landscape : Beneficiary and victim of a century of \nanthropogenic trophic change . In Inter-sectoral governance of inland fisheries. TBTI \nPublication Series,  E-01/2017. Too Big To Ignore -WorldFish, St. John’s, Canada.  Edited \nby A.M. Song, S.D. Bower, P. Onyango, S.J. Cooke, R. Chuenpagdee. pp. 32–47. \nBaer, J., Kugler, M., Schubert, M., Schotzko, N., Rösch, R., Vonlanthen, P., & DeWeber, J. T. (2023). \nA matter of time —Efficacy of whitefish stocking in a large pre -alpine lake. Fisheries \nManagement and Ecology, 30(6), 615–626. \nBaer, J., Schliewen, U. K., Schedel, F. D. B., Straube, N., Roch, S., & Brinker, A. (2022). Cryptic \npersistence and loss of local endemism in Lake Constance charr subject to anthropogenic \ndisturbance. Ecological Applications: A Publication of the Ecological Society of America , \ne2773. \nBaer, J., Spiessl, C., Auerswald, K., Geist, J., & Brinker, A. (2022). Signs of the times: Isotopic \nsignature changes in several fish species following invasion of Lake Constance by quagga \nmussels. Journal of Great Lakes Research, 48(3), 746–755. \nBaken, E. K., Collyer, M. L., Kaliontzopoulou, A., & Adams, D. C. (2021). geomorph v4.0 and gmShiny: \nEnhanced analytics and a new graphical interface for a comprehensive morphometric \nexperience.  Methods in Ecology and Evolution, 12(12), 2355–2363. \nBarbato, M., Orozco -terWengel, P., Tapio, M., & Bruford, M. W. (2015). SNeP: a tool to estimate \ntrends in recent effective population size trajectories using genome-wide SNP data. Frontiers \nin Genetics, 6, 109. \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 27 \nBeemelmanns, A., Bouchard, R., Michaelides, S., Normandeau, E., Jeon, H.-B., Chamlian, B., Babin, \nC., Hénault, P., Perrot, O., Harris, L. N., Zhu, X., Fraser, D., Bernatchez, L., & Moore, J.-S. (2024). \nDevelopment of SNP panels from low -coverage Whole Geno me Sequencing (lcWGS) to \nsupport Indigenous fisheries for three Salmonid species in Northern Canada. Molecular \nEcology Resources, e14040. \nBenestan, L., Gosselin, T., Perrier, C., Sainte -Marie, B., Rochette, R., & Bernatchez, L. (2015). RAD \ngenotyping reveals fine -scale genetic structuring and provides powerful population \nassignment in a widely distributed marine species, the American lobster  (Homarus \namericanus). Molecular Ecology, 24(13), 3299–3315. \nBernatchez, L., Wellenreuther, M., Araneda, C., Ashton, D. T., Barth, J. M. I., Beacham, T. D., Maes, \nG. E., Martinsohn, J. T., Miller, K. M., Naish, K. A., Ovenden, J. R., Primmer, C. R., Young Suk, H., \nTherkildsen, N. O., & Withler, R. E. (2017). Harness ing the power of genomics to secure the \nfuture of seafood. Trends in Ecology & Evolution, 32(9), 665–680. \nBernos, T. A., Gibelli, J., Michaelides, S., Won, H., Jeon, H. -B., Marin, K., Boguski, D. A., Janjua, M. \nY., Gallagher, C. P., Howland, K. L., & Fraser, D. J. (2024). Widespread admixture blurs \npopulation structure and confounds Lake Trout (Salvelinus namaycush) conservation even in \nthe genomic era. Scientific Reports, 14(1), 30838. \nBlain, S. A., Justen, H. C., Easton, W., & Delmore, K. E. (2024). Reduced hybrid survival in a migratory \ndivide between songbirds. Ecology Letters, 27(4), e14420. \nChen, S., Zhou, Y., Chen, Y., & Gu, J. (2018). fastp: an ultra -fast all-in-one FASTQ preprocessor. \nBioinformatics (Oxford, England), 34(17), i884–i890. \nCollyer, M. L., & Adams, D. C. (2018). RRPP: An r package for fitting linear models to high -\ndimensional data using residual randomization. Methods in Ecology and Evolution , 9(7), \n1772–1779. \nDahms, C., Roch, S., Elmer, K. R., Ros, A., Brinker, A., & Jacobs, A. (2024). Intra -lake origin and \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 28 \nrapid expansion of invasive pelagic three -spined stickleback in Lake Constance. NeoBiota \n92: 259-280. \nDe-Kayne, R., Selz, O. M., Marques, D. A., Frei, D., Seehausen, O., & Feulner, P. G. D. (2022). Genomic \narchitecture of adaptive radiation and hybridization in Alpine whitefish. Nature \nCommunications, 13(1), 1–13. \nDe-Kayne, R., Zoller, S., & Feulner, P. G. D. (2020). A de novo chromosome-level genome assembly \nof Coregonus sp.“Balchen”: One representative of the Swiss Alpine whitefish radiation. \nMolecular Ecology Resources, 20(4), 1093–1109. \nDeSaix, M. G., Anderson, E. C., Bossu, C. M., Rayne, C. E., Schweizer, T. M., Bayly, N. J., Narang, D. \nS., Hagelin, J. C., Gibbs, H. L., Saracco, J. F., Sherry, T. W., Webster, M. S., Smith, T. B., Marra, \nP. P., & Ruegg, K. C. (2023). Low -coverage whole ge nome sequencing for highly accurate \npopulation assignment: Mapping migratory connectivity in the American Redstart (Setophaga \nruticilla). Molecular Ecology, 32(20), 5528–5540. \nDeSaix, M. G., Bulluck, L. P., Eckert, A. J., Viverette, C. B., Boves, T. J., Reese, J. A., Tonra, C. M., & \nDyer, R. J. (2019). Population assignment reveals low migratory connectivity in a weakly \nstructured songbird. Molecular Ecology, 28(9), 2122–2135. \nDeSaix, M. G., Rodriguez, M. D., Ruegg, K. C., & Anderson, E. C. (2024). Population assignment from \ngenotype likelihoods for low -coverage whole-genome sequencing data. Methods in Ecology \nand Evolution, 15(3), 493–510. \nEckmann, R., & Rösch, R. (1998). Lake Constance fisheries and fish ecology . https://kops.uni-\nkonstanz.de/entities/publication/c6eca024-23d7-42ed-8819-8ea0c84e8495 \nEtheridge, E. C., Adams, C. E., Bean, C. W., Durie, N. C., Gowans, A. R. D., Harrod, C., Lyle, A. A., \nMaitland, P. S., & Winfield, I. J. (2012). Are phenotypic traits useful for differentiating among \na priori Coregonus taxa? Journal of Fish Biology, 80(2), 387–407. \nExposito-Alonso, M., Booker, T. R., Czech, L., Gillespie, L., Hateley, S., Kyriazis, C. C., Lang, P. L. M., \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 29 \nLeventhal, L., Nogues-Bravo, D., Pagowski, V., Ruffley, M., Spence, J. P., Toro Arana, S. E., Weiß, \nC. L., & Zess, E. (2022). Genetic diversity loss in the Anthropocene. Science (New York, N.Y.), \n377(6613), 1431–1435. \nFeulner, P. G. D., & Seehausen, O. (2019). Genomic insights into the vulnerability of sympatric \nwhitefish species flocks. Molecular Ecology, 28(3), 615–629. \nFox, E. A., Wright, A. E., Fumagalli, M., & Vieira, F. G. (2019). ngsLD: evaluating linkage \ndisequilibrium using genotype likelihoods. Bioinformatics , 35(19), 3855–3856. \nFrei, D., De-Kayne, R., Selz, O. M., Seehausen, O., & Feulner, P. G. D. (2022). Genomic variation from \nan extinct species is retained in the extant radiation following speciation reversal. Nature \nEcology & Evolution, 1–8. \nFrei, D., Mwaiko, S., Seehausen, O., & Feulner, P. G. D. (2023). Ecological disturbance reduces \ngenomic diversity across an Alpine whitefish adaptive radiation. Evolutionary Applications. \nhttps://doi.org/10.1111/eva.13617 \nFrei, D., Reichlin, P., Seehausen, O., & Feulner, P. G. D. (2023). Introgression from extinct species \nfacilitates adaptation to its vacated niche. Molecular Ecology, 32(4), 841–853. \nFriedman, K. J., Bartley, D. M., Rodríguez -Ezpeleta, N., Beveridge, M., Carolsfeld, J., Carvalho, G., \nCowx, I., Dean, G., & Glazov, E. (2022). Current and future genetic technologies for fisheries \nand aquaculture: implications for the work of FAO (No. No. 1387). FOOD AND AGRICULTURE \nORGANIZATION OF THE UNITED NATIONS.   \nGaio, D., Anantanawat, K., To, J., Liu, M., Monahan, L., & Darling, A. E. (2022). Hackflex: low -cost, \nhigh-throughput, Illumina Nextera Flex library construction. Microbial Genomics, 8(1). \nGarner, B. A., Hand, B. K., Amish, S. J., Bernatchez, L., Foster, J. T., Miller, K. M., Morin, P. A., Narum, \nS. R., O’Brien, S. J., Roffler, G., Templin, W. D., Sunnucks, P., Strait, J., Warheit, K. I., Seamons, \nT. R., Wenburg, J., Olsen, J., & Luikart, G. (2016). Genomics in conservation: Case studies and \nbridging the gap between data and application. Trends in Ecology & Evolution, 31(2), 81–83. \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 30 \nGum, B., Geist, J., Eckenfels, S., & Brinker, A. (2014). Genetic diversity of upper Lake Constance \nwhitefish Coregonus spp. under the influence of fisheries: a DNA study based on archived \nscale samples from 1932, 1975 and 2006. Journal of Fish Biology, 84(6), 1721–1739. \nHaase, S., Baer, J., Berg, C. W., DeWeber, J. T., Roch, S., Zimmermann, C., & Brinker, A. (2025). New \ninsights into the collapse of Upper Lake Constance whitefish (Coregonus wartmanni) and the \nimpact of commercial fisheries management. Canadian Journal of Fisheries and Aquatic \nSciences, 82, 1–10. \nHan, K. Y., Brennan, R. S., Monk, C. T., Jentoft, S., Helmerson, C., Dierking, J., Hüssy, K., Kokubun, É. \nE., Fuss, J., Krause-Kyora, B., Thomsen, T. B., Heredia, B. D., & Reusch, T. B. H. (2025). Genomic \nevidence for fisheries -induced evolution in Eastern  Baltic cod. Science Advances , 11(26), \neadr9889. \nHarrod, C., Mallela, J., & Kahilainen, K. K. (2010). Phenotype-environment correlations in a putative \nwhitefish adaptive radiation: Phenotype-environment correlations in whitefish. The Journal of \nAnimal Ecology, 79(5), 1057–1068. \nHartmann, J., & Knöpfler, G. (1977). Zur Unterscheidung von Blaufelchen und Gangfisch des \nBodensees. Fischwirt. 27, 49-50. \nHendricks, S., Anderson, E. C., Antao, T., Bernatchez, L., Forester, B. R., Garner, B., Hand, B. K., \nHohenlohe, P. A., Kardos, M., Koop, B., Sethuraman, A., Waples, R. S., & Luikart, G. (2018). \nRecent advances in conservation and population genomics data a nalysis. Evolutionary \nApplications, 11(8), 1197–1211. \nHirsch, P. E., Eckmann, R., Oppelt, C., & Behrmann -Godel, J. (2013). Phenotypic and genetic \ndivergence within a single whitefish form --detecting the potential for future divergence. \nEvolutionary Applications, 6(8), 1119–1132. \nJacobs, A., Carruthers, M., Eckmann, R., Yohannes, E., Adams, C. E., Behrmann -Godel, J., & Elmer, \nK. R. (2019). Rapid niche expansion by selection on functional genomic variation after \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 31 \necosystem recovery. Nature Ecology & Evolution, 3(1), 77–86. \nJacobs, A., De Noia, M., Praebel, K., Kanstad -Hanssen, Ø., Paterno, M., Jackson, D., McGinnity, P., \nSturm, A., Elmer, K. R., & Llewellyn, M. S. (2018). Genetic fingerprinting of salmon louse \n(Lepeophtheirus salmonis) populations in the North -East Atlantic using a random forest \nclassification approach. Scientific Reports, 8(1), 1203. \nKardos, M., Armstrong, E. E., Fitzpatrick, S. W., Hauser, S., Hedrick, P. W., Miller, J. M., Tallmon, D. \nA., & Funk, W. C. (2021). The crucial role of genome -wide genetic variation in conservation. \nProceedings of the National Academy of Sciences of the United States of America , 118(48), \ne2104642118. \nKefford, B. J., Nichols, S. J., & Duncan, R. P. (2023). The cumulative impacts of anthropogenic \nstressors vary markedly along environmental gradients. Global Change Biology, 29(3), 590–\n602. \nKorneliussen, T. S., Albrechtsen, A., & Nielsen, R. (2014). ANGSD: Analysis of Next Generation \nSequencing Data. BMC Bioinformatics, 15, 356. \nLeitwein, M., Duranton, M., Rougemont, Q., Gagnaire, P. -A., & Bernatchez, L. (2020). Using \nHaplotype Information for Conservation Genomics. Trends in Ecology & Evolution , 35(3), \n245–258. \nLi, H., & Durbin, R. (2009). Fast and accurate short read alignment with Burrows -Wheeler \ntransform. Bioinformatics (Oxford, England), 25(14), 1754–1760. \nLi, H., & Ralph, P. (2019). Local PCA Shows How the Effect of Population Structure Differs Along \nthe Genome. Genetics, 211(1), 289–304. \nLou, R. N., Jacobs, A., Wilder, A. P., & Therkildsen, N. O. (2021). A beginner’s guide to low-coverage \nwhole genome sequencing for population genomics. Molecular Ecology, 30(23), 5966–5993. \nLou, R. N., & Therkildsen, N. O. (2022). Batch effects in population genomic studies with low -\ncoverage whole genome sequencing data: Causes, detection and mitigation. Molecular \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 32 \nEcology Resources. https://onlinelibrary.wiley.com/doi/abs/10.1111/1755-0998.13559 \nMeisner, J., & Albrechtsen, A. (2018). Inferring Population Structure and Admixture Proportions in \nLow-Depth NGS Data. Genetics, 210(2), 719–731. \nMérot, C., Berdan, E. L., Cayuela, H., Djambazian, H., Ferchaud, A. -L., Laporte, M., Normandeau, E., \nRagoussis, J., Wellenreuther, M., & Bernatchez, L. (2021). Locally Adaptive Inversions \nModulate Genetic Variation at Different Geographic Scales in a Seawe ed Fly. Molecular \nBiology and Evolution, 38(9), 3953–3971. \nMoser, D., Frey, A., & Berner, D. (2016). Fitness differences between parapatric lake and stream \nstickleback revealed by a field transplant. Journal of Evolutionary Biology, 29(4), 711–719. \nNumann, W. (1978). VERSUCH EINER KLASSIFIZIERUNG DER BODENSEECOREGONEN DURCH \nVERGLEICH KOMBINIERTER MERKMALE UND DIE EINGLIEDERUNG DES SOGENANNTEN \nBRAUNFELCHEN.   \nNumann, W. (1986). Bastardierungsvorgänge bei Felchen des Bodensee-Untersees vor 20 Jahren. \nArchiv Für Hydrobiologie, 105(3), 285–288. \nPearse, D. E. (2016). Saving the spandrels? Adaptive genomic variation in conservation and \nfisheries management: saving the spandrels. Journal of Fish Biology, 89(6), 2697–2716. \nPedersen, B. S., & Quinlan, A. R. (2018). Mosdepth: quick coverage calculation for genomes and \nexomes. Bioinformatics (Oxford, England), 34(5), 867–868. \nPinsky, M. L., & Palumbi, S. R. (2014). Meta -analysis reveals lower genetic diversity in overfished \npopulations. Molecular Ecology, 23(1), 29–39. \nQuintela, M., Kvamme, C., Bekkevold, D., Nash, R. D. M., Jansson, E., Sørvik, A. G., Taggart, J. B., \nSkaala, Ø., Dahle, G., & Glover, K. A. (2020). Genetic analysis redraws the management \nboundaries for the European sprat. Evolutionary Applications, 13(8), 1906–1922. \nReid, A. J., Carlson, A. K., Creed, I. F., Eliason, E. J., Gell, P. A., Johnson, P. T. J., Kidd, K. A., \nMacCormack, T. J., Olden, J. D., Ormerod, S. J., Smol, J. P., Taylor, W. W., Tockner, K., Vermaire, \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 33 \nJ. C., Dudgeon, D., & Cooke, S. J. (2019). Emerging threats and persistent conservation \nchallenges for freshwater biodiversity: Emerging threats to freshwater biodiversity. Biological \nReviews of the Cambridge Philosophical Society, 94(3), 849–873. \nRobinson, J. A., Ortega -Del Vecchyo, D., Fan, Z., Kim, B. Y., vonHoldt, B. M., Marsden, C. D., \nLohmueller, K. E., & Wayne, R. K. (2016). Genomic flatlining in the endangered island fox. \nCurrent Biology. 26(9), 1183–1189. \nRobledo, D., Palaiokostas, C., Bargelloni, L., Martínez, P., & Houston, R. (2018). Applications of \ngenotyping by sequencing in aquaculture breeding and genetics. Reviews in Aquaculture , \n10(3), 670–682. \nRos, A., Dunst, J., Gugele, S., & Brinker, A. (2019). Anti -predator mechanisms in evolutionarily \npredator-naïve vs. adapted fish larvae. Ecosphere (Washington, D.C), 10(4), e02699. \nRösch, R., Baer, J., & Brinker, A. (2018). Impact of the invasive three -spined stickleback \n(Gasterosteus aculeatus) on relative abundance and growth of native pelagic whitefish \n(Coregonus wartmanni) in Upper Lake Constance. Hydrobiologia, 824(1), 243–254. \nSchluter, D., Veen, T., Thompson, K. A., Owens, G. L., & Rennison, D. J. (2025). Coexistence in \nsympatry with gene flow before speciation has completed. Molecular Ecology, e70079. \nSchweizer, M. (1894). Die Felchen des Bodensee und ihre natürliche und künstliche Vermehrung. \n– Mitteilungen der Thurgauischen Naturforschenden Gesellschaft. Mitteilungen Der \nThurgauischen Naturforschenden Gesellschaft, 11, 13–26. \nSeljestad, G. W., Quintela, M., Bekkevold, D., Pampoulie, C., Farrell, E. D., Kvamme, C., Slotte , A., \nDahle, G., Sørvik, A. G., Pettersson, M. E., Andersson, L., Folkvord, A., Glover, K. A., & Berg, F. \n(2024). Genetic stock identification reveals mismatches between management areas and \npopulation genetic structure in a migratory pelagic fish. Evolutionary Applications, 17(10), \ne70030. \nShpak, M., Lawrence, K. N., & Pool, J. E. (2025). The precision and power of population branch \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 34 \nstatistics in identifying the genomic signatures of local adaptation. Genome Biology and \nEvolution, 17(5), evaf080. \nŠlenker, M., Koutecký , P., & Marhold, K. (2022). MorphoTools2: an R package for multivariate \nmorphometric analysis. Bioinformatics (Oxford, England), 38(10), 2954–2955. \nSteffen, W., Richardson, K., Rockström , J., Cornell, S. E., Fetzer, I., Bennett, E. M., Biggs, R., \nCarpenter, S. R., de Vries, W., de Wit, C. A., Folke, C., Gerten, D., Heinke, J., Mace, G. M., Persson, \nL. M., Ramanathan, V., Reyers, B., & Sörlin, S. (2015). Sustainability. Planetary boundarie s: \nguiding human development on a changing planet. Science (New York, N.Y.) , 347(6223), \n1259855. \nSteinmann, P. (1950). Monographie der schweizerischen Koregonen: Beitrag zum Problem der \nEntstehung neuer Arten. Schweizerische Zeitschrift für Hydrologie, 12(2), 340–491. \nTarasov, A., Vilella, A. J., Cuppen, E., Nijman, I. J., & Prins, P. (2015). Sambamba: fast processing of \nNGS alignment formats. Bioinformatics (Oxford, England), 31(12), 2032–2034. \nTengstedt, A. N. B., Liu, S., Jacobsen, M. W., Gundlund, C., Møller, P. R., Berg, S., Bekkevold, D., & \nHansen, M. M. (2025). Genomic insights on conservation priorities for North Sea houting and \nEuropean lake whitefish (Coregonus spp.). Molecular Ecology, 34(23), e17367. \nTheissinger, K., Fernandes, C., Formenti, G., Bista, I., Berg, P. R., Bleidorn, C., Bombarely, A., Crottini, \nA., Gallo, G. R., Godoy, J. A., Jentoft, S., Malukiewicz, J., Mouton, A., Oomen, R. A., Paez, S., \nPalsbøll, P. J., Pampoulie, C., Ruiz-López, M. J., Secomandi, S., … European Reference Genome \nAtlas Consortium. (2023). How genomics can help biodiversity conservation. Trends in \nGenetics: TIG, 39(7), 545–559. \nTherkildsen, N. O., & Palumbi, S. R. (2017). Practical low -coverage genomewide sequencing of \nhundreds of individually barcoded samples for population and evolutionary genomics in \nnonmodel species. Molecular Ecology Resources, 17(2), 194–208. \nThomas, G., Rösch, R., & Eckmann, R. (2010). Seasonal and long-term changes in fishing depth of \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint \n\n 35 \nLake Constance whitefish: CHANGES IN FISHING DEPTH OF LAKE CONSTANCE \nWHITEFISH. Fisheries Management and Ecology, 17(5), 386–393. \nTickner, D., Opperman, J. J., Abell, R., Acreman, M., Arthington, A. H., Bunn, S. E., Cooke, S. J., Dalton, \nJ., Darwall, W., Edwards, G., Harrison, I., Hughes, K., Jones, T., Leclère, D., Lynch, A. J., Leonard, \nP., McClain, M. E., Muruven, D., Olden, J. D., … Young, L. (2020). Bending the Curve of Global \nFreshwater Biodiversity Loss: An Emergency Recovery Plan. BioScience, 70(4), 330–342. \nVardakas, L., Perdikaris, C., Freyhof, J., Zimmerman, B., Ford, M., Vlachopoulos, K., Koutsikos, N., \nKaraouzas, I., Chamoglou, M., & Kalogianni, E. (2025). Global patterns and drivers of \nfreshwater fish extinctions: Can we learn from our losses? Global Change Biology , 31(5), \ne70244. \nVerta, J.-P., & Jones, F. C. (2019). Predominance of cis -regulatory changes in parallel expression \ndivergence of sticklebacks. eLife, 8. \nVonlanthen, P., Bittner, D., Hudson, A. G., Young, K. A., Müller, R., Lundsgaard -Hansen, B., Roy, D., \nDi Piazza, S., Largiader, C. R., & Seehausen, O. (2012). Eutrophication causes speciation \nreversal in whitefish adaptive radiations. Nature, 482(7385), 357–362. \n \n \n.CC-BY 4.0 International licenseavailable under a \n(which 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 \nThe copyright holder for this preprintthis version posted January 18, 2026. ; https://doi.org/10.64898/2026.01.18.700192doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}